Open main menu
TWA Flight 800 investigation
Portal:TWA Flight 800 investigation

NTSB Board Meeting on TWA 800
August 22, 2000

Day 1 of 2
Morning Session

Jim Hall: Good morning and welcome to this meeting of the National Transportation Safety Board. This morning's agenda item is an aviation accident report, carried as Board notation 6788G, an in-flight break-up over the Atlantic Ocean of Trans World Airlines Boeing 747-131, Registration N93119 that occurred near East Moriches, New York on July 17, 1996.

Under the Government in Sunshine Act, multi-member federal agencies, such as the Safety Board, conduct much of their business in open session. Therefore Board meetings are often called "Sunshine meetings." While the public is invited to observe today's meeting, only the Board Members and NTSB staff will participate in discussions. Today's meeting is also being simulcast to a worldwide audience on our website at

A handout is available to our guests at the entrance to the Boardroom reiterating the information about the investigative process. Copies of a pamphlet that explains the Board and its work in some detail are also available there, and general information is also available on the Board's website.

During this meeting, the Board Members will discuss a draft report that has been prepared by staff. Because it is at this point just a draft, it is not available to the public. We will discuss the report section by section, soliciting staff comments and explanations on many points. Once we have reviewed all the issues, we will consider the staff's draft conclusions, probable cause determination, and specific safety recommendations. We will then determine if we should approve the draft with any revisions that we have discussed.

Sometimes all or part of a draft conclusion, probable cause, or recommendation is revised or rejected by the Board Members. That is because you are reviewing the Board's actual deliberations over this document. That is the purpose of the Sunshine Act - to provide the public with a window into the decision-making process.

Now, a couple of brief administrative announcements for those in our audience. In the event of an emergency, such as a fire, the building alarm system will activate and a voice message will instruct that we vacate the building. You should proceed to the nearest exit. There are emergency exits up here to the left and right of the platform and at the back of the room.

Restrooms are located in the foyer on the left as you exit this room, and on the promenade level above us. You may use the phones in the foyer for local and credit card calls. Cell phones will work if you walk outside this room. Most do not work in here, nor do most pagers. There are many eating establishments on the promenade level. Please understand that food and drink of any kind may not be brought into the meeting room.

There are many members of the NTSB staff here today. I know they will be glad to assist you in any way they can. Please do not hesitate to ask them for help if any of us can provide assistance.

Ladies and gentlemen, almost 1,500 days have passed since that terrible day in July 1996 when TWA Flight 800 crashed off the coast of Long Island, New York. It was a tragic event that stirred strong emotions and feelings throughout this country and throughout the world.

We had an airline of world renown, a category of aircraft, the Boeing 747, that had compiled an outstanding safety record in some three decades of service - and yet, 230 individuals lost their lives in a very few stark moments.

The crash of Flight 800 graphically demonstrates that even in one of the safest transportation systems in the world, things can go horribly wrong. It should stand as a reminder to all of us of the need for diligence and aggressive action in identifying and eliminating potential safety problems.

I would like to welcome at this time and acknowledge the presence here today of many of the 800 family members. These next two days, as the last four years have been, may be difficult for you. But I do hope that you take some comfort from seeing the great amount of work that has gone into this investigation. I want you to know that all our efforts have been aimed at preventing similar tragedies in the future.

I would also again express my appreciation to the authorities in New York - to the police, the divers, the fire rescue and emergency assistance units, the Red Cross, the Salvation Army, as well as the many private citizens - who made valiant efforts in the immediate hours and days after the aircraft went down. I would also like to thank the Coast Guard, the Navy, the FBI, NOAA [National Oceanographic and Atmospheric Administration], and the many other state and federal agencies that assumed major roles in the search and recovery effort.

I would like to note the encouragement and support we have received from the White House and Congress in providing the resources needed to conduct what has become the most extensive, complex, and expensive investigation in the Safety Board's 33-year history.

From the beginning, the scope and dimensions of this investigation have been extraordinary. The salvage effort organized by the Navy, one of the largest diver-assisted salvage operations ever conducted, extended from July to November 1996. The Navy divers worked in very difficult and dangerous conditions, and for a time their efforts had to be halted because of the onset of the Atlantic hurricane season. When the diving operations were completed, there followed months of work by contracted fishing trawlers that scoured hundreds of miles of the ocean floor. In the end, we recovered the remains of all 230 victims and more than 95 percent of the aircraft.

The reconstruction of a 93-foot segment of the aircraft fuselage, including the center wing fuel tank, was unique both in size and scope. More than 30 people worked meticulously for many months to sort through innumerable pieces of wreckage and assemble the reconstruction in an effort to better understand what happened to Flight 800.

The number of organizations, public and private, that played a significant role in this investigation is extensive. I'd like to pause for a few minutes, so you can see the almost 500 names of those entities and individuals that contributed to the investigative process. I direct your attention to the screens in front of you. The Safety Board staff and various government and private research organizations, under contract to the National Transportation Safety Board, undertook an unprecedented amount of research and testing that was paid for by the American taxpayers.

For example, Safety Board staff leased a Boeing 747 to study the temperatures and environment inside the aircraft's center wing tank. We also conducted extensive research into the composition and explosive characteristics of Jet A fuel. In addition, we conducted test and computer simulation work to study flame and pressure propagation in the center wing tank. Early on in the process, investigators began looking at what role electromagnetic interference from external emitters or sources internal to the aircraft may have played in the crash.

The investigation also included the most extensive radar data study in the Board's history, including a review of several hundred thousand radar returns from nine radar locations in five states. The investigative team also spent a great deal of time organizing and carefully analyzing the summaries of witness interviews the Federal Bureau of Investigation provided to the Board. We will be reviewing the work done by the witness group, and many of the others, during the course of this meeting.

All of the investigative work undertaken as part of this investigation was extremely complex. Because of the need for precision and, in some cases, the danger posed to those performing the tests, the work had to be painstakingly done to make sure that it was done properly, safely, and accurately. And, of course, it was not inexpensive.

We were fortunate to secure the assistance of a broad array of institutions, including the Department of Defense laboratories at Wright-Patterson Air Force Base and the Navy's China Lake and Patuxent River facilities. Important work was also done at NASA's Langley Research Center and the Sandia Laboratories, among others.

We also contracted with private institutions, such as the California Institute of Technology and the University of Nevada, Reno, and various specialties to conduct research. Experts from other countries, including the United Kingdom, Norway, and Canada, also assisted us, and the French aviation authorities participated under the terms of the Convention on International Civil Aviation.

Much has been learned over the course of the past four years, and the five Board Members seated before you will be examining and discussing the results of the staff's work during this Sunshine meeting. I must emphasize that over the next two days you will observe some extremely technical discussions about the issues raised in the investigation. In preparation for this meeting, the Board Members each read the 684-page report and the 177 pages of information that were provided in party submissions. The extensive record of this investigation now approaches some 15,000 pages and is available to everyone in the Board's public docket. The investigative groups' factual reports can also be found on our web page. The other supporting documentation is available in CD-ROM format.

During the course of this investigation, the Board received a great number of suggestions and comments from many individuals and organizations on possible causes of the crash of Flight 800 and recommendations for possible lines of investigation. Much of this commentary has been well informed, and we appreciate receiving it. Safety Board staff has reviewed all of this material, and took those ideas that appeared to have a scientific basis and offered a reasonable line of inquiry into account as the accident investigation progressed.

In the early months of the investigation, it became clear that an explosion of flammable vapors in the aircraft's center wing tank initiated the break-up and subsequent crash of Flight 800. In December 1996, based on the Board's conclusion that heated, flammable vapors in the aircraft fuel tank pose a serious risk to safe flight, the Board recommended that the Federal Aviation Administration study design changes to deal with this problem and that, in the interim, they require operational changes to enhance safety. In April 1998, the Board issued another set of recommendations focused on aircraft wiring and the fuel quantity indication system. During this meeting, we will be assessing what has been done in response to those recommendations, as well as what as what remains to be accomplished.

More broadly, the Flight 800 investigation has uncovered and focused the attention of the aviation community on some very important safety issues - fuel tank protection, the vulnerability of aircraft wiring, and a number of aging aircraft issues. We will pursue each of these items in some detail over the next two days. This is a lot of ground to cover, but before moving ahead, I would like to make one additional comment.

I know that at the outset many believed that the crash of Flight 800 was caused by a criminal act. And for many the events of the times - the ongoing court trials in the aftermath of the World Trade Center bombings in New York, and the heightened concern about terrorism at the 1996 Olympic Games in Atlanta - seemed to lend a certain credence to the notion. Certainly, the nature of the event and its rarity led some to question whether the crash of Flight 800 was really an accident.

As many of you know, a substantial law enforcement investigation was conducted in parallel with the Safety Board's investigation. After conducting a thorough investigation, the FBI suspended its investigation in November 1997, indicating that no evidence had been found to indicate that a criminal act was the cause of the tragedy of TWA Flight 800.

Despite this finding by our nation's law enforcement agency, the Federal Bureau of Investigation, some have urged the Safety Board to assume, in effect, a law enforcement role to prove or disprove their assertion that the crash of Flight 800 was the result of a bomb or a missile. That is beyond this agency's mandate and authority. Our focus is safety. Our people are aviators, engineers, and scientists - I believe, some of the best in the world - but they are not criminal investigators.

However, even though our employees are not law enforcement personnel, they examined every piece of wreckage for any physical evidence that the crash of Flight 800 had been caused by a bomb or missile. Had we found such evidence, we would have immediately referred the matter back to the appropriate law enforcement agencies for their action. Let me state unequivocally, the Safety Board has found no evidence.

To the families of Flight 800, I would like to add this comment: It is unfortunate that a small number of people, assuming their own agendas, have persisted in making unfounded charges of a government cover-up in this investigation. These people do a grievous injustice to the many dedicated individuals, civilian and military, who have been involved in this investigation. Some 75 NTSB members have participated in this investigation. I pause while their names are listed on the screens in front of you.

These individuals, collectively, have more than 1,000 years of government and aviation industry experience. Many of them have served in the military, including service in Vietnam and the Gulf War. These men and women, in my opinion, represent the very best in United States government service. They are public servants all of us can be proud of.

I recognize that this TWA 800 investigation is technically complex, and that knowledgeable people can disagree over some substantive matters. But I take exception to those who consistently distort the record and persist in making unfounded charges of a cover-up. They do a disservice to all of us - but most especially to you, the families of the TWA 800 victims, who have suffered so much in this tragedy. And for that I am very sorry.

In its 33-year history, the Safety Board has earned a well-deserved reputation for independence, impartiality, honesty, and diligence. We have adhered to those values during this investigation, as we do in each investigation. The NTSB staff has the highest personal and professional integrity, and I assure you that we have done our very best to find the cause of this accident and to make recommendations that will prevent similar accidents from occurring in the future.

With that, Mr. Campbell, would you please introduce the staff for today's meeting.

Daniel Campbell: Good morning, Mr. Chairman, Members of the Board. I'll introduce staff at the front table and staff behind me will be introduced as they take part in the presentations. Working from my immediate right is Ronald Battocchi, agency General Counsel; Bernard Loeb, Director of the Office of Aviation Safety; Alfred Dickinson, who is the Investigator-in-Charge in the accident; James Wildey, who is the Metallurgy and Sequencing Group Chairman; Joseph Kolly, who did the Fire and Explosion Group work; Robert Swaim, who was Systems Group Chairman.

Dr. Loeb will begin this morning's presentations with an opening statement. Dr. Loeb has been employed at NTSB for 22 years and among assignments before becoming Director of Aviation Safety he served as the Director of the Office of Research and Engineering and as the Acting Director of our previous Bureau of Accident Investigation. He also has 15 years of prior private sector and government research experience, an undergraduate degree from Maryland University and his doctorate from George Washington in engineering and science. Dr. Loeb has twice received the presidential Distinguished Rank Award for his work as a senior government executive. Dr. Loeb.

Bernard Loeb: Thank you. Good morning, Mr. Chairman, Members of the Board. As you have just said, Mr. Chairman, the investigation into the crash of TWA Flight 800 has been the most extensive and encompassing aviation accident investigation the Safety Board has ever undertaken. It has truly been a monumental effort for everyone involved. Most people are aware of the lengthy underwater recovery operations and the large-scale reconstruction of the airplane fuselage and parts of the cabin interior that took place on scene in Long Island - and indeed, that massive effort symbolizes the extent to which this investigation has gone in leaving no stone unturned. Al Dickinson, the Investigator-In-Charge, will be discussing the on-scene portion of the investigation during his opening remarks. So I will not go into detail about that at this time.

What I am going to do is to summarize the significant findings of our investigation. This will just be an overview - more detailed explanations will be provided by the investigators during their individual presentations over the next two days. But I think an overall summary at this point would be valuable to put things in context.

First, we knew almost immediately after the accident that TWA Flight 800 had experienced an in-flight breakup. This was strongly suggested by the radar data - there was a loss of transponder returns and the primary radar returns indicated that pieces had departed the airplane and were fairly widely dispersed in the ocean. The wreckage recovery locations made it evident relatively early in the investigation that the in-flight break-up was initiated by an event in the area of the fuselage near the forward part of the center wing tank.

Specifically, pieces from the forward part of the center wing tank and adjacent areas of fuselage were recovered from the westernmost portion of the wreckage field (the portion of the wreckage field closest to JFK Airport from where Flight 800 took off). This first wreckage area is referred to as the "red zone." The recovery of the pieces from the red zone indicated that they were the first pieces to separate from the airplane. The nose portion of the airplane was found farther to the east, in what was labeled the "yellow zone," indicating that this portion of the airplane separated later in the breakup sequence. And most of the remaining wreckage was found in the easternmost portion of the wreckage field, farthest from JFK, which was labeled the "green zone."

This basic evidence - the radar data and the wreckage recovery locations - indicated that the airplane broke up in flight, and that the break-up initiated in the area of the fuselage near the forward part of the center wing tank.

On the basis of this initial information, we considered several possible causes for the initiation of the in-flight break-up:

· a structural failure and decompression;

· a detonation of a high-energy explosive device, such as a bomb or missile warhead; and

· a fuel air vapor explosion in the center wing tank.

We found no evidence that a structural failure and decompression initiated the break-up. A thorough examination of the wreckage by our engineers and metallurgists did not reveal any evidence of fatigue, corrosion, or any other structural fault that could have led to the break-up.

As a side note, I would like to mention that there was absolutely no evidence of an in-flight separation of the forward cargo door - one of the many theories suggested to us by the members of the public. The physical evidence demonstrated that the forward cargo door was closed and latched at water impact.

We also considered the possibility of a bomb or missile. However, high-energy explosions leave distinctive damage signatures on the airplane's structure, such as severe pitting, cratering, hot gas washing, and petaling. No such damage was found on any portion of the recovered airplane structure, and as you know, more than 95 percent of the airplane was recovered. Our investigators, together with many outside participants from the parties to the investigation, closely examined every piece of recovered wreckage. All of the participants agreed that none of the wreckage exhibited any of the damage characteristics of a high-energy explosion - that is, of a bomb or a missile.

Further, no missing portions of fuselage were large enough to represent the entry of a missile. You may have noticed that some of the photographs of the reconstruction show what appear to be several large missing areas, such as those that are shown on the screen now. However, almost all of the fuselage structure in these areas is actually attached to the adjacent pieces, but has been folded back or crushed in such a way that it does not cover its original area. Therefore, these large gaps that appear to exist in the reconstructed fuselage do not represent areas of damage that could have been caused by a missile.

In addition, we found no localized area of severe thermal or fragmentation pieces and no localized severe damage or fragmentation of the seats, such as would be expected if a high-energy explosive device had detonated inside the airplane. The injuries to the occupants and the damage to the airplane were fully consistent with an in-flight break-up and subsequent water impact. In light of all this evidence, a bomb or missile strike has been ruled out as an initiating event of the in-flight break-up.

The FBI did find trace amounts of explosive residue on three pieces of the wreckage. However, these three pieces contain no evidence of pitting, cratering, hot gas washing, or petaling, which would have been there had these trace amounts resulted from a bomb or missile. Further, these trace amounts could have been transferred to these pieces in various ways. For example, in connecting with ferrying troops during the Gulf War or during dog-training explosive detection exercises that were conducted on the accident airplane about one month before the accident. There is also the possibility that the explosive residues could have been deposited on the wreckage during or after the recovery operations as a result of contact with the military personnel, ships, and vehicles used during those operations. We don't know exactly how the explosive residues got there - but we do know from the physical evidence I've just discussed that the residues were not the result of the detonation of a bomb.

Unlike the other two scenarios I've just mentioned (a structural failure or a high-energy explosive), the third scenario we considered - a Jet A fuel/air explosion in the center wing tank - was consistent with the physical evidence. Specifically, as I've already mentioned, the wreckage recovery locations indicated that the first pieces to depart the airplane were from in and around the front of the center wing tank.

Based on these recovery locations and damage characteristics, the investigative group led by Jim Wildey (known as the Metallurgy and Structures/Sequencing Group) determined that the earliest event in the break-up sequence was an overpressure inside the center wing tank that caused structural failure of its forward section. This overpressure event started the break-up sequence that ultimately resulted in the destruction of the airplane.

I would like to emphasize that all of the parties to the investigation, as well as numerous outside experts and researchers, have agreed with the findings of the Sequencing Group.

Jim Wildey will be explaining the break-up sequence a little later today. The point I would like to make now is simply that the initial break-up sequence and early departure of pieces from in and around the center wing tank clearly indicate that the break-up was initiated by an overpressure inside the center wing tank. Given that there was no high-energy explosion in this (or any other) area, this overpressure must have been caused by a fuel/air explosion inside the center wing tank.

However, questions were raised early in the investigation about whether the conditions necessary for a fuel/air explosion could have existed inside the accident airplane's center wing tank, and also whether a Jet A fuel/air explosion could generate sufficient pressure to break apart the fuel tank and destroy the airplane.

To address the first issue, the Safety Board conducted flight tests at JFK in July 1997 using a 747 leased from Evergreen Airlines. Several test flights were conducted under conditions similar to those experienced by Flight 800. The fuel/air vapor inside the center wing tank was measured at various locations during the flights. The temperatures inside the center wing tank at the altitude at which the accident occurred (approximately 13,800 feet) ranged between 101 and 127 degrees Fahrenheit.

Extensive work done by scientists at the California Institute of Technology showed that Jet A fuel under the conditions experienced by Flight 800 would be flammable at these temperatures. In fact, their work demonstrated that fuel vapors under those conditions may have been flammable at temperatures as low as 96 degrees. Dr. Joseph Kolly will be talking more about this research later today.

The second issue - whether an explosion of Jet A fuel could generate sufficient pressure to break apart the fuel tank and destroy the airplane - was also put to rest in the investigation. Laboratory tests and quarter-scale tests under the direction of scientists at the California Institute of Technology demonstrated that pressures exceeding the structural limitations of the forward portion of the center wing tank were produced from the combustion of a Jet A fuel/air mixture similar to the one that existed in the center wing tank of TWA Flight 800.

Further, computer modeling of combustion in a full-scale tank confirmed that a localized ignition could generate pressure levels that would cause the damage we saw in the wreckage of the center wing tank from Flight 800. Finally, previous fuel/air explosions in the center wing tanks of commercial airliners that contained Jet A fuel have confirmed that a center wing tank explosion involving Jet A fuel can result in the destruction of an airplane. Specifically, I am referring to the November 1989 accident involving a Boeing 727 operated by the Colombian airline Avianca that occurred during the climb after takeoff, and the May 1990 accident involving a Boeing 737 operated by Philippine Air Lines that occurred on the ground at the airport.

The bottom line is that our investigation confirmed that the fuel/air vapor in the center wing tank was flammable at the time of the accident and that a fuel/air explosion with Jet A fuel was more than capable of generating the pressure needed to break apart the center wing tank and destroy the airplane. Together with the other physical evidence I have already mentioned, this leads to the inescapable conclusion that the cause of the in-flight break-up of TWA Flight 800 was a fuel/air explosion inside the center wing tank.

The next obvious question is: What ignited the flammable vapor inside the center wing tank and caused the explosion? Our Systems Group Chairman, Bob Swaim, has devoted the past four years to this question. He and his group considered every conceivable potential ignition source. They did exhaustive research and testing - much of it with help and participation from outside experts - to better understand some of these potential ignition mechanisms. As a result of this work, we determined that a number of possible ignition sources were very unlikely in this case.

It's important to note that ignition sources that were deemed unlikely under the circumstances of this case could be ignition sources under other circumstances. Bob Swaim and others will be prepared to further discuss these ignition sources later today if the Board has any questions about why they were deemed unlikely.

One ignition source that we could not deem unlikely, however, was that a short circuit involving electrical wiring outside the center wing tank somehow transferred excess voltage to fuel quantity indication system wiring leading to the center wing tank. Although the voltage in the fuel quantity indication system line is limited by design to a very low level, a short circuit from higher-voltage wires could allow excessive voltage to be transferred to the fuel quantity indication system wires and enter the fuel tank.

We cannot be certain if this in fact occurred, but of all the ignition scenarios that we considered, this scenario is the most likely. As I said, Bob Swaim will be telling us much more about this part of the investigation later today.

The accident airplane was 25 years old at the time of the accident. The electrical wiring that was recovered showed definite signs of deterioration and damage. One of the things that Bob Swaim and his Systems Group did was examine the condition of electrical wiring in numerous transport category airplanes of various ages from a variety of different operators. What they found was that cracked and damaged wire insulation, contamination of electrical wiring, and noncompliant wiring repairs could be found throughout the transport fleet, but were especially common in older airplanes. Therefore, the condition of the accident airplane was not atypical for an airplane of its age.

Although these types of conditions are common, they obviously increase the potential for short circuits to occur and, therefore, are a cause for concern. It became clear from our investigation that current maintenance practices do not adequately protect aircraft electrical wiring, especially with regard to older airplanes. In April of 1998, the Board issued six recommendations to the FAA, aimed at correcting several of the potentially hazardous conditions that we found on aircraft wiring. I am pleased to say that the FAA has taken action towards addressing all of those recommendations.

The potentially hazardous wiring conditions discovered in the context of this accident investigation have focused attention on the need to change maintenance practices, and to ensure that the integrity and safety of aging airplane systems is maintained. The FAA has begun to address these issues by developing an Aging Transport Non-Structural Systems Plan, and establishing a rulemaking advisory committee. These initiatives are still ongoing, so we do not yet know the final results of these programs. Our draft report outlines our concerns in this area and includes a recommendation to ensure that all of the issues identified by the FAA's plan are addressed. Of course, we will be following the FAA's progress on these issues with great interest.

Bob Swaim will be talking more about this part of our investigation tomorrow.

Our examination of the numerous potential ignition sources - and the recognition that many of these ignition sources cannot be reliably eliminated - led us to another concern: the adequacy of the current certification philosophy, which assumes that a flammable fuel/air mixture exists in fuel tanks at all times and attempts to preclude fuel tank explosions solely by eliminating all ignition sources.

We concluded that this approach is seriously flawed because experience has demonstrated that all possible ignition sources cannot be reliably eliminated and, further, it is not rational to believe that we can predict all possible ignition sources. Therefore, the most effective approach to preventing fuel tank explosions is to eliminate flammable vapors inside fuel tanks in addition to attempting to eliminate ignition sources.

The results of the flight tests and the flammability research I discussed earlier are especially significant because they indicate that many commercial aircraft may routinely operate with flammable fuel/air vapor inside their fuel tanks. In particular, airplanes that have air conditioning packs located directly beneath their center wing tanks are especially likely to operate with flammable vapor in those tanks because of the large amount of heat generated by these packs.

The FAA has recognized this, and has proposed rulemaking that would preclude the use of such designs in the future, unless the design includes a means for reducing the transfer of heat to the fuel tank. One of the concerns we have, however, is that not enough has been done to reduce flammability in existing designs in the current fleet.

This accident would not have occurred but for the flammable vapor in the center wing tank, and we believe that perhaps the most valuable lesson that can be learned from this accident, and our best hope for preventing similar accidents in the future, lies in recognizing this fact.

The draft report that you have before you does not propose any new recommendations to address fuel tank flammability. But as you know, the Safety Board already recommended in December 1996 that the FAA preclude the operation of transport-category airplanes with explosive fuel/air mixtures in fuel tanks. The Board specifically asked the FAA to consider airplane design modifications, such as fuel tank inerting systems that would make the fuel/air mixture nonflammable. In addition to this long-term recommendation, the Board also asked the FAA to require more immediate, short-term changes - specifically, to require modifications in operational procedures that would reduce the potential for flammable fuel/air mixtures in fuel tanks.

As I just mentioned, the FAA has initiated rulemaking aimed at minimizing fuel tank flammability in newly designed airplanes. In addition, the FAA is also evaluating the use of directed ventilation to cool the center wing tank area, the use of ground-conditioned air instead of air conditioning packs when an airplane's on the ground, and fuel tank inerting systems both on-board for future design and ground-based for the existing fleet. Each of these approaches has some benefit to the existing fleet. But at this time, fuel tank inerting appears to be the most promising method for dramatically reducing fuel tank flammability in both future designs and in the existing fleet.

In addition to our concerns about fuel tank flammability, this investigation and several others have brought to light some broader issues regarding aircraft certification. For example, there are questions about the adequacy of the risk analyses that are used as the basis for demonstrating compliance with many certification requirements. We believe that the certification approach could be improved by requiring a reliable, independent means for overcoming or counteracting any potential failure that could cause catastrophic results, regardless of the calculated probability of that failure.

Although those issues are raised in the draft report on this accident, they are not examined in depth. However, we believe these would be appropriate issues to explore in depth as part of the Board's upcoming safety study on aircraft certification.

The final topic that we will address is the reported witness observations. In particular, we will discuss the 258 witnesses who reported seeing a streak of light. There has been a persistent belief by some (outside the investigation) that the streak of light reported by these witnesses was a missile attacking the airplane. As I have already explained, the physical evidence indicated indisputably that a missile did not strike the airplane.

Nonetheless, because of the media attention and public interest in these witness reports, we analyzed all of the witness documents in great detail in an attempt to understand what the witnesses may have seen. We studied all 736 witnesses for whom we had documentation, including those who reported seeing a streak of light.

As Dr. David Meyer will explain in more detail during his presentation tomorrow, almost all of the witness accounts are consistent with their having observed some portion of the accident sequence. In particular, after the center wing tank explosion and the separation of the nose portion a few seconds later, the airplane continued in its crippled flight for roughly 30 seconds or more, during which time burning fuel from the damaged airplane likely appeared as a streak of light. A small portion of the reported witness observations (56 of them) were not completely consistent with the airplane's flight path. However, these accounts can also be explained in a number of different ways and, as I said, Dr. David Meyer will be discussing all of this in much greater detail tomorrow.

In closing, I would like to reiterate that the physical evidence irrefutably indicated that the first pieces to depart the airplane were from the forward part of the center wing tank; that there was no physical evidence of a bomb or missile strike, but rather of an overpressure event inside the center wing tank; and that research, tests, and previous accidents demonstrate beyond any doubt that the overpressure was the result of a Jet A fuel/air vapor explosion in the center wing tank.

The safety issues in the accident relate to fuel tank flammability, potential ignition sources, and maintenance and aging of aircraft electrical wiring. I think aviation safety will best be served if we can focus our attention primarily on those issues.

This concludes my remarks. Mr. Al Dickinson will address the on-scene portion of the investigation, followed immediately by Mr. Jim Wildey, who will be discussing the break-up sequence of the airplane, and then staff will be prepared to answer any questions that the Board Members may have.

Jim Hall: Before you introduce Mr. Dickinson, let me mention to the Board that we all have in front of us a proposed agenda. The agenda includes Mr. Dickinson's presentation and a presentation on the break-up sequence by Mr. Wildey that will be followed by questions and answers on that issue as well as a presentation on fuel tank flammability and potential fuel tank ignition sources by Mr. Kolly then Mr. Swaim.

We hope to complete those presentations today. That will call for a discussion of the maintenance and aging of aircraft systems, design and certification issues, and the report of witness observations, followed by the Board's consideration of the conclusions, probable cause, and recommendations that are proposed by staff.

Are there any objections, additions, or deletions from the Board Members to this agenda?

We will then proceed, and I'll ask Dr. Loeb if he will introduce Mr. Dickinson for his presentation.

Dr. Bernard Loeb: Mr. Dickinson has served as IIC and as a U.S.-accredited representative in numerous aviation accidents and incidents. He has been a member of the Army Reserves and National Guard for more than 20 years. Mr. Dickinson holds a commercial pilot's license and has logged more than 5,000 hours flight time in both airplanes and rotary wing aircraft. He is a recipient of the Aviation Week and Space Technology Laurel Award from February 1997 granted to the investigation team in recognition of their work on TWA 800. He has an aerospace engineering degree from the University of Southern California.

Jim Hall: IIC, for those who may not be familiar with that term, is our Investigator-in-Charge of this particular investigation. Please proceed.

Al Dickinson: Good morning Mr. Chairman, Board Members, and staff. Today I will discuss the launch of the Go-Team, the Safety Board personnel involved in the investigation, the Navy involvement in the recovery, the mock-ups and partial reconstruction of the wreckage, and the parties involved in the investigation of TWA Flight 800.

On July 17, 1996, at dusk, TWA Flight 800, a Boeing 747-131, crashed into the Atlantic Ocean near East Moriches, New York. The aircraft was a scheduled air carrier flight operated under Title XIV Code of Federal Regulations, Part 121, from New York to Paris. All 230 people on board were killed in the accident.

The flight departed the John F. Kennedy International Airport at 8:19 PM from runway 22-Right. Visual meteorological conditions prevailed, and an instrument flight rulesflight plan was filed. Air traffic control communications with Flight 800 were routine, and the last transmission from Flight 800 was at 8:30 and 19 seconds, when the flight crew acknowledged a clearance to 15,000 feet. About one minute later, Flight 800 disappeared from radar.

I was notified that the airplane was missing about 8:50 PM. While the Go-Team was being assembled in Washington, investigators from the NTSB's regional office in New Jersey went immediately to the scene of the accident. The headquarters Go-Team arrived on-scene early the next morning. The initial Go-Team consisted of myself, a deputy senior IIC, and investigators in the areas of systems, structures, powerplants, survival factors, and air traffic control. The team was accompanied by former Safety Board Vice Chairman Robert Francis, as well as members of the Safety Board's Office of Government and Public Affairs.

We arrived at the Coast Guard station in East Moriches, shown above, at about 7:30 on July 18, 1996. Normally, the Coast Guard station is manned by about five people on duty; however, 48 hours after the accident, over 30 agencies and 2,500 people had converged on the facility.

Due to the magnitude of this investigation, more than one NTSB investigator was assigned to many of the groups, and as the investigation progressed, several new groups were formed - to date, over 22 groups have participated, by far the most groups ever to participate in an investigation in the Safety Board's history. Safety Board staff assigned to the investigation included

I was designated as the Investigator-in-Charge. Deputy IIC - Bob Benzon; Systems - Bob Swaim; Electromagnetic Induction - Scott Warren; Structures - Deepak Joshi, assisted by Alex Lemishko; Airplane Interior Documentation - Hank Hughes; Witness Groups - David Meyer, Dana Sanzo, Norm Wiemeyer, Doug Brazy, and Heather Knapp; Data Management - Deborah Bruce and David Meyer; Radar - Charlie Pereira and John Schade; Flight Data Recorder - Dennis Grossi; Cockpit Voice Recorder - Jim Cash; Metallurgy and Structures/Sequencing - Jim Wildey assisted by Frank Zakar; Reconstruction - Larry Jackson; Meteorology - Greg Salottolo; Hazardous Materials and Security - Tom Lasseigne; Medical Forensic - Burt Simon; Fire and Explosion - Joseph Kolly and Merritt Birky; Powerplants - Jim Hookey; Air Traffic Control - Al Lebo; Operations - Norm Wiemeyer; Aircraft Performance - Dennis Crider; Airport Security - Larry Roman; Trawling - Charlie Pereira and Doug Brazy; Flight Test - Robert Benzon and Dan Bower; Visibility - David Meyer assisted by Dana Sanzo; Report Writers - Jodi Moffett and Karen Bury; and Editor - Kristen Sears.

These are just a few of the various groups formed to investigate different aspects of the crash. Additionally, many other Safety Board staff members provided support to the investigation.

The Safety Board requested the assistance of the Supervisor of Salvage of the U.S. Navy for the recovery of the victims and the aircraft wreckage. The Navy recovery vessels were on-scene within two days, and by the time they completed the effort, over 95 percent of the 400,000 pound aircraft and remains of all of the 230 people aboard had been recovered. The Navy was assisted by the U.S. Coast Guard, Oceaneering, Underwater Search and Survey, the National Guard, and the National Oceanographic and Atmospheric Administration, as well as dive teams from Suffolk County, New York City and State Police, Fire Department personnel from both Suffolk County and New York City, and the FBI.

The cockpit voice recorder and flight data recorder were recovered by Navy divers on July 24, 1996. Both contained good quality data and revealed a routine flight until ending within a fraction of a second of one another at about 8:31 and 12 seconds.

After the aircraft wreckage was recovered from the ocean, it was transported to an abandoned Navy facility in Calverton, New York. The wreckage was documented and thoroughly examined and tested for chemical residues by the FBI - both on the recovery ships, at the dock, and again in the hangar. The hangar floor was marked and the wreckage laid out as to its position on the aircraft.

As pieces of wreckage were identified as being recovered from areas closest to JFK, and therefore assumed to have departed the airplane the earliest in the break-up sequence, they were placed on an initial mock-up, which is shown above. To further understand the accident, a three-dimensional reconstruction including the structure around the center wing tank from just after the cockpit to the forward portion of the aft cargo department was started.

The planning started in September of 1996 for the actual truss being constructed in February of 1997 and finishing over two months later in April. The 93-foot reconstruction - the largest wide body in the world - took over two months to construct, and contains over 870 pieces of wreckage weighing over 60,000 pounds. Metallurgists from the Sequencing Group thoroughly investigated each piece of aircraft, examining holes and penetrations and conducting a sequence study to determine the sequence in which the pieces came off the aircraft. Additionally, the Fire and Explosion Group used the reconstruction to map soot and fire patterns.

While the wreckage was being recovered, the Maintenance Group assembled in Kansas City, Missouri, to review the maintenance records of the aircraft. The airplane, which was manufactured in July of 1971, was purchased new from the Boeing Company by TWA. The Maintenance Group reviewed all maintenance records from the date of manufacture until July 17, 1996. The records indicated that TWA had accomplished mandatory directives, maintained scheduled maintenance, and maintained a continuous airworthiness maintenance program on the accident aircraft. The records indicated that the airplane was in compliance with all applicable airworthiness directives. Prior to the accident flight, routine periodic service was accomplished at JFK International Airport.

The Safety Board leverages its technical expertise with assistance from parties to the investigation. The parties to the TWA 800 investigation were: the Federal Aviation Administration; the Boeing Commercial Airplane Group; Trans World Airlines; the International Association of Machinists, Aerospace Workers, and Flight Attendants; the Air Line Pilots Associations; the National Air Traffic Controllers Association; Pratt and Whitney; Honeywell; and the Crane Company, Hydro-Aire.

During this extended investigation, there have been countless meetings and weekly telephone conference calls to provide for an open exchange of information and ideas and to keep all of the parties informed as to the progress of the investigative groups. During all of these discussions, the parties were asked to provide their comments on the scope of the investigation. Additionally, as normal Safety Board practice, following the technical review, the parties were asked to provide submissions to the Safety Board on our analysis of the factual evidence, findings, probable cause, and recommendations.

Submissions were received from: the Boeing Commercial Airplane Group; Trans World Airlines; the International Association of Machinists, Aerospace Workers, and Flight Attendants; the Air Line Pilots Association; and the Crane Company, Hydro-Aire. All of the parties have agreed with the findings in our report that the explosion initiated within the center wing tank of the accident airplane. Although the IAM agrees with our findings, they feel that the explosion of the center wing tank was preceded by an unspecified event.

Additionally, the investigation has utilized a variety of resources, including: NASA; Sandia National Laboratories; the University of Nevada, Reno; Applied Research Associates in Denver; Brookhaven National Laboratories; the California Institute of Technology; Wright Laboratory at Wright-Patterson Air Force Base; the Naval Research Laboratory; the United States Naval Air Warfare Center in China Lake, California; Britain's Defense Evaluation and Research Agency; and the Christian Michelson Research Institute in Norway. Combustion Dynamics, Ltd. was also included.

Under the rules of International Civil Aviation Organization, air safety investigators from the United Kingdom, France, Singapore, Australia, Canada, and New Zealand participated in the investigation as technical observers.

Mr. Chairman, this concludes my statement. I will be followed by Mr. Jim Wildey, who will talk about the extensive work done to determine the sequence of events.

Jim Hall: Let's proceed then with Mr. Wildey's presentation and then we'll have a short break before we take questions from the Board on the two presentations.

Dr. Loeb: Mr. Wildey has been employed at the Safety Board in the Materials Laboratory Division for about 25 years. He has been Chief of the Laboratory for a little more than two years. Mr. Wildey has participated in many of the major accident investigations involving component or structural failures investigated by the Safety Board. Mr. Wildey was involved in the 1985 Indian Airlines Boeing 737 bombing, the 1988 Aloha Airlines 737 structural failure, and the 1988 Pan Am 103 bombing in Lockerbee, Scotland, just to mention a few. Mr. Wildey has been the recipient of the NTSB Chairman's Award and the recipient of the Aviation Week and Space Technology Laurel Award, in February 1998, in recognition of his analysis of the break-up of the TWA 800 airplane. Mr. Wildey possesses a bachelor's degree in Metallurgical Engineering from Virginia Polytechnic Institute and State University. Mr. Wildey.

Jim Wildey: Thank you. Good morning, Mr. Chairman, Members of the Board. My presentation is on the in-flight break-up of the TWA Flight 800 airplane, and how it was determined that the break-up initiated from an overpressure event within the center wing tank.

As part of an introduction, I will discuss the formation of the Metallurgy and Structures/ Sequencing Group and how we did our work and will give a description of the wing center section in the center wing tank structure. The overall break-up sequence will then be presented, with detailed information on specific portions of the break-up sequence as well as the evidence that led to the conclusion that the break-up was initiated from an overpressure event. My presentation will finish with a short video of the reconstructed portion of the accident airplane, showing the break-up sequence.

To address the question of how the airplane broke up, the Board formed what was titled the Metallurgy and Structures/Sequencing Group, with group members from the major parties to the investigation. Representatives from the FBI also monitored the Group's investigation, but did not participate in the generation of our reports. Our task was to find out how the airplane broke apart and where the break-up initiated so that investigation efforts could concentrate on the cause of the break-up.

Starting in December of 1996, the Group spent many weeks examining the structural pieces after they were recovered from the ocean and transported to the facility at Calverton. As each piece of the recovered wreckage was brought to the hangar facility, it was initially examined by bomb explosion experts with the FBI and ATF [Bureau of Alcohol, Tobacco, and Firearms] as well as by Safety Board and FBI metallurgists for characteristic evidence of the detonation of a bomb or missile warhead. Simply stated, none of the pieces had this evidence.

All of the major structural pieces, and many of the smaller pieces, were labeled with their tag number and recovery area and examined by the Structures Group to determine what part of the airplane they came from. Once they were identified, the pieces were laid out on a two-dimensional grid on the hangar floor.

The photograph we are looking at now shows most of the fuselage section laid out on the two-dimensional grid. The nose of the airplane extends toward the hangar door at the top of the picture, and the tail is toward the lower left. The inside surfaces of the pieces are facing up. Not shown in this picture are the areas containing the layouts for the wings, for the tail section, for the engines, and for the cabin interior. Also visible here are some of the smaller-scale, three-dimensional reconstructions that were made of portions of the wing center section.

The Sequencing Group did a large portion of its examinations on the structure as it was located on the two-dimensional grid and on the smaller-scale three-dimensional reconstructions.

As Mr. Dickinson discussed, in the spring of 1997, investigators assembled a 93-foot long, three-dimensional reconstruction of the center portion of the airplane's fuselage from just after the cockpit to within the aft cargo compartment. The reconstruction included all of the wing center section, the main landing gear bay, and the furthest in-board pieces of the wing. All of the fuselage and wing center section pieces that were recovered from the red zone were a part of this reconstruction.

This photograph shows an overall view of the right side of the reconstructed portion of the airplane. The structural pieces of the airplane were attached to an iron beam truss that extended in the main cabin space between two support columns. The Sequencing Group used this reconstruction to verify earlier conclusions and to more easily compare fire and soot accumulation patterns. The initial examination showed that there was a relatively clear demarcation between the pieces that were found in the red, yellow, and green zones.

This model gives a rough idea of the recovery locations of the various portions of the airplane. The first structural pieces found along the flight path were the pieces recovered from the red zone. These pieces included a ring of fuselage structure from in front of the wing center section, structure from the aft end of the forward cargo compartment, and pieces of the wing center section itself.

Other items recovered from the red zone included: the forward portion of the keel beam, which is located under the wing center section; main cabin floor beams; and the two forward air conditioning packs. Any viable break-up sequence had to account for the early release of these red zone parts from the remainder of the structure.

At the conclusion of our efforts, all the Sequencing Group members agreed that the physical evidence contained on the recovered pieces showed that the break-up initiated within the wing center section, at spanwise beam 3. This beam is the most forward boundary member of the center wing tank. The pieces ejected as part of the initial overpressure event contained at most minor soot accumulation, indicating that some type of combustion was associated with the initial overpressure event. However, the lack of significant soot on these red zone pieces shows that there was no severe fire before the overpressure event.

I would like to spend a few minutes to describe the important elements that make up the wing center section on a 747-100 series airplane. The wing center section of the airplane visible here in the center of the model is designed to carry the wing loads through the fuselage and to support the fuselage on the wings.

This is a model of the wing center section, oriented the same way as the airplane shown on the previous slide. The wing center section is a large box structure with a footprint about the size of a two-car garage. The wing center section extends between the front spar and the rear spar. Also found here are the main cabin floor beams, which serve to stiffen the upper skin panel of the wing center section. Internally, the wing center section is divided into compartments by a series of lateral, or spanwise, beams. The center wing fuel tank occupies most of the wing center section, extending from the rear spar to spanwise beam 3, which is just behind the front spar.

I'll also mention some of the other structural members and systems associated with the wing center section. The keel beam is located along the bottom center line of the airplane. This beam extends from the aft end of the forward cargo compartment under the wing's center section, through the landing gear bays, and to the forward end of the aft cargo compartment. The keel beam acts to transfer fuselage loads under the wing center section. Also, the 747 has three air conditioning packs that are located below the wings. Dr. Kolly will discuss these packs and their relationship to heating of the center wing tank.

This is a drawing of the wing center section viewed from the right side with the upper skin removed. The structural members of the wing center section are identified, including: the rear spar, spanwise beam one, the mid spar, spanwise beam two, spanwise beam three, and the front spar. There is also a center line rib that extends between the rear spar and the mid spar. These internal beams and ribs provide stiffening and reinforcement to the upper and lower skin panels of the wing center section. As I previously indicated, the center wing tank, here shaded darker, occupies most of the wing center section on a Boeing 747-100 series airplane. The forward portion of the wing center section, between spanwise beam 3 and the front spar, is an unpressurized dry bay that does not carry fuel.

As a final item, please note the location of the large storage bottles for drinking water carried on the airplane. These bottles are located at the center of the forward side of the front spar.

This photograph shows a view of the right side of a reconstructed portion of the airplane. Coloration has been added to correspond to the recovery positions of the structure from the red, yellow, and green zones You can see in this photograph a distinct band of red zone fuselage parts.

This photograph shows a closer view of the red zone area on the right side of the reconstructed portion of the airplane. Most of these red zone pieces, which were the first pieces to depart the structure, were located in front of the front spar, whose location is indicated.

How did our group determine the break-up sequence? We began by examining portions of the structure in great detail. Localized, sequenced segments were then developed based on the observable features in each portion. Eventually, individual sequence segments were combined until a cohesive and comprehensive break-up sequence, consistent with the overall body of evidence, was generated. We also used stress analysis to provide confidence that proposed scenarios were consistent with structural properties and expected failure modes.

Based on our detailed examinations of the structure, and supported by the results of stress calculations, the Sequencing Group determined that the break-up of the airplane initiated with the fracture of spanwise beam 3 at its upper end. We found that the separated beam rotated forward and impacted the front spar.

This drawing shows an overall right side view of the wing center section, depicting how spanwise beam 3 rotated forward. Also, the various members of the wing center section have been color-coded to reflect their recovery positions. Note that most of spanwise beam 3 and the front spar are red, indicating their early release from the structure.

As spanwise beam 3 rotated forward, it created impact marks in two locations: first, on an upper skin panel stiffener, located just forward of the upper end of the beam; and second, on the aft side of the front spar. The locations of the two sets of impact marks will be illustrated on this drawing, which shows the forward portion of the wing center section as viewed from the right side. The location of the front spar is shown in its normal position. Spanwise beam 3 is shown both in its normal position with dash lines and in a rotated position with solid lines. This red arrow indicates the location of one set of impact marks on the aft side of a stiffener for the upper skin panel. This second red arrow indicates the set of impact marks on the aft side of the front spar. The fracture and the resulting forward rotation of spanwise beam 3 were consistent with overpressure on the aft side of spanwise beam 3 within the center wing tank.

I also have a video describing the fracture of spanwise beam 3 and the creation of the impact marks on the aft side of the front spar. The video was taken from the right side of the reconstructed portion of the airplane, where we'll be looking first into the fuel tank bay behind spanwise beam 3, then into the dry bay between spanwise beam 3 and the front spar. Forward will be to the right as we view the video.

"I'm standing in the fuel tank of the 747, just behind spanwise beam 3. This beam is the most forward member of the fuel tank. The very first event identified by the Sequencing Group was an explosion of the fuel air vapor inside the tank. This explosion forced spanwise beam 3 forward, fracturing the stiffeners on the back side of the beam and across the top of the beam. Spanwise beam 3 rotated forward and impacted another stiffener just in front of it as the roof lifted up, and then spanwise beam 3 continued to rotate forward and impacted the aft side of the front spar just in front of this.

"This is a dry bay between spanwise beam 3 and the front spar. After spanwise beam 3 separated across the top, pressure behind spanwise beam 3 forced it forward and it rotated. The top end of spanwise beam 3 struck the aft side of the front spar and it left behind these very distinctive witness marks, and those marks extended pretty much all the way across the front spar, indicating that all of spanwise beam 3 was rotating forward."

I have described so far the fracture of spanwise beam 3 and its impact on the front spar. Following the impact from spanwise beam 3, the front spar fractured at its upper end, cracking progressed down the front spar, and fuselage cracking initiated at stringer 40 right.

The cracking in the lower fuselage is illustrated in this schematic drawing. We are looking down and aft on the lower internal portion of the fuselage skin in the area forward of the front spar. The grid pattern represents the location of the internal structure. The initiation area of the fuselage skin cracking, which is along stringer 40 right, is indicated. After cracking initiated in the fuselage at this location, it quickly spread through the lower fuselage along the positions indicated by the dark red lines drawn on the structure.

Based on deformation patterns, the fracture, as indicated by the dark red lines, occurred earlier than any other fuselage fractures in the entire airplane. The direction of cracking along these early fractures was determined by specific features associated with the rivet-to-rivet fracture pattern of the fuselage skin in these areas. This rivet-to-rivet fracture pattern was established by detailed examination of every inch of these early fuselage cracks.

The blue arrows on the red fractures show how these early fractures stemmed from the initiation site at stringer 40 right adjacent to the front spar. The early fuselage cracking shown here propagated primarily under the normal pressure differential between the interior of the fuselage and the outside air at 13,800 feet. These early fractures generated a large hole in the belly of the airplane, through which structure and interior components would have been injected.

At this point in the sequence, spanwise beam 3 has fractured and impacted the front spar, and the front spar cracking has spread through the forward of the front spar. After loss of the belly structure, fractures progressed up the sides and across the top of the fuselage, until the nose portion of the airplane was completely separated.

The damage that I have described so far is the extent of the damage created as a result of the initial overpressure event, and I would like to provide some overall perspective regarding the speed of this portion of the sequence. Even though it has taken me several minutes to describe it, I would emphasize that this portion of the sequence occurred very quickly, within only a few seconds.

Before presenting the remainder of the sequence, which occurred over many more seconds, I would like to summarize the evidence that led to the conclusion that the break-up of the airplane was initiated by an overpressure event. First of all, we determined that spanwise beam 3 fractured at its top and rotated forward. These features are indicative of excessive pressure on the aft side of spanwise beam 3. Calculations indicate that a pressure differential of about 25 pounds per square inch across this beam would initiate the beam's fracture. In addition, the calculations show that spanwise beam 3 is the weakest of the boundary members of the center wing tank.

We also found that the upper skin panel was bulging upward as spanwise beam 3 was separating. This upward bulging was determined by the pattern of the impact mark created by the upper end of spanwise beam 3 as it struck the stiffener immediately forward of the beam's upper end. This upward lifting of the skin is also indicative of excessive pressure within the center wing tank.

We found that the front spar bulged forward in two lobes, restricted by the inertial resistance of the water bottles mounted at the center of the spar on its front side. This bulging is an indication of the escaping overpressure within the center wing tank acting on the aft side of the front spar.

The group also determined that a downward pressure load on the lower skin panel of the wing center section reacted through the keel beam and was the primary source of the stress that initiated cracking in the fuselage at stringer 40 right.

The examinations of the structure showed that the break-up began with the fracture of spanwise beam 3, and the features I have just mentioned clearly show that the fracture of spanwise beam 3 initiated from an overpressure event within the center wing tank of the 747 airplane.

I will now continue with the break-up sequence after separation of the red zone pieces, as previously described.

This photograph is a repeat of an earlier slide. The nose portion, here on the right and colored yellow, separated as a result of the loss of initial pieces, here colored red. After separating from the remainder of the airplane, this nose portion remained intact all of the way to water impact. I would like to point out the heavy compression, crushing, and break-up damage found on the lower and right sides of the nose portion. The location of this heavy damage indicates that the nose portion impacted the water relatively flat, but rolled slightly to the right. The forward cargo door, located on the lower right side of the nose portion, contained damage very similar to the neighboring pieces of the fuselage. This is a clear indication that the door was in place when the nose portion impacted the water.

The major portion of the airplane, here on the left colored green, also remained intact for a period of time after separation of the red zone pieces and loss of the nose. This portion of the airplane contained both wings, most of the wing center section, which extends between the arrows, as well as a limited amount of fuselage structure in front of the wing's center section.

Aerodynamic considerations and radar returns indicate that the major portion of the airplane climbed and rolled, then began a steep descent to the water. As speeds and loads built during the descent, the wings separated at the outboard engines and the wing center section failed adjacent to the left wing. Fire and soot patterns were concentrated on the inboard end of the right wing and on structure that remained attached to the right wing during the descent.

Fuel escaping from the right wing was a major source of the fuel for this fire. These fire and soot patterns assisted in the determination of portions of the sequence.

This completes my description of the break-up sequence. To summarize, I have a short video presentation. This video again was shot at the reconstructed portion of the airplane, and hopefully the video will be the next best thing to having the reconstruction here with us. The video shows the right side of the airplane, and forward is to the right.

"The break-up sequence of the TWA Flight 800 airplane began with the explosion of fuel/air mixture within the wing center section fuel tank. Pressure from this explosion fractured the front member of the fuel tank, spanwise beam 3, along the top. Spanwise beam 3 rotated forward and impacted the aft side of the front spar. Cracking in the front spar came downward and entered into the fuselage. This fuselage cracking progressed around three sides of a large piece of belly structure. This created a hole through which pieces of the front spar, spanwise beam 3, and one piece of spanwise beam 2, were free to be ejected from the airplane.

"Fuselage cracking progressed upward on both sides of the hole, and very quickly, a ring of fuselage material completely separated from the airplane. This allowed the nose portion to fall and hit the ocean surface in one piece. The bulk of the airplane, the major portion of the airplane, remained intact for a period of time, eventually breaking apart further downrange, and fell into the ocean in several pieces."

In conclusion, I can say that the Sequencing Group found that the break-up of the TWA Flight 800 airplane initiated with the fracture of the forward boundary member of the center wing fuel tank as a result of an overpressure event within the tank. Furthermore, the vast majority of the features documented by the Sequencing Group were not consistent with any other proposed scenario for the break-up of the airplane.

Jim Hall: Thank you, Mr. Wildey. What I would propose to the Board, since we have been sitting here along with the audience for approximately an hour and a half and I know we have many questions for Mr. Wildey and Mr. Dickinson, I suggest we take a 15-minute break and then we will reconvene this Board meeting.

After Break:

Jim Hall: We will reconvene this meeting of the National Transportation Safety Board. The Board is in the midst of our deliberations and discussion of an aviation accident report that is carried as Board's notation 6788G, an in-flight break-up over the Atlantic Ocean of Trans World Airlines Boeing 747, that occurred off the coast of East Moriches, New York, on July 17, 1996.

The Board's program has included an introduction and general overview by Dr. Bernard Loeb, and then a discussion of the on-scene accident investigation recovery by Mr. Dickinson, followed by a discussion of the break-up sequence by Mr. Wildey. These Board proceedings are being carried worldwide on the Board's Internet site,, as well as on C-SPAN.

In addition, I would like to thank the Department of State for providing two individuals who are interpreters for our French families. You may see them in the booth that is indicated to my left to the rear of the Boardroom. I spoke with the interpreters on the break and they requested that, to the extent possible, if all of us could speak a little slower, it would assist them in being sure that the interpretation for the French families is accurate and that they get all of the information that we are trying to cover in this material.

Mr. Wildey, I would ask at the beginning if you would please stand up so that the individuals in the room could see how tall you are. How tall are you, Jim? Six feet seven. I wanted to be sure to put in perspective the film and video you saw of Mr. Wildey in the center tank and in the video.

Mr. Wildey, I think it would be helpful if you could quickly give us again the relationship of the recovery positions of the pieces and the determination of the sequence to set us up for our questioning.

Jim Wildey: Well, we certainly used the recovery positions of pieces from the ocean to initiate our group. Having said that, however, the recovery positions and the determination of the sequence are really independent. We would have come up with this break-up sequence regardless of where the pieces were found. That's based on the damage patterns and the types of fractures and all the witness marks and things like that that were documented. So even though we started with the recovery positions and the assumption that these red zone pieces must have been the first pieces to come out, the conclusions that we reached were really independent of the fact that we started with that as a baseline.

Jim Hall: I ask that because there have been some allegations made that the recovery pieces were not properly identified and marked. Mr. Dickinson, could you tell us a little bit about how we went with the FBI in putting together the recovery from the ocean of this very, very large aircraft and, obviously, if you could just sketch that for us and what was done in the early days of the investigation?

Al Dickinson: Yes, sir. We coordinated with SUPSAL, the Navy contingent; actually it was Captain Chip McCord, who coordinated with me the first day that we got there. And I remember him telling me that it would take a couple of days to get the assets there and first he was going to survey the whole area, which involved getting all the other boats that were in the area out of the area because they were towing a side scan sonar over the whole area to determine where the wreckage was. After they did that for a couple of days, they also attempted to locate the CDR and the FDR using the pinger system, but it was not recovered in that effort. Apparently, those transmitters were shielded by the wreckage. The Navy stayed on-site for three months after we initially found all the wreckage in the different zones, and we organized a system where we recovered pieces and tagged them with different locations and brought them back into Calverton for the reconstruction.

Jim Hall: The first part of the wreckage that was recovered, obviously, was floating. Do we have any idea how much percentage-wise of the aircraft was taken off of the surface of the water initially?

Al Dickinson: It was a small percentage, Mr. Chairman. We don't know the exact amount.

Jim Hall: Some of that was brought and turned over to the FBI, and it was clearly floating, and we couldn't identify that to a specific zone until possibly a best guess situation.

Al Dickinson: Exactly.

Jim Hall: Mr. Wildey, there's also been some question as to the airplane itself when the nose departed. Why did it pitch up? Why did the airplane go up?

Jim Wildey: That's an aerodynamic and a performance consideration. I guess Mr. Crider would be best for that.

Bernard Loeb: Let me try to answer that. The airplane pitched up simply because gravity shifted aft when the nose came off. It's just a simple matter of physics. When the nose came off, the weight from the front end of the airplane and the CG, the center of gravity, shifted aft. It would be just a natural thing. It's physics. There's nothing exotic about it or mysterious.

Jim Hall: The aircraft, it was pointed out, was 25 years old. Was there any fatigue, cracking that contributed to the break-up of the aircraft. If this event had occurred on a newer aircraft, do we think we would have had the same sequence? I understand that requires some speculation on your part, but I'm trying to get into some basic, common-sense questions that some of the family members and others have.

Jim Wildey: There were some fatigue cracks that were found on the airplane, and these have been documented and are contained in our report. However, it's clear and the Sequencing Group determined that the location of these fatigue cracks and their presence played no role in either the location of the initial fractures or the sequence of the break-up.

Jim Hall: If we could put up one of the side views of the aircraft that shows those large holes on the side of the aircraft.

Jim Wildey: That could be one of Dr. Loeb's, or my number 10 would be a good one.

Jim Hall: That's fine. That's a little distant view. But as we look at that, Mr. Wildey, there's obviously pieces of the structure missing. We know that we were able to only recover, despite all of our efforts, approximately 95 percent of the aircraft, which I may say is probably the largest under-ocean recovery of an aircraft in aviation history. Is it possible that an evidence of a bomb is only on the small portions of the structure that was not recovered or identified?

Jim Wildey: First of all, I will point out that, as Dr. Loeb said, the holes that you see here aren't really holes. The structure, especially the two big ones you see on the right side and then center left - those holes, if you call them that, are totally filled by structure that's folded in or crushed. As for the possibility that the evidence of an explosive device might have been contained on the missing structure, experience has shown clearly that the evidence of an explosion is contained through a larger volume than just a relatively small area. They've done the Lockerbee history and other explosions that the Safety Board has been involved in. So therefore it's very, very unlikely that the small amount of structure that's missing, which is from various areas spread throughout the airplane, could have been the only pieces that could have contained the evidence of a bomb or a missile explosion.

Jim Hall: On page 185, again, being slightly redundant on the subject of the report, you speak on lines 5 and 6 about the apparent hole that was 2 to 3 feet longitudinally and about 5 feet circumferentially. Would you again tell me how you know that you had most of that fuselage skin there, in terms of that particular hole, the largest hole on the aircraft?

Jim Wildey: Actually, I don't believe that's the largest one. In fact, we don't have a view of that. It's on the left side and most of our views are from the right side. And that hole is not untypical of some of the other holes. The structure is actually there. It's folded and bent and deformed inward in various ways. This particular hole that you mention in the report is from the left wing upper side at the aft end of the wing, where that structure is connected to the wing's center section. We determined that one of the events that occurred was the fracture of the wing center section adjacent to the left side, adjacent to the left wing, and as part of that motion that created this hole as it came upward and inward. It would have been where the hole is now. So we believe that that was the failure of the left wing as it came upward and into that hole.

Jim Hall: Mr. Wildey, who else looked at these holes, and is this the opinion of basically the Sequencing Group report that was placed in the docket?

Jim Wildey: It is the opinion of the Sequencing Group, and it is a factor that was signed off on by the entire Sequencing Group. But we discussed this with a large number of people and there were all kinds of people there who would bring various features to the Sequencing Group's attention and we tried to address every single one of these as they were brought forward and I think we did that quite successfully.

Jim Hall: Do you have rough guesstimate of how much time you spent up in Calverton looking at the metal and the structure of that particular aircraft?

Jim Wildey: I spent a total of 90 days on the accident investigation last time I counted. I will say that not all of that was part of the Sequencing Group but the vast majority of it was.

Jim Hall: And do you think that was an adequate amount of time for you to form your opinions?

Jim Wildey: Yes, I do.

Jim Hall: Member Hammerschmidt.

John Hammerschmidt: First of all I'd like to thank the staff for their very excellent and professional presentations this morning. A great job. And also for a good and comprehensive job on this report, as voluminous as it was to try to digest and work through for the past several weeks. As we've been talking about the procedural aspects of this investigation, it occurred to me that in many respects the way we approach this investigation in terms of the material failure analysis or the structural failure analysis from an in-flight break-up that ended up in the ocean is in many respects similar to the work we did on the space shuttle Challenger investigation where we started out with two-dimensional layouts of much of the wreckage and the fuel tanks of that space launch vehicle and then we completed a three-dimensional mock-up of the orbiter and came to conclusions in terms of break-up sequence as we're terming it in this report. So that struck me as somewhat analogous to the procedures that we employed in this investigation and when we're talking about structural failure analysis we're really talking about, essentially, accident investigation 101 in terms of witness marks and what hit what first and how material breaks apart and what that shows us, especially when we put under the scrutiny of an electron microscope.

The factual portion of these two discussion items is quite a few pages but I would like to select out just a few questions that I have, miscellaneous in nature, beginning at the front of the report and going back into the report. The first question is on page 124 and 125 where you make reference to some of the wreckage that was recovered from the green zone as we've termed out. At the bottom of page 124, we make reference to the tires of the aircraft. We indicate that the 16 main landing gear tires were recovered from the green zone and examined at the hangar in Calverton, We go on to describe what that examination revealed. Could staff please elaborate on what the evidence showed us from the tires that were uncovered.

Robert Swaim: The tires that were recovered were examined with the help of representatives from the tire company, I believe it was Goodyear, TWA, and the other parties helped so we did have a number of tire specialists with us. We also had some material folks with us. The bursts were consistent with impact-type damage or sharp object-type damage where we had fairly classic chevrons and so forth. Some of the tires had a light evidence of burning on the surface near structure that had a flow pattern of burning.

John Hammerschmidt: Did we find any evidence of a tire that had exploded in flight?

Robert Swaim: No sir. Tires in main landing gears that do explode, there's a lot of force in them and they leave a lot of damage. No we did not.

John Hammerschmidt: Thank you. Moving to the next question, also back in the factual section of the report. We begin a section on page 132 which we entitle "Brown Splatter Material on Air Conditioning Ducts." We go on to describe what was found in considerable detail. Would someone please describe what that refers to and what the significance or nonsignificance of that might be?

Jim Hall: Mr. Kolly, we've got a big room with a lot of people who want to hear, so please speak up and speak slowly.

Joseph Kolly: The brown splatter material that was found was an elastomeric???? material, and it was tested by the FBI, DERA (the Defense Evaluation Research Administration), ARTEC ??, and NASA.

Jim Hall: DERA?

Speaker: The Defense Evaluation Research Administration. It's a new one. They helped us throughout this investigation.

Jim Hall: I understand. We use a lot of abbreviations in government, and we've got a lot of people here that aren't in government that are interested and that we need to communicate with.

Joseph Kolly: The results indicated that this was consistent with a melting of a polyurethane foam that had covered the duct that was in that area.

John Hammerschmidt: Where I was going with that question was, we discovered these brown splatters and at first they raised question marks. And then we delved further into them, and we did an elaborate evaluation of them and determined that they really weren't that significant.

Bernard Loeb: There was an issue raised fairly early in the investigation by one or two of the investigators about what this brown splatter meant, this dark splatter. And there was at one point a suggestion that it meant that there may have been a fire that initiated outside the fuel tank - it's part of the insulation, the duct insulation - to melt the polyurethane and get it splattered onto the various members, including spanwise beam 3, and inside some of the ducts. The issue was raised, how could that have happened in the fuel tank explosion. And we've gone through that in great detail. There was also a report that there were fiber materials in the splatter, which we have identified as materials from the carpeting in the airplane. We went into other airplanes and found the same kind of fiber materials throughout some of the same locations and did an extensive analysis and determined that all of that was consistent with the break-up as we see it of the fuel tank and with the scenario we have given you.

John Hammerschmidt: Thank you. Because we're dealing with the section entitled On-Scene Accident Investigation and Recovery, I was going to ask about the nature of the recovered electrical components and wiring from the vicinity of the center wing tank.

Bernard Loeb: Bob Swaim is going to have a presentation that will talk about that, and we'll go into that in great detail at that time.

Jim Hall: Okay, thank you. You were referring to a section of the factual portion of the report entitled "Information Regarding Certain Primary Radar Targets Recorded by the Islip, New York Radar Site." The factual begins on page 155. And we also make reference to that issue, I believe, in the analysis on page 456. Since that hasn't been addressed yet, and it does fall within this discussion category of the break-up sequence, would someone wish to elaborate on what we learned in that area and what significance it might have to our understanding of this accident?

Charles Pereira: There were numerous primary radar returns recorded by six primary radars covering the accident. This particular section of the report covers the Islip primary radar den. There were several sequences of anomalous radar returns well away from the accident area that we went ahead and reviewed anyway to try to develop an explanation for them.

On this slide, in the lower left-hand corner, this is one of the sequences that you're looking at here. In the center there's a yellow dot that represents the last secondary return for TWA 800. The small white triangles throughout the plot represent primary radar returns and reflections of energy. It's not the airplane talking back to the radar; it's just a reflection; a bird or a building would create such a thing. The blue dots that are there represent secondary returns, the position of the other airplane talking back to a radar and the transponder. So on the lower left-hand corner there, you've got some primaries that appear and disappear and have randomly varying signal strengths. They only occur on the Islip radar, out of all the six primary radar systems looking at this area, Islip was the only one to show these and they only occurred on the 150 to 160-degree ?? And they occurred several times, both before and after the accident, always in the same area.

So we discussed these with FAA radar specialists and read various literature regarding false primary returns, and we also reviewed the Islip radar area for building structures that might cause such reflections. And we and the FAA unanimously came to the conclusion that if these were false primary radar returns then aircraft in other areas of Islip's airspace that were being reflected off some building structure and showing up in this area in a random fashion.

So this particular section of the report that you raised questions about, that's what we're discussing. If you have any detailed questions on that I'd be happy to answer them.

John Hammerschmidt: No, I was just hoping to get a brief explanation of that and that was just right. Thank you, Mr. Pereira. Continuing on with the report, on page 167, line 4, in the part of the report where we're talking about the break-up sequencing, we state that the timing of the wing tip and wing center section failure were based largely on witness statements and therefore are not precise to the second. Could staff elaborate on that sentence, that reference?

Bernard Loeb: Yes, the sentence is essentially what it says: We have no other basis for determining when the outer sections of the wing failed, other than the witnesses, and we got that information based on the description of the fireball and the developing progress of the fireball and the further eruptions that occurred in their descriptions. We coordinated that with certain other events that we did have in a fixed time, for example, the captain of the Eastwind flight who had seen this - he was about 20 miles away or so- he'd seen what he believed to be landing lights of an airplane approaching him coming out of JFK. In fact, he flicked his landing lights to indicate to the airplane that he was where he was and to give them a warning. He was the first one to report that there was an accident, and he reported that to air traffic control.

So we had the timing of that event, and he gave us some information that indicated when this thing developed and when the outer portions of the wings must have broken, because that was how fuel got out before the airplane fully came apart. Later on the left wing broke and then the right wing broke off from the remaining pieces of fuselage. Prior to that we had to have a source of fuel for the fire that was being seen and was reported by the captain of the Eastwind as this great explosion. And so that's how we timed the outer portions of the wings, and there was no precision to that; we were within a few seconds, but that's the best that we could do.

John Hammerschmidt: Thank you for that explanation. The reason I brought that up is because when you say in your report in witness statements, oftentimes someone thinks of people standing on the ground looking a great distance, but this was in fact the pilot of another airline.

Bernard Loeb: That's correct. And a pilot who had reported into air traffic control, so that we could actually time when that occurred.

John Hammerschmidt: Thank you. Proceeding on in the factual, about page 203, we have a table in which we describe the 747-100 wing center section, center wing tank component failure strengths - minimums and maximums for the different sections within this center wing tank - such as spanwise beam 3, spanwise beam 2, etc. And we have described also, in the presentation thus far, that we think that this in-flight break-up initiated with an overpressure event. In terms of this Table 3, we have data that's in PSI, pounds per square inch. Do we have any way of calculating what we think the pressure of the overpressure event was in this accident sequence?

Speaker: The next presentation will be Dr. Joseph Kolly, who will go into that end of it, in other words, what we believe the pressures that were developed would have been. If your question is more to the structural capacity of these members, then Jim Wildey could answer that for you now.

John Hammerschmidt: I'm jumping ahead on that a little bit. But speaking of these beams and spars in the center wing tank, Mr. Wildey indicated in his presentation that they were primarily designed that way for structural reasons. Were there any considerations given to their location having to do with protection against fuel ignition in the center wing tank or am I jumping ahead on that as well?

Jim Wildey: I don't think that the tank is designed to withstand the stresses generated by an explosion of the fuel/air vapor mixture. The loads, primarily of the internal components, are loads associated with sloshing of the fuel and things like that, which are relatively minor. So the structural strength is also necessary to carry the stiffness requirements and the reinforcement requirements for the upper and lower skin panels. I believe these are the major criteria that are used to design the sizes of these internal beams.

John Hammerschmidt: Well speaking of sloshing, was any consideration given in this design in reference to, say, sloshing to prevent that because sloshing might aerate some fuel to perhaps generate a greater fuel/air mixture than would otherwise be present?

Jim Wildey: The sloshing loads are mainly.... If you didn't have any internal beams, for example, as you stopped on the runway, the fuel in this large tank could force itself against the forward boundary member, so one of the things that the internal members do is that they meter the fuel, if you will, they slow it down so you don't get this large surge against any of the members as you're turning or decelerating, and that is a consideration in the design and the strength of these beams.

John Hammerschmidt: Right. Thank you. That's what I was trying to get to. Concerning some work that we reference in the section beginning on page 233, "A Study of Computer Model Calculations of Full-Scale Center Wing Tank Combustion"....

Jim Wildey: Again, Dr. Kolly's going to cover this in his next presentation.

John Hammerschmidt: In that case we'll just wait until then. As I say, those were excellent presentations and when you have great presentations like that it cuts down on the questioning. That's all I have.

Jim Hall: Very well. We will turn to Member Goglia

John Goglia: Thank you, Mr. Chairman. I just have a few questions and I guess I'll take them in sequence. Dr. Loeb, in your opening statement you mentioned the seats, the work that was done with the seats in the cabin interior. In this airplane accident investigation and the reconstruction effort, we did something that was quite unusual inasmuch as we put the entire interior of the airplane back together as best we could. I wonder if you would take a moment to explain just why we did it and what we gained from that effort.

Bernard Loeb: I'll try to do that. Early in the investigation, of course, there was a lot of issues raised about the possibility that there was some sort of a high-energy explosive detonation in the airplane caused by a bomb or a missile, a bomb had exploded or a missile warhead had exploded inside the airplane. For that reason, and other reasons, we attempted to reconstruct the seats and the interior of the airplane to see whether there was any evidence in the cabin area that would point to the location of an explosive device or to give us any information about whether there was in fact evidence of a high-energy explosion. So we gathered as many of the seats and also galley equipment that we could and they tried to reconstruct to the best of their ability. One of the difficulties is of course trying to determine where the seats go and there's numbering issues that make it very difficult, although a number of these seats did have numbers on them so it did make it possible for us to reconstruct essentially what would have been the cabin interior in terms of the seats. Our examination of that determined that there was, as I indicated, no area where locally the damage was a high degree of fragmentation and a high degree of thermal damage, so that's one of the major things we got out of the seat locations. Also, there was an examination forensically of the occupants to see if there was any evidence - the same kind of evidence - on the passengers. One of the difficulties there is that passengers don't always sit in their assigned seats so it's very hard to draw strong conclusions from that, but because we were able to get the seats and pretty much locate them what we found out was that the damage and in fact the burn patterns were very consistent with what we would expect from first of all the explosion in the center wing tank and second of all the water impact.

John Goglia: Thank you. Mr. Dickinson, I have a question that goes to the tagging. We've talked a lot initially here about the three zones, the red, the yellow, and the green zones, and the tagging of those parts. Someone told me once that there was a million pieces that were recovered. I don't know if that's an accurate number or an inaccurate number. It really doesn't matter. There were a considerable number of pieces found, which we labeled red, green, or yellow. Human beings being what they are, there were probably mistakes that were made in that. Are you comfortable in your own mind that we have factored in the possibility of those mistakes and that there's very little impact to the actual investigation if there were any mistakes?

Al Dickinson: Yes, I am comfortable. You have to realize that it was the first full recovery of an aircraft from 120 feet below the ocean surface that we had ever attempted and we had some situations where pieces that were recovered from certain areas had fallen off from various things as far as mis-tagging but we adjusted and we went back and looked into every piece and have fully identified as far as we can tell and we're fully satisfied that the tagging system is accurate.

John Goglia: I would like, if you wouldn't mind, for Dr. Meyer, since he was in charge of the tagging and the database project, to perhaps say a word or two about the tagging process.

Jim Hall: Well, in leading into that, let me just make an observation. We had participated with the supervisor of salvage I think twice before this particular accident, once on the recovery of an aircraft door in the Pacific, a cargo door, and the other was on the Bergen?? Air crash. There may have been other instances ... and the AeroPeru, and Valujet has assisted us in the Everglades. But this was the first time in this depth of water. Since that, of course, we have regrettably had the SwissAir accident, which is investigated by the Transportation Safety Board of Canada, as well as Alaska Air and EgyptAir. We have learned a lot and things have refined but Dr. Meyer, if you would please go into that answer to Member Goglia's question, but I think it's very important that people following this investigation are aware that this is the first time that an entire aircraft like this has been recovered and there have been attempts to document the recovery of all the various pieces of the aircraft. Doctor.

David Meyer: Mr. Chairman, early on in the investigation it was recognized that there was a need to track evidence both for purposes of criminal investigation and also for purposes of the accident investigation. It was very important to know where pieces were recovered to enable a number of studies that we were certain would follow up on the recovery process. The way this done is that as pieces were recovered at sea they were tagged with a color-coded tag, but that's not the only information we got about pieces. In addition to the color-coded tags that were attached directly - aluminum tags like these - paperwork was filled out and documented that took a separate path back to the hangar.

What occurred is that the pieces were transported back to the hangar with their tags on them, but the paperwork documenting where those pieces were recovered took a different path and were transported also back to the hangar. We set up a database staff to read each sheet of paper that came in - there were thousands and thousands of sheets of paper that documented the wreckage. On the sheets of paper we didn't just document whether it cam back from the red, yellow, or green zone, we documented the specific latitude and longitude from which the piece was recovered. We used those pieces of paper to validate every single piece that was tagged at the scene to determine whether or not the tag color was correct based on the latitude and the longitude.

We also had another built-in check that we were able to use. A lot of pieces were recovered by mobile dive teams, in other words, sonar operations in the first week or so of the search operation identified thousands of targets and those targets were dived on during the coming months in the recovery process. As each target was dived on, it was assigned to a mobile dive team. The mobile dive team went to the assigned latitude and longitude, the assigned location of the piece, recovered the piece, and wrote down taking a separate GPS (global positioning system) reading where they were as the pieces were recovered.

All of that information came back to our data center in the hangar. We worked through each of these pieces to determine if the pieces that were tagged at sea were tagged with the appropriate colors based on where the diving operation was assigned and where the dives were accomplished. We spent months reconciling that information and I am confident that we have a good product that documents the recovery positions of all the ship-tagged items.

John Goglia: Thank you.

Jim Hall: If I may interject one more time, I apologize. We also had video, did we not, of a lot of the underwater recovery?

David Meyer: That's a good point, Mr. Chairman. We had underwater video that was taken by the ROVs, the remotely operated vehicles, the submarines, and we also had, in addition to the sonar images, we had laser line scan images. It's sort of an underwater television technology. We had all of that information available to us as we checked those pieces. In fact I've seen some members of my team in the audience here today who spent literally hours accomplishing that work.

Jim Hall: And I want to ask, Dr. Meyer, and Mr. Dickinson, and I apologize in advance for having to ask this question, but do you know of anyone who intentionally mistagged a piece of wreckage so we wouldn't have the factual information in order to proceed with this investigation?

David Meyer: No sir.

Al Dickinson: No sir, I do not.

Jim Hall: Thank you.

John Goglia: I have a couple of questions for Mr. Wildey and I'm trying to put them in order. Mr. Wildey, you mentioned the water bottles. For the benefit of the folks that were not in the hangar and did not receive the in-depth briefing that you provided to several of us Board Members, which we greatly appreciated, would you just go into a little bit of detail and explain the significance of those bottles? What were the size? You mentioned the location, and in fact it's on the screen now. And why were they important to the investigation?

Jim Wildey: I don't have the weight, but they are very heavy. They are large water bottles about 2 feet in diameter and maybe 5 to 6 feet tall, and there are two of them and they are the drinking water that's carried in the airplane for the passengers to use. Their significance was that they provide a large inertial resistance to motion of the front spar and during the break-up sequence as spanwise beam 3 fractured, the releasing overpressure from the wing center section began to act on the aft side of the front spar. When we reconstructed the front spar we found out that the spar had come kind of bulged forward in two lobes, one on either side of the center, and our conclusion was that these water bottles acted to restrain the center of the beam of the front spar, and that these two lobes bulging forward on either side, were a sign of the pressure acting on the aft side of the front spar.

John Goglia: Thank you. Also, I wonder if you could put the section of the fuel tank up that shows the keel beam. Now I'm going to take a little liberty here and call the keel beam the backbone of the airplane for the benefit of folks here who are not aeronautical. The backbone of the airplane, in this case the keel beam, only extends to the rear of the forward cargo compartment. But in order to hang such a large structure, we still need the equivalent of a backbone to carry up to the front of the airplane - for example, to take the loads from the nose landing gear and just the weight of the rest of the airplane, which is pretty big in the front. To accomplish that, that backbone then changes form and it changes to a structural member that actually represents the floor of the cargo compartment.

In your description, you went over that a little bit lightly - the significance of that large piece of fuselage that failed very early in the sequence. I wonder if you would go back and try to explain for the benefit of others here just what the significance of that event is.

Jim Wildey: The keel beam broke approximately in the middle, underneath the wing center section. I don't think we have a good picture of that, but about halfway back through the tank is where it fractured. The forward end of the keel beam came out early on in the sequence, with the red zone pieces, and it was found in the red zone. The aft end of the beam remained with the rest of the structure and was found way down range in the green zone. So it's clear that the forward part of the keel beam came out early on.

Now what our group determined was that the piece created in the fuselage, which is connected strongly to the bottom of the keel beam, once it fractured around three sides of this piece of the fuselage structure, there was sufficient loads downward on this structure forward of the keel beam to physically pull it out from the bottom of the skin - the lower skin of this wing center section - break bolts that attach it to the wing center section, and then fracture that beam about halfway back. So it was part of the overall sequence and we did analyze, using stress analysis, how this could occur, and developed a rationale for the fracturing and release of the forward portion of the keel beam.

John Goglia: Would you move forward now to the portion underneath the cargo compartment floor?

Jim Wildey: That would be number 16, I would think.

John Goglia: I think many in this room would benefit from the significance of this forward area and what it means to the forward portion of the airplane.

Jim Wildey: The significance is as high you can get. If you lose the lower portion of the fuselage in this area, you have no capacity to take the bending loads from the nose section and at least one of the features that we saw after loss of the structure from window belt to window belt as a part of the initial event, then the nose simply started to fold down because there's no capacity for taking the bending of the nose down. And we saw this compression damage extend from the window belts up towards the top of the airplane. Once this hole is created across the bottom, the nose portion is going to come off.

John Goglia: Thank you. That wasn't clear in your initial presentation. Another piece of significant work that you and your team have accomplished that you went over rather lightly, I thought, was the rivet-to-rivet analysis that you did of the fracture. The rivet's facing is about 3/8 of an inch?

Jim Wildey: I believe it's about one inch.

John Goglia: So this jigsaw puzzle of fractures that are out there, your team analyzed inch by inch. That's a significant amount of work.

Jim Wildey: I would agree with that, yes sir.

John Goglia: I don't believe there was enough recognition of that given in the presentation. You didn't just take for granted anything as you went over the fracture.

And one last comment I've got was the speed of travel. You said that whole event took seconds. I recall that someone said to us early on that these fractures travel at about 6,000 feet per second, and since the airplane's only 225 feet in total length, 6,000 feet per second means that these fractures probably only took a second.

Jim Wildey: Yes, I think I can address that a little bit. There's kind of two phases of the early portion of the break-up. The earliest portion, the first part of it, would include a very, very rapid propagation of these early fractures in the fuselage, and that in fact occurs so rapidly that it gets out ahead of the escaping overpressure to a large degree. However, after these early fractures are created then things slow down a little bit and it takes several more seconds for the nose to come off as the fractures are tearing - not in a rapid manner but more of a tearing manner - across the top of the airplane. So the several seconds refers to the total separation of the nose, and certainly there would be something less than a second, I'm sure. I don't have the number for you, but it would be less than a second for the initial fractures of spanwise beam 3, the front spar, and these early fractures that we were looking at in the lower portion of the fuselage.

John Goglia: One last piece of this. And again it's for the benefit of those that may not understand this. I wonder if you could just explore briefly the shearing and tearing and the differences it meant to you as a metallurgist and to your team.

Jim Wildey: The fuselage contained - I can't put a number on it - a large amount of skin fractures, and one of our tasks was to look at all of these fractures and try to characterize them in terms of the sequence. There was an initial assessment that anything that fractured along the rivet line, for example, that would have been from an overpressure event. We quickly determined that that's not true. You can fracture along the rivet line from out-of-point stresses. For example, if the pieces on either side of the fracture are bending or deforming in any way next to each other, then we determine that there must have been other fractures somewhere else to allow these pieces to bend, relative to each other, before the fracture occurred.

That's distinctly different from these early fractures we saw, again maybe on slide 16, if we could put that back up there. These early fractures are distinctly different in that the pieces on either side of the breaks here pulled apart either in direct tension, with no deformation, or else in an unzippering effect. And that was the distinguishing characteristic that we used to distinguish these early fractures from the vast multitude of fractures throughout the red zone area and the rest of the airplane that occurred later because the pieces were moving relative to each other. I hope that answered your question.

John Goglia: I hope that it put some at ease that they now understand what we went through. One last question. I've had people ask me about Section 41 repairs - and I know you're familiar with that term. Section 41, for the benefit of others, is the forward section of the airplane, just after the main entrance door from the nose back. There have been some AD notes issued against this airplane - required inspections and repair from the FAA, that's what an AD note means. Did you find any evidence in your work to indicate that any of the repairs, or any of the issues in the Section 41 AD note was a factor in the break-up of this airplane?

Jim Wildey: We found that there were no issues related to that, we found no fatigue cracks in that area addressed by the airworthiness directives for the upper lobe, I believe it is. And also in that area, we looked at the Section 41, 42 joint, found no evidence of a preimpact with the water for that joint there at Section 41, 42.

John Goglia: Thank you, Mr. Chairman. No further questions.

Jim Hall: Member Black.

George Black: The wonderful thing about being this order in the sequence - there's not much left. Jim, would you talk a little about - a lot of this initial break-up came around the area where there was a pairing for the wing, was there not? That pairing is fiberglass, is it not?

Jim Wildey: It's a honeycomb and fiberglass. It's not designed for structure; it's just designed to smooth the transition from the wings to the fuselage, and things like that.

George Black: And that material, for the most part, was part of what the chairman was talking about earlier about floating material that was found early on, and we could not place it.

Jim Wildey: And also we did find a large portion of the right wing that was floating. That was, I believe, the largest structure.

George Black: I guess the point I'm trying to make is a lot of this pairing would have come off and been floating would have actually been red zone material because it goes around the area that trailed. In connection with that, could you talk a little bit about the effects that airflow would have had on this belly skin area as it was coming loose? In other words, there's still a slipstream of close to 400 knots there. Would you talk a little bit about how that would have helped unravel that area around that ultimately ran to the front section departing the airplane?

Jim Wildey: We talked about this a little bit before. The early fractures that I discussed happened so fast that the air stream flow doesn't really have time to really act on them before they're created in the present. After they're created, the piece from the lower portion of the fuselage is pushed downward by the pressure differential inside the airplane, and that piece then would get into the air stream. Again, though, this is happening very dynamically, and I believe that a lot of the fractures in even of the keel beam are happening probably before the air stream has a chance to really play a major role in the break-up.

George Black: The only party that dissented to some degree to our report, in the conclusions during the factuals, raised this issue about the area around, I believe it was, the number 3 left door. Could you talk a little about what they're talking about in their party's submission there? Or is that to be covered later?

Jim Wildey: I'm not sure what that issue is. Perhaps someone else can address that.

Bernard Loeb: Member Black, the truth of the matter is, I'm not quite certain what the IAM is trying to say in that submission regarding L3 door, but it appears that they're trying to indicate that there must have been some sort of an event in that vicinity that led to a structural failure initially that then led to the explosion in the center wing tank. What I can tell you, and I think Jim Wildey can address further, is that we looked at those areas very closely, saw no evidence, none whatsoever, that there was an event that resulted in a structural failure in that area prior to the CWT explosion. Jim, do you want to talk about that a little bit?

Jim Wildey: Dr. Loeb phrased it perfectly. There's just no evidence that there's an initiating event of any kind associated with any of the doors, including the L3 door.

George Black: I'm glad you got a chance to respond to that because they had also linked that to the splatter study we talked about earlier. I think that in the factual document that was a little late in being produced and I think still went very well. And I just wanted to close out that particular issue since it was the only disagreement that we had that I saw in party submissions.

Speaker: You mentioned CWT, and that was the first time that we used that acronym.

Bernard Loeb: Center wing tank. Sorry about that.

George Black: I believe that's all I have, Mr. Chairman.

Jim Hall: Alright. In order of seniority, Member Carmody.

Carol Carmody: I want to add my thanks to others who have expressed them to the staff and to military who worked on this issue. I think the presentations this morning were wonderful. I think the report, which I read several weeks ago for the first time, was comprehensive, factual, clear. My compliments to all of you.

I have a couple of questions being less than others, perhaps. First, for Mr. Dickinson, I understand the recovery lasted about 10 months, although I believe you said that a great portion of the plane was recovered reasonably soon. I wonder if you'd tell us how long you stayed on the site and what your activities were during the time you were there.

Al Dickinson: Since the investigation extended for a prolonged period of time, we actually set up rotation schedules, where a lot of our group chairmen, including myself, had a deputy IIC standing for me and we rotated back and forth. Overall, though, we were there for probably more than a year altogether, and that includes all the groups - all the party members. They formed different groups and rotated back and forth because of the extended time.

Carol Carmody: Thank you. My next question relates to the FBI, and I don't remember frankly whether it was Dr. Loeb or Mr. Dickinson who mentioned this, but whoever is appropriate to respond, please do. You mentioned the FBI was testing items as they were recovered several times, on the ship, the recovery vessel, I assume at Calverton and elsewhere. Could you describe that process a little bit for us?

Al Dickinson: Yes, they had equipment brought in to Calverton, where they did some testing for explosive residues on specifically the floorboards within the cabin area and a lot of other selected pieces. And in addition, when they'd find something that they didn't understand, they would send it back to the FBI headquarters in Washington.

Carol Carmody: Did they also do some testing on-site if they recovered material?

Al Dickinson: Not right on the boats. As far as I understand, it had to be brought in to Calverton. If it was just visual examination, there were certain concerns of pieces of wreckage brought in that they thought indicated something and when it was brought in we discussed it.

Carol Carmody: I also saw reference in the report to the Bureau of Alcohol, Tobacco, and Firearms - they're the explosives people, I guess. What did they do? How did their work relate to the FBI, and what did they do in this process in terms of testing parts of the aircraft?

Al Dickinson: They were actually underneath the FBI. They were responsible to the FBI, although they did participate in our progress meetings. We had progress meetings every day for months and months into the investigation. And they would go around and look at every piece of wreckage and then eventually put together a report that indicated what they found, which basically wasn't anything.

Bernard Loeb: Let me just clarify that. I don't think it was nothing. They did conclude that there was no evidence of a bomb having exploded within the airplane.

Carol Carmody: Exactly. That's what I was getting at. It wasn't just our experts, it was a lot of other experts.

Bernard Loeb: No. In fact, as the chairman put up earlier on the slide, there were a great number of outside organizations that participated in the explosives end of it. A large number of different organizations, within the government and outside the government, who looked at the metal all came to the same conclusion that there was no physical evidence of a bomb or a missile warhead exploding.

Carol Carmody: Thank you. I was going to ask Mr. Wildey to describe his work on the metal, but it's already been asked and answered. I think that was very important. I was fortunate to have a briefing, as many of us were, at Calverton, and I was especially impressed at the way you were able to analyze the different tearing patterns and hydraulic damage to the metal. So thank you. That's all I have for now.

Jim Hall: Mr. Wildey, I want to be sure we've covered all the various issues that could be discussed. Do you think we've covered the discussion on the forward cargo door adequately in your presentation and your response to questions and answers? Or is there anything else you want to add in that area?

Jim Wildey: I would only add that in a certain way we've really hit it too hard. The evidence is very clear and it's overwhelming. We could hit it again, but it's clear that the door was closed, attached to the rest of the structure, and a part of the nose portion when it hit the water.

Jim Hall: As you're aware, I get and try to read every piece of correspondence that's come into my office on TWA 800, and at the beginning we had a whole lot of correspondence from several individuals about the Section 41, 42. Is there anything else that needs to be said about Section 41, 42.

Jim Wildey: I don't believe so. It's clear that that did not play any role in the break-up sequence.

Jim Hall: Again, is there anything else that needs to be said about fatigue cracking - the age of the aircraft or anything else that could have caused this particular break-up that you have not covered in some detail?

Jim Wildey: I don't think so.

Jim Hall: I guess I spurred Member Goglia's mind. He's got a comment.

John Goglia: Also, Mr. Wildey, after the Aloha accident, we had a considerable effort in the pilot industry to deal with aging structural issues. Did you see any issues in this airplane that would indicate that they have overlooked anything in their work?

Jim Wildey: From a structural standpoint, absolutely not. We did find, as I mentioned, a few small fatigue cracks, but they did not play any role in the break-up sequence. Also, the amount of corrosion was minimal, even having considered it was in seawater. The airplane structure was in good condition and did not play any role in the initiation of the break-up.

John Goglia: I would like to note that that work that flowed out of that Aloha accident was the result of some work that was done here by NTSB staff, and it's a credit to them and to the FAA who followed up on that to make sure that we didn't have any additional problems. Thank you.

Jim Hall: Again, as I said, keep it in mind that this is the very first time that either the FBI or the National Transportation Safety Board were charged with recovering an aircraft from the ocean and documenting the location of every piece of the aircraft. Mr. Dickinson, was there a system in place, prior to TWA 800, to retrieve from the sea and document every piece of wreckage, or did that system have to be developed as the recovery operation proceeded?

Al Dickinson: As far as I know, Mr. Chairman, we had to do develop it as we went.

Jim Hall: And the FBI, I know, was very interested in that because there was a possibility that one of this piece of the aircraft might be used as evidence in a criminal prosecution and the trail of that piece of wreckage, was it extremely important?

Al Dickinson: Yes, sir. They were very interested in chain of custody issues.

Jim Hall: And how many people, can you give me a guess, were actually involved in the retrieval of the wreckage after the initial response in which - as I acknowledged and thanked in my opening statement - there were, I would say, tens if not hundreds of small craft that went out attempting to locate and recover, most importantly, the loved ones off this flight and then did bring in pieces of floating wreckage from the aircraft?

Al Dickinson: Mr. Chairman, I know there were over 100 divers involved, and all the ships that were involved - there were at least five different Navy ships that were involved and various others that we contracted out - so you're talking hundreds of people altogether that were involved in the recovery.

Jim Hall: And how was it structured in an attempt to supervise that particular operation inthe recovery and documentation of the wreckage?

Al Dickinson: The Navy actually had a command ship out at the accident site that supervised all the Navy assets out at the site, and at Coast Guard Moriches, they set up their headquarters there that we were involved in every day. We oversaw the whole operation.

Jim Hall: Did we have someone on each ship at all times? Or did the FBI?

Al Dickinson: Yes, it was either us or the FBI. When we got into the trawling, it was a six-month period over the winter months, we did not have NTSB people there, but the FBI had agents on each one of the ships.

Jim Hall: And these FBI agents that were out there on those ships, because I was out there, were extremely cold and they did hard work and we should thank them for their efforts.

Al Dickinson: Yes, sir.

Jim Hall: Mr. Wildey, you have gone through this in great detail, and you're still going to be at the table through the whole two days, and we may come back and ask you additional questions in regard to this part of the investigation. But is there anything else that you think needs to be added at this time?

Jim Wildey: No, sir.

Jim Hall: Mr. Dickinson, again, you're going to be with us through the two days. Is there anything else that you think needs to be added at this time?

Al Dickinson: No, sir. Thank you.

Jim Hall: If not, we will stand, we will take a lunch recess. We will return at 1:30 sharp to rebegin these proceedings. The National Transportation Safety Board stands in recess.

8/22 - Afternoon Session, Part 1

Jim Hall: We will reconvene this meeting of the National Transportation Safety Board. The Board is in the midst of our consideration of items involving the in-flight break-up over the Atlantic Ocean of Boeing 747-131, known as Trans World Airlines Flight 800 that occurred near East Moriches, New York on July 17, 1996. I apologize - it has been pointed out to me that I inadvertently said this hearing was going four days, rather than two, and I did not want to put anyone in a state of shock, but it is two days. And the proceedings are being covered worldwide, live on the National Transportation Safety Board's Internet site,

We are now ready for the next item on the agenda, which is a discussion of fuel tank flammability, and I will ask Dr. Loeb to introduce our presenter.

Bernard Loeb: Thank you, Mr. Chairman. Dr. Joseph Kolly will be presenting the section on flammability. Dr. Kolly is a Fire and Explosion Investigator in the Office of Research and Engineering, where he responsible for fire and explosion accident investigation and research in all modes of transportation. He joined the Safety Board in 1998, but before joining the Safety Board he worked for six years as a Senior Research Scientist at Calspan-University at Buffalo Research Center, where he was responsible for conducting basic research in fluid and thermal sciences of high-speed flows. He also held the position of Operations Manager of the Large Energy National Shock Tunnel at Calspan.

Dr. Kolly earned a BS degree, with high honors, in Mechanical Engineering from the State University of New York at Binghamton in 1988, and he earned a Ph.D. in Mechanical Engineering from the State University of New York at Buffalo in 1996. Dr. Kolly?

Jim Hall: Dr. Kolly, before you begin, again, may I remind our presenters that these proceedings are being translated for the benefit of some of our French families and French media, and if you would speak slowly and clearly into the microphone, it would be most appreciated.

Joseph Kolly: Good afternoon, Mr. Chairman, and Members of the Board. I would like to present the results of four years of investigation, testing, research, and analysis that were conducted to investigate the explosion of the center wing tank on board TWA Flight 800.

Jim Hall: Can everyone hear Dr. Kolly alright? If you could speak up just a little. These microphones you just have to get very close to.

Joseph Kolly: My presentation will emphasize two key findings. First, that a flammable condition was present inside the center wing tank of TWA Flight 800 at the time of the accident. Second, that the ignition and combustion of this flammable mixture of Jet A vapor can generate sufficient pressure to break apart the structure of the center wing tank.

When we began our investigation, it became evident that much of the basic information we needed to establish these findings did not exist. Therefore, an international team of some of the world's leading fuel and combustion experts was assembled to develop this information. In the course of this presentation, I will explain how this information was developed and tell how the results are consistent with an accident that resulted from the combustion of a flammable mixture of Jet A vapor inside the center wing tank on board Flight 800. Following this discussion, I will present a summary of the computer modeling effort that was undertaken to simulate combustion in a full-scale center wing tank.

To fully understand the work and findings of this team, you will need to be somewhat familiar with the construction of a 747 center wing fuel tank. This simplified drawing shows the salient features of the center wing tank. Note that the tank is divided by sparring beams into seven individual compartments referred to as bays. The forward most bay is referred to as a dry bay and does not hold fuel. The remaining six bays are intended to hold fuel. The fuel-containing bays are interconnected by a series of openings and passageways in the spars and beams and by a set of vents, which vent the tank to the wing tips and then to the atmosphere outside the airplane.

The type of fuel that this aircraft uses is known as Jet A fuel. Flight 800 took off for Paris as planned, with a nearly empty center wing tank. The tank contained approximately 50 gallons of residual Jet A fuel. This amount of fuel would form a small pool only a few inches deep, its location depending upon the attitude of the airplane. The remaining volume of the center wing tank is occupied by fuel vapor and air. The space occupied by fuel vapor and air, and not the liquid fuel, is known as the ullage of the tank.

As you can see, the center wing tank has a complex configuration. Little was known about the environment found in and around this configuration before the accident and, therefore, no credible means are readily available to assess the flammability of the accident airplane's center wing tank. Therefore, the Safety Board developed an experimental flight test program to obtain scientific data concerning the thermal and vapor environment within and around the center wing tank.

In the summer of 1997, the Safety Board, with assistance from Evergreen Air and Boeing, conducted an experimental flight test program in which the temperatures and fuel vapor concentrations were measured and recorded during different flight and ground conditions of a Boeing 747-100 series airplane. Nine flight tests were performed. One test, known as the emulation flight test, was designed to replicate the configuration, ground operation, and flight profile of Flight 800 as closely as possible.

Measurement censors were installed in the flight test airplane within and around its center wing tank, as illustrated by the yellow circles in this simplified figure. Temperature, pressure, and vibration measurements were continuously taken during the entire period of ground and flight operations at nearly 200 different locations. Additionally, fuel vapor samples were taken from within the tank. This flight test program marked the first time that the thermal and vapor environment inside an in-flight aircraft fuel tank was measured.

The results presented here are from the emulation flight test, and represent the condition on board Flight 800 as closely as possible. This graph shows some of the temperatures of the ullage, or vapor space above the fuel, measured in various locations of the center wing tank. These temperatures are shown as vertical bars. They were measured at the time in flight that the test airplane reached 13,800 feet, the altitude at which Flight 800 was at the time of the accident. The temperatures of the ullage had increased from their initial preflight temperatures, which averaged less than 80 degrees Fahrenheit, to an average high temperature of about 120 degrees Fahrenheit.

Shown on this graph is a red horizontal line at the equivalent flash point of Jet A fuel at this altitude. Note that all the fuel tank temperatures are higher than the flash point temperature. As a general approximation, when temperatures of fuel exceed their flash point temperature, the fuel tank is considered flammable. These measurements indicate that a flammable condition was present in the center wing tank of Flight 800 at the time of the accident.

The Safety Board sought a more definitive measure of flammability to accurately determine the conditions on board Flight 800. This was accomplished by sampling and analyzing the fuel vapors from within the center wing tank on board the test airplane to directly measure the fuel vapor concentration and to accurately assess its flammability.

The Safety Board contracted with the Desert Research Institute of Reno, Nevada to design a vapor sampling system for installation on board the test airplane and to analyze the vapor samples that were collected from the flight test. The analysis was meant to determine the chemical composition and quantity of the fuel vapors.

Shown in the figure is the sample collection device installed on board the airplane. The stainless steel canisters and the metal container are where the vapor samples were collected. Following the flight, the sample canisters were shipped to the laboratory for analysis. The analysis of these samples showed that the fuel vapor inside the center wing tank of the test airplane was flammable at an altitude of 13,800 feet. This indicates that a flammable condition was present in the center wing tank of Flight 800 at the time of the accident. This finding is consistent with the previous comparisons of temperatures to flash point temperature.

A third, independent analysis of flammability was provided by the experts at the University of Nevada at Reno's Center for Environmental Sciences and Engineering. They performed research to determine the chemical composition of many different sources of liquid Jet A fuel and its vapor. This provided us with a method to predict the flammability of Jet A fuel at a variety of simulated flight conditions. The results of this analysis independently confirmed that the fuel air vapor inside the center wing tank of Flight 800 was flammable.

Based upon the analysis of the emulation flight test measurements and the supporting laboratory research, we note this significant finding: that three independent flammability assessment methods were used which confirm that the center wing tank on board Flight 800 at the time of the accident was flammable. On the basis of this information, investigators looked closely at the 747 center wing tank and its relationship to other components of the airplane to determine how and why the temperatures in the center wing tank can rise to such high levels during operation of the airplane.

This graphic shows a frontal view of the configuration of the center wing tank and the air conditioning packs. The center wing tank is shown as a rectangular, tan-colored area below the passenger cabin floor. Beneath the center wing tank is an enclosed space, called the pack bay, which contains the air conditioning units called packs. Two of the three packs are visible in this view as the dark-colored units below the center wing tank. To either side of the center wing tank are the in-board main fuel tanks, shown as the dark area inside each wing. The arrows represent the main flow of heat.

When the air conditioning packs are in use, waste heat from these packs raises the temperature of the air in the pack bay beneath the center wing tank. In the cast of Flight 800, the surface temperatures of some pack components may have exceeded 300 degrees Fahrenheit, and the air space in portions of the pack bay may have exceeded 200 degrees Fahrenheit before takeoff. These elevated temperatures drive heat into the center wing tank and raise the temperature of the fuel vapor within the center wing tank.

Heat leaves the center wing tank from the tops and sides of the tank. A portion travels into the in-board main fuel tanks on either side, which is then transferred to the air outside the airplane through the wings. This process of heat removal is slowed when the airplane remains on the ground and in a hot weather environment for extended periods of time, as was the case for Flight 800, which remained at the gate for nearly three hours with the air conditioning system operating. Once the aircraft begins its flight, the colder air at high altitudes helps to remove heat from the center wing tank and the inboard mains.

The FAA has made estimates of the flammability exposure time or percentage of time that an aircraft is operating with a flammable condition within the center wing tank. For the entire fleet of large transport aircraft that had this configuration, the flammability exposure time is estimated to be 30 percent.

To summarize, the center wing tank can be heated to temperatures much higher than the outside air temperatures, which increases the flammability of the tank. A major source of heat that causes this temperature increase is the air conditioning packs located beneath the center wing tank.

Having established that a flammable condition existed on board Flight 800 at the time of the accident, we then needed to determine the conditions for ignition of this vapor, the combustion pressures that could develop, and the rate at which such pressures could develop.

When we began the investigation, sufficient technical information required to answer these questions did not exist in the available literature. Therefore, the Safety Board sought to develop an extensive research program of Jet A fuel flammability and combustion behavior. This research was contracted to the California Institute of Technology's Graduate Aeronautical Laboratory and the experts from its Explosion Dynamics Laboratory.

This work explored many of the important elements of the combustion process relevant to the Flight 800 accident and took over four years to complete. This work represents a significant portion of what the world now knows about Jet A fuel flammability and combustion.

One area of research was aimed to determine the energy required to ignite Jet A fuel vapor. To achieve this, specialized combustion test chambers, like the one seen here, were constructed. Electronic circuits were devised to precisely measure the energy required to ignite the fuel vapors by a small spark. Several hundred experiments were run to statistically determine Jet A fuel's ignition limits. For the conditions of Flight 800, the spark ignition energy requirement was determined to be between one-half and 500 millijoules. For reference, the energy in a static electricity spark that can jump from your fingertip to a doorknob can be in the range of one to 10 millijoules. Following my presentation, Mr. Swaim will illustrate examples of spark ignition mechanisms that have energies that greatly exceed these values.

Another area of Cal Tech's research was aimed to determine the combustion behavior of Jet A fuel. This involved hundreds of experiments in large combustion chambers for a wide variety of conditions. These experiments measured the maximum overpressures that result from the combustion of Jet A fuel vapors and the speed at which this occurs. The results of this research indicate that for the conditions on board Flight 800, peak explosion pressures ranged from 39 to 52 pounds per square inch higher than the initial pressure of the tank. This range of peak pressure is greater than the pressures required to break the 747's center wing tank structure.

As a result of this research, three important findings were reached concerning the conditions on board Flight 800 at the time of the accident: first, that fuel/air mixture in the center wing tank was flammable and, on the basis of laboratory tests that simulated these conditions, was consistently ignitable. Second, this fuel/air mixture could have been ignited by spark ignition energies between one-half and 500 millijoules. And third, that peak explosion pressures were sufficient to break apart the structure of the center wing tank.

The previously presented combustion research was conducted in large laboratory test chambers. The Safety Board determined that combustion behavior also needed to be investigated in chambers more representative of the center wing tank that contained intercompartment connections and vents. The primary purpose of these tests was to provide a definitive set of data for the development of state-of-the-art computer models of center wing tank combustion.

With this objective, the Safety Board contracted with the same Cal Tech explosion experts and also with Applied Research Associates, experts in the field of large-scale explosion testing. A three-year test campaign was conducted in which a quarter-scale model of the center wing tank was constructed and a total of 72 explosion experiments were performed.

The first phase of testing used a fuel that simulated the behavior of Jet A fuel, without needing to recreate the temperature and altitude conditions of Flight 800 at the time of the accident. The second and third phases of testing used Jet A fuel at the same temperature and altitude of Flight 800 at the time of the accident. The quarter-scale model shown here is essentially one-quarter the length, width, and height of the actual full-scale center wing tank. The model incorporates the important features of the actual tank, such as vents and passageways; however, the finer details of the geometry are simplified.

This photograph shows the initial installation of the quarter-scale model at the outdoor test site in Colorado. The red outline you see in the picture is looking through a heavy window that comprises the side of the tank. The heavy steel structure surrounding the red outline is the support structure which strengthens and stabilizes the model during testing. Temperature and pressure sensors are installed throughout the model, which record these signals during each experiment. The model is housed in a small building, which allows the model to be climate controlled. This was necessary to simulate the temperatures and altitude conditions of Flight 800.

Phases 2 and 3 of testing conducted 28 tests using Jet A fuel vapor at the conditions of Flight 800. In every test, the Jet A vapors ignited and combusted. Analysis of the peak explosion pressures measured during these tests indicated that pressure levels can well exceed those required to break apart the full-scale center wing tank.

I will play three video clips of one of the quarter-scale tests. This test was conducted using Jet A fuel at the temperature and altitude conditions, and equivalent fuel load, of Flight 800's center wing tank at the time of the accident.

The first video clip will show an external view of the test rig building. Within the black window frame on the side of the building, you will see the orange glow from the fire of an explosion. On the left side of the building, you will see a metal panel of the quarter-scale structure being ejected, along with flames and fire. This panel failed when high pressures had developed to break it free from the tank, which simulates the failure of the front of the center wing tank in Flight 800. This clip will be shown twice in succession, first at normal speed and second at slower speed.

This next video clip is of the same test shot, this time with a close-up of the rear of the building and the opening where the quarter-scale model panel is ejected. This clip will also be shown twice.

The final clip is also of the same test shot, this time showing a close-up view through a heavy Lexan window on the side of the quarter-scale model tank. The diagonal striping you see is used to aid visual measurements taken from the video. Labeled are four rectangular compartments representing the bays of the center wing tank. This test shot will reveal the flash of a hot filament near of the bottom of bay 2 used to ignite the vapors. Following this, there is a growth of a nearly invisible flame front that burns the fuel vapor in bay 2. This process takes less than half a second.

By the time the flame front reaches the passageways and vents connecting the neighboring bays, the neighboring bays have been pressurized and the ullage space is very turbulent. Because of this, all remaining bays ignite very quickly - in fact, in less than one-tenth of a second. This clip will also be shown twice.

In addition to providing data toward the development and validation of computer models, two important findings were made on the basis of these tests. First, in every test using Jet A fuel at the temperature and altitude conditions of Flight 800's center wing tank, combustion had occurred. Second, that peak explosion pressures were measured that exceeded those calculated to cause structural failure of the Boeing 747 center wing tank. In the development of these pressures, it was demonstrated that combustion can travel from bay to bay through the passageways and vents, as we've seen in the video clip.

On the basis of these research, testing, analysis, and investigation efforts, we conclude the following: first, that the center wing tank, at the time of the accident, was flammable. This was supported by the flight test temperature measurements that exceeded the flash point of the fuel, by the analysis of the flight test vapor samples that determined that the vapor concentration levels inside the tank would have been flammable, by a chemical analysis of Jet A fuel that simulated those flight conditions, and by flammability and combustion experiments.

We also conclude that ignition and combustion of Jet A vapor at the conditions on board Flight 800 can generate pressures that exceed the pressures necessary to fail the 747's center wing tank. This is supported by laboratory and quarter-scale model experiments that measured peak explosion pressures that exceeded the calculated failure pressures of the 747's center wing tank.

I am now going to summarize the computer modeling simulation effort that was developed to attempt to determine the location of the ignition of the center wing tank for Flight 800. There was hope that if the primary ignition location could be found, the ignition source might be identified. For the past three years, an extensive research effort has been undertaken to develop computer models capable of accurately simulating the complex phenomena of Jet A fuel combustion within the intricate configuration of the center wing tank.

Because of this test's enormous technical challenge, the Safety Board assembled an international team of the world's leading experts in large-scale combustion dynamics modeling. Experts from Christian Michelsen Research from Bergen, Norway, were selected, bringing with them their vast experience in large-scale fire and explosion models, such as their computer modeling of off-shore oil rig explosions. Additionally, experts from Sandia National Laboratories in Albuquerque, New Mexico, were selected, also bringing with them a great deal of large-scale explosion modeling experience, such as their computer modeling of the gun turret explosion on board the U.S.S. Iowa.

Each of these groups developed independent computer models that employed different modeling techniques. These codes complemented each other and provided a sort of check-and-balance system. Both models were developed and validated with extensive support from the Cal Tech experts and the results from their quarter-scale experiments. Thousands of simulations were performed throughout the development cycle. A final set of full-scale simulations was conducted, each simulation representing a different set of conditions. We call these "simulation scenarios."

This research revealed that the combustion process can be dependent upon where in the tank the ignition begins, and that once ignited the flame may not travel into all of the bays within the center wing tank. This indicated that it may be possible to analyze the results from all different combinations of computer simulation scenarios to determine the probable ignition of the explosion onboard Flight 800.

This task was contracted to Combustion Dynamics, Ltd. of Nova Scotia, Canada, leaders in the field of explosion and combustion dynamics research. Combustion Dynamics developed an analysis method that was used to determine the structural damage that would be predicted by each computer model simulation scenario and compare that to the damages observed in the wreckage of Flight 800. Using this approach, the search for the probable ignition location was pushed to the limits of current technology.

An accounting of the scientific uncertainties was meticulously maintained throughout the entire experimental, computational, and analytic processes. In the end, the uncertainties were too great to permit the identification of the probable location of ignition. However, the analysis did show that these simulations were consistent with the damages observed in the wreckage of Flight 800 and with the structural failure calculations of the 747 center wing tank.

In summary, two important findings were made that support the conclusions of the research presented earlier. First is the confirmation that the pressure levels caused by the combustion in the Jet A vapor in a full-scale center wing tank configuration can exceed the structural limitations of the center wing tank. Second, that an internal ignition and combustion of Jet A vapor is consistent with the damages observed in the wreckage of Flight 800 and with structural failure calculations of the 747 center wing tank.

I will conclude my presentation by reiterating the conclusions presented to you today.

First, that the fuel air vapor inside the center wing tank of Flight 800 was flammable. And second, that the ignition and combustion of this flammable mixture of Jet A vapor can generate sufficient pressure to break apart the structure of the center wing tank. This concludes my presentation.

Jim Hall: Thank you Dr. Kolly. I want touch on a couple of questions here and then we'll pass it around to the other board members. First of all, that was a very informative presentation. Maybe either you or Dr. Loeb can indicate why we selected California Institute of Technology, the Sandia Labs, Christian Michelson Institute, and the Combustion Dynamics group, briefly?

Bernard Loeb: Okay, I'll take a shot at this first. We identified these groups through a variety of means early in the investigation when we determined that we were going to need to do a good bit of work on both Jet A fuel and determining the characteristics of Jet A fuel. And also we had hoped that we could possibly help to pinpoint the ignition source by pinpointing the ignition location in the fuel tank, and therefore we wanted to find two things. One is those folks who had the best research capability and the most knowledge in fuels and especially the ability to work with Jet A fuel, and second of all, those people who had the best capabilities in terms of computer fluid dynamics modeling. And our search led us to Dr. Joseph Shepherd at the California Institute of Technology and also the folks at Sandia, and from there we learned about the people at Christian Michelson and Combustion Dynamics.

Jim Hall: Thank you. I wanted to put that on the record, particularly for the families, because I want everyone who's followed this investigation to know a great deal of thought went into trying to do testing with outside experts to try to pinpoint the ignition source.

Bernard Loeb: I want to make it clear that we considered virtually all of the research facilities that existed in this country and elsewhere in the world in this search, and what we wanted to do was to locate those research organizations that simply had the best capabilities no matter where they were located and that's why we settled on those organizations that we did.

Jim Hall: I want to talk a little bit about fuel, Dr. Kolly. Can you explain to us a little

about Jet A fuel and if that fuel has been used in the operation of jet engines since the inception of flight? Or maybe Dr. Loeb would like to get into that, and then I'd like to talk a little bit about what difference, if any, there was between the Jet A fuel that was put aboard the aircraft in Athens and the fuel that was added in New York. I believe that's covered basically on page 211 of the report, lines 11 through 15.

Bernard Loeb: To answer your first question, Jet A was not the first fuel that was used in jet transport category aircraft. There's a variety of different jet fuels, derivatives of kerosenes that are used in jet aircraft of a variety of types. Jet A came into being a number of years ago in use in the jet transport fleet because it was a fuel that was less likely to be flammable than the fuels that were used at that time. And there was concern about flammability and Jet A tended to provide the best set of characteristics in terms of the compromises that were needed.

First, you need a fuel that of course will burn and burn well in a jet engine and produce the thrust and the efficiencies, but also will start and operate under a variety of conditions and at the same time not be as volatile as other fuels. And Jet A is a less volatile fuel than the fuels that used to be used in the jet transport fleet, and the military in fact uses it in many of their aircraft, including their fighter aircraft. Jet A has been used probably for the last 15 years or so.

Jim Hall: I know maybe I am being redundant in this, but there have been some observers of this investigation and there were some statements made early on that the Jet A fuel isn't flammable. What, Dr. Kolly, have we done to establish the flammability of the Jet A fuel?

Joseph Kolly: The researchers at Cal Tech performed extensive studies to determine the ignition limits and the flammability limits of Jet A, and I can say that between the testing done at Cal Tech and the quarter-scale testing, there were literally thousands of tests performed of Jet A fuel vapors at the conditions of Flight 800, and they routinely and consistently ignited.

Jim Hall: I'm referring now to page 240 in our report. I believe that there had been some studies done by Boeing for the United States Air Force in 1979 and 1980 on this subject as well. Could someone briefly summarize that work, and was that done with Jet A fuel?

Robert Swaim: The Air Force research was done...I don't remember the fuel type. It wasn't to measure the fuel or the ullage characteristics, the vapor characteristics, it was to measure how much heat was transferred into the tanks from the packs while the airplane sat on the ground, and this was following some heating problems that were reported by some airlines.

Jim Hall: So that report did not get into the subject of flammability?

Robert Swaim: It did not get into flammability. It was looking at the heating of the fuel for the purpose of pumping the fuel.

Jim Hall: I wanted to make sure of that. Now what, Mr. Kolly, can we do to reduce fuel tank flammability?

Joseph Kolly: There are several things that can be done. Essentially, to reduce flammability, two main methods are to reduce the fuel vapors that are generated within the tank, or to inert those fuel vapors so that they're not flammable. Now, when you reduce the amount of fuel vapors you're essentially talking about reducing the temperatures of the fuel or changing the fuel properties by, say, raising flash point. When you inert the fuel tank, you add an inerting gas so that it does not matter the quantity of fuel vapors. They will be rendered nonflammable.

Jim Hall: I believe the FAA is looking at various options, and is that going to be discussed later or is this an appropriate place for that discussion?

Bernard Loeb: This is an appropriate place for the discussion. Of course, one of the options that had been considered a while ago was moving to a fuel that had a higher flash point, such as JP 5. The problem is that there's a huge infrastructure out there that's producing the Jet A, and the effects that JP 5 may have on the existing fleet, and so forth. The FAA was looking into that. They have not indicated to us that that's still under active consideration.

But there are a number of other things that they're looking at right now, including using nitrogen inerting for both an on-ground system that would essentially put nitrogen in the tank while the airplane was on ground. They are also looking at the possibility of an airborne system that would continue to produce nitrogen and keep the tanks inert. They are also looking at ways of trying keep the heat out of the tanks, or reduce the heat that goes into the tanks, and one of those ways is to use ground-conditioned air while the airplane is on the ground rather than using the air conditioning packs. So there's a number of different things that are being looked at right now, and we hope that in short order they'll pick some solutions that make sense. But we've been working with the FAA on this process, and Joe Kolly has been a major part of it.

Jim Hall: I believe we covered this in our hearing in Baltimore, but what does the military do in inerting, and what is their philosophy on the flammability. Bob, do you want to talk a little bit about that?

Robert Swaim: Certainly. The military has several options that they follow in the tactical, the fighter type of world, the ground attack world, where they know they have a hostile environment, people will be shooting at them. Some of those airplanes do have means of making nitrogen on board and putting it into the fuel tank. As far as the military equivalent or use of the 747, the presidential airplane, the alert airplanes, E4Bs, they typically stay with what is in the commercial world. They stay with the equivalent of commercial maintenance, use airworthiness its directives, and so forth.

Jim Hall: Okay, I've got a few more questions, but I imagine some of the other Members may cover them. If they don't, I'll get back to them. Member Hammerschmidt.

John Hammerschmidt: Considering the fact that this 747 had a fuel/air mixture in the center wing tank that was capable of creating this overpressure event, how many other aircraft in the fleet have this same condition during their daily flights? How prevalent is this condition, where a fuel/air mixture, capable of sustaining an overpressure event, would occur?

Bernard Loeb: The FAA has had some research done in this area and that research tends to indicate that about 30 percent of the time, if you take an overall view of the fleet at all times of the year, about 30 percent of the time airplanes are flammable, the tanks are flammable. Now let me put that in perspective a little bit, and that is, there may be some airplanes that fly and the fuel tank, the center wing tank, is virtually never flammable, and there may be conditions under which virtually all of every flight of a certain type is flammable. So this is an overall average that they're looking at.

John Hammerschmidt: Does a flight crew have any capability of being aware of the flammable situation in a fuel tank?

Bernard Loeb: In the center wing tank in these airplanes the answer is no. There are indicators in some airplanes, in some wing tanks, but in the 747 center wing tank, there is no indicator in the cockpit to indicate whether that tank is hot enough for it to be flammable.

John Hammerschmidt: What is our stated position in terms of giving the flight crew better information in that regard?

Bernard Loeb: We made a recommendation a while ago, asking that there be an indicator, a censor in the center wing tank and an indicator in the cockpit to allow the flight crew to know the condition of the center wing tank.

John Hammerschmidt: And we made that recommendation in December of 1996?

Bernard Loeb: That is correct.

John Hammerschmidt: And what is the status of the response by the FAA to that recommendation?

Bernard Loeb: They have not agreed with us on that recommendation at this point.

John Hammerschmidt: And in the report on that subject - and I'm on page 356, beginning on line 14 - we reference a November 3, 1999, letter, in which the FAA indicated that the work of the Fuel Tank Harmonization Working Group and researchers at CIT was complete. Then the FAA stated that it had concluded that no practical means for reducing fuel temperatures in fuel tanks existed, but that it was doing further work in that area. Is that the FAA's current position?

Bernard Loeb: I can't speak for the FAA in answer to that question directly, Member Hammerschmidt. What I can say is that my discussions with counterparts at the FAA indicate that the FAA is working now on looking at a number of means to remediate the flammability, and included in that is the possibility of the things that we've talked about - inerting and perhaps the use of ground air conditioning when the airplane's on the ground and some other things.

We are still interested in seeing some more work being done on the possibility of directed ventilation, of somehow cooling the air conditioning pack bays. As you know, as an airplane rises in altitude, the outside temperature drops significantly - if there was some way to take advantage of that to cool the center tank. The FAA and Boeing have done work on that to some extent and indicate that there's not much benefit in that. However, I'm not certain that we're as absolutely convinced that there is no way, perhaps using insulation and ventilation both, to try to keep the heat down in the center wing tank. So there are some things that we're still encouraging them to look at.

John Hammerschmidt: One final question, and I may back up into Mr. Wildey's subject area, but I was asked this question during the lunch break by someone in the audience, an airline pilot, and it was part of the factual report, actually. What role, if any, did the access doors and the spanwise beams and the mid spar play in the break-up sequence or in the overpressure event.

Bernard Loeb: They helped us to understand it better. But I'll let Jim Wildey explain that to you.

Jim Wildey: The access doors didn't really affect the break-up sequence. The initial fracture was in spanwise beam 3 at the top of the beam; it wasn't through the door. We did find in spanwise beam 2 that one of manufacturing access doors was released also, but again this was not the initial event, this was a follow-on event. So the doors themselves didn't really play a role in the break-up.

Bernard Loeb: But they did help us to understand the overpressure a little bit, and maybe you can explain that a little bit.

Jim Wildey: We found some sign of a direction of overpressure in spanwise beam 1, near the aft end of the tank, that gave an indication of an aft-reacting overpressure. In spanwise beam 2, the manufacturing access door, after it broke, was flung forward and up. Again, this is consistent with the pressure in the bay in front of that as spanwise beam 3 fractured. The bottom line on it with most of the doors is we weren't really able to use the signs of the pressure deferential that much in the determination of the break-up sequence.

John Hammerschmidt: Thank you. That's all I have.

Jim Hall: Member Goglia.

John Goglia: Yes, sir. Thank you. Mr. Kolly, I have a couple of questions around this whole area of fuel. I've written them down and I don't have them in order, so excuse me a minute. I'd like to go back to some of the early testing that we did. My recollection is that during some of the early testing, we used another fuel in the tanks. We used another fuel to simulate the conditions that we believed existed in the tank. I wonder if you could explain to us all the process and why we used an alternative type of fuel.

Joseph Kolly: The initial phase, as I said, the quarter-scale tests, were conducted in three phases. And they were conducted primarily to develop a set of data to develop the computer codes which we could use for a validation.

Phase one of the tests used a fuel to simulate Jet A fuel, without the need of having to also raise the temperature of the test fixture to the temperatures that were on board Flight 800, and also we did not require the need to reduce the atmosphere to the pressures associated with the altitude of Flight 800. And so a similar fuel was used in phase one, which was developed in the laboratories at Cal Tech to duplicate the peak pressure rise of Jet A fuel at those conditions. However, it allowed us to conduct those conditions at the outdoor test site at atmospheric conditions at the site, which facilitated the phase one testing - allowed us to concentrate on other areas of the testing to develop that data.

In phase two and three we upgraded the model to incorporate the heating and the pressure reduction, or the simulation of altitude. When you start to do that, the testing starts to become much more complex, and you start to get less and less experiments done in a given time. So it was appropriate to first use a fuel to simulate so that we could get the appropriate amount of data to start the development of the computer models.

John Goglia: And did that phase one testing commence before we had the data from the flight tests?

Joseph Kolly: No it did not. We needed the data from the flight test in order to understand what temperatures we were dealing with and how Jet A behaves in that environment, so that a simulant fuel could be developed to mimic that behavior.

John Goglia: You mentioned a second ago about altitude. Can you explain to us the effects of altitude on these fuel/air explosions?

Joseph Kolly: Yes. I have a slide that I'd like to go to. Slide number 12 of the supplementals, please. Here is a simplified drawing of fuel tanks. The one at the left is at sea level and the one on the right is at 13,800 feet or at some altitude. The red circles indicate fuel vapor molecules and the blue circles represent air. When we're at sea level the air is essentially thicker, and so you have more air per fuel molecule, so the fuel/air ratio is a lot smaller or lesser at sea level. If temperatures are low enough, this could essentially be nonflammable. However, when you go up to altitude, you have the same amount of fuel, but less air. So the fuel/air ratio increases, and you can be in a position to move from being under the lower flammability limit to above the lower flammability limit, to move yourself into a condition of a flammable vapor.

John Goglia: What effect do these changes - namely, the lower and higher flammability limit changes - have on ignition sources energy levels?

Joseph Kolly: There is an optimal level between the lower and upper flammability limit in which the smallest amount of energy is required. As you move to one extent or the other - as you get more towards the lower or more towards the upper flammability limit - the energy required to ignite increases. And as I explained in my presentation, when you went from 50 degrees Fahrenheit fuel, it required only one-half a millijoule energy; when we went to 40 degrees, it was closer to the lower flammability limit, and it required now 500 millijoules for ignition.

John Goglia: You answered two questions with that one. What was the lower flammability level of Jet A, was it 100 degrees? I'm thinking flash point. I'm sorry.

Joseph Kolly: The flash point of TWA 800 was approximately 114 degrees Fahrenheit. At altitude, that temperature would reduce to approximately 96 degrees Fahrenheit, because of the altitude effect.

John Goglia: Now early on, in part because of some of the work that was done at the Board here and some of the disclosures at Baltimore, there were some discussions about changing the fuel to a JP 5, which is an alternative. I wonder if you could explain to the Board the differences between JP 1 and JP 5 - some of disadvantages as well as the advantages.

Joseph Kolly: Essentially, while flammability point of view is what I can discuss here, Jet A fuel is what's used in the American commercial fleet, and it has a flash point of approximately 115 to 120 degrees average. It's specified as a minimum of 100. JP 5, which is used by the military on Navy aircraft, has a minimum flash point of 140 degrees. So if you're comparing the same temperatures of Jet A fuel and the same temperatures of JP 5, you will get less fuel vapors generated for the same temperature of JP 5, and therefore JP 5 can be raised to a higher temperature before it becomes flammable.

John Goglia: As we talked about the flammability limits, we didn't talk about the effects of changing the jet fuel on engine performance and, in particular, the ability to relight an engine in flight. Are there any negatives associated with changing the fuel in this area?

Bernard Loeb: That's one of the issues that needs to be considered. There are some concerns. You can't just simply change from one fuel to another without doing some additional work, and so it's one of the issues that needed to be considered. As I said, I don't know whether the FAA is still looking at that issue or not. I'm not aware that they are.

John Goglia: Okay. One question that I have gets to some of the modeling work, and it was something that I picked up in reading through this report and elsewhere. We talk a lot about the modeling, the work that we did, and the folks that reached out to help us do this modeling. When you read some of the things that are out in the public from other sources about this accident, they sometimes talk about tank explosions and reference work done mostly by people in the oil industry who have had tank explosions in the past and have some experience with that. And I took note of the fact that you mentioned that we reached out to some of those same people in this work. I wonder if you'd take a moment to explain to all of us what makes the tank on an airplane different from the tank that throws out fuel in these tank farms that we see around the country?

Joseph Kolly: Essentially we're talking about the type of fuel that it holds and also the fact that aircraft fuel tanks will operate in reduced atmospheres. So if you have an equivalent tank at sea level, it may in fact not be flammable. And if you took that same tank and went up 15 or 30 thousand feet, it may in fact be flammable. So from a flammability standpoint, it's not only the type of fuel and the temperature of fuel, but jet aircraft also have the environment in which the pressure is reduced. And that is one of the main factors in assessing flammability.

John Goglia: When I looked at the pictures today - which only refreshed my memory, since I've been inside these tanks - the mid-span ??? or baffles if you will that exist in the tanks do not exist in other large storage containers. Did those in fact have a role to play here in what we've observed?

Joseph Kolly: They were very important in the modeling aspect, as far as how fast the combustion process occurred, because our modeling showed that when ignition had occurred in one bay, it started to pressurize the remaining bays, essentially accelerating the combustion throughout any remaining bays. But the bays, as Mr. Wildey discussed, I believe are there primarily for holding the fuel in place on board the aircraft.

John Goglia: The infrastructural integrity of the airplane. We had a lot of talk around Jet A. As I travel around the country, and as I recall past life on airports, you see trucks that have Jet A on the side of them, you see trucks that have Jet A1 on the side of them. What are the differences?

Joseph Kolly: From a flammability standpoint, there's essentially no difference, as long as you compare similar flash point fuels. Jet A1 is commonly used in Europe and specifications are such that it has a slightly lower freezing point and it also has anti-static additives to it. But from a flammability standpoint, they're essentially identical.

John Goglia: And one additional question gets to the fuel on this airplane, and I don't know if you're the right person, but maybe Bernie could help here. The airplane was lost in the ocean. It stayed down there a length of time. So we didn't have the ability to get a sample of fuel from the center tank. Can you explain for a minute how we derived the fact that we had Jet A in the tank, that we didn't have a mixture of some other fuels, which you might get outside of the U.S. where quality control may not be the same, and to what lengths we went to determine that?

Bernard Loeb: We did get fuel samples from Athens, from the same source that the airplane had been fueled from.

John Goglia: And how long after the accident was that?

Bernard Loeb: We got samples fairly quickly. I don't remember the exact time. Bob, do you?

Robert Swaim: We did several things in this aspect. We were extremely concerned about what was the fuel in the airplane, so we got the Athens authorities - and I don't remember whether it was the next day or the day after that, but it was very soon after - they, I believe, actually locked a truck down for us and took samples. We also were concerned about what was in JFK. We locked that truck down and took samples. And then we went to the records of this aircraft, the accident aircraft, to follow the flight path backwards to know what airports they'd been at previously. And then, finally, we went to another airplane, another 747 that was flying the same route and took samples from that airplane so we'd have samples of how the fuel had weathered in flight. We did quite a bit of research into that.

John Goglia: And were we able to get any samples from the fuel filters from the engines? Were they contaminated?

Robert Swaim: I believe that would be Mr. Hookey.

Dr. Loeb: Mr. Hookey was our Powerplants Group Chairman.

Jim Hall: Mr. Hookey, get yourself a seat.

Bernard Loeb: Before Mr. Hookey starts, I think one thing that we need to make clear is that of course this airplane was not burning the fuel that was in the center wing tank. The fuel that it was burning was from JFK from where it had taken off. We were just interested in knowing what the composition of the fuel in the center wing tank was, and that was the leftover fuel from Athens. So I think that point needs to be made clear.

Jim Hookey: For the fuel, we did find fuel in all four engines, when we were dismantling the engines at Calverton. This was in the dead space in the fuel manifolds between the P and D (pressurizing and dump) valve and the fuel nozzles. It was collected and stored and we took it to a lab in New Jersey, Saybolt Lab, and it was tested and found to conform to Jet A. It was above the minimum flash point. I think it was about, as Joe Kolly said, 117 degrees flash point.

John Goglia: Thank you. One last question back to Mr. Kolly. Fuel comes from different sources, and sometimes it's called napta-based (?) as well as petroleum-based. Do we think that if any of this fuel was in fact derived from some of these alternative sources it would have had any sort of an impact different from what we've tested for?

Joseph Kolly: I can't say. I would defer that to the fuel experts.

John Goglia. Okay, thank you. I have no further questions.

Jim Hall: Member Black.

George Black: This issue has been brought up once or twice, it's not an issue, but the fact that this airplane had baffles in the center tanks, which are really the structural components of the center wing section of the airplane. Were those sections, those compartments and such, one of the reasons that we could not simulate the ignition location because of the quenching effect? Would you talk a little bit about that?

Joseph Kolly: The geometry of the center wing tank is very complex, as I showed, and simulating the combustion in that is a very difficult process. Not only do you have to simulate the chemical combustion, but also the fluid dynamics of fuel vapors being pressurized and passed from one bay to another. Combining all that is a very difficult task, and it leads to a level of uncertainty that you must carry on through the calculations. Quenching is a phenomenon in which a flame front, when passing - in this case, when passing from one bay through an opening into another bay - may actually extinguish. That is an area in which not a lot of research has been done and not a lot is known about it. Therefore, we had to carry a certain amount of uncertainty through our calculations as well to account for that, and that was one of the factors that led to our inability to determine the ignition location.

George Black: You mentioned early on, and I think it's worth mentioning it again, is basically what the Board and the people who helped us with this have done - basic research on fuel/air chemistry. Would you talk a little bit, or if you know, indeed, the answer - probably Bernie does because he's older - about why this had never been done before. Why we did not know these numbers.

Joseph Kolly: I'm not exactly sure why it hasn't been done. But I do know, in talking with all our researchers, primarily basic research is done on more simple or more elementary research elements. What I mean by that is if you did flammability research you'd essentially be doing it on single-component fuels. Jet A is so entirely complex, just the chemical make-up of it - and it has variation from season to season and from refinery to refinery - that researchers primarily didn't focus on the details that we needed in our investigation. And when we looked for those answers, looking in the open literature, there really was none that existed, only the more fundamental, single-component fuels.

Bernard Loeb: Member Black, let me try to elaborate a bit. Jet A fuel, as you're aware, is a fuel that's the cleanest in the aviation industry and their jet air transports. And the certification philosophy of the FAA was to assume that the vapors in the head space above the liquid fuel was in fact flammable. The assumption was that it was flammable. As a result of that, there was no drive, no impetus, to go ahead and do research on a very complex fuel, as Dr. Kolly has just indicated. Most of that research is on much less complex fuels, and Jet A is a very complex fuel. So there was simply no impetus. The assumption was, it's flammable, we're not going to worry about it, let it be flammable. And therefore when we went to study it, or when the argument started as to whether it was flammable and will it sustain combustion, and so forth, there was simply nothing in our literature to support that, and that's why had to do this research.

George Black: That's more the area I was looking for. One more just separate question. Joe, does the energy of an explosion resulting from a certain mixture at a certain temperature, does the temperature affect the energy available?

Joseph Kolly: Yes, because the temperature affects the amount of fuel that's vaporized, and so if you had more fuel you'd have more energy.

George Black: So knowing these temperatures, like we gained from the Evergreen airplane, we basically learned quite a bit about not only the chemistry but also the explosive effects of something once it did happen as a result of that testing.

Joseph Kolly: That's correct.

George Black: I've had pilots ask me this questions, and it's in one of the party submissions. If we do have these temperature gauges on airplanes - and I've actually flown on an airplane that had one, the question is what does the pilot do about it? Worry more? What's the result of knowing this, having this information known to the pilot right now? If we had the gauges out, what would he or she do?

Bernard Loeb: One of the more important uses of the gauge would be on the ground at the airport before you take off. If you know at that time that you have a temperature in that tank that is likely to be near or around the flammable limit or above it, you know very well that it is very likely that as you increase in altitude for a while it is going to definitely be above it, and so you don't take off under those conditions. And that's one of the main reasons that we recommended such an indicator.

George Black: So this would be a preventative measure. A pilot would refuse to take an airplane because...

Bernard Loeb: The issue is there a way to find a way to cool it and keep it below that temperature, and if what's happened is that it's a real hot day and you've been running the air conditioning packs on the ground, very quickly you're going to approach the flash point. And one of the things you might want to do is to shut down the air conditioning and let the thing cool down for a while. Now that presents some operational problems to the air carrier. On the other hand, it would tend to reduce the problem of the potential for an explosion.

George Black: Thank you.

Jim Hall: Member Carmody.

Carol Carmody: Thank you. I just have one question on inerting. I noticed from reading the report that on May 17 of this year the FAA sent a letter indicating that their research on ground-based inerting had reduced the potential for CWT flammability to 2 percent. I think the FAA also indicated in that letter that they were going to do more research and convene an ARAC (?) subgroup. That was three months ago. I'm wondering if any of you at the staff level have perhaps been in touch with the FAA and know what their progress and timing on this is. Thank you.

Bernard Loeb: Yes, we have been in touch with the FAA. In fact, we've spoken to them very recently. Dr. Kolly has been talking to his counterparts in Seattle. The FAA is going to attempt to do some tests in an airplane. There are tests they'd hoped would get underway in August. They have been postponed for a while longer, and the FAA does not now know when these tests may take place, but they're the first major step forward in this process, and we're looking forward to that very much.

Carol Carmody: Thank you.

Jim Hall: I believe Member Goglia has some additional questions, and there may be some additional questions from some of the other Members.

John Goglia: I have just one additional comment, Mr. Chairman. I have been very active in watching what the FAA has done in this area, in particular, the ground-based inerting process. I've been monitoring it very closely. It's made me happy to see the progress that the FAA has made. They haven't ignored what we uncovered and the work that they've done with us. In fact, just as recently as last week, the folks in Atlantic City, their research facility, in working with ground-based inerting, have also found it to be cost-effective, but have also found that they may be able to make it cost-effective to use it in flight. The spin-off of that would be to use that same nitrogen inerting system to combat fires through cargo bins as well. So there has been substantial progress in this area, and it is because of some of the work that the FAA has done - especially some of the folks in Atlantic City.

Jim Hall: Mr. Kolly, you have talked about flammability, and I want to acknowledge at this point the work of Dr. Merritt Birkey, who is now retired. One of the family members asked me about Dr. Birkey, and he of course made presentations at the hearing in Baltimore. He did quite a bit of work. Could you briefly tell us how you and Dr. Birkey overlapped in your work in this area?

Joseph Kolly: About two years ago, I began working with Dr. Birkey directly on this project. We worked hand in hand every day on it, and when Dr. Birkey retired at the beginning of this year, I completed the work.

Bernard Loeb: I'd just like to weigh in on this a little bit, since I've been here long enough to see them all, including Merritt, who did work for me, and Merritt had some unique experience and qualifications. We've been unable to find a single individual with all the experience that he had, including his fire and explosions capabilities, and he was for a long time the only person here who had any real knowledge in toxicology. He was also a physical chemist and possessed a lot of capabilities. But I do want to tell you this. I've been watching Dr. Kolly for the last two years, and he has picked up admirably. I was very concerned at one point about how we would continue this without Dr. Birkey here, but Dr. Kolly has filled in well, and I think has done very well today, and we look forward to him doing very well in the future for us.

Jim Hall: Dr. Kolly, I appreciate your presentation, and I thought it was important since Dr. Birkey had done a considerable amount of work that we acknowledge his work. Anytime you have a four-year investigation, there are obviously a lot of life experiences that all of us, as an organization and individually, go through during that four-year period. And one of them, of course, was Dr. Birkey's retirement, which he had actually planned, and had looked at retiring, and when we then had the ValuJet accident and TWA, I was very appreciative that he was willing to stay an extended period of time to try to assist with both of those important investigations.

We talked about the fuel, and I think we discussed briefly the different types of fuel, that the technology has changed. There's a slight difference, I guess, in the fuel that is used in Europe and in the United States, and that's well discussed in the report. There's also the issue of flammability, and of course the military fuel does have a higher flash point from some of the fuel they use, and of course they're very concerned, obviously, about this particular issue, because of the hostile environment they sometimes operate in. And I think that is well discussed in this report. I was pleased that Mr. Hookey came up and explained how we were able to secure some of the fuel of the aircraft for testing. I know we went through extensive work to get the fuel from Athens so we could compare the fuels, and I think we've discussed that, and that's documented in this report.

Is there anything else in regard to the fuel of the aircraft that we think needs to be put on the public record, other than obviously the discussions in the report?

Joseph Kolly: I don't think so, Mr. Chairman.

Jim Hall: I know that the industry and the FAA have looked at this subject, at our recommendation. There are a number of operational issues and economic issues that went into the decision. Now, let's talk about the air conditioning equipment. That obviously tends to heat the fuel in this model of aircraft. I guess that was looked at as early as the Air Force study I mentioned. Are there any recommendations that we have, Dr. Loeb, in trying to address this? As I understand it - all of us, I imagine, in this audience are familiar with sitting on the ground in an aircraft for an extended period time waiting for a variety of reasons that are sometimes and sometimes not reported. But we looked basically at the air conditioning system. What recommendations have we made in that area, and how does this model aircraft differ from any other model aircraft?

Bernard Loeb: The Boeing airplanes, excluding some of the McDonnell-Douglas portions of Boeing, have center wing tanks with the environmental conditioning systems underneath the center wing tanks. Airbus airplanes also generally have that. One of the differences in the Airbus airplanes is they do have some ventilation. The ventilation is not, as we understand it, there to try to keep the fuel in the fuel tank from getting hot, but nevertheless, it does help to cool the area directly under the center wing tank. Some of the McDonnell-Douglas airplanes do not have the air conditioning packs under or near a fuel tank. Some are in the nose area, some in the tail. So there are a variety of them. But most of the airplanes flying do have these air conditioning packs under center wing tanks. So it's a pretty pervasive issue.

We have not made a recommendation directly aimed at that. Our recommendation was to simply not allow airplanes to fly with a flammable mixture in the center wing tank, and of course one of the ways to prevent that would be to not have the air conditioning packs under the center wing tank of the airplane. As I mentioned earlier, the FAA does have a rulemaking out, an NPRM, in which they have proposed the possibility of, in the future, not allowing the air conditioning packs to be placed under a fuel tank like that unless there are means to cool and keep the heat from getting into the tank. We did not see right now a need to generate or develop a new recommendation aimed at that issue, but rather to wait and see and work with the FAA and see how things will turn out. But I think they're in agreement now that putting air conditioning packs under a fuel tank without means to keep the heat from getting into the fuel tank is not a good idea.

Jim Hall: I know that the FAA has made itself available to brief the family members, and I think has proactively been briefing the media in our country on the things that they have done, and I think that's quite appropriate. This was a tremendous tragedy. A lot of time and effort has gone into this investigation by this Board, as well as a lot of time and effort by the men and women of the Federal Aviation Administration in trying to assist us in the investigation as well as pursue our recommendations, in terms of trying to come to the one ultimate goal we're all interested in - that we do not have a recurrence of this type of event.

Specifically, as you mentioned, they are looking at mandating airplane designs to minimize the existence of flammable vapors. They are looking at ground-conditioned air - not having the air conditioning units run, but have another source of cooling. And they're, most importantly, looking at the whole issue of ground-inerting, which I assume we'll get into in more detail in future presentations, and is certainly significant.

Dr. Kolly, obviously you're not excused until the end of the two days, but this investigation has taken four years. I know the beginning of this was with Dr. Birkey, but you've done an able job of taking it over. Is there anything we have not covered for the public record that you think ought to be covered before we leave this subject and start looking more closely at the ignition sources and the wiring in Mr. Swaim's presentation?

Joseph Kolly: No, sir.

Jim Hall: Dr. Loeb?

Bernard Loeb: I don't believe there's anything else that we need to cover.

Jim Hall: Very well. Do any of the Members have anything else on this issue? If not, we will take a break and we will promptly put the gavel down at 1:30. We'll take a little extended break here so everybody can get in and out of the room and get some coffee. The gavel will come down again at 1:30.

Afternoon Session

Jim Hall: We will reconvene this meeting of the National Transportation Safety Board. The Board is in the process of reviewing the presentation from staff on Board notation concerning the in-flight break-up over the Atlantic Ocean of Boeing 747-131, near East Moriches, New York, July 17, 1996. We have basically had overview and presentation by Dr. Loeb, the on-scene accident investigation recovery was reviewed by Al Dickinson, the break-up sequence by Mr. Jim Wildey, and we just concluded a presentation on fuel tank flammability by Dr. Joseph Kolly. We will now turn to potential fuel tank ignition sources, a presentation by Mr. Robert Swaim. I'll ask Dr. Loeb if he would please introduce Mr. Swaim for his presentation.

Bernard Loeb: I'd be pleased to do so, Mr. Chairman. Bob Swaim has been working as a Systems Engineer for the Safety Board for 12 years. He's worked on some of the more significant investigations that we've worked on, including the ValuJet accident in Miami, and of course for the last four years he has been dedicated solely to TWA Flight 800. Prior to his employment with the Safety Board, he worked for approximately 20 years or more in various capacities in industry. He has a bachelor's degree in Industrial Education from the University of Maryland and has had extensive studies in aerospace engineering. Bob?

Robert Swaim: Thank you, Dr. Loeb. Good afternoon, Mr. Chairman and Members of the Board. The object of my presentation is to review the research relating to our search for the ignition source of the center wing tank explosion of the TWA 800 center wing fuel tank and the findings of the Systems Group. I'll also explain how staff concluded that the most likely source of ignition energy was a short circuit outside of the center wing fuel tank that was transferred into the tank through electrical wiring for the fuel quantity indication system, or FQIS. For simplicity, I'll usually refer to the FQIS and attached parts as the fuel gauge wiring.

In order for the fuel gauge wiring to have played a role in igniting the flammable vapors in the center wing fuel tank, two things must have occurred. First, energy much greater than that required for normal fuel gauge operation had to be transferred into the wires and carried into the tank. Second, that energy had to be released inside of the tank. All potential ignition sources were considered and pursued. These included: main landing gear or wheel well fire or explosions; fuel pump failures; an uncontained engine failure; fire migration through the vent system; auto-ignition or hot surface ignition; a malfunction involving the fuel quantity indication system, or again the fuel gauges; static electricity of lightening; a small explosive charge; a missile fragment; and we even considered a meteorite strike.

Evidence in this accident made almost all of these unlikely. Before I address the ignition scenario we found was the most likely, I would like to discuss a couple of the potential ignition sources that were deemed unlikely. Even though they were deemed unlikely in this case, each has the potential to ignite flammable vapors under different circumstances.

For example, an electrical or mechanical malfunction could have caused ignition at the scavenge fuel pump, and this was thoroughly considered. Although the pump itself was not recovered for examination, other evidence made this possibility seem unlikely. For example, the CVR (cockpit voice recorder) did not reveal a cockpit conversation regarding activation of the pump. The switch for this pump was found in the off position in the cockpit wreckage, and testing found that the flame protection features of the pump were effective.

In another example, a possible discharge of electricity from unbonded parts inside the center wing tank could not be ruled out as a possible ignition source in other circumstances. The reason was that the ungrounded metal parts were found in the tanks of other airplanes, and in the wreckage of TWA. However, in a laboratory setting, with conditions designed to mirror the center wing tank of TWA, the Systems Group was unable to generate sparks or even sufficient energy to be a hazard.

Another potential ignition source deemed unlikely in this case was energy from electromagnetic induction, which is sometimes referred to as interference, or EMI. The electromagnetic energy received from external emitters in the vicinity of the explosion, such as surface-based radar systems that included Navy ships, was not enough to achieve ignition. In addition, testing showed that the energy emitted from other airplane systems, or personal electronics, such as cell phones and laptop computers, did not cause an ignition threat.

I'll turn now to the ignition scenario that was found most likely. As the investigation proceeded, potential ignition sources unrelated to the fuel gauges became more unlikely. At the same time, the possibility of an electrical malfunction that possibly involved fuel gauge wiring became more evident.

The fuel gauge system in the center wing tank includes seven probes, a compensator, and a series of wires. The probes go from the bottom to the top of the tank and the compensator is a shorter probe in the lowest point of the tank that can be seen in this photo from an intact 747. Near the top of each probe is a terminal block that the wires attach to. I get asked if this electrical wiring inside the tank is exposed to the fuel and vapors - and the answer is that it is.

Wires connect the probes in each of the fuel tanks, including the center tank, through a connector on the rear spar. From the rear spar, a set of wires connect the tank components to the fuel gauge in the cockpit. Although the fuel gauge uses only a low level of power to operate, fuel gauge wires are routed in bundles with high-voltage wires that power aircraft systems. A short circuit involving high-powered wires and fuel measuring system wiring could inadvertently transfer this higher voltage into the lower-voltage fuel gauge wires. I misspoke on one thing - some tanks do have the connector on the forward spar.

This illustration explains how this energy transfer, or switching, could occur through a short circuit between the blue low-voltage wiring and the high-voltage wiring. If there is damage to each of the wires that exposes the copper core, a short circuit can occur, transferring high voltage into the low-voltage wire. As I will discuss later, such damaged wire insulation is common.

As this illustration shows, we further found that the wiring associated with almost any tank could reach the center wing fuel tank through ground refueling system wiring that is in the wings. Therefore, for a short circuit to have affected the center wing tank, it did not necessarily have to involve the center wing tank fuel gauge wiring. A short circuit through the wires for other tanks could have affected the center tank.

A short circuit can result in electrical arcs that melt the conductor of an electrical wire. The ends of the wires were found molten in two places that contained fuel gauge wires. The illustration shows one of these wires near the right wing route, where the right wing fuel gauge wires are routed with numerous other wire bundles. The second location was beneath the cabin and forward of the center tank. This location is almost beneath a galley that maintenance records reported had repetitive water leaks.

Each area contained fire damage, so it is not possible to determine if arcing took place or preceded the fuel tank explosion. Additionally, since the group was only able to identify about half of the original amount of fuel gauge wiring, there easily may have been other locations where a short circuit could have taken place.

The red fasteners in this photo show where a structural doubler had been added in the avionics compartment. Two and a half minutes before the end of the flight, the cockpit voice recorder captured a comment about a "crazy fuel flow" indication. This location was slightly over a foot from where the fuel gauge-related wires are in a common bundle with wires from the fuel flow indication system wires.

Repairs and structural modifications, such as the ones shown in this picture, require an extensive amount of drilling that would have created drill shavings for installation of the rivets. The Maintenance Records Group found that, rather than removing drill shavings, maintenance people sometimes used compressed air to simply blow shavings off the avionics in the compartment. The Systems Group also found drill shavings on a cabin floor beam fragment that had been about two inches from the route of the center wing tank fuel gauge wires.

The shavings raised the possibility that similar shavings had gotten into the wire bundle and damaged the wires. The number of repairs found throughout the wreckage indicated that there may have been numerous likely locations for a possible short circuit on TWA 800. After finding these shavings, the group began the investigation into the wiring of other airplanes. Those findings will be in a separate presentation of tomorrow.

The Safety Board contracted the Lectromechanical Design Company laboratory to evaluate the possibility that drill savings could create a short circuit. The upper photo shows a test setup in which a shaving was placed in the small bundle. One end of the wire bundle was then moved back and forth. Lectromec found that shavings could cut through the wiring insulation. When the shavings contacted a conductor, a short circuit was created from high-voltage to low-voltage wires. Much of the wire that was recovered from the TWA 800 wreckage had damaged insulation. The wire was also stiff and had cracks in the insulation, many of which exposed the copper core.

Chafes and gouges were very common in the wiring of other airplanes we inspected, as well as a smaller proportion of cracks. In other airplanes, it was extremely hard to tell if the cracks were only in the top insulation layer or if they penetrated through both insulation layers to the core. Again, the other airplanes will be discussed tomorrow in more depth, but the next photograph shows wiring along the floor of the cockpit behind the flight engineer's station in one of the 747s that the group inspected. This was a 100,000-hour airplane that had been recently retired, so it should be regarded as an extreme case. But note that the photo shows fluid stains near the route of fuel quantity wires that are routed down to the avionics compartment.

These are the wires that would pass the riveted doubler that I showed you previously in other repairs. Mixed into the same bundles are numerous 115-volt AC wires, and it is a few feet beneath this that Boeing drawings show 350-volt lighting wires join the bundle. Now you can see why investigators were concerned about the potential for water and fluids to reach the exposed conductors and create short circuits.

The Safety Board again contracted the Lectromec lab to perform short circuit tests with water and blue laboratory fluid, dripped onto the intentionally damaged wiring from the 1970 vintage 747. The test dripped fluids onto the small horizontal bundle of wires that are connected to the right clips at the lower left of this photograph. The fluid drips are coming from the red clip at the top of the picture and an aluminum shield is behind where various small flashes will take place between cuts that expose the copper. In the upper right is a round photo inset that gives a close-up of the bundle to show the two wires that have razor cuts in the insulation.

The following video contains three clips of about 10 seconds each. You'll see a clock in the upper right to show the elapsed time as the test progresses. The first clip shows development of very brief short circuits called scintillations. These appear as just small white sparkles on the surface of the wire. The test has now been running for nearly 11 minutes and the flashes are starting to appear. Notice that the time is now more than 20 minutes. The increasing intensity of the flashing will finally trip a circuit breaker to end this particular test. Lectromec performed similar tests that did last longer.

As you have just seen, fluid dripping onto damaged wires could create short circuits that lasted for many minutes. Longer than the flight of TWA 800 in fact. Further, the energy that was transferred during flashes like the one in this photograph was far in excess of what Dr. Kolly explained was necessary to ignite flammable tank vapors. This slide shows wiring that had been short-circuiting in flashes for five minutes. The scale at the bottom is in inches. Notice how little evidence is left from even the worst of these short circuits.

Now that you have seen the potential for a short circuit to have taken place, the next slide contains a list of related potential evidence that indicated to us that a short circuit immediately before the explosion may have affected the fuel gauge wiring in the TWA accident. Preceding the explosion, a cockpit crew member mentioned a "crazy fuel flow" indication. We previously noted that the fuel flow lighting and fuel gauge are routed together. There were drop-outs from the background noise of the cockpit voice recording slightly less than a second before the end of the recording.

The electromagnetic test that I mentioned earlier had found that the CVR background noise could be affected by simulated short circuits on co-routed wires. CVR wires were routed with lighting and left wing fuel gauge wires. The center wing tank fuel gauge indicated more than double the amount of fuel in the tank before take-off. Testing found that short circuits could change the display. We also learned from the Maintenance Records Group that in the last weeks of operation there were repetitive problems with the cabin lighting and ground refueling systems, which share wire routes with each other and the CVR.

We have now discussed several occasions at which energy could have entered the center wing fuel gauge wiring through a short circuit, and the probability that a short circuit should have been created by direct contact between conductors or by an indirect contact through metal shavings or fluid. Now I will present how this energy might have been released inside the tank.

First, it is important to understand that little was recovered from the center wing tank fuel gauge wires. The small amount of wiring attached to this connector was the only wiring we know to have been from the center wing fuel tank. Pre-accident damage to the exposed copper core of wires was found in wing tanks at TWA and also the fuel tanks of other airplanes. As I have already explained, voltage from exposed conductors can arc to other metal, such as other damaged wires, or to the airplane structure.

The group also found fine wire filaments caught in the pump inlet of another 747 during an inspection. These wires look like thin safety wire scraps. Scraps of wire could bridge the fuel probes or wiring to a ground and conduct electricity.

I will slow down for the interpreters. I will restart this slide.

Pre-accident damage that exposed the copper core of the wires was found in the wing fuel tanks of TWA 800 and also from the fuel tanks of other airplanes. As I have already explained, voltage from exposed conductors can arc to other metal, such as other damaged wires, or to the airplane structure.

The group also found fine wire filaments caught in the pump inlet of another 747 during inspection. These wire scraps look like thin safety wire. Scraps of wire could bridge fuel probes or wiring to a ground and conduct electricity. Anybody who has touched an incandescent light bulb can agree that heat is released by the application of power to thin pieces of wire. The FAA also mentioned filaments as a potential fuel system hazard in a proposed new regulation - NPRM 99-18.

The group found a thin film of gray deposits on fuel gauge parts from the accident airplane fuel tank. Similar deposits were also found on fuel gauge parts from the fuel tanks of other airplanes, such as the 747 terminal strip in this photo. The dark deposits that the arrows are pointing to are called sulfidation, created by an interaction of sulfur contaminants in the fuel with the silver-plated fuel gauge wires. Electrically, the deposits can be semi-conductive. We learned that application of power to sulfide deposits may ignite fuel vapors.

The Air Force Research Laboratory, or AFRL, had found that application of electricity to similar deposits in fuel probes from a military airplane could cause the deposits to burn off in a flash. The AFRL attached test wires to small areas of the deposits that you can see in this photograph. The deposits burned off as a loud pop and a bright flash when power was applied.

Due to these findings, the Safety Board recommended in April 1998 that the FAA further investigate sulfidation. As part of the FAA work, a researcher at the University of Dayton Research Institute working under a contract found a way to create sulfide deposits in the laboratory. Although the similarity of laboratory deposits to natural deposits is still being determined, the laboratory deposits have been used to ignite fuel vapors.

The staff learned that nickel-plated fuel gauge wires are now in use in some airplanes, and that the nickel will not lead to silver sulfide deposits. Staff believes that even though the chemistry of forming the deposits and electrical breakdown could be further researched, a potential hazard has been documented and an answer to this potential source of ignition has already been proven successful.

The draft report contains a recommendation pertaining to this. Throughout this report, that is, presentation, I have shown that there are two events needed for the ignition of the center wing fuel tank. The first event needed is that energy must enter the tank through the fuel gauge wires. I've discussed how short-circuiting, such as through drill shavings or fluid drops, could have short-circuited higher-than-intended energy onto the center wing tank fuel gauge wires. I've also shown evidence that such an electrical event may have taken place.

The second necessary event is the energy must be released inside of the tank. Arcing from exposed conductors, resistance heating, and power application to sulfide deposits are possible means that we found. Although we were not able to recover and examine sufficient evidence from the wreckage of TWA 800 to determine how the energy was released in this case, the summary of our testing and research is that the most likely initial ignition event was a short circuit outside of the center wing fuel tank that entered the tank through electrical wiring associated with the fuel quantity indication system.

Thank you for your attention, and I stand ready for your questions, sir.

Jim Hall: Thank you, Mr. Swaim. I know you have basically lived with this investigation for four years, and I appreciate that. Have there been other fuel tanks that have been ignited by short circuits in the fuel gauge wiring?

Robert Swaim: Yes, sir. Matter of fact, one was not too far from here in Eastern Maryland - Cambridge, Maryland - and it was a Navy C-130 in 1972.

Jim Hall: This accident airplane had 93,000 hours on it. Do you know whether there was any of the wiring to the fuel quantity indication system that had been replaced or whether there was a requirement for it to be replaced?

Robert Swaim: The wiring in the airplanes is replaced on condition of a failure. As far as whether we have a record of such replacement, we could not find any. We have no record to indicate that this was anything other than the original wiring installed in the airplane from the factory.

Jim Hall: The wiring showed us on the screen where the deposits were, would the type of wiring, if it was detected, be replaced, or should it be replaced?

Robert Swaim: The only time these wires would be replaced, fixed, or receive any attention is if they have a problem in service other than the general inspections of a compartment that the wiring is in, but, like I say, those are general compartment inspections.

Jim Hall: Is aircraft wiring different from that that you would have in your home or in your automobile?

Robert Swaim: Yes, sir. It is. The wiring in an airplane like this, or later, is a very thin walled wiring which has very thin insulation, and it's a bit tougher. Beyond that, Boeing estimated that they save 400 to 600 pounds per airplane by use of this thinner insulation wiring.

Jim Hall: The wiring that is put in planes now - maybe I'd like you to get into how it's bundled and put together. In fact, I wish we had brought - I don't know whether we have a photograph of a wire bundled.

Robert Swaim: If I may have slide 36, please. The wire bundles containing the fuel quantity indication system are shown in different locations in this set of photos as they proceed through the routing. Using the photo in the lower left, the red-bordered photograph, the fuel quantity indication system wire would be where the dotted blue line is in a larger bundle. That bundle is about as big as a person's wrist or a woman's ankle. It contains many hundreds of wires - it depends on the person's wrist - and it contains a variety of types of power or ground. As mentioned, we could have typically 115-volt power running in the same bundle with single wires. Wires are tied into smaller bundles and the smaller bundles are tied into large bundles as you see here.

Jim Hall: At the time that this aircraft was manufactured, was there a separation of the fuel quantity indication wires from other wires that might carry more power?

Robert Swaim: No. In fact, we did find a document from Boeing. They designed an engineering type of document that stated specifically that the fuel quantity indication system wires required no separation other than for EMI.

Jim Hall: Has that been changed, or is that being addressed or looked at by the FAA or Boeing since this accident?

Robert Swaim: Yes, sir. We made two safety recommendations in April 1998. One was for separation of the fuel quantity wires from other wiring. The other recommendation was to put a surge suppressor, also called a transient suppression device or TSD, at the wing spar where the wires go into the tank. It was a belt-and-suspenders type of approach to make actually sure we don't get power into fuel tanks from an arcing adjacent wire - in the case of the separation - or from another source, such as in an indicator or another attached electrical device.

Jim Hall: And what's the status of those recommendations?

Robert Swaim: The FAA has now issued airworthiness directives, ADs, requiring the protection of the fuel gauge wiresby separation in both the 747 and the 737 fleet. The other thing that I have in tomorrow's presentation, but can bring it up now - I'm really pleased to be able to say that we now have transient suppression devices designed, certificated, and available for airlines that they can use at the wing spars to keep the energy out of the fuel tank.

Jim Hall: I think we get into the aging wiring tomorrow in more detail, so I won't venture into that area right now. On EMI, under what conditions would EMI be a problem?

Robert Swaim: For that question, I would like to defer to the person who headed the subgroup on the electromagnetic testing we did, and that would be Mr. Scott Warren.

Jim Hall: I'm looking at page 502, Scott, lines 1 and 2.

Scott Warren: The reason that we used the language in the report was because EMI and how it affects an airplane is very much a distance-related phenomenon. The closer you are to the source or to a transmitter of an electromagnetic wave, the more power would be absorbed into the airplane. In this case, in TWA 800, we were several miles away from the transmitters that could have been a potential problem, if an airplane was substantially closer to a transmitter than TWA 800 was, that could be a problem for other aircraft.

Jim Hall: When we say on page 555, line 12-14, or in findings, do we need to have any additional research in this area? The reason I'm bringing this up, of course, everybody's aware of all the new equipment that's being invented all the time - all the things people carry on to airplanes, all the things that are transmitted through the airwaves - and I would certainly want to be sure that we don't end up in a situation where we're trying to track something down after an event. Maybe this is something we could be looking at before that were to happen.

Scott Warren: Absolutely. We looked at several different mechanisms for electromagnetic interference in this accident, and in all those cases and in all the research programs that we re-used - for example with Patuxent River and with NASA Langley and with the Joint Spectrum Center in Annapolis, Maryland. All those research programs concluded that there was substantial margin between the levels that we had in this accident and what it would have taken to ignite the fuel/air mixture in the center wing tank. So no, I do not think we need additional research in EMI as far as this accident investigation.

Jim Hall: I think Member Goglia wants to interject on EMI, since we're on the subject.

John Goglia: What about more modern airplanes with much greater use of composites? And the reason why, for the audience, is that metal airplanes tend to insulate the interior wires from some of this EMI phenomenon, but composite airplanes don't offer us that added protection.

Scott Warren: That's absolutely true that the composite materials will not shield as well as the aluminum that we have in the aircraft now. What I would suggest to that is that, again, you've got a distance factor to take into account, and you would have to be extremely close to your source, even without the shielding from the fuselage, in order to create a problem for this aircraft as far as igniting fuel vapors.

John Goglia: The FAA, though the ARAC (?) process, has been working on EMI for I bet close to 10 years. Did we look at the work that was done there? And have they reached any outcomes and what effect that had on any of the work that you did?

Scott Warren: The work that's been done in the past on EMI is really oriented more with interference with aircraft systems, such as navigation, communication, and the computers on board the airplane. Our work was entirely different from that, and it was really the first in its category that I could find where it was solely oriented toward igniting the fuel air vapors in the tank. The two efforts were really very different.

John Goglia: Would it be safe to say that the first sign of anything in an airplane of something happening because of EMI would be in those other systems before it would be seen in induced voltages?

Scott Warren: I think that would be a very good statement. In general, the other systems are much more susceptible to EMI than a fuel tank would be. That's correct.

John Goglia: Thank you, Mr. Warren.

Jim Hall: Again, you're saying a significantly stronger electromagnetic environment (on page 502) than was present at the time of TWA Flight 800 might present an ignition hazard. I'm trying understand - is that a correct statement? Why have we made that statement? And is there anything else that we need to do if there other environments that would present an ignition hazard, or is that issue being addressed? Or am I just thick on this particular subject?

Scott Warren: No, it's certainly not a straightforward subject at all. I think the issue is being addressed by the FAA in their continuing HIRF (high-intensity radiated fields) work that the FAA has as an ongoing project. That effort, I think, will be sufficient to examine this problem in the future.

Jim Hall: We don't feel that a Board recommendation is needed, then?

Scott Warren: That's correct.

Jim Hall: I've got some other issues, but I want to defer to my other Members, and if those aren't covered, then I'll try to get to them later. Member Hammerschmidt.

John Hammerschmidt: Thank you. Just as was the case with Dr. Kolly's excellent presentation, as I was sitting here, I was ticking off all my questions as he was answering them in his presentation. The same is true for Dr., or Mr., Swaim's presentation. I do have three miscellaneous questions.

Jim Hall: Mr. Swaim has been so busy he hasn't had time to get his doctoral degree yet.

John Hammerschmidt: Let me ask you about something you mentioned in your presentation. We reference, of course, in the factual portion of the report and in the CVR transcript, the captain stating, at 20:29 and 15 seconds, "Look at that crazy fuel flow indicator there on number four. See that?" "See that" is in the form of a question - so we have it in the report. Again, what significance do you attach to that statement in terms of when it occurred and what we know about the wiring routing. I know you went into that, but is there anything more you can elaborate to help us understand that statement better?

Robert Swaim: We have a graphic of the wiring routes in the airplane. The wire coming up from the center wing tank comes over the top of the cockpit, and this view is looking at the right side of the nose of the airplane. The wire coming from the tank to the flight engineer's station is shown as a light blue wire, and it attaches behind the flight engineer's panel to wires going to other users of the signals from the fuel tank. The lower portion shown is located right behind the flight engineer's panel where the wire goes to another gauge called the fuel totalizer. From a splice right in that area, a wire goes down behind the nose landing gear to an avionics box, located in the avionics compartment.

This is where you have the fuel flow box, which is an avionics computer, right next to the ground refueling computer, The ground refueling system, as you might remember, gave a problem on this airplane before this flight. What is shown here as faded red coming into that box from the left is a major wiring raceway coming from the right wing. That brings in the signals from the right most engine with the fuel flow. From the boxes up to the flight engineer's station in the bright red wire path that I just showed is where you have the fuel flow and the fuel quantity wires actually tied into a common wire bundle. This can be seen in the photo 35 also, where they're running up between the windows and up into the floor of the flight engineer's station. The upper photo is where the wires go up into the flight compartment, and you can see that those wires are actually tied together into a common route. So if you have something affect one one wire bundle, it may affect the others.

Jim Hall: I wanted to ask John if it would be okay to ask a question on this. How many wires are in a bundle? When we look at that bundle, how many different wires are you talking about?

Robert Swaim: It depends on several factors - not only the thickness of the wire, but the flatness of the bundle. A bundle can be a three-wire set, just three wires alone, or it literally can be as thick as your wrist. And when we have many of those thick wire bundles routed together, they're separated slightly in what is called a raceway.

John Hammerschmidt: Just a couple of other questions pertaining to proposed findings. In the findings, we have a proposed finding number 10 that says that "Boeing's design practice of permitting parts less than 3 inches in any direction to be electrically unbonded may not provide adequate protection against potential ignition hazards as a result of static electricity generated by lightning and other high-energy discharges." Would you please explain the rationale behind that proposed conclusion or finding?

Robert Swaim: Certainly. That really comes out of a couple of places in research and previous documents. After a pair of 1970 Boeing 727 static ignition fuel tank explosions in Minneapolis that happened as a result of ground refueling, there was some testing done, and it was found that even small clamps could achieve some very, very high levels of voltage and energy that could lead to a static discharge.

However, in design specifications, we found that, to this day, small metal parts are allowed, such as the same type of clamps. So there is a disparity there, and what we're trying to do is see if the previous test results would be more appropriate and these clamps should be removed. This is also, by the way, mentioned in the lightning protection handbooks of both the FAA and of NASA - or if we should continue to allow small metal parts to be ungrounded in the fuel tanks. Did I answer your question directly?

John Hammerschmidt: That's good. And my other question goes to our proposed finding number 18. I might as well just read it. It says that "the ignition energy for the center wing tank explosion most likely entered the center wing tank through the fuel quantity indication system wiring and, although it is possible that the release of ignition energy inside the center wing tank was facilitated by the existence of silver sulfide deposits on a fuel quantity indication system component, neither the energy release mechanism nor the location of ignition inside the center wing tank could be determined from the available evidence." I know we've basically explained that in your presentation, but is there anything else, Dr. Loeb, we would wish to add to the public record about how we've arrived at that conclusion?

Bernard Loeb: Of course, the way we would have arrived at the conclusion was covered by Bob Swaim in his presentation and discussions as he went through each individual potential ignition source thoroughly and carefully, spending a lot of time looking at each one. Bob and some other folks - Dr. Joseph Leonard, who had been with the Naval Research Laboratory, and people at Wright-Patterson Air Force Base at the Air Force Research Laboratories - worked diligently for months and months on looking at static electricity.

We looked at the possibility of meteorites, of missile fragmentation, small explosive charges, and we were able to determine that the vast majority of them were highly unlikely or very unlikely. And we were not able to reach that same conclusion with the potential for the short circuit from the higher-voltage wire into the FQIS system and a subsequent discharge of energy inside the tank. We know that there are a number of possible ways in which that could happen. And so that's how that we arrived at that.

Now, I think it needs to be well understood that it is not surprising that we cannot necessarily pinpoint the ignition source because in an explosion of this nature - and the Board has dealt with many of them over the years, as you well know Member Hammerschmidt, you've been there and heard many of them, especially those in the Marine environment - the accident itself, the explosion itself, obliterates, eliminates much of the physical evidence that you need.

And so it's not surprising that we cannot pinpoint it. But there's clear evidence to suggest that a short circuit from a higher-voltage system in the ship's wiring to the FQIS (or fuel quantity information system) wiring and subsequent discharge of energy in the fuel tank is the most likely. And we indicate that the silver sulfide deposits were a very real possibility for the discharge inside the tank, but cannot say that to the level of probability that we would normally want to do so.

One thing I would like to add is, all of these other possibilities could be ignition sources in other cases, which I think reinforces the absolute need to have this dual approach - and that is, one, to continue to try to eliminate ignition sources, but more importantly, to eliminate or reduce to the maximum extent possible the flammability. The very fact that there are so many possible ignition sources indicates just how difficult it is to do the one side of it and not the other side and believe you're going to be successful. So it speaks volumes toward the need to eliminate the flammability or reduce it to the maximum extent possible.

John Hammerschmidt:: Thank you very much, Dr. Loeb. And that's all I have.

Jim Hall: Member Goglia.

John Goglia: Thank you, Mr. Chairman. Just a couple of questions. I'd like to go back to wire bundles for a second. As I was listening to everybody's questions and some of the comments, there may be a misunderstanding of the purposes of wire bundles. Mr. Swaim, do you know if it's a certification requirement under the FARs to have wires protected from abrasion from surrounding structure?

Robert Swaim: The FARs functionally work through advisory circulars and the advisory circulars do address abrasion or protection against abrasion. We will return to that in the Aging Systems presentation of tomorrow, where we did find some. But it is addressed and it is required. It's something that every basic A&P learns in school. I think it's necessary to emphasize that in a large airplane you have multiple systems, and they're segregated and separated to some extent, but fuel quantity wires, like I said previously, don't have that one requirement in these older airplanes.

Bernard Loeb: In fact, we are proposing a recommendation in the draft report for the FAA to take a second look at the issue of separation, and there will be more discussion on that tomorrow in Bob's presentation.

John Goglia: However, even with separation, we will continue to have wire bundles, we will still have them traveling with standoffs, to keep them away from fuselages, from structure, from abrasion, from harm's way, hopefully. And that's the main purpose for all those wire bundles.

Bernard Loeb: Absolutely. We're not going to be able to do away with wire bundles, but the FAA's looking at the issue and looking at other inspection techniques and so forth, and there will be more, again, said on this tomorrow.

John Goglia: Mr. Swaim, a couple of questions on the fuel quantity indication system. What's the voltage in the system?

Robert Swaim: The fuel quantity indication system in a 74 uses a variety of voltages for different purposes. The internal lighting is one thing. The compensator in the probes in the tank, the maximum they reach is a little over 25 volts. But it's a very, very low amperage.

John Goglia: And the fuel quantity indicators in the refueling station in the cockpit - what is their voltage, what drives them?

Robert Swaim: They receive 115 volt ac power from the aircraft, and they have an internal transformer to step down the voltage to what is required for each of the functional uses in the system. The transformer has different leads that go to the different subcomponents.

John Goglia: And did we recover those indicators?

Robert Swaim: We recovered one. We recovered the one from the cockpit. Yes.

John Goglia: And did we do any testing to that indicator?

Robert Swaim: Yes, sir. We found that the center wing tank indicator - actually, after we examined it for damage and documented its condition - we used the electrical portions of the gauge, replaced the parts that were mechanically broken during impact, and, after those parts had been replaced, most of the gauge was still original and it did work.

John Goglia: And we found no evidence of short circuit inside the indicator, or someplace where we could get cross voltage?

Robert Swaim: In the indicator, no we did not. We really cannot determine the condition of the connector from behind the indicator.

John Goglia: And the wire bundle that was showing in both blue and red as it comes up through the system, what's the voltage there? Is that 25 volts as well?

Robert Swaim: That wire bundle I know contains 115 volts and the possibility of higher voltage exists, but I know where that bundle comes up under the cockpit floor. We also have the cabin lights come into that bundle and that's where 350 volts is in the bundle.

John Goglia: And the fuel quantity indication system wires that are in that bundle, what's the voltage there? Is that 115 volts at that point?

Robert Swaim: No, they would be low-voltage, low-current.

John Goglia: And you mentioned rather pointedly that the totalizer wires run together. What's the voltage from the fuel totalizer system?

Robert Swaim: The totalizer takes the signals from the fuel tank indicator, so it would be the same as the center wing tank gauge. But again, like the center wing tank gauge, it utilizes a single connector in the back of the gauge, and that gauge connector has both the low-voltage, low-current fuel tank wires and the 115-volt power wire.

John Goglia: Many of my questions deal with your presentation of tomorrow. So I will hold them all until then, although I would like to correct one oversight that I made earlier when I commented and commended the FAA tech center in New Jersey for the work that they've done. One of our staff people correctly pointed out to me that there's also some fine work done by other folks in the FAA here in headquarters and in Seattle and other places. So my remarks were a little too narrow in singling out the FAA folks in Atlantic City. So for the FAA folks that are here, please accept my apologies, and that's all, Mr. Chairman..

Jim Hall: Member Black.

George Black: This analysis of the last tenth of a second or so of the cockpit voice recorder. Are we going to discuss that tomorrow in connection with this, or is that additional information?

Robert Swaim: The cockpit voice recorder drop-outs were at about seven-tenths of a second from the end of the recording, a little less than one second. Tomorrow's presentation is more on just generic aging and other airplanes that we looked at.

George Black: You consider that to be further evidence of something going on in the electrical system, do you not?

Robert Swaim: Potentially, potentially. The cockpit voice recorder wires are routed in common clamps and common bundles with fuel quantity wires going to the left wing of the airplane. The left wing fuel quantity wires and the lighting wires are also routed with the fuel flow wires on the right side. So it is potentially an indicator.

George Black: The report goes into this in some depth, and I think it adds a little bit more to this circumstantial case, which you're building for the source of ignition. I thought we should bring it out - since Mr. Cash worked so hard on it.

Bernard Loeb: You're speaking to the 400 hertz and the higher drop-outs. Bob, why don't you expand on that just a bit?

Robert Swaim: We found that the electrical power system of the 747 is, in engineering terms, a fairly noisy system. The cockpit voice recorder, for example, when it records the sounds in the cockpit there's a background humming noise. A lot of this humming noise comes from other electrical items that have wires that are cobundled, actually routed with the wires to the cockpit voice recorder. The captain's channel of the cockpit voice recorder, and none of the other channels in the last second, had two periods where all of a sudden we had no noise.

We still had the basic electrical signal of 400 hertz and a little bit of some of the noise in the 800 harmonic, but almost all of that energy went somewhere. So when I say it's a potential indication of something happening in the electrical system, like I say, that energy had to go somewhere not to be transmitting and received and captured by the cockpit voice recorder. So that's one of our thoughts - that maybe we did have a short circuit in this airplane.

Jim Hall: Member Carmody.

Carol Carmody: Mr. Swaim, with respect to the silver sulfide deposits, I notice we're asking the FAA to take corrective actions to eliminate them. What would those actions be? How would you clean them, or what's the way that you would eliminate the silver sulfide deposits?

Robert Swaim: We just had a discussion about this. So far, the ability to mechanically clean the sulfide deposits or chemically clean the sulfide deposits has been not totally successful, and once they are removed, they come back very rapidly. One possibility that the airlines have found is when they get an indication air problem, they replace the harnesses, the wiring in the fuel tank.

That obviously gets very expensive and it's a major task. However, after the C5 Galaxy fuel quantity indication errors of 1990, and in that time frame generally, and after Boeing found in the late 1980s and early 1990s that they were having service problems from airlines where the fuel quantity indication system was chronically a problem, chronically giving errors, Boeing and B.F. Goodrich, separately and unbeknownst to the other, developed essentially the same solution. They took the silver component out of the fuel tanks, and rather than using silver-plated wiring in fuel tanks, they used nickel-plated wiring. It is harder to attach in manufacturing and repairs, but by removing the silver, you remove the possibility of developing these deposits.

There is another possibility that Boeing has been exploring, and that is to simply fill the ends of all the connectors and termination in all the wiring in the fuel tank and, like I say, that's a process they're still working on and exploring.

Carol Carmody: Thank you. On wiring inspection, other than the manual and visual inspection, what inspection methods for wiring exist today?

Robert Swaim: This is a little more into tomorrow's presentation. Currently, the primary inspection of wiring is visual, and it's on condition. We also have the detailed inspections of compartments, where you go through everything that's in a compartment. Now when I say "a compartment," the center wing fuel tank in a 747 is regarded as one compartment, even though it's essentially seven small rooms that are collectively the size of a two-car garage. So it's a big compartment. But we do have that inspection requirement, to go through the compartment and inspect everything and ascertain its connection and its proper installation. The industry is now looking at automated test equipment. Automated test equipment has been found to be much more effective than visual inspection. Lectromec did some tests for the Navy that I'm aware of, where they found several times more defects electrically than they would visually. I'm not saying we need to discard or discount the visual inspections - there's a purpose for them and a place for them - but we're finding that the automated tests are quickly gaining and able to do things that the visual inspections cannot do, such as inspect wires in the middle of one of these bundles, which is thick as a person's wrist.

Carol Carmody: Thank you.

Jim Hall: Mr. Swaim, I would imagine that every Board Member and all of the families probably read the cockpit voice recorder transcript numerous times, trying to understand what happened. I keep coming back to this "crazy fuel flow indicator," trying to understand - where was that in the sequence, how long before the event? Less than 10 minutes before the event. Yet the report states that this was not an uncommon occurrence. If the lights were flickering in my house that often and there were that many reports of it, why does staff believe that was not something that should have been of concern?

Robert Swaim: In this case, you're absolutely right. In a 747 or a 737 of the older vintage, they are electromechanical devices, they have a part that rotates out at the engine and the system captures the rotational speed and reports it to the instrument in the cockpit as how fast the fuel is flowing. In any electromechanical device you're more prone to have failures than in what we now have in digital devices. However, what caught our eye in this was the timing of it. The same with the CVR drop-outs that I mentioned with Member Black. Yes, it and the CVR drop-outs could have been flukes on their own. But when you start to take all these individual straws together, they add up to a more likely probability that you have something in common between them - especially since they're all just within the last couple of minutes of flight.

George Black: Mr. Chairman, could I just add something to this?

Jim Hall: Surely.

George Black: Another thing struck me as I went back that wasn't real clear in the report. The captain who made this comment, who was getting his initial operating experience, even though he was just being checked out as a captain, had put 4 or 5 thousand hours in this airplane, as a copilot, a first officer. And the very fact that he would comment - they were beginning a climb, and the pilot naturally looks to the engine instruments, especially on an older airplane, and it would have been natural for his attention to have been drawn, because they were adding climb power, or they had been climbing earlier, and you watch the engine instruments. The fact that with that many hours in the airplane he should raise that issue also got my attention. Just another bit of information.

Robert Swaim: You're absolutely right. There was the fuel flow indicator - and it's in our draft report - having been changed. But again, it was a matter of a year or two before, it wasn't two and a half minutes before. And that's what I say about coincidence in time.

Jim Hall: I am again referring to Appendix E, which is the accident airplane's maintenance records, in which there were 25 fuel system-related maintenance write-ups two years before the accident.

Robert Swaim: I believe that's the correct number.

Jim Hall: Is that common to the fleet? Is that related to the age of the aircraft? Is that not an indication of any type of problem that should have been explored further?

Robert Swaim: There were several questions there, sir.

Jim Hall: I'm not as technical as you are, so I have to kind of throw a huge net out and hope that you'll pick the important part out of it.

Robert Swaim: These airplanes do have anomalies, and we have found that as they increase in age, the anomalies increase. Something for tomorrow, in the Aging Aircraft presentation, Dr. Chris Smith of the FAA's technical center has done a study of service difficulty reports and found that in the aging aircraft, a sample of five years of aged aircraft versus new aircraft, the aged aircraft have a higher percentage of anomalies. So, to sit here and say, this is typical for a 747, we'd have to study the 747 fleet of that age.

Now the other question is, is it typical for an airplane in general? This was a high-time airplane, and I just experience-wise would expect more write-ups on a higher-time airplane, and Dr. Smith's work seems to verify that.

Jim Hall: We're going to get into that more tomorrow because I really want to get into that issue but I don't want to preclude that discussion tomorrow. Member Goglia wanted to contribute on this as well.

John Goglia: I don't believe that on the surface just the fluctuating indication system would cause someone a great amount of concern given the number of times that we see that. I know from my past experience that it was just a very common event across all types of aircraft, and you were correct in your statement saying that at some point, usually during a scheduled heavy maintenance visit, which occurs on a year and a half cycle probably, that we would go in and replace the harnesses so if there was enough of these repetitive complaints about fuel quantity indicators. You mentioned the fuel flow system. Oftentimes when the engines would have fuel flow or other indication problems, those problems were corrected because we would replace the engines and the wiring that gets so much vibration located on the engines and those connectors that get all that vibration on engine-mounted components are then replaced at the same time as part of the package. So those problems tend to go away but from an operating point of view, from the pilot's point of view, and from the people that are responsible for dispatching this airplane every time it comes through a maintenance station, that would not raise any eyebrows or any serious concern on anybody's part.

Robert Swaim: That one by itself, you're absolutely right.

One of the things that is in the report along that line is that in the last, I believe it's six weeks of operation, this airplane all of a sudden had a whole lot - probably 10 or 12 - ground refueling write-ups. Ground refueling in these airplanes is common over time but not to have such a cluster, and especially to have that cluster right before the accident. That was eye-catching.

Jim Hall: Well, I raised all those issues because, as you know, in-flight fires continue to be a problem. The Board just launched a go-team to an Air Tran electrical fire event that I believe is described in an article in one of the morning newspapers today. I've been on this Board six years and I have seen a number of accidents in which we've had some sort of electrical problem or fire problem, and are we doing the type of job we should be?

Obviously two issues - the issue tomorrow of aging aircraft because obviously the flying public in our country has no idea how old an aircraft is before we get on it, and secondly with just the fleet, period. That's why the things that make common sense to me are the questions I try to ask and I still say that as long as we end up with these in-flight fires of unknown origins, we need to continue. I know our folks at the Canadian Safety Board, and I'm grateful we have five of their representatives here observing these proceedings, are working very diligently on the SwissAir accident. This might be a good time again to acknowledge and thank them for all of their assistance with us on this aircraft.

Now, back to sulfide deposits. You discussed in response to Member Carmody's question about B.F. Goodrich and Boeing's work in that area, and I'm referring now, Members, to page 304. One of the challenges all of us have as Members is trying to read everything to prepare ourselves for these meetings, and of course the International Aviation Industry just put out an aircraft fuel system safety program report that we'll probably get into tomorrow. But you say B.F. Goodrich started on the basis of Air Force Research Laboratory? Let me just read this, and you explain.

"According to B.F. Goodrich representatives at the November 9, 1999, meeting, on the basis of AFRL..." which is what, Bob?

Robert Swaim: Air Force Research Laboratory.

Jim Hall: "...Air Force Laboratory findings of the early 1990s, B.F. Goodrich had developed a means of improving the accuracy and reliability of fuel quantity indication system wiring used in military applications through design changes that began in 1993 and involved the use of nickel-plated wire, gold-plated ring connectors, sealant and shrink tubes, and separate inner and outer layers of shrink tube. The B.F. Goodrich representative stated that previously reported inaccuracies in the fuel quantity indication system had resulted largely from current leakages through the sulfide deposits, and the design change resulted in a large reduction in reported FQIS problems." And you go on further to say, "According to Boeing personnel, Boeing uses nickel-plated now instead of silver-plated wiring in its newly manufactured 777 and 737 NG aircraft. However, in a letter dated December 7, 1999, Boeing wrote, `Overall, the wholesale replacement of FQIS bundles in the tank is not recommended.'"

My question here is, did the Air Force retrofit their fleet? Do we know - based on the work B.F. Goodrich did?

Robert Swaim: I believe it was through attrition. It wasn't a sweep-the-fleet campaign. I don't remember that positively, but I'm pretty sure it was through attrition.

Jim Hall: I would appreciate it if you could find that information out and add it to the factual report.

Robert Swaim: We will do so.

Jim Hall: I've got other things, but I think I'm going to drift into tomorrow's discussion. Do other Board Members - Mr. Hammerschmidt, do you have any other items to pursue?

John Hammerschmidt: Not at this point. Thank you.

Jim Hall: Member Goglia?

John Goglia: Yes, sir, I do. Just one thought, quickly. You mentioned about the short circuit because of the findings on the voice recorder and that energy having to go someplace else. One would, in the simplest terms, expect the circuit breakers to do their job, if there was, in fact, a sudden drain on some circuit and interrupt that circuit. That's what they're designed to do. But I know that in much of the work that you have done, you have found that that is not always the case. I wonder if you would talk about that for a minute, and maybe talk a little bit about emerging technology for circuit breakers, such as the arc-fault interrupter.

Robert Swaim: Certainly, sir. We did a couple of types of research into circuit breakers and how long it takes for the energy going through a wire to trip a circuit breaker. One was a historical basis and another path that we looked was the specifications. Now, if you look at the specifications for the typical circuit breaker on a 747, it takes at least - for a brand new breaker, performing to specifications - it takes at least 60 milliseconds and I believe it's 1,000 percent over current to trip that fast.

That's because what happens in a circuit breaker, in simple terms, is the electricity going through a wire heats up a little heater, and when the heater heats up, it actually bends and opens the mechanical part of the circuit breaker, and that's how it functions to disconnect the power. It takes a certain amount of time for that mechanical movement and heating to take place. That one flash that I put up the illustration of - from Lectromec and the wet short testing - that flash transferred eight kilowatts in about one millisecond.

Now in human terms, a night light for a bathroom or a child's room is about 10 watts. That flash was 8,000 watts. So there's a lot of power that can go through that circuit breaker before it has the ability to stop the circuit, to open the circuit. There are new technologies being developed, they're actually in use on the ground for houses, called arc-fault circuit interrupters. There is some information presented in the draft report from Underwriter's Laboratory concerning these, and they're now being developed for aircraft use. The FAA, the military, Boeing - they're all working very hard to develop this technology. And what it will do is interrupt the circuit in a matter of milliseconds rather than letting a heater warm up, which, actually, at a lower energy short circuit, can take minutes. This will happen in milliseconds rather than minutes.

John Goglia: Thank you.

Jim Hall: Member Black and Member Carmody? I apologize. I've got so many books moving up here. I'm trying to find the one on the submissions.

I want to return to the submission of the International Association of Machinists and Aerospace Workers. They conclude that the existing wiring recovered from the Flight 800 wreckage does not exhibit any evidence or improper maintenance or any malfunction that can lead to a spark or other discrepancy. The examination indicates that the wiring was airworthy and safe for flight. Is that the wiring that was recovered, that you showed us, that they're referring to?

Robert Swaim: I understand that is in their submission. I do have a slide in tomorrow morning's presentation that addresses that directly. We did find some maintenance problems in this airplane. In a problem of this size and this age, I'd be surprised not to. And at the Baltimore public hearing we showed some photographs of some wiring problems, and so I'm not sure that I would agree with that statement.

John Goglia: It actually goes one step further, Bob. Even in an airplane of this age, not all wiring is created equal. In different places wiring looks different. There are sections of the airplane where you can look at 25-year-old wire and it looks every bit as good as brand new wire, and there are other sections - and I can take you right out here to National Airport and get airplanes that are only a few years old, and they're certain areas where most maintenance people know where to go, and I know where some of them are from my past experience - and I can go and show you airplanes, again, that are not very old, but the wiring looks pretty thick.

Robert Swaim: We definitely will be hitting this very hard in tomorrow morning's presentation.

Jim Hall: What I'm getting on is that I can't understand basically the conclusions and how they match with their recommendations - which we'll get into tomorrow. Their recommendations are that "we need improved inspection and maintenance practices to reflect current and future industry standards. We need the adoption of aviation electrical, electronic standards throughout the aviation industry. We need inspection standards established in concert with maintenance representatives from officially recognized maintenance groups. We need better and more concise communication between the engineering field and the maintenance field. This must be exercised to the floor where the maintenance is accomplished. We need realistic time periods to accomplish maintenance and inspection procedures in the field and in the hangar. Time studies produced by manufacturers are unrealistic in that all environments are not considered. We need better access to items requiring inspection and maintenance once the aircraft has been put in service." In addition, they say, we need a further study of aviation fuels and a look at the destructive properties of currently used hydraulic fluids.

I don't find the support for that in some of their findings, but I certainly understand and welcome their recommendations.

Now, what we've been on is ignition sources, right Dr. Loeb? We've been trying to wander into tomorrow's presentation, and I think all the Members are guilty of that. What I'm looking at is trying to find that list you put up that listed all the ignition sources. Did we go to that slide? There were two slides. There's one in Dr. Loeb's presentation that presented the same information in different orders than yours.

But all of these items, Mr. Swaim, do you think we have discussed in enough detail for the record here today?

Robert Swaim: Absolutely, sir.

Jim Hall: Is there any additional information that we need to know about or discuss? I notice in terms of the small explosive charges and the missile fragments, did we get into the cockpit voice recorder work that Mr. Cash did in that area?

Robert Swaim: Yes, we did. I think we've probably adequately covered that.

Jim Hall: Is Mr. Cash here?

Robert Swaim: Yes, he is.

Jim Hall: Why don't we let Mr. Cash, if he could, come up and replace Mr. Sweedler (?) and briefly cover that if he could. Mr. Sweedler (?), we promise to ask you a question tomorrow. Please go ahead, Mr. Cash.

Jim Cash: I'm not sure what the question was, sir.

Jim Hall: The question was, the spectrum work you did on the cockpit voice recorder, in terms of what type of sound it was. Was it an explosive device, or decompression in the aircraft, or whatever your work was in that area?

Jim Cash: We went through a lot of tests to try to see if we could get some more information from the sounds that were on the cockpit voice recorder. Bob Swaim alluded to some of the work we had done with the spectrum, looking at the abnormalities that we did find. You brought up about the fuel flow gauge, which the crew kind of dismissed as normal operation, even though they did make a comment about it. The work we had done in the ?? tests, to try to determine what kind of signatures we were looking for, and we pretty much did all that we could possibly think of, and all that the parties could possibly think of, to try to glean whatever information we could out of the cockpit voice recorder.

John Goglia: Mr. Cash, you said a very important word there, very offhandedly - "everything that the parties could think of as well." All of us at this agency are guilty of that - that we go over that once over lightly, because we do it as a matter of course. But every single thought that every single person has on our teams gets explored. It is not rejected out of hand. We explore every single one of them. I sat as a party to these investigations for a long time before I came here, and I think that that's missed on the part of the general public - that we don't overlook anything, that we do reach out.

Oftentimes, a team decision is far better than any individual. And the thoughts of everybody that was on your team was, in fact, explored. I read all the details of the work that you did. I visited you in the lab from time to time asking questions. And never once did I leave with the impression that you have not turned over every single stone possible. That's something that we do. We take it so lightly, but the other folks in the world need to know that - that we don't overlook those facts.

Jim Hall: I'm referring now to Dr. Loeb's presentation - that's the one we found when Member Goglia was speaking. We talked about a lightning strike, and we just had a possible accident we launched on where there was an indication that there might be a lightning strike. The meteorite - we got a lot of correspondence on meteorites, and as you know we had someone testify at the public hearing on the subject of meteorites.

The missile fragment? Is there anything else we need to say on the missile fragment? I guess we looked at a missile, we looked at a missile exploding near the aircraft.

Bernard Loeb: We looked at the possibility that a missile had been fired but did not strike the airplane. After a certain period of time in-flight warheads will self-destruct. We considered the possibility that a warhead might have self-destructed in the vicinity of the center wing tank and put some pieces through the center wing tank that could have been the source of ignition - a very high velocity piece, fragment coming from a warhead, penetrating the tank, the high-velocity going through the tank would heat up the area right around it to a temperature that could cause ignition (we know that from some work that has been done), and we looked at the possibility that self-destruct - we can go into that in detail if you wish but suffice it to say that there was no position in which we could get a warhead to self-destruct and get a particle or fragment through the tank at a sufficient speed and not pepper the airplane with other fragments, with two or three hundred other fragments. And we have looked at all of the wreckage and seen no such evidence whatsoever. So we did look at that and consider it very seriously.

Jim Hall: Auto-ignition? Hot surface ignition?

Bernard Loeb: We also looked at the issue of auto-ignition and hot surface ignition, and for a variety of reasons, the temperature in the auto-ignition case that was required, the entire fuel tank, or basically the entire fuel tank, to reach a temperature level that was just not possible from other evidence. And the same thing with the hot surface ignition that would require a temperature even greater for a lesser area, but we simply couldn't find any evidence to corroborate and, in fact, there's a lot of evidence to suggest that that did not happen.

Jim Hall: Fire through the vent system?

Bernard Loeb: That was a theory that came up, I think, a half a year or so into the investigation. We examined that very thoroughly, and there's a whole lot of reasons why we do not believe that happened, but the most important one is that the flight tests that were done at JFK with the Evergreen airplane, the temperature measurements we took, the possibility of a flame going back in through the vent stringer system into the center tank - the temperatures were such that the flow would simply not sustain the combustion. It would not be flammable. It would be quenched and could not happen.

Jim Hall: Uncontained engine failure? I assume each one of these items is well documented in the report.

Bernard Loeb: Yes, it is. And by different groups and different individuals working in each of these areas.

Jim Hall: Air conditioning turbine burst?

Bernard Loeb: Again, not very much like the engine. We simply saw absolutely no evidence of that whatsoever.

Jim Hall: The jettison override pump malfunction?

Bernard Loeb: We had the jettison pumps. They were disassembled by Bob Swaim. He can go into that in more detail. But there is simply insufficient evidence to suggest that they were a player. I would like to add though, again, that some of these could be sources for ignition in other cases. We're not saying that you cannot ignite the vapors with any of these. You can. We're just saying that in this case, the physical evidence wasn't there.

Jim Hall: You went through a process of elimination?

Bernard Loeb: Yes, sir. That's what took Bob four years. He was not on vacation most of that time. He was working all that time, and it took him four years to systematically - with a variety of different people on his groups and consultants that were helping us in each and every one of these areas - they systematically went through them, and that's what took the four years.

Jim Hall: The scavenge pump, as we all know, has been identified as not recovered?

Bernard Loeb: We did not find the major parts of the scavenge pump; however, with the pieces that we had, there was no evidence that the scavenge pump was used. We do have the switch that indicated that. We also have a mounting plate from the rear wall of the rear spar. But, more importantly, we've done a number of tests to indicate, with these pumps, that if you did get some sort of vapor in the motor that was ignited, the arresting systems and suppression systems would quench the flame, and you would not get a flame getting into the tank.

Again, I would like to reiterate: That is not to say that it is not possible for that to happen. We just have no evidence in this particular case.

Jim Hall: Static electricity?

Bernard Loeb: Static electricity. As you well know, Mr. Swaim spent months and months and months and some more months after that, and a few more even after that working on this issue with some of the smartest people in this country and the most knowledgeable people in this country on that issue. Again, static electricity has been demonstrated in the past to be a source of ignition. In this particular case, there is no evidence that that is the case. In fact, we were unable to find a way in which it could have, again in this particular situation.

Jim Hall: Now, electromagnetic interference?

Bernard Loeb: You've just heard about that from Scott Warren. We had Langley Research, we had their Naval Air Weapons Center at Patuxent River, the Joint Spectrum Center working on doing research for us. All of them agreed that EMI was not likely the cause of the ignition in this case.

Jim Hall: Mr. Swaim, after four years of working on this, I assume that you would have liked to have been able to come here today in front of the family members and tell them what the ignition source was.

Robert Swaim: Absolutely. Sir, I would just love to be able to walk in here with a molten piece of wire and say, here it is. If we wanted to overanalyze, we could say in the report we at least got down to a single wire bundle - but, again, I don't think that would be beneficial, because this airplane had the interior changed, including the lighting systems, at least twice in its service life. So we don't truly know what the condition or routing of all the wires were in this airplane. We find the airplanes are quite different over time. So, from the half of the wiring from this airplane we had, we have some solid records and indicators. But no sir, I cannot walk up and say, this is the piece.

Jim Hall: Very well. And tomorrow, Mr. Swaim, we'll continue to hear about your four years of work in the morning, when we begin. And I assume we will get into the 40 directives and rules that the FAA has looked at and the various work of Boeing to address some of the ignition sources and recommendations.

Bernard Loeb: Yes, we will sir. I would like to make one correction. I inadvertently credited with Mr. Bob Swaim for one part of the work that actually his boss, John Delisi, had done, and that was the fuel pump air downs. I had forgotten. My age got to me and I was reminded. And I'm sorry, John. You should take the credit for it.

Jim Hall: Mr. Delisi, we'll try to find a question for you tomorrow, too. Tomorrow we're going to discuss the maintenance and aging of the aircraft systems, the design and certification issues, and the reported witness observations, in addition to any other topics or issues that our five Board Members may want to explore out of this report. We will also follow that with a reading of the proposed findings, probable cause, and recommendations. And I would like to lay on the record tomorrow, Dr. Loeb, at some point, the status of the recommendations that we had made in 1996 and 1998. I know we have additional recommendations that are going to be made in this report, but I would like to be sure that the American people and particularly the families are provided with a status report on what has been done on the recommendations that the Board has made previously.

Bernard Loeb: Okay, and I think that would be an excellent thing for Mr.Sweedler to be able to handle as his going away present.

Jim Hall: We are all teasing Mr. Sweedler just a little bit because he has just completed 31 years of government surface. He has been the head of our Office of Safety Recommendations and Accomplishments, and as soon as this meeting concludes, he is heading for his retirement home in San Francisco, California, and he will report to us tomorrow - be prepared, Mr. Sweedler, if you'd please provide a report for us on the status of the recommendations.

Are there any other items that need to be discussed this afternoon? If not, we will reconvene this meeting of the National Transportation Safety Board at 9:30 in the morning.