On behalf of the Safety Board I would like to thank the National Defense Industrial Association for sponsoring this meeting, and for inviting us to participate. We hope that the critical focus that forums like this bring to bear on the problems of fuel flammability will lead to advances in the safety of air transportation.

As you know, it was concluded very early in the investigation of the crash of TWA Flight 800 on July 17, 1996, that an explosion occurred in the nearly empty center wing tank, located under the passenger cabin between the wings of the Boeing 747-100 aircraft. In our search for an explanation of this explosion, the National Transportation Safety Board has enlisted the assistance of parties to the investigation, as well as some of the world’s premier research organizations. We have initiated a number of programs of tests and research to understand the hazards associated with Jet A fuel vapors in aircraft fuel tanks, and to seek the source of the ignition of the explosion in the TWA800 center wing tank. While we have not yet concluded the investigation, I would like to provide a brief overview of the scope of the investigative activities that have been pursued during the past two years, and that are still on going. The Board’s docket on the TWA 800 investigation was opened at our public hearing in Baltimore in December of 1997, and we continue to add to the public record as the investigation proceeds.

Wreckage Reconstruction and Analysis of Breakup Sequence

One of the most publicized aspects of the investigation was the recovery of wreckage from the floor of the Atlantic Ocean and the reassembly of a significant portion of the Boeing 747-100 aircraft in a hanger near Calverton, New York. The success of the recovery effort was truly spectacular with a very large proportion of the aircraft being not only recovered from the ocean, but identified and fitted to a 90 foot long reconstruction. Careful inspection of this wreckage, both before and after the assembly of the reconstruction, provided the crucial evidence of the explosion of the center wing tank. A very extensive breakup sequence analysis, conducted by a group comprised of all of the parties to the investigation, documented both the nature and the time sequence of the events begun at the time of the CWT explosion and continuing through to the final breakup of the aircraft about a minute later. Study of these damage patterns continues as we search for clues of the specific location of ignition within the CWT.

Discovery that a CWT explosion precipitated the catastrophic breakup of TWA800 has, of course, directed our attention to both the mechanisms that could have ignited such an explosion; and the conditions that produced explosive vapors in the empty tank. I would like to briefly review the major areas of testing, research and investigation that have consumed our attention for the last two years, and that continue to do so.

Search for Ignition Sources

The Systems Group, supported by the Maintenance Records Group and other investigators, continues to examine wreckage from TWA800, as well as to inspect other aircraft, to seek possible sources of a spark, a short, or any energy that could have caused the ignition of vapors in the CWT. They have examined the Fuel Quantity Indication System and other wiring systems on the accident aircraft and other airplanes. They have considered static discharges, electromagnetic interference, and a wide variety of other potential ignition sources. This work is continuing.

Laboratory Studies of Jet A Fuel

The Safety Board contracted with the California Institute of Technology very early in the investigation of TWA-800 to perform a series of laboratory studies of Jet A fuel. The objective of these studies has been to determine the flammability characteristics of Jet A vapors at a number of temperatures and vapor concentrations, with emphasis on measuring basic explosion parameters, including: vapor pressure, flammability limits, peak explosion pressure, and pressure as a function of time. A study of the minimum ignition energy as a function of temperature and vapor concentration was also conducted as a part of this program. This program of practical research has provided important information on the properties of Jet A and its ignitability, and the research is continuing. These laboratory studies have established that 50 gallons of fuel in CWT is sufficient to create a flammable mixture; that the required ignition energy is greatly reduced because of the high temperatures in the tank; and that a resulting explosion produces sufficient over-pressure to create the damages observed in the TWA800 wreckage. Reports on Jet A flammability and minimum ignition energies are contained in the TWA800 public docket.

Chemical Characterization of Jet A Vapors

An important set of supporting tests and research has been carried out for the Safety Board at the University of Nevada, Reno in the Desert Research Institute and the Center for Environmental Sciences and Engineering. The work performed by the Center for Environmental Sciences and Engineering laboratory has focused on determining the vapor composition and vapor pressure of a variety of samples of Jet A fuel as a function of temperature and the ratio of ullage to liquid volume, using automated headspace gas chromatography. Samples of Jet A obtained from a variety of sources, including aged Athens fuel and fuels from the JFK flight tests have been examined with these methods.

Work conducted at the University’s Desert Research Institute has focused on the speciation of the samples taken from the ullage during the JFK flight test effort, as well as the Marana ground test. Work at both of these laboratories is also supporting the quarter scale explosion test program.

Flight and Ground Tests with Boeing 747/100 Aircraft

A series of flight tests were conducted from JFK Airport in New York City, in July, 1997 to determine the environmental conditions in the center wing tank under conditions similar to those at the time of the explosion on TWA flight 800, as well as under a variety of other scenarios. The flight ascent profile of TWA Flight 800 was determined in detail and the conditions of this flight were approximated as closely as possible in each of the test flights.

A series of nine test flights were conducted over a period of several days. One flight test was dedicated to replicating the factors associated with TWA Flight 800 as closely as possible. Fuel was loaded at the same time of day (including 50 gallons of Jet A fuel from Athens in the CWT), AC packs 1 and 3 were run during pre-flight, take-off was conducted at the same time of day (and under remarkably similar temperatures), the TWA-800 pre-flight procedure was followed, and the test flight flew the TWA 800 ascent profile.

Other test flights were conducted under other conditions, including one test that loaded 12,000 pounds of fuel in the CWT one-half hour before taxi.

The B747-100 aircraft was equipped with a very extensive array of temperature, pressure and vibration measurement sensors. Vapor samples were collected from a single location in the CWT on three of the test flights (including the TWA-800 emulation flight).

The factual report of this flight test is now contained in the TWA800 docket. The tests demonstrated very high temperatures in the CWT on the ground and during the flight profile, as well as considerable reduction in ullage temperatures with the addition of 12,000 pounds of fuel in the CWT.

An additional set of ground tests were also conducted this past May, in Marana, Arizona to collect additional data on the thermal properties of the CWT as a function of the operation of the air conditioning packs. An even larger array of instrumentation was used on this aircraft and additional vapor sampling was conducted. Data from these tests are still being analyzed. Mechanisms of heat transfer from the air conditioning pack bays to the CWT are an important focus of these analyses.

Quarter Scale Explosion Testing

The Cal. Tech. laboratory work on the flammability and explosive behavior of Jet A, that was referred to previously, has been done in relatively simple, single compartment test apparatus. The CWT from the B747 aircraft is, on the other hand a very large, geometrically complex structure, and it is believed to be very important to systematically deal with this complexity in our explosion testing. A quarter scale test program was, in consequence, initiated through contracts between the Safety Board and the California Institute of Technology and the Applied Research Associates. The goal of this program is to determine if a probable ignition location can be identified through combustion testing, and to determine the effects of various tank features such as compartments, vents, and partition failures that could not be examined in laboratory scale testing. We are also hopeful that this effort will help us determine whether large-scale (airplane or intact fuel tank) tests will be required.

A quarter scale model of the 747 CWT was constructed to carry out explosion testing in a tank consisting of multiple compartments, with restricted communication between these compartments. The main objectives of this test program are to determine the effects of ignition location, source of ignition, and flame front communications between compartments. A series of 30 explosion trials was carried out in a facility managed by ARA near Denver, Co. in the Fall of 1997. This first series of quarter scale tests used a simulant fuel, matched to the combustion properties of Jet A (at 14,000 ft. and 50 degrees C) in laboratory tests. A second series of quarter scale tests, using Jet A in a modified test fixture (providing for heating of up to 60 degrees C and pressurization to an effective 14,000 ft. altitude) is currently in progress.

Computer Modeling of Explosion Behavior

An activity that is very closely related to the quarter scale test program, is the computer modeling of explosion behavior through the application of computational fluid dynamics models. Two separate modeling activities, one by Sandia National Laboratories in Albuquerque, and the other by Christen Michelsen Research Institute, in Norway, are being carried out simultaneously with the experimental quarter scale work. The objective of this modeling effort is to predict the effects of changing ignition location, the effects of communication between compartments, and the expected over pressure. This work is continuing.

The feasibility of coupling the CFD models to structural models of the failure of the tank is also under investigation. We are being assisted in this effort by Combustion Dynamics Limited of Nova Scotia, Canada.

In summary, a very considerable amount of work has been generated by the TWA800 investigation, but a great deal more remains to be done. The issues that you are dealing with here today are clearly relevant to our specific investigative purposes, as well as to the broader goal of reducing the probability of fuel tank explosion. We certainly support this effort.