mustangdriver wrote:
The NTSB tried to conduct a similat test, but they used propane in the fuel tank, because they could not get the jet fuel to explode! True story.
Yes, they used a simulant fuel for some of the tests for research purposes. But saying what you stated above is not correct and used out of context

. Below is extracted from the NTSB report of TWA 800:
....Seventy-two tests were conducted, in three separate test phases, over a 2-year period. The initial phase of testing was conducted using a simulant fuel (268). The second and third phases of testing were conducted using Jet A fuel. Variations in model configurations, ignition locations, and fuel vapor conditions were examined. Because the primary intent of these tests was to provide a validation database for computer model development (which required that the essential physical phenomena of the CWT combustion event be represented), researchers did not require experimental duplication of all of the conditions on board the accident airplane. Important distinctions between the 1/4-scale test model and the full-scale CWT and the limitations of the testing were the following:
The complexities of the full-scale CWT were simplified in the 1/4-scale model. The features believed to be most significant-bay volumes and lengths and vent and passageway areas-were scaled from the full-scale CWT, with some simplifications. For example, variations in tank height and the finer geometric details of stiffeners and stringers were not included in the model.
It is not clear if the results of the 1/4-scale model testing can scale directly to replicate full-scale results. The effects of scaling on some features (such as flow turbulence and flame quenching) are not well understood.
The effects of temperature and fuel vapor concentration variations within a single bay and between bays could not be examined in this limited experimental program (269).
Other simplifications of the 1/4-scale model included the even distribution of fuel between the bays and a smooth floor geometry. Thus, the role of liquid layer participation may not be accurately demonstrated.
The testing did not attempt to simulate partition failures, representative of fullscale CWT structural dynamic behavior during the combustion event, and the effects of such failures upon the combustion dynamics. Because of the differences between the 1/4-scale test model and the full-scale CWT, the interpretations that could be made by directly comparing these experimental results to a full-scale explosion were limited. Within these limitations, however, analysis of the experimental results led the researchers to reach the following conclusions regarding direct application of the 1/4-scale test results to TWA flight 800:
Jet A fuel ignited and combusted during every test using conditions approximating those that existed in the accident airplane's CWT at the time of the accident.
The ignition of Jet A fuel in one bay of the 1/4-scale model resulted in transmission of the flame through the bay passageways and vent stringers and ignition in neighboring bays, illustrating the behavior of multicompartment flame propagation. Flamefront quenching was also observed to be a characteristic of flame propagation.
After the combustion flamefront propagated from the ignition bay to a neighboring bay, the combustion process dramatically accelerated, allowing explosion pressures to increase rapidly.
In certain tests, pressure levels in bay 1 (the bay between SWB2 and SWB3) of the 1/4-scale model exceeded those needed to fail SWB (as indicated by Boeing's structural analysis of the full-scale geometry).
268 - The simulant fuel (1.4 percent propane, 7 percent hydrogen, and 91.6 percent air) was developed and used in this test series to permit the observation of multicompartment fuel vapor explosions at the local ambient temperature and pressure conditions of the test site (near Denver, Colorado), simulating the combustion behavior of the Jet A fuel vapor in the accident airplane's CWT at the conditions that existed at the time of the accident. Using the simulant fuel greatly reduced the experimental effort required (compared to using Jet A fuel). Using Jet A fuel during the second and third phases of tests required significant modifications to the test model to elevate the temperature to between 104° and 122° F at a reduced pressure equivalent to an altitude of 13,800 feet msl.
269 - Temperature and vapor concentration variations within a single bay and between bays were observed in the JFK flight test results. Subsequent ground tests were conducted in Marana, Arizona, in which CWT temperatures and vapor concentrations were measured with higher resolutions than those measured during the JFK flight tests. These tests confirmed the existence of temperature and vapor concentration variations within a single bay and between bays. For additional information, see Summary Data Report: B-747-100 Center Wing Tank Ground Testing at Marana, Arizona, dated January 20, 2000, and the Flight Test Group Chairman's Factual Report, dated November 19, 1997.
Find the entire report here (Note it is a 8MB .pdf file):
http://www.ntsb.gov/Publictn/2000/AAR0003.pdf