Results of 1/4-Scale Experiments. Vapor Simulant And Liquid Jet A Tests

Abstract

A quarter-scale engineering model of the center wing tank (CWT) of a 747-100 was constructed. This engineering model replicated the compartmentalization, passageways, and venting to the atmosphere. The model was designed to scale the fluid dynamical and combustion aspects of the explosion, not the structural failure of the beams or spars. The effect of structural failure on combustion was examined by using model beams and spars with deliberately engineered weak connections to the main tank structure. The model was filled with a simulant fuel (a mixture of propane and hydrogen) and ignited with a hot wire. The simulant fuel was chosen on the basis of laboratory testing to model the combustion characteristics (pressure rise and flame speed) of Jet A vapor created by a Jet A liquid layer at 50C at an altitude of 13.8 kft. A series of experiments was carried out in this model in order to: (a) investigate combustion in a CWT geometry; and (b) provide guidance to the TWA 800 crash investigation. The results of the experiments were observed with high-speed film, video, and still cameras, fast and slow pressure sensors, thermocouples, photodetectors, and motion sensors. A special pseudo-schlieren system was used to visualize flame propagation within the tank. This report describes the test program, facility, instrumentation, the first 30 experiments, comparisons between experiments, and performance of the instrumentation; then examines the significance of these results to the TWA 800 crash investigation. The key results of this study are: Flame Motion: The motion of flame was dominated by the effects of turbulence created by jetting through the passageways and vent stringers. A very rapid combustion event (lasting 10 to 20 ms) occurred once the flame traveled outside of the ignition bay and interacted with the turbulent flow. Most of the gas within the tank was burned during this rapid event. Compartments: The combustion time decreased with an increasing number of compartments (bays) within the tank. With six bays, combustion took only 100 to 150 ms to be completed from the time of ignition until the end of the rapid combustion phase. The total combustion event was three to four times shorter with compartments than without. Venting: Venting to the outside of the tank through the model vent stringers had a negligible effect on the combustion progress or on the peak pressure reached at the end of the burn. Ignition Location: Variation of the ignition location produced distinctive pressure loads on the structural components. Liquid Fuel: Lofting of a cold liquid fuel layer was produced by the combustion-induced gas motion. Although this spray of liquid eventually ignited and burned, it did not contribute to the pressure loading. Structural Failure: Structural failure resulted in flame acceleration, decreasing the overall combustion time. TWA 800 Investigation: The pressure loads were sufficiently high, up to 4 bar, and the combustion events were sufficiently short, that the forward portion (spanwise beam 3, front spar) of the CWT structure would fail as a direct consequence of the explosion. A combination of pressure loads was produced in some tests consistent with the TWA 800 wreckage. Replica tests, structural modeling, and sensitivity studies on fuel concentration are needed before any conclusions can be drawn about probable ignition locations. Cargo Bay: Tests with a simplified model of a half-full cargo bay indicated that repeated pressure waves with an amplitude of 1 bar or less are produced when an explosion scenario similar to TWA 800 is tested. Future Testing: Future studies should include replica tests, tests with Jet A vapor and warm liquid Jet A layers, and sensitivity tests to examine ignition location, fuel concentration, and vent area perturbations. Summary: Explosion tests in a 747-100 CWT model reveal that a very complex pattern of combustion occurs due the interaction of the flame and the flow-generated turbulence. A wide range of structural load patterns occur, depending on the location of the ignition source. Some of these load patterns are consistent with damage believed to be associated with the initial explosion event in TWA 800. Sensitivity of the loading to the ignition location indicates that narrowing down the ignition location in TWA 800 may be possible. However, the complexity of the combustion and structural failure processes in the actual center wing tank mandates extremely careful consideration of the uncertainties that enter into this process.

Files

Additional details