Inadequate quality control of fire extinguishing and suppression agents may affect airworthiness through a reduction in fire protection capability, or pose a hazard to personnel where contaminated extingusihants are toxic.

Transport Canada, the Federal Aviation Administration, and the United Kingdom Civil Aviation Authority requested a study be carried out to review the processes used in North America and Europe for the quality control of agents in fire extinguishers and fire suppression systems. This report reflects the outcome of the study and contains recommendations for optimized processes for consideration by the airworthiness authorities and industry.

This report summarizes the research effort undertaken by the Federal Aviation Administration to develop a laboratory-scale flammability test for magnesium alloys used in the fabrication of aircraft seat structure. During the initial phase, a laboratory-scale test rig was constructed to allow flame exposure to various magnesium alloy bars as they were suspended over a small steel pan. An oil-fired burner configured in accordance with Title 14 Code of Federal Regulations Part 25.853(c) Appendix F Part II was used to simulate the fire. Test samples representing a variety of magnesium alloy combinations were evaluated, including two new-generation alloys containing rare earth elements. The test samples were subjected to the burner flames for various durations. In most cases, the alloys melted, depositing pieces of molten material into the catch pan below. Subsequent to the melting event, the materials would typically ignite, emitting an intense light during ignition. Measurements were made of the flaming duration and the amount of material consumed during each test. From the initial tests, it was determined that several rare-earth-containing alloys showed increased flammability resistance when compared to a legacy magnesium alloy, such as AZ31. Two new generation alloy materials, Elektron®WE43 and Elektron®21, self-extinguished shortly after removing the fire source. By comparison, the AZ31 magnesium alloy configurations did not self-extinguish and continued to burn, sometimes until completely consumed.

Subsequent full-scale tests were conducted with a large external fuel fire adjacent to an aircraft fuselage, simulating a severe, but survivable, accident in which the fire entered the cabin through a simulated fuselage rupture. The tests determined that no significant change to survivability (based on the survivability model) resulted when using seat frames constructed of the new generation WE43 magnesium alloy in the primary components, when compared to identical tests in which the standard aluminum alloys were used. The primary seat frame components included the legs, the cross tubes, and the spreaders. Two types of magnesium alloys were used in separate tests: a well-performing alloy (WE43) and a poor-performing alloy (AZ31).

During the final phase of work, a flammability test for magnesium alloys was developed based on the findings of the realistic full-scale tests. The intent was to expose an appropriately-sized test sample to the flames of an oil-fired burner for a period of time that allowed the test sample to melt, as initial tests had indicated the magnesium alloys would not ignite until melting had occurred. Numerous test sample shapes, sizes, and exposure levels were trialed in an effort to replicate the outcome of the full-scale tests, namely, the amount of time required to melt and ignite a sample and the approximate amount of time required for the sample to self-extinguish. The final configuration used a 0.25-inch-thick by 1.5-inch-wide by 20-inch-long horizontally-oriented bar test sample that was exposed to the oil burner for a period of 4 minutes. A passing sample is not permitted to ignite prior to 2 minutes and must also self-extinguish within 3 minutes of the burner being turned off (7 minutes from the start of the test). In addition, the sample must not lose more than 10% of its initial weight.

Lithium-metal and lithium-ion batteries power many consumer electronic devices. There have been incidents in which lithium batteries have overheated, creating either a fire, an explosion, or both. Federal Aviation Administration tests have shown that when a single cell in a battery pack undergoes thermal runaway, its heat causes adjacent cells to do likewise. The propagation of thermal runaway can be prevented and the resultant fire extinguished if the correct extinguishing agent is used.

The objective of this study was to compare the effectiveness of fire extinguishing agents for suppressing lithium-metal and lithium-ion battery fires and preventing thermal runaway propagation.

Tests were performed in a 64-cubic-foot test chamber with a sealable door. First, quantitative tests were done to compare the capacity of extinguishing agents to cool a hot plate; water and other aqueous extinguishing agents were the most effective coolants and nonaqueous agents were the least effective. Next, qualitative demonstration tests were performed with lithium batteries to verify the hot plate results. These tests also showed that aqueous extinguishing agents were most effective.

The lithium-metal cells showed various behaviors while in thermal runaway, such as the creation of alternate vent holes and the ejection of internal contents. The hazards of lithium-metal cells in thermal runaway varied significantly during replicate tests.

Extinguishing agents that contained water were the most effective and their effectiveness increased with greater volumes. The gaseous streaming agents were less effective and exhibited a relatively small increase in effectiveness with increased volume.

This study was commissioned by the Federal Aviation Administration to analyze accident data to large transport airplanes registered in the United States of America and worldwide. It assessed trends in airplane safety in terms of number of accidents, accident rates, number of fatalities, fatality rates, the probability of an accident being survivable, and the probability of death in a survivable accident.

Over the study period, there has been a marked reduction in the total accident rate, both for the world fleet and the U.S. fleet. This reduction is apparent when the accident rate is measured on a per-flight, per-passenger, or per-revenue-passenger-mile basis.

The survivability of accidents has also shown a marked improvement over the study period with a greater proportion of accidents being survivable and the proportion of occupants surviving an accident increasing. These improvements are apparent in both the world fleet and the U.S. fleet.

It would seem that fatalities attributable to impact represent a larger proportion of the total number of fatalities in survivable accidents than those that are caused by fire. However, the extent to which the number of fatalities attributable to each of these two areas might be reduced is beyond the scope of this study.

Several material fire properties were evaluated as sole explanatory variables for two pass/fail flammability tests using two different probabilistic models. The fire behavior and properties investigated span the range of commercial polymers and flame-retardant plastics. The flammability tests included a fire test used by the Federal Aviation Administration for regulatory purposes and a flame test used by Underwriters Laboratories, Inc., both with categorical outcomes. The probabilistic analysis involved nonlinear least squares regression of the fire test data coded as a binary (pass/fail) outcome on each of the continuous explanatory variables using the logistic response function and a new (phlogistic) response function. The chi-square test showed that both probability models were appropriate for describing the likelihood of passing fire and flame tests using the heat release capacity, heat of combustion, volatile fuel fraction, and heat release parameter as the sole explanatory variable, but neither model was appropriate for thermal decomposition temperature as an explanatory variable. The overall efficacy of the thermal and combustion properties for predicting flammability test results based on the correlation coefficient and mean deviation of the grouped data from the fitted functions is: heat release capacity ≈ heat of combustion > heat release parameter > volatile fuel fraction > thermal decomposition temperature. Both probabilistic models are suitable for determining the likelihood of passing a flammability test using the thermal and combustion properties investigated, but the phlogistic model provides a better fit of the data when the likelihood of passing the test approaches unity.