UC Berkeley

Electrical wires are potential sources of fire safety issues and fire ignition in electrical systems for structural, transportation, and space applications. Electrical wires acting as fire hazards become especially important when considering the current global-scale transition from fossil fuel energy technologies towards increasing use of electrically driven energy technologies, especially transportation and heating, fueled by renewable energy sources. Additionally, NASA’s next generation of spacecrafts are planned to operate with reduced pressure and elevated oxygen concentrations within the cabins, causing an increased risk for fire hazards in such environments. In combination, these factors make fire safety in electrically powered systems increasingly important. Thus, it is of interest to understand the burning behavior of electrical wires in different environments, particularly in space exploration atmospheres. This understanding can be improved by obtaining results which provide further insight into the complex mechanisms present in flame spread along electrical wiring. Future analogous experiments planned to take place on the International Space Station (ISS) can also be compared to this work for increased understanding of this problem and improved predictive capabilities of models of wire burning in spacecraft.In this work, simulated electrical wires were burned horizontally subject to various forced flow, ignition, ambient pressure, and oxygen concentration conditions. The wire samples consisted of cores 125 mm in length surrounded by insulation sheaths 100 mm in length, with these lengths being determined by the available experimental apparatus. The cores were made of either solid copper rods with diameters of 0.64 mm, 1.8 mm, or 2.5 mm, nichrome rods with diameters of 0.64 mm, or stainless-steel tubes with outer diameters of 2.4 mm. The surrounding insulation was composed of low-density polyethylene (LDPE) with an outer diameter of 3 mm for wires with core diameters of 1.8 mm or an outer diameter of 4 mm for all other wire sample types. The cores and insulation geometries were selected to match those of experiments to be conducted in the ISS. Each of the environmental variables were tested in different combinations, but the overall ranges of each parameter were as follows. The flow varied from a no forced flow condition up to 0.3 m/s in either an opposed or concurrent configuration relative to the direction of flame spread. The ignition time was increased as a means to test the effect of excess heat in the wire. The ambient pressure ranged from 40 kPa to 100 kPa. Finally, the oxygen concentration was varied from 18% to 27%. For each combination of conditions tested, the flame spread rate over the surfaces of these wires was measured to characterize their burning behavior. Dripping of molten insulation was also observed, and both the frequency and the total mass loss by dripping were tracked and reported as well. Results showed that in experiments with variation in flow velocity, the flame spread rate was found to increase linearly with the flow velocity for concurrent flame spread but to decrease for opposed flame spread. The mass loss due to dripping was found to remain approximately constant for all airflow velocities. These trends were observed for all tested wire types, except for thick copper wire samples, which showed no spread in the opposed flow regime due to the core’s heat sink effect on the flame in such an environment. When varying the ignition condition, in a 100 kPa environment, it was found that for wire samples with either thin copper rod cores or other less conductive cores, the length of igniter exposure had very little effect on the flame spread rate. For more conductive wire samples, a slight effect was observed in which longer lengths of exposure to the igniter produced faster flame spread along the wires. For the highly conductive, thick copper wire samples, a much more exaggerated form of the same trend was observed. Repeating these igniter exposure experiments in a low-pressure environment caused delays in ignition and enhanced the effect of igniter exposure length on flame spread rate. As with the atmospheric-pressure environment, tests in a low-pressure, 60 kPa, environment showed samples with low-conductivity cores had negligible changes in flame spread rate as the length of igniter exposure increased. The more conductive wire samples showed similar trends to one another which included a drastic increase in flame spread rate with increased exposure to the igniter. Results from experiments which kept the igniter exposure time constant and varied pressure showed that flame spread rate increases with pressure. Melted and burning insulation left behind by flame dripped with a frequency that increased with pressure, and the total mass of insulation dripped decreased with pressure. Coincidingly, as the mass of dripped insulation increased, the flame spread rate decreased. Comparison of present results with those from previous analogous studies with different wire samples show that the effect of environmental parameters on flame spread and insulation dripping depends strongly on core conductivity and core and insulation diameters. Results from further experiments which varied both ambient pressure and oxygen concentration as well as forced flow velocity showed that the flame spread rate along these horizontal simulated electrical wires tends to increase with increasing oxygen concentration. It was also found that this increase in flame spread rate for increasing oxygen concentration occurred for all tested forced flow velocities and pressures. The limiting oxygen concentration for the tested wires was identified to be either slightly below or between the range of 18 to 21% oxygen concentration. Finally, the possible observation of elevated oxygen concentrations allowing for increased flame spread rates even at lower pressures compared to atmospheric, sea-level conditions was unable to be confirmed due to disruptions in testing due to the COVID-19 pandemic and a resulting incomplete dataset. In additional to the experimental work completed for this analysis, further understanding of the flame spread over electrical wire problem was achieved through use of an artificial neural network (ANN). This ANN was trained to predict the flame spread rate along simulated electrical wires of different sizes and compositions while exposed to different ambient conditions. The wire core materials used to train the ANN included solid copper, nichrome, and iron and stainless-steel tubing. The wire insulation material used to train the ANN included high-density polyethylene (HDPE), LDPE, and ethylene tetrafluoroethylene (ETFE). Finally, the conditions used to train the ANN included varying forced flows, ambient pressure, oxygen concentration, wire orientation, and gravitational strength. To facilitate the training of the ANN which allowed it to make flame spread rate predictions, a comprehensive data base of 1200 data points was created by incorporating flame spread rate results from both the data presented in this work as well external experiments from other sources. After this training, predictions from the ANN show that it is possible to merge together various data sets, including results from horizontal, inclined, vertical, and microgravity experiments, and obtain unified results. While these initial results are very encouraging with an overall average error rate of 14%, they also show that future improvements to the ANN could still be made to increase prediction accuracy. ANN predictions in the form of parametric trends were also compared with experimental flame spread rate results both from the present work and from the literature. These predictions alongside experimental results confirmed that the effect of environmental parameters on flame spread rate depends strongly on core conductivity, insulation diameters, and insulation dripping. Consequently, care should be taken in extending results obtained from specific wire tests to other wires without justification, especially if there was variation in gravitational strength across the experiments.