Single compartment fire risk analysis using a fuzzy neural network

A fuzzy neural network enhanced with evolutionary algorithms, based on the GRNNFA, is proposed that is able to accurately predict the effects of a single compartment fire, based on experimental data. This system is shown to make predictions with within 5% accuracy, thus demonstrating that it can learn the non-linear nature of fluid dynamics. Because of its speed it is able to quickly generate views of the nature of the fire, enabling users to interrogate it and gain intelligence as to what compartment geometries lead to greater fire hazards.