High rising building fire risk assessment based on the artificial neural network

High building fire is one of the most important aspect in city safety. Fire prection based on pre-formance is becoming a primary way in building fire prection design. The difference between new design method and the traditional is performing fire risk assessment to decide the fire prelection equipment, so that design can meet the needs of building safety by the lowest economic cost. Fuzzy Comprehensive Assessment is a method widely used in safety assessmnet, but it can not non-linear issues very well. Artificial Neural Network is compared with Fuzzy Comprehensive Assessment approach in this paper. The result of comparison shows the former is more suitable. A new risk assessment model for building fire based on the Artificial Neural Network and Genetic Algorithm is established after analyzing the limitation of the Artificial Neural Network such as its searching and optimizing method of weight. An example is given and the results shows this model is reliable. 1fig. ,3tabs. ,11refs.