Fire risk evaluation model of high-rise buildings based on multilevel BP neural network

On the basis of establishing the fire risk evaluation index system and their standard values of high-rise buildings, a multilevel BP neural network model is developed. The model can solve the problem of uncertainty, fuzziness and dynamic complexiiy well in the process of fire risk evaluation, which makes evaluation result more practicable and scientific. Exemplified case study has shown that the reliable evaluation result could be obtained in shorter time as long as the values of related fire risk evaluation indexes are input into the model. And the evaluation result accords with the actual situation and that of fuzzy evaluation by fire experts.