Public Building BIM Safety Early Warning Algorithm Based on Improved Cyclic Wavelet Neural Network

Abstract

This paper uses an improved cyclic wavelet neural network algorithm to predict the safety of public buildings in BIM. First, by introducing the BIM safety warning model, the feasibility of the BIM model in the safety warning of public buildings is analyzed. Then, this paper proposes an improved cyclic wavelet neural network training algorithm, which composes the parameters of the wavelet neural network into a multi-dimensional vector, which is used as the particles in the algorithm to evolve. The BIM module extracts 4M1E basic factor information, combines the cyclic wavelet neural network algorithm to establish a safety prediction model, and adjusts the unsafe behavior and equipment in the BIM model through the prediction results. The prediction results show that the algorithm can effectively predict the safety problems of public buildings