Researches On The Safety Evaluation Method Of Multi-ribbed Composite Wall Construction Based On The Neural Network

As a new type of green prefabricated structure system,multi-ribbed composite wall structure has integrated ecological environment protection,energy saving and earthquake resistance.Due to its remarkable economic and social benefits,it has already been applied in the pilot project and the safety evaluation of its construction process plays a significant role in engineering According to the building construction characteristics of multi-ribbed composite wall structure,building construction safety evaluation index system is proposed based on the theory of system engineering and engineering ergonomics,and the prediction model of construction safety evaluation is established based as well on neural networks,which solves the problem of safety prediction of multi-ribbed composite wall construction.The main work s of this article are as follows:(1)According to the characteristics of the construction of multi-ribbed composite wall structure,the safety factors of construction safety are expounded comprehensively combined with the theory of construction safety and efficiency from the man-machine environment system.On the basis of this,five factors influencing the construction of multi-ribbed composite wall structure are constructed,namely: human factors,material factors,programs and technical factors,management factors and environmental factors.(2)According to the principle of establishing of evaluation index system,the index system of five modules is refined decomposition based on the relevant literature and expert investigation.This paper establishes the safety evaluation index system for the construction of the multi-ribbed composite wall structure by the induction and screening of index system.The index system are analyzed and discussed as well.(3)Based on investigation and expert advice on dangerous sources of building construction with multi-ribbed composite wall structure,BP neural network and wavelet neural network model for safety evaluation of multi ribbed composite wall structure is established.The two neural networks are compared and analyzed,and the superiority of wavelet neural network in comprehensive evaluation is demonstrated.(4)In order to avoid the limitation of considering the comprehensive evaluation of neural network and ignoring the single factor evaluation to make safety evaluation.In this paper,a safety evaluation model based on the single factor and multi factor comprehensive evaluation model is proposed and the reliability of the model is established,and the experiments validate the reliability of the established evaluation model.