基于ICUOWGA-RBF神经网络的装配式建筑施工安全评价

Safety Evaluation of Prefabricated Building Construction Based on ICUOWGA-RBF Neural Network

为解决装配式建筑施工安全评价指标的模糊不确定性、随机性和指标间非线性关系造成专家难以洞察指标全部信息,从而降低评价结果科学性的难题,提出了基于ICUOWGA-RBF神经网络的装配式建筑施工安全评价方法。该方法首先根据事故致因理论,从人员-物-管理-技术-环境5个维度构建装配式建筑施工安全评价指标体系;然后引入模糊语义量化算子对传统CUOWGA算子进行改进得到ICUOWGA算子,并利用ICUOWGA算子计算评价指标的权重,进一步消除运算过程的主观性;最后利用RBF神经网络训练样本数据,规避模拟过程陷入局部最小值的缺陷,优化学习效率,提高收敛速度。通过将该方法运用于郑州某装配式建筑工程施工项目的安全评价,结果表明该装配式建筑施工的安全等级较高,并指出影响安全施工的主要因素,为装配式建筑施工安全管理提供了管理思路。

In order to solve the fuzzy uncertainty, randomness and nonlinear relationship of the safety evaluation index of the prefabricated building construction,it is difficult for experts to gain insight into all the information of the index and reduce the scientific difficulty of the evaluation results.This paper presents an assembly construction safety evaluation model based on ICUOWGA-RBF neural network. Firstly, according to the accident cause theory,the paper constructs the evaluation index system from the five dimensions of personnel-material-management-technical-environment, and introduces the fuzzy semantic quantization operator to improve the traditional CUOWGA operator to obtain the ICUOWGA operator and eliminate the subjectivity of the operation process. Finallys the paper applies the RBF neural network to training the sample data to avoid the defects of the simulation process falling into the local minimum, optimizing the learning efficiency and improving the convergence speed. Finally,the paper applies the model to the safety evaluation of a prefabricated building construction in Zhengzhou. The results show that the construction safety grade of the prefabricated building is high and the main factors affecting the safety construction are pointed out,which provides management ideas for the safety management of the prefabricated building construction.