基于BP神经网络的推料机结构可靠性分析

Structural Reliability Analysis of Pusher Based on BP Neural Network

为分析推料机在实际载荷工况条件下的工作可靠性,提出一种基于BP神经网络的推料机结构可靠性评估方法。首先,构建推料机有限元模型,并分析其在实际载荷工况条件下的最大应力响应;其次,结合APDL语言,构建推料机参数化分析模型,并对其开展试验设计,从而获取推料机不确定性因素与相应结构响应样本数据,在此基础上,确定基础BP网络结构,进而拟合推料机不确定性因素与响应间的信息传递关系,完成推料机BP网络模型的建立;最后,结合结构可靠性分析原理,构建推料机结构可靠性指标评估模型,完成推料机结构可靠性评估。分析结果表明:构建的BP网络模型能够合理表征推料机不确定性因素与最大应力响应间的内在关系;由于参数不确定性的影响,推料机最大应力响应呈现出一定的随机性并存在失效的可能,该结论可为推料机进一步优化设计提供一定参考价值。

In order to analyze the working reliability of pusher under the actual load conditions,a structural reliability evaluation method of pusher based on BP neural network is proposed.Firstly,the finite element model of the pusher is constructed and its maximum stress response under the actual load condition is analyzed.Secondly,combined with APDL language,the parametric analysis model of the pusher is constructed and the experimental design is carried out,so as to obtain the sample data of the uncertainty factors and the corresponding structural response of the pusher.On this basis,the basic BP network structure is determined,and then the information transmission relationship between the uncertainty factors and the response of the pusher is fitted to complete the establishment of the BP network model of the pusher.Finally,combined with the principle of structural reliability analysis,the structural reliability index evaluation model of pusher is constructed to complete the structural reliability evaluation of pusher.The analysis results show that the BP network model can reasonably characterize the internal relationship between the uncertainty factors of the pusher and the maximum stress response.Due to the influence of parameter uncertainty,the maximum stress response of the pusher presents certain randomness and there is the possibility of failure.This conclusion can provide a certain reference value for further optimization design of the pusher.