基于神经网络的门机金属结构有限元模型修正

Portal Crane Metal Structure Finite Element Model Updating Based on Neural Network

传统的有限元模型修正方法计算量大,计算过程复杂、求解困难,为提高有限元模型的修正效率,提出了一种基于遗传算法优化的BP神经网络(GA-BP)有限元模型修正方法,针对BP神经网络的固有缺陷,引入遗传算法优化BP神经网络的参数,以结构响应为网络输入,结构参数为网络输出,构建了适用于起重机金属结构有限元模型修正的神经网络模型结构,最后通过所建立的GA-BP方法完成了门机金属结构的有限元模型修正。研究结果表明,与初始模型相比,修正后的有限元模型误差明显减少,精确性更高,可以保证结构有限元模型与实际结构的吻合度,为后续基于有限元模型的研究工作提供保障。

The traditional finite element model updating method requires a large amount of calculation,and the calculation process is complex and difficult to solve.In order to improve the updating efficiency of finite element model,a BP neural network(GA-BP)finite element model updating method based on genetic algorithm optimization is proposed.Aiming at the inherent defects of the BP neural network,genetic algorithm is introduced to optimize the parameters of the BP neural network.With the structural response as network input and the structural parameters as network output,a neural network model suitable for the finite element model updating of crane metal structure is constructed.Finally,GA-BP method was used to complete the finite element model updating of portal crane metal structure.The results show that compared with the initial model,the error of the modified finite element model is significantly reduced and the accuracy is higher,which can ensure the coincidence between the structural finite element model and the actual structure,and provide guarantee for the subsequent research work based on the finite element model.