摘要
针对挖掘机动臂结构优化过程中需要反复调用ANSYS有限元软件,导致应力强度控制难度大、优化过程繁琐且效率低下的问题,提出一种基于BP神经网络预测动臂应力的方法。通过在挖掘机动臂结构优化设计软件中设定截面选取规则,选取应力普查的截面导出截面应力样本,并在MATLAB中运用BP神经网络建立挖掘机动臂应力预测模型。以中小型挖掘机动臂为例,建立基于动臂应力普查的约束,获取应力预测样本,运用建立的神经网络预测模型对动臂应力进行预测。结果表明,网络预测应力值与实验数据吻合且误差小于6.80%,建立的预测模型能提高动臂结构的优化效率。
The finite element software ANSYS needs to be run repeatedly in the excavator boom structure optimization process, making the optimization process cumbersome and inefficient. To address this problem, an intelligent optimization model for four typical stress conditions of the excavator boom was proposed. Stress census section was determined through setting rules in opti- mal design software of excavator boom and establishing a stress prediction model for the excavator boom based on BP network. The small and medium excavator booms were used as examples and stress prediction models were established to improve the opti- mization efficiency of the boom structure. The results show that predict stress was in consistent with the experimental data with error less than 6.08 %.
出处
《现代制造工程》
CSCD
北大核心
2018年第1期59-62,103,共5页
Modern Manufacturing Engineering
基金
国家自然科学基金青年基金项目(51405085)
关键词
挖掘机动臂
应力普查
应力特征截面
BP神经网络
应力预测
excavator boom
stress census
stress characteristic section
BP neural network
stress prediction
