液压支架顶梁疲劳寿命的改进神经网络预测

Fatigue Life Prediction of the Hydraulic Support Roof Beam by Improved BP Neural Network

对液压支架顶梁疲劳寿命计算量大,耗时多而导致的顶梁疲劳寿命很难使用智能寻优算法进行优化的问题进行了研究。提出使用遗传算法优化了的BP神经网络对液压支架顶梁疲劳寿命进行预测的方法。首先选取对顶梁疲劳寿命影响较大的5个设计参数,使用ANSYS计算其80组参数水平下的疲劳寿命。再用遗传算法优化了的BP神经网络对前70组数据进行训练建立神经网络模型。最后用10组数据验证建立的BP神经网络疲劳寿命模型的预测精度。结果表明,遗传算法优化了的神经网络能快速估算出顶梁疲劳寿命,并且估算的疲劳寿命平均相对误差较低,仅为4.04%,完全满足工程实际要求。

The fatigue life of the hydraulic support roof beam is hard to optimize by intelligent optimization algorithm because of large computation and time consuming.According to this problem,a research was put forward.BP neural network optimized by genetic algorithms was proposed to predict the fatigue life of hydraulic support roof beam.Firstly,five design parameters which had great impact on the fatigue life of roof beam were selected and ANSYS was used to calculate each fatigue life under 80 groups of parameters.Secondly,the former 70 groups of parameters were trained by the improved BP neural network,and the models were established.Lastly,the last 10 groups of data were used to verify the prediction accuracy of the BP neural network model for fatigue life.The results showed that the improved BP neural network based on genetic algorithms could quickly estimate the fatigue life of roof beam,and the average relative error of the fatigue life was only 4.04%,which fully met the engineering requirements.