基于神经网络的连续球压痕法预测钢轨钢拉伸力学性能

Prediction of Tensile Mechanical Property of Rail Steel by Continuous Ball indentation Method Based on NN

神经网络在预测中具有灵活性和高效性,是当下机器学习领域的热门技术。强大的非线性建模能力使神经网络适用于各种复杂的数据模式。基于神经网络,采用连续球压痕法对RAT钢轨钢拉伸力学性能进行预测。以MATLAB软件和ANSYS软件联合仿真方式建立数据库,训练获得神经网络,在载荷位移和应力应变双重优化的前提下,通过遗传算法优化目标函数,得到预测结果。预测结果与拉伸试验结果误差较小,表明所提出的方法能够较好地预测钢轨钢的拉伸力学性能。

Neural network is a popular technology in the field of machine learning due to its flexibility and efficiency in prediction.Powerful nonlinear modeling capability makes neural network suitable for a wide range of complex data patterns.Based on the neural network,the continuous ball indentation method was used to predict the tensile mechanical property of RAT rail steel.The database was established by joint simulation of MATLAB software and ANSYS software,the neural network was trained,and the objective function was optimized by genetic algorithm under the premise of dual optimization of load displacement and stress strain,and the prediction result was obtained.The error between the prediction result and the tensile test result is small,which indicates that the proposed method can better predict the tensile mechanical property of rail steel.