利用BP神经网络预测材料温锻流动应力

Prediction of Material Flow Stress in Warm Forging with BP Neural Network

采用 BP神经网络算法对 0 8F钢和 40 Cr在温锻温度范围内流动应力的实验数据进行处理 ,建立起预测材料温锻流动应力的 BP网络模型 .同传统回归方法相比 ,该方法不需事先判断变形温度 t、应变速率 ε·-、应变 ε与流动应力 σ的复杂关系 ,而是通过对大量离散样本进行多次训练 ,找出蕴含在 t、ε·-、ε与σ之间的本质联系 .结果表明 ,训练好的神经网络模型能够比较准确地描述温锻成形时 t、ε·-、ε与 σ的关系 ,从而预测出材料在一定变形条件下的

A BP neural network model used for predicting material flow stress in warming forging was set up on the base of dealing with the experimental data of 08F steel and 40Cr flow stress. Compared with the traditional regression method, the method of BP neural network does not need to analyse the complex relations between the temperature, strain rate, strain and the flow stress. By training many discrete samples, it can find out the potential relations between the temperature, strain rate, strain and the flow stress. The research result shows that the well trained BP neural network model can accurately describe the influence of the temperature, strain rate, strain on the flow stress, and as a result, it can predict the material flow stress at certain forming conditions precisely.