基于BP神经网络的金属薄板包辛格效应预测

Prediction of Bauschinger effect for metal thin sheet based on BP neural network

针对在测量薄板材料包辛格效应时极易失稳的问题,提出了利用BP网络实现薄板材料包辛格效应预测的方法。首先,构建了X80管线钢的预拉伸+三点弯曲连续变形过程的仿真模型,并设计试验验证了模型的可靠性。其次,通过仿真计算构建了宏观力学响应(载荷-位移曲线)和包辛格效应的映射关系。最后,以映射关系为样本点,采用BP神经网络构建了包辛格效应预测模型,并通过单轴拉压试验对模型进行了验证。结果表明:该模型对X80管线钢包辛格效应的预测误差为3%,为薄板包辛格效应测定提供了新方法。

Aiming at the problem of instability when the Bauschinger effect of thin sheet materials is measured,the method for Bauschinger effect prediction of thin sheet using BP neural network was proposed.Firstly,the simulation model of continuous deformation process of pre-stretching and three-point bending for X80 pipeline steel was built,and tests were designed to verify the reliability of model.Secondly,the mapping relationship between the macro mechanical response(load-displacement curve)and the Bauschinger effect was established by simulation calculation.Finally,the mapping relationship was taken as the sample point and the BP neural network was used to construct the Bauschinger effect prediction model,and the model was verified by uniaxial tension and compression test.The results show that the prediction error of the model for Bauschinger effect of X80 pipeline steel is 3%,which provides a new method for the determination of Bauschinger effect of thin sheet.