基于人工神经网络的导管数控弯曲成形工艺参数预测方法

Prediction of NC tube bending process parameters based on artificial neural network

阐述了导管数控弯曲成形过程的要素,分析主要工艺参数对导管弯曲成形质量的影响,并建立预测工艺参数的BP(Back Propagation)人工神经网络模型。选取实验数据作为样本,采用LM(Levenberg_Marquardt)贝叶斯正则化算法对该模型进行训练,确定模型的主要参数。通过实例预测并与实验数据进行比较,验证该方法的有效性。与其他BP训练算法进行比较,结果表明,该算法收敛速度快、预测精度高、稳定性好。

Elements of NC tube bending process are described. The influence of the primary process parameters on the quality of formed tube is analyzed. The prediction model for NC tube bending parameters based on BP （ Back Propagation） artificial neural network is built. Experimental data of NC tube bending is selected as samples, and LM （Levenberg_Marquardt）Bayesian regularization algorithm is used to train the prediction model ; then the primary parameters are educed. The validity of the model is proved by comparing the prediction result with the experimental data of NC tube bending examples. The LM Bayesian regularization algorithm is compared with other BP training algorithms ; and it is shown that the algorithm adopted by this has fast convergence rate and high prediction precision and stability.