基于RBF神经网络的强力旋压连杆衬套力学性能预测研究  

Study on mechanical property prediction of power spinning connecting rod bushing based on RBF neural network

作　　者：佘勇[1] 樊文欣[1] 陈东宝[1] 曹存存[1] 

机构地区：[1]中北大学机械与动力工程学院,山西太原030051

出　　处：《锻压技术》2016年第6期128-132,145,共6页Forging & Stamping Technology

基　　金：山西省自然科学基金资助项目(2012011023-2);山西省高校高新技术产业化项目(20120021)

摘　　要：针对难以运用公式来表达强力旋压连杆衬套工艺参数与力学性能之间的复杂关系问题,建立了旋压工艺参数(减薄率、热处理温度、进给比)与力学性能(布氏硬度、伸长率、屈服强度、抗拉强度)之间的径向基函数(RBF)神经网络模型。用实验所得的数据对RBF神经网络进行训练,再用训练好的RBF神经网络对成形件的力学性能进行预测,通过与实测值对比分析,并与用BP神经网络所建模型的预测结果进行比较,发现RBF神经网络模型具有较BP神经网络更优的预测性能。RBF神经网络模型预测能力强、建模时间短、能有效提高连杆衬套工艺的设计效率和降低实际实验的所需成本。For the difficulty of expressing the complex relationships between process parameters and mechanical property of connecting rod bushing formed in power spinning using a formula, it was established the model of RBF neural network between the spinning process pa- rameters （thinning ratio, heat treatment temperature and feed ratio） and mechanical property （brinell hardness, elongation, yield strength and tensile strength）. Then, the RBF neural network was trained by experimental data, and the mechanical property of formed part was predicted by the above trained RBF neural network. Comparing the experimental data with the predicted results of BP neural network, it is found that the RBF neural network has better prediction performance than BP neural network. Therefore, the RBF neural network model is of high prediction performance and short modeling time. Thus, it can effectively improve the design efficiency of connecting rod bushing process and reduce the cost in actual experiment.

关 键 词：强力旋压 旋压工艺参数 力学性能 RBF神经网络 

分 类 号：TG306[金属学及工艺—金属压力加工] TP183[自动化与计算机技术—控制理论与控制工程]