摘要
针对半固态挤压复合材料棒材成形时工艺参数难于选取、试验工作量大的问题 ,采用人工神经网络技术与试验相结合的方法 ,通过对样本处理、神经网络模型参数及收敛性等进行分析 ,建立了工艺参数ANN预测模型 ,可以对复合材料半固态挤压成形的关键工艺参数进行预测 ,预测值与试验值吻合较好 ,最大误差不超过 0 72 % 。
It is difficult to determine the process parameters for forming composite bar products by semi\|solid state extrusion and generally a lot of experiments are required.For solving this problem,the artificial neural network forecasting model of the process parameters has been established by combining with experiment method in this paper.In the same time,some techniques including disposal of sample data,selection of neural network model parameters and convergent profile have been investigated.By the established model,the key process parameters for extruding composite bar products have been forecasted.The results are well coincident with the experimental values,and the error is not larger than 0 72%,which prove the method and established model are efficient and practical.
出处
《塑性工程学报》
CAS
CSCD
2003年第1期20-24,共5页
Journal of Plasticity Engineering
基金
国家自然科学基金资助项目 (5 0 175 0 91)
国防基金资助项目 (5 1412 0 5 0 10 1HK0 3 3 6)
西北工业大学博士论文创新基金资助项目
