基于正交试验和BP网络的2524铝合金热处理工艺优化及预测

轻有色合金 | 浏览量 : 0 下载量: 247 CSCD: 8

基于正交试验和BP网络的2524铝合金热处理工艺优化及预测
Optimization and Prediction of Heat Treatment Process for 2524Aluminum Alloy Based on Orthogonal Experiment-Artificial Neural Network Model
2014年34卷第5期页码：549-551
纸质出版日期： 2014 ，
DOI： 10.15980/j.tzzz.2014.05.022

扫描看全文

[1]宋新华,陈胜迁,陈立,宋斌,袁江,周金鑫.基于正交试验和BP网络的2524铝合金热处理工艺优化及预测[J].特种铸造及有色合金,2014,34(05):549-551.

Song Xinhua, Chen Shengqian, Chen Li, et al. Optimization and Prediction of Heat Treatment Process for 2524Aluminum Alloy Based on Orthogonal Experiment-Artificial Neural Network Model[J]. Special Casting & Nonferrous Alloys, 2014,34(5):549-551.
[1]宋新华,陈胜迁,陈立,宋斌,袁江,周金鑫.基于正交试验和BP网络的2524铝合金热处理工艺优化及预测[J].特种铸造及有色合金,2014,34(05):549-551. DOI： 10.15980/j.tzzz.2014.05.022.

Song Xinhua, Chen Shengqian, Chen Li, et al. Optimization and Prediction of Heat Treatment Process for 2524Aluminum Alloy Based on Orthogonal Experiment-Artificial Neural Network Model[J]. Special Casting & Nonferrous Alloys, 2014,34(5):549-551. DOI： 10.15980/j.tzzz.2014.05.022.

对2524铝合金进行热处理正交试验

结合电导率测试

对正交试验结果进行极差分析和方差分析

并且通过人工BP神经网络模型优化

得到2524铝合金的最佳热处理制度:固溶温度为500℃

固溶时间为11h

时效温度为185℃

时效时间为5h。结果表明

该模型能较好的反映热处理工艺参数与铝合金电导率之间的内在规律

BP网络预测最大相对误差为2.19%

表明该神经网络模型具有较高的预测精度。

The orthogonal experiment of heat treatment process was designed for 2524aluminum alloy. Combined with the conductivity testing

range analysis and variance analysis of orthogonal experiment were carried out.The process was optimized by artificial BP neural network model

and the desirable heat treatment parameters of aluminum alloy were presented as follows:solid solution treatment at 515 ℃for 11h

and aging treatment at 185℃for 5h.The results indicate that

this model can better reflect the relationship between heat treatment process parameters and electric conductivity of aluminum alloy

and the maximum relative error of BP network forecasting is 2.19%

showing a higher prediction accuracy.

正交试验BP网络铝合金电导率

文章被引用时，请邮件提醒。

提交