作者:崔桂梅1, 刘伟1, 张帅1, 王磊2
作者单位:1. 内蒙古科技大学信息工程学院,内蒙古 包头 014010;
2. 内蒙古金属制造有限责任公司,内蒙古 包头 014010
关键词:轧制力预测;核函数;支持向量机;差分进化
摘要:
板带材热连轧轧制过程中,轧制力的精确控制对改善带钢板形性能有至关重要的作用。针对B钢厂2250热连轧轧线轧制力模型计算值与实际值误差较大(±10%)的问题,建立基于线性(Linear)核函数、多项式(Poly)核函数、高斯(RBF)核函数3种支持向量机回归预测模型,分析3种核函数支持向量机预测模型,选用拟合效果最好的RBF核函数支持向量机(RBF_SVM)为基础模型。通过差分进化算法对模型的惩罚系数和核函数参数进行最优参数搜索,提高预测模型准确度。实验结果表明,模型轧制力预测值与轧制力实际值误差在±5%的准确率为99.16%,可解决实际轧制过程问题,具有广阔的工程应用前景。
Rolling force prediction based on differential evolution support vector machine
CUI Guimei1, LIU Wei1, ZHANG Shuai1, WANG Lei2
1. School of Information Engineering, Inner Mongolia University of Science and Technology, Baotou 014010, China;
2. Inner Mongolia Metal Manufacturing Co., Ltd., Baotou 014010, China
Abstract: In the process of hot continuous rolling of strip steel, the precise control of rolling force plays a vital role in improving strip shape performance. In view of the large error (±10%) between the calculated value and the actual value of the rolling force model of the 2250 hot rolling mill in steel B, we proposed a support vector machine regression prediction model based on the Linear kernel function(Linear), the Polynomial function(Poly) and the Gauss kernel function(RBF). Analyze three kernel function support vector machine prediction models, and select the RBF kernel function support vector machine (RBF_SVM) with the best fitting effect as the basic model. This paper uses differential evolution algorithm to search for the optimal parameters of the model's penalty coefficient C and the kernel function parameter gamma to improve the accuracy of the prediction model. The experimental results show that our model has an accuracy of 99.16% when the error between the predicted value of the rolling force and the actual value of the rolling force is within ±5%. It can solve the actual rolling process problems and has broad engineering application prospects.
Keywords: rolling force prediction;kernel function;support vector machine;differential evolution
2021, 47(8):83-88 收稿日期: 2020-09-26;收到修改稿日期: 2020-11-03
基金项目: 国家自然科学基金(61763039)
作者简介: 崔桂梅(1963-),女,河北保定市人,教授,博士,研究方向为复杂过程系统的建模及运行优化控制研究
参考文献
[1] PENG K X, ZHONG H, ZHAO L, et al. Strip shape modeling and its setup strategy in hot strip mill process[J]. The International Journal of Advanced Manufacturing Technology, 2014, 72(5-8): 589-605
[2] 孙伟, 张炜星, 张慧星, 等. 60Si2Mn弹簧钢连续冷却组织调控研究[J]. 中国测试, 2020, 46(2): 137-142
[3] CAO J G, XIONG H T, HUANG X H, et al. Work roll shifting strategy of uneven "cat ear" wear control for profile and flatness of electrical steel in schedule free rolling[J]. Steel Research International, 2020, 91(9): 25-40
[4] LIU X H, ZHAO Q L, HUANG Z Y. Prospect of artificial intelligent application in rolling[J]. Steel Rolling, 2017, 34(4): 1-5
[5] DENG J, SUN J, PENG W, et al. Application of neural networks for predicting hot-rolled strip crown[J]. Applied Soft Computing, 2019(78): 119-131
[6] XU Z W, LIU X M, ZHANG K. Mechanical properties prediction for hot rolled alloy steel using convolutional neural network[J]. IEEE Access, 2019(7): 47068-47078
[7] AGARWAL K, SHIVPURI R. On line prediction of surface defects in hot bar rolling based on Bayesian hierarchical modeling[J]. Journal of Intelligent Manufacturing, 2015, 26(4): 785-800
[8] 马威, 李维刚, 赵云涛,等. 基于深度学习的热连轧轧制力预测[J]. 钢铁研究学报, 2019, 31(9): 805-815
[9] 曹卫华, 李熙, 吴敏,等. 基于极限学习机的热轧薄板轧制力预测模型[J]. 信息与控制, 2014, 43(3): 270-275
[10] 章顺虎, 姜兴睿, 尤凤翔,等. 融合工业大数据的热轧厚板轧制力模型研究[J]. 精密成形工程, 2020, 12(2): 8-14
[11] 王前锋. 基于改进型支持向量机算法的轧机轧制力预测[J]. 锻压技术, 2019, 44(4): 131-137
[12] 魏立新, 魏新宇, 孙浩,等. 基于改进遗传算法优化SVM的轧制力预报[C]//中国自动化学会控制理论专业委员会. 第37届中国控制会议论文集(B), 2018: 198-202.
[13] FUREY T, CRISTIANINI N, DUFFY N, et al. Support vector machine classification and validation of cancer tissue samples using microarray exptession data[J]. Bioinformaics, 2000, 16(10): 906-914
[14] CORTES C, VAPNIK V. Support vector network[J]. Machine Learning, 1995(20): 273-297
[15] 姜红, 朱晓晗, 何歆沂, 等. 差分拉曼光谱法对塑料瓶的鉴别分析[J]. 中国测试, 2020, 46(8): 76-79, 100