基于QCSSA-LSTSVR的氧化铝质量预测模型  

Alumina Quality Prediction Model Based on QCSSA-LSTSVR

作　　者：徐辰华 陈瑞[2] 宋海鹰 程若军 何俊隆 宋绍剑 XU Chen-hua;CHEN Rui;SONG Hai-ying;CHENG Ruo-jun;HE Jun-long;SONG Shao-jian(School of Automation,Guangdong Polytechnic Normal University,Guangzhou 510080,China;School of Electrical Engineering,Guangxi University,Nanning 530004,China)

机构地区：[1]广东技术师范大学自动化学院,广东广州510080 [2]广西大学电气工程学院,广西南宁530004

出　　处：《控制工程》2022年第10期1857-1865,共9页Control Engineering of China

基　　金：校级科研项目人才专项项目(2021SDKYA118);广西自然科学基金资助项目(2017GXNSFAA198225)。

摘　　要：针对氧化铝焙烧过程氧化铝质量指标检测滞后的问题,提出融合最小二乘孪生支持向量机(LSTSVR)与量子混沌樽海鞘算法(QCSSA)方法,建立一种氧化铝焙烧过程的质量指标预测模型。首先,利用LSTSVR建立氧化铝质量指标预测模型;其次,针对LSTSVR模型中核宽度系数和惩罚因子选取困难的问题,采用QCSSA进行LSTSVR模型结构参数寻优,利用Logistic混沌策略和量子局部搜索策略来提高SSA的全局寻优能力;最后,利用实际生产数据对所提方法进行实验验证。仿真结果表明,QCSSA优化LSTSVR的方法具有较好的预测效果。To address the problem of the lag in the detection of alumina quality index of the alumina sintering process, a model combining least square twin support vector regression(LSTSVR) with quantum chaotic salp swarm algorithm(QCSSA) is proposed, for predicting the quality of alumina sintering process. Firstly, the prediction model of alumina quality index is established by using LSTSVR. Secondly, for the problem of difficult selection of kernel width coefficients and penalty factors in the LSTSVR model, QCSSA is applied to optimize the structure parameters of LSTSVR model. Logistic chaos strategy and quantum local search strategy are adopted to improve the global optimization ability of SSA. Finally, the proposed method is confirmed by actual production data, and the simulation results show that the method of QCSSA optimizing LSTSVR has good prediction accuracy.

关 键 词：质量指标预测 樽海鞘算法 最小二乘孪生支持向量机 Logistics混沌映射 量子局部搜索 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]