工程科学学报，第39卷.第1期：39-47,2017年1月 Chinese Journal of Engineering,Vol.39,No.1:39-47,January 2017 D0L:10.13374/j.issn2095-9389.2017.01.005;htp://journals.usth.edu.cm 基于全局优化支持向量机的多类别高炉故障诊断 张海刚12)，张森12)四，尹怡欣12) 1)北京科技大学自动化学院，北京1000832)北京科技大学钢铁流程先进控制教育部重点实验室，北京100083 ☒通信作者，E-mail:zhangsen(@usth.cdu.cn 摘要针对高炉故障诊断系统快速性和准确性的要求，提出基于全局优化最小二乘支持向量机的策略.首先，采用变尺度 离散粒子群对最小二乘支持向量机的参数和故障特征的选取进行优化：然后，利用核主元分析法对选取的特征向量进行压缩 整理：最后，构造了以Fisher线性判别率为标准的启发式纠错输出编码.仿真结果表明，通过对故障训练样本有意义地分割重 组，用较少的最小二乘支持向量机分类器，得到较高的故障判断准确率且增强了整个系统的实时性 关键词高炉：故障诊断：最小二乘分析：支持向量机：全局优化 分类号T549 Multi-class fault diagnosis of BF based on global optimization LS-SVM ZHANG Hai-gang2),ZHANG Sen)YIN Yi-xin2) 1)School of Automation and Electrical Engineering,University of Science and Technology Beijing,Beijing 100083.China 2)Key Laboratory of Advanced Control of Iron and Steel Process (Ministry of Education),University of Science and Technology Beijing,Beijing 100083,China Corresponding author,E-mail:zhangsen@ustb.edu.cn ABSTRACT Aiming at the requirement of high speed and precision in blast furnace fault diagnosis systems,a new strategy based on global optimization least-squares support vector machines (LS-SVM)was proposed to solve this problem.Firstly,the variable metric discrete particle swarm optimization algorithm was employed to optimize the feature selection and LS-SVM parameters.Secondly,the feature vector was compressed by kernel principal component analysis.Finally,the heuristic error correcting output codes were con- structed on the basis of Fisher linear discriminate rate.In the fault diagnosis scheme,fewer IS-SVM classifiers were applied through meaningful partitions and recombination of fault training samples.Simulation results show that the proposed fault diagnosis method can not only improve the fault detection accurate rate,but also enhance the timeliness of the entire system. KEY WORDS blast furnaces;fault diagnosis;least-squares analysis;support vector machines;global optimization 钢铁工业是我国国民经济的基础产业和支柱产 到节能减排的目的.高炉炉况故障诊断在高炉自动化 业).高炉炼铁在钢铁工业中处于举足轻重的地位. 控制的研究中一直是热点话题，准确及时的故障诊断 高炉炼铁系统生产设备繁多，具有多耦合、大延时、非 技术能够确定高炉的稳定生产，降低故障所带来的经 线性等特点.尽管在高炉本体上安装了很多自动化的 济损失[2) 检测装置，然而由于高炉运行炉况复杂，无法建立准确 在高炉生产中，由于缺少准确的机理模型，往往从 的机理模型，在高炉自动控制决策过程中，仍将其当为 数据驱动角度建立高炉故障诊断模型).基于专家系 “黑箱”系统进行处理.高炉生产追求稳定，稳定炉况 统的故障诊断方法，计算机模仿专家经验，进行故障决 不仅能够保证铁水质量，而且能够提高煤气利用率，达 策[.虽然引进国外的专家系统有一定的效果，但是 收稿日期：2016-03-16 基金项目：国家自然科学基金资助项目(61333002,61673056)工程科学学报,第 39 卷,第 1 期:39鄄鄄47,2017 年 1 月 Chinese Journal of Engineering, Vol. 39, No. 1: 39鄄鄄47, January 2017 DOI: 10. 13374 / j. issn2095鄄鄄9389. 2017. 01. 005; http: / / journals. ustb. edu. cn 基于全局优化支持向量机的多类别高炉故障诊断 张海刚1,2) , 张 森1,2) 苣 , 尹怡欣1,2) 1) 北京科技大学自动化学院, 北京 100083 2) 北京科技大学钢铁流程先进控制教育部重点实验室, 北京 100083 苣 通信作者, E鄄mail: zhangsen@ ustb. edu. cn 摘 要 针对高炉故障诊断系统快速性和准确性的要求,提出基于全局优化最小二乘支持向量机的策略. 首先,采用变尺度 离散粒子群对最小二乘支持向量机的参数和故障特征的选取进行优化;然后,利用核主元分析法对选取的特征向量进行压缩 整理;最后,构造了以 Fisher 线性判别率为标准的启发式纠错输出编码. 仿真结果表明,通过对故障训练样本有意义地分割重 组,用较少的最小二乘支持向量机分类器,得到较高的故障判断准确率且增强了整个系统的实时性. 关键词 高炉; 故障诊断; 最小二乘分析; 支持向量机; 全局优化 分类号 TF549 Multi鄄class fault diagnosis of BF based on global optimization LS鄄SVM ZHANG Hai鄄gang 1,2) , ZHANG Sen 1,2) 苣 , YIN Yi鄄xin 1,2) 1) School of Automation and Electrical Engineering, University of Science and Technology Beijing, Beijing 100083, China 2) Key Laboratory of Advanced Control of Iron and Steel Process ( Ministry of Education), University of Science and Technology Beijing, Beijing 100083, China 苣 Corresponding author, E鄄mail: zhangsen@ ustb. edu. cn ABSTRACT Aiming at the requirement of high speed and precision in blast furnace fault diagnosis systems, a new strategy based on global optimization least鄄squares support vector machines (LS鄄SVM) was proposed to solve this problem. Firstly, the variable metric discrete particle swarm optimization algorithm was employed to optimize the feature selection and LS鄄SVM parameters. Secondly, the feature vector was compressed by kernel principal component analysis. Finally, the heuristic error correcting output codes were con鄄 structed on the basis of Fisher linear discriminate rate. In the fault diagnosis scheme, fewer LS鄄SVM classifiers were applied through meaningful partitions and recombination of fault training samples. Simulation results show that the proposed fault diagnosis method can not only improve the fault detection accurate rate, but also enhance the timeliness of the entire system. KEY WORDS blast furnaces; fault diagnosis; least鄄squares analysis; support vector machines; global optimization 收稿日期: 2016鄄鄄03鄄鄄16 基金项目: 国家自然科学基金资助项目(61333002, 61673056) 钢铁工业是我国国民经济的基础产业和支柱产 业[1] . 高炉炼铁在钢铁工业中处于举足轻重的地位. 高炉炼铁系统生产设备繁多,具有多耦合、大延时、非 线性等特点. 尽管在高炉本体上安装了很多自动化的 检测装置,然而由于高炉运行炉况复杂,无法建立准确 的机理模型,在高炉自动控制决策过程中,仍将其当为 “黑箱冶系统进行处理. 高炉生产追求稳定,稳定炉况 不仅能够保证铁水质量,而且能够提高煤气利用率,达 到节能减排的目的. 高炉炉况故障诊断在高炉自动化 控制的研究中一直是热点话题,准确及时的故障诊断 技术能够确定高炉的稳定生产,降低故障所带来的经 济损失[2] . 在高炉生产中,由于缺少准确的机理模型,往往从 数据驱动角度建立高炉故障诊断模型[3] . 基于专家系 统的故障诊断方法,计算机模仿专家经验,进行故障决 策[4] . 虽然引进国外的专家系统有一定的效果,但是