作者：金星, 邵珠超, 王盛慧
作者单位：长春工业大学电气与电子工程学院, 吉林 长春 130012
关键词：水泥回转窑;故障诊断;主成分分析;支持向量机;改进的灰狼算法
摘要：
为实现水泥回转窑故障的精确诊断，提出一种基于主成分分析（PCA）和支持向量机（SVM）的回转窑故障诊断模型。通过引入差分进化（DE）算法的变异、交叉、选择操作来维持种群的多样性，克服灰狼算法易早熟收敛的缺陷，然后采用这种改进的灰狼算法（IGWO）对SVM的惩罚因子c和核函数参数g进行动态的寻优。运用PCA对采集数据进行降维处理，消除非相关因素，降低数据处理难度，然后将特征提取后的数据作为输入建立故障诊断模型，并与普通的SVM建模方法进行比较。实例表明：在有用信息量损失较小的前提下，分类准确率达到96.153 8%，模型构建时间为2.972 0 s，从而验证模型的准确性和高效性。
Research of fault diagnosis of cement rotary kiln based on PCA and IGWO-SVM
JIN Xing, SHAO Zhuchao, WANG Shenghui
College of Electrical and Electronic Engineering, Changchun University of Technology, Changchun 130012, China
Abstract: In order to precisely diagnose the faults of cement rotary kiln,a fault diagnosis model based on principal component analysis (PCA) and support vector machine (SVM) is proposed.The variation,crossover and selection operation of differential evolution (DE) are used to maintain the diversity of the population and are introduced into GWO to avoid premature convergence.Improved gray wolf optimizer (IGWO) is used to dynamically optimize the penalty factor (c) and the kernel function parameter (g) of SVM model.PCA is used to reduce the dimension of the collected data,eliminate the irrelevant factors and reduce the difficulty of data processing.Then the data after feature extraction are used as the inputs to establish the fault diagnosis model and the built IGWO-SVM model is compared with the general SVM model.The experiment shows that under the condition of less useful information loss,the classification precision is 96.153 8% and the model building time reaches 2.972 0 s,which verifies the accuracy and high efficiency of the IGWO-SVM model.
Keywords: cement rotary kiln;fault diagnosis;PCA;SVM;IGWO
2017, 43(10): 92-96 收稿日期: 2016-12-08;收到修改稿日期: 2017-01-23
基金项目: 吉林省科学技术厅计划项目（20150203003SF）
作者简介: 金星(1976-),男,吉林长春市人,副教授,硕导,研究方向为测控技术与智能系统。
参考文献
[1] ABDELAZIZ E A, SAIDUR R. A review on energy saving strategies in industrial sector[J]. Renewable and Sustainable Energy Reviews,2011,15(1):150-168.
[2] 李轩. 基于贝叶斯网络的水泥回转窑故障诊断模型的研究[D]. 秦皇岛:燕山大学,2014.
[3] 王兰军. 水泥回转窑故障诊断系统的研究[D]. 杭州:浙江大学,2004.
[4] 高绪伟. 核PCA特征提取方法及其应用研究[D]. 南京:南京航空航天大学,2009.
[5] 钟小凤,贺德强,苗剑. 基于PCA-LSSVM的机车走行部滚动轴承故障诊断研究[J]. 广西大学学报(自然科学版),2014,39(2):328-333.
[6] SAHIN F, YAVUZ M C, ARNAVUT Z, et al. Fault diagnosis for airplane engines using Bayesian networks and distributed particle swarm optimization[J]. Parallel Computing,2007,33(2):124-143.
[7] BACHA K, SOUAHLIA S, GOSSA M. Power transformer fault diagnosis based on dissolved gas analysis by support vector machine[J]. Electric Power System Research,2012,83(1):73-79.
[8] 吴翊,李永乐. 应用数理统计[M]. 长沙:国防科技大学出版社,2012:290-292.
[9] WIDODO A, YANG B S. Support vector machine in machine condition monitoring and fault diagnosis[J]. Mechanical Systems and Signal Processing,2007,21(6):2560-2574.
[10] MIRJALILI S, MIRJALILI S M, LEWIS A. Grey wolf optimizer[J]. Advances in Engineering Software,2014, 69(7):46-61.
[11] 罗佳,唐斌. 基于收敛因子非线性动态变化的灰狼优算法[J]. 中国科技论文,2016,11(17):1991-1997.
[12] 姚鹏,王宏伦. 基于改进流体扰动算法与灰狼优化的无人机三维航路规划[J]. 控制与决策,2014,31(4):701-708.
[13] 邱东,张楠,赵晨旭,等. 基于灰色神经网络的中低碳烙铁终点硫含量预报模型研究[J]. 中国测试,2014,40(4):68-70.
