基于S-PSO分类算法的故障诊断方法
收稿日期: 2014-12-15
修回日期: 2015-01-13
网络出版日期: 2015-01-23
基金资助
国家自然科学基金民航联合研究基金(U1233202);中国民用航空飞行学院青年基金(Q2013-049)
Fault diagnosis method based on S-PSO classification algorithm
Received date: 2014-12-15
Revised date: 2015-01-13
Online published: 2015-01-23
Supported by
Joint Fund for Civil Aviation Research of National Natural Science Foundation of China (U1233202);Youth Foundation of Civil Aviation Flight University of China (Q2013-049)
将监控数据的已知状态作为先验类别标签,构造出新的有监督的粒子群优化(S-PSO)分类算法,并对设备进行故障诊断。为提高故障诊断的准确率,降低随机性对分类算法的影响,提出了新的基于动态邻域的自适应探测更新(ADU-DN)的干预更新策略来拓展粒子搜索整个解空间的能力,引导粒子自适应地跳出局部最优区域,确保获得全局最优解;同时设计出基于最小类内距离、最大类间距离和训练样本最大分类精度的适应度函数,使得输出的最优类别中心兼顾了这3个因素,增强了分类算法在故障诊断中的通用性和容错性,提高了测试样本的分类精度。S-PSO分类算法有效克服了聚类算法只考虑数据间相似性特征、不考虑数据蕴含的物理意义以及不能很好指导样本分类的缺陷。对GE90发动机孔探图像纹理特征分类进行了对比研究,研究数据表明:S-PSO分类算法表现出了较强的鲁棒性,在故障诊断中的分类精度高于支持向量机(SVM)和常用神经网络模型。
关键词： 监督的粒子群优化分类算法; 动态邻域; 自适应探测更新; 适应度函数; 故障诊断
郑波 , 高峰 . 基于S-PSO分类算法的故障诊断方法[J]. 航空学报, 2015 , 36(11) : 3640 -3651 . DOI: 10.7527/S1000-6893.2015.0020
On the basis of taking the known states of monitoring data as prior class labels, a novel supervised particle swarm optimization (S-PSO) classification algorithm is constructed and used in fault diagnosis. In order to improve the accuracy of fault diagnosis and reduce the influence of randomness on the classification algorithm, a novel intervention updating strategy named an adaptive detecting updating based on dynamic neighborhood (ADU-DN) is proposed to expand the particles' ability of searching the entire solution space and guide the particles' adaptively jumping out of the local optimal region, ensuring to obtain the global optimal solution. Meanwhile, a fitness function based on minimum distance of intra-class, maximum distance of inter-class and maximum classification precision of train samples is designed to make these three factors constraint the output optimal class centers, enhance classification algorithm's generality and fault-tolerant ability in fault diagnosis and increase the classification precision of test samples. The S-PSO classification algorithm overcomes the defects of the clustering algorithms that only consider the similarity features of data but not the physical meanings, and do not guide the classification of samples well. A comparison study on the GE90 engine borescope image texture feature classification is carried out and the research data show that S-PSO classification algorithm has a strong robustness and the classification precision is higher than the support vector machine (SVM) and the common neural network model.
[1] Xiong Z, Shao H, Hua B, et al. A new fault-tolerant federated filter without fault isolation[J]. Acta Aeronautica et Astronautica Sinica, 2014, 35(5):918-926(in Chinese).熊智,邵慧,华冰,等.改进故障隔离的容错联邦滤波新方法[J].航空学报, 2014, 35(5):918-926.
[2] Mangat V, Vig R. Novel associative classifier based on dynamic adaptive PSO:Application to determining candidates for thoracic surgery[J]. Expert Systems with Applications, 2014, 41(18):8234-8244.
[3] Martinez-Soto R, Castillo O, Aguilar L. Type-1 and type-2 fuzzy logic controller design using a hybrid PSO-GA optimization method[J]. Information Sciences, 2014, 285:34-49.
[4] Kibria S, Islam M T, Yatim B, et al. A modified PSO technique using heterogeneous boundary conditions for broadband compact microstrip antenna designing[J]. Annals of Telecommunications-Annales Des Telecommunications, 2014, 69(9):509-541.
[5] Ghanei A, Assareh E, Biglari M, et al. Thermal-economic multi-objective optimization of shell and tube heat exchanger using particle swarm optimization (PSO)[J]. Heatand Mass Transfer, 2014, 50(10):1375-1384.
[6] Ma A X, Li Y J, Cao Y Y, et al. Intelligent diagnosis for aircraft engine wear failure based on immune theroy[J]. Acta Aeronautica et Astronautica Sinica, 2014, 35(11):1718-1727(in Chinese).马安详,李艳军,曹愈远,等.基于免疫理论的航空发动机磨损故障智能诊断[J].航空学报, 2014, 35(11):1718-1727.
[7] Cagnina L, Errecalde M, Ingaramo D, et al. An efficient particle swarm optimization approach to cluster short texts[J]. Information Sciences, 2014, 265:36-49.
[8] Lam Y K, Tsang P W M, Leung C S. PSO-based K-means clustering with enhanced cluster matching for gene expression data[J]. Neural Computing & Applications, 2013, 22(7-8):1349-1355.
[9] Avanija J, Ramar K. A hybrid approach using PSO and K-means for semantic clustering of WEB documents[J]. Journal of WEB Engineering, 2013, 12(3-4):249-264.
[10] Zhang J, Wang Y P, Feng J H. A hybrid clustering algorithm based on PSO with dynamic crossover[J]. Soft Computing, 2014, 18(5):961-979.
[11] Fathi V, Montazer G A. An improvement in RBF learning algorithm based on PSO for real time applications[J]. Neurocomputing, 2013, 111:169-176.
[12] Guraksin G E, Hakli H, Uguz H. Support vector machines classification based on particle swarm optimization for bone age determination[J]. Applied Soft Computing, 2014, 24:597-602
[13] Zheng B. Investigation on aeroengine maintenance level decision based on PSO-SVM[J]. Journal of Propulsion Technology, 2013, 34(5):687-692(in Chinese).郑波.基于PSO-SVM的民航发动机送修等级决策研究[J].推进技术, 2013, 34(5):687-692.
[14] Li Y B, Li Q H, Huang X H, et al. Research on gas fault fusion diagnosis of aero-engine component[J]. Acta Aeronautica et Astronautica Sinica, 2014, 35(6):1614-1622(in Chinese).李业波,李秋红,黄向华,等.航空发动机气路部件故障融合诊断方法研究[J].航空学报, 2014, 35(6):1614-1622.
[15] Tang Y. Research on aero-engine interior damage evaluation and maintenance decisions based on internet[D]. Nanjing:Nangjing University of Aeronautics and Astronautics, 2007(in Chinese).汤洋.基于Internet的民航发动机内部损伤评估与维修决策研究[D].南京:南京航空航天大学, 2007.
[16] Bergh V D. An analysis of particle swarm optimizers[D]. Pretoria:University of Pretoria, 2002.
[17] Zheng B, Gao F. Research on the prediction of aeroengine wear based on IPSO-SVR[J]. Lubrication Engineering, 2014, 39(11):81-87(in Chinese).郑波,高峰.基于IPSO-SVR的航空发动机磨损预测研究[J].润滑与密封, 2014, 39(11):81-87.
[18] Jagannath S, Ganapati P. Automatic clustering algorithm based on multi-objective immunized PSO to classify actions of 3D human models[J]. Engineering Applications of Artificial Intelligence, 2013, 26(5-6):1429-1411.
[19] Zhao J. Research on some issues of quantum-behaved particle swarm optimization algorithm and its application[D]. Wuxi:Jiangnan University, 2013(in Chinese).赵晶.量子行为粒子群优化算法及其应用问题中的若干问题研究[D].无锡:江南大学, 2013.
[20] Kiranyaz S, Ince T, Yildirim A. Evolutionary artificial neural networks by multi-dimensional particle swarm optimization[J]. Neural Networks, 2009, 22(10):1448-1462.
/
| 〈 |
|
〉 |
CJA