Application of Optimized Parameters SVM Based on Photoacoustic Spectroscopy Method in Fault Diagnosis of Power Transformer
ZHANG Yu-xin1, CHENG Zhi-feng2, XU Zheng-ping2, BAI Jing1*
1. School of Electronic and Information Engineering, Beihua University, Jilin 132021, China 2. Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, Changchun 130033, China
Abstract：In order to solve the problems such as complex operation, consumption for the carrier gas and long test period in traditional power transformer fault diagnosis approach based on dissolved gas analysis(DGA) , this paper proposes a new method which is detecting 5 types of characteristic gas content in transformer oil such as CH4, C2H2, C2H4, C2H6 and H2 based on photoacoustic spectroscopy and C2H2/C2H4, CH4/H2, C2H4/C2H6 three-ratios data are calculated. The support vector machine model was constructed using cross validation method under five support vector machine functions and four kernel functions, heuristic algorithms were used in parameter optimization for penalty factor c and g，which to establish the best SVM model for the highest fault diagnosis accuracy and the fast computing speed. Particles swarm optimization and genetic algorithm two types of heuristic algorithms were comparative studied in this paper for accuracy and speed in optimization. The simulation result shows that SVM model composed of C-SVC, RBF kernel functions and genetic algorithm obtain 97.5% accuracy in test sample set and 98.333 3% accuracy in train sample set, and genetic algorithm was about two times faster than particles swarm optimization in computing speed. The methods described in this paper has many advantages such as simple operation, non-contact measurement, no consumption for the carrier gas, long test period, high stability and sensitivity, the result shows that the methods described in this paper can instead of the traditional transformer fault diagnosis by gas chromatography and meets the actual project needs in transformer fault diagnosis.
[1] CHEN Wei-gen, YUN Yu-xin, PAN Chong, et al(陈伟根，云玉新，潘 翀，等). Automation of Electric Power Systems(电力系统自动化)，2007，31(15)：94. [2] YUN Yu-xin, ZHAO Xiao-xiao, CHEN Wei-gen, et al(云玉新，赵笑笑，陈伟根，等). High Voltage Engineering(高电压技术)，2009，35(9)：2156. [3] CHEN Wei-gen, ZHOU Heng-yi, HUANG Hui-xian, et al(陈伟根，周恒逸，黄会贤，等). Chinese Jouranl of Scientific Instrument(仪器仪表学报)，2010，31(3)：665. [4] ZHANG Yu-xin, LIU Yu(张玉欣，刘 宇). Applied Mechanics and Materials, 2014, 448-453: 3605. [5] DUAN Hui-da, TIAN Yan-tao, LI Jin-song, et al(段慧达，田彦涛，李津淞，等). Control and Decision(控制与决策)，2012, 27(2): 216. [6] LIU De-jun, TIAN Yan-tao, ZHANG Lei(刘德君，田彦涛，张 雷). Control and Decision(控制与决策)，2012, 27(12): 1890. [7] LIU De-jun, TIAN Yan-tao, ZHANG Lei(刘德君，田彦涛，张 雷). Journal of Mechanical Engineering(机械工程学报)，2011, 47(19)：14. [8] CHEN Ming(陈 明). Neural Networks 43 Case Studies Based On MATLAB (Neural Network Theory and Examples Based on MATLAB(MATLAB神经网络原理与实例精解). Beijing：QingHua University Press(北京：清华大学出版社), 2013. 10. [9] WANG Xiao-chuan, SHI Feng, YU Lei, et al(王小川，史 峰，郁 磊，等). Neural Networks 43 Case Studies Based on MATLAB (MATLAB神经网络43个案例分析). Beijing：Tsinghua University Press(北京：北京航空航天大学出版社), 2013. 102. [10] DU Yan-li, TANG Zhi-guo, LI Yuan-chun(杜艳丽，唐志国，李元春). Journal of Jilin University (Engineering and Technology Edition)(吉林大学学报·工学版), 2013, 43(2): 466.
