Quantitative Recognizing Dissolved Hydrocarbons with Genetic Algorithm-Support Vector Regression

Online monitoring of dissolved fault characteristic hydrocarbon gases, such as methane, ethane, ethylene and acetylene in power transformer oil has significant meaning for condition assessment of transformer. Recently, semiconductor tin oxide based gas sensor array has been widely applied in online monitoring apparatus, while cross sensitivity of the gas sensor array is inevitable due to same compositions and similar structures among the four hydrocarbon gases. Based on support vector regression (SVR) with genetic algorithm (GA), a new pattern recognition method was proposed to reduce the cross sensitivity of the gas sensor array and further quantitatively recognize the concentration of dissolved hydrocarbon gases. The experimental data from a certain online monitoring device in China is used to illustrate the performance of the proposed GA-SVR model. Experimental results indicate that the GA-SVR method can effectively decrease the cross sensitivity and the regressed data is much more closed to the real values. DOI:  http://dx.doi.org/10.11591/telkomnika.v11i9.2776

[1]  Mingcang Zhu,et al.  Housing price forecasting based on genetic algorithm and support vector machine , 2011, Expert Syst. Appl..

[2]  Fan Yang,et al.  Support vector machine with genetic algorithm for machinery fault diagnosis of high voltage circuit breaker , 2011 .

[3]  Mingjun Wang,et al.  Particle swarm optimization-based support vector machine for forecasting dissolved gases content in power transformer oil , 2009 .

[4]  Jin Luo,et al.  Pattern recognition for sensor array signals using Fuzzy ARTMAP , 2009 .

[5]  Minhat Ade Erawan,et al.  Comparison Studies of Electrical Discharge Machining (EDM) Process Model for Low Gap Current , 2012 .

[6]  Ning Wang,et al.  Design of On-Line Monitoring System for Transformer Oil Chromatogram and its Data Analysis , 2012 .

[7]  K. D. Srivastava,et al.  Review of condition assessment of power transformers in service , 2002 .

[8]  Francesco Tortorella,et al.  Classification of food, beverages and perfumes by WO3 thin-film sensors array and pattern recognition techniques , 2001 .

[9]  Tao Huang,et al.  Patent classification system using a new hybrid genetic algorithm support vector machine , 2010, Appl. Soft Comput..

[10]  S Singh,et al.  Dissolved gas analysis technique for incipient fault diagnosis in power transformers: A bibliographic survey , 2010, IEEE Electrical Insulation Magazine.

[11]  Jin Luo,et al.  Nanoparticle-structured sensing array materials and pattern recognition for VOC detection , 2005 .

[12]  Gregory W. Kauffman,et al.  Pattern recognition analysis of optical sensor array data to detect nitroaromatic compound vapors , 2001 .

[13]  E Gockenbach,et al.  Intelligent agent-based system using dissolved gas analysis to detect incipient faults in power transformers , 2010, IEEE Electrical Insulation Magazine.

[14]  Jin Luo,et al.  A multi-module artificial neural network approach to pattern recognition with optimized nanostructured sensor array , 2006 .

[15]  Sun Cai-xin,et al.  Analysis of Infrared Absorption Properties of Dissolved Gases in Transformer Oil , 2008 .

[16]  Chia-Jung Lu,et al.  Utilizing diversified properties of monolayer protected gold nano-clusters to construct a hybrid sensor array for organic vapor detection , 2012 .

[17]  Francesco Tortorella,et al.  Gas recognition by activated WO3 thin-film sensors array , 2001 .

[18]  Gregory P. Harmer,et al.  Semiconducting metal oxide sensor array for the selective detection of combustion gases , 2003 .

[19]  R. Penner,et al.  Trace detection of dissolved hydrogen gas in oil using a palladium nanowire array. , 2011, Analytical chemistry.