A multiclass SVM-based classifier for transformer fault diagnosis using a particle swarm optimizer with time-varying acceleration coefficients

Abstract

SUMMARY A multiclass support vector machine (SVM) classifier based upon particle swarm optimization (PSO) with time‐varying acceleration coefficients for fault diagnosis of power transformers is proposed in this paper. The one‐against‐one combination scheme is adopted to extend SVM for settling the multiclass classification problem. The algorithm of PSO with time‐varying acceleration coefficients (PSO‐TVAC) is employed to optimize the parameters for SVM. The results show that the convergence of the PSO‐TVAC algorithm is relatively faster and much more precise than that of the classical PSO. Moreover, compared with other reference diagnosis approaches, the results demonstrate the improved classification accuracy based upon the proposed approach and show it can be used as an effective tool for fault diagnosis of power transformers. Copyright © 2011 John Wiley & Sons, Ltd.

A New Support Vector Machine Model Based on Improved Imperialist Competitive Algorithm for Fault Diagnosis of Oil-immersed Transformers

“…The data in the training set can be considered as a separable set only when the training set is bound to satisfy the formula (4) [19].…”

“…Fault diagnosis of oilimmersed transformers based on DGA is often considered as a multiple class problem [18,19], therefore the proposed IICASVM algorithm described in section 3 was explored for fault diagnosis of transformers in the paper. This paper collects DGA data from several electric power companies to build fault types of models.…”

“…Those methods include IEC criteria [6], RBFNN, SVM, PSOSVM [19] and IICASVM. Since OAO scheme is proofed to be the best multiple classification scheme in section 4, the OAO scheme is applied in PSOSVM and IICASVM.…”

“…The main drawback of these approaches is that the observed gas ratios do not match all the fault types [9,10]. To improve the drawback and to dig the potential fault law from the gases content, some new fault diagnosis methods, including fuzzy logic [11], Dempster Shafer theory [12], grey clustering [13] and rough set [14], self-adaptive RBF neural network [15], auto-associative neural networks and mean shift [16], evolutionary wavelet neural network based on genetic algorithm [17], and support vector machine (SVM) based methods [18][19][20], have been proposed in the recent years. In these methods, SVM is an effective method and is wildly applied in classification problems (fault diagnosis of transformers based on DGA is often considered as a classification problem) [21][22] because its advantages of small sample learning, global optimization, and structural risk minimization [23][24].…”

“…In general, the performance of the support vector machine (SVM) have a significant impact factors, including the kernel function, the punishment coefficient C, and kernel parameter γ, which are uncertain and should be optimized by artificial intelligence (AI) approaches. Some AI methods such as clonal selection algorithms [18], PSO [19], GA [20], were used to optimize the two parameters in the classification problems and receive a good performance.…”