Prediction Model for Dissolved Gas Concentration in Transformer Oil Based on Modified Grey Wolf Optimizer and LSSVM with Grey Relational Analysis and Empirical Mode Decomposition

Abstract

Oil-immersed transformer is one of the most important components in the power system. The dissolved gas concentration prediction in oil is vital for early incipient fault detection of transformer. In this paper, a model for predicting the dissolved gas concentration in power transformer based on the modified grey wolf optimizer and least squares support vector machine (MGWO-LSSVM) with grey relational analysis (GRA) and empirical mode decomposition (EMD) is proposed, in which the influence of transformer load, oil temperature and ambient temperature on gas concentration is taken into consideration. Firstly, GRA is used to analyze the correlation between dissolved gas concentration and transformer load, oil temperature and ambient temperature, and the optimal feature set affecting gas concentration is extracted and selected as the input of the prediction model. Then, EMD is used to decompose the non-stationary series data of dissolved gas concentration into stationary subsequences with different scales. Finally, the MGWO-LSSVM is used to predict each subsequence, and the prediction values of all subsequences are combined to get the final result. DGA samples from two transformers are used to verify the proposed method, which shows high prediction accuracy, stronger generalization ability and robustness by comparing with LSSVM, particle swarm optimization (PSO)-LSSVM, GWO-LSSVM, MGWO-LSSVM, EMD-PSO-LSSVM, EMD-GWO-LSSVM, EMD-MGWO-LSSVM, GRA-EMD-PSO-LSSVM and GRA-EMD-GWO-LSSVM.

Determinación de fallas en transformadores de potencia inmersos en aceite mineral aislante basándose exclusivamente en el DGA y mediante la evaluación del NEI

“…Para el caso del acetileno se observa que los transformadores 6, 7, 8 y 24 se encuentran en la condición 2 con concentraciones en el rango de 2-9 ppm, mientras que los transformadores [1][2][3][4][5] y [9][10][11][12][13][14][15][16][17][18][19][20][21][22][23] se encuentran en la condición 1 con menos de 1 ppm. Para etileno, los transformadores 6, 7, 8 y 24 presentan elevación en su concentración, donde el 6,7 y 8 están en la condición 2 en el rango de 51-100 ppm, mientras que el 24 se encuentra en la condición 3 en el rango de 101-200 ppm.…”

“…En caso del etano, se observa que el transformador 7 se ubica en la condición 2 en el rango de 66-100 ppm, el 6, 8 y 24 se encuentran en la condición 3 en el rango de 101-150 ppm, de igual manera, se observa un incremento significativo en el transformador 15 con concentración mayor a 150 ppm, ubicándose en la condición 4. Los transformadores [1][2][3][4][5], [9][10][11][12][13][14] y [16][17][18][19][20][21][22][23] están en la condición 1 con menos de 65 ppm. Se observa que, para el metano, los transformadores 7 y 15 presentan concentraciones entre 121-400 ppm lo que los ubica en la condición 2 y los transformadores [1][2][3][4][5][6], [8][9][10][11][12][13][14] y [16][17][18][19][20][21][22][23][24] se ubican en la condición 1 con concentraciones menores a 1 ppm.…”

“…Una falla catastrófica en un transformador puede terminar en un accidente, el sistema puede ser dañado generando grandes pérdidas económicas y daños sociales, por lo que resulta importante detectar las fallas potenciales en esos equipos [1].…”

Review of Fiber Optic Diagnostic Techniques for Power Transformers

“…Recently the modified grey wolf optimizer and least squares support vector machine (MGWO-LSSVM) with grey relational analysis (GRA) and empirical mode decomposition (EMD) [72].…”

An Interval Forecasting Model Based on Phase Space Reconstruction and Weighted Least Squares Support Vector Machine for Time Series of Dissolved Gas Content in Transformer Oil

“…In terms of data preprocessing, some scholars decompose the original sequence based on the nonlinear and nonstationary characteristics of the time series. Zeng et al [52] used the empirical mode decomposition (EMD) method to process the DGA data and decompose nonstationary signals into characteristic frequency function components of different frequencies, and modeled each subsequence component separately, reconstructing its prediction by superposition. As a result, the combined prediction results, which meet the accuracy requirements, were obtained.…”