Time-series prediction technology plays a significant role in evaluating the health status of power transformers and forecasting inchoate operation failure. This study presents a variational mode decomposition (VMD) model with the application of autoregressive integrated moving average (ARIMA) to develop a prediction technology for the dissolved gas analysis (DGA) time series. In addition, the two-step stationary test method regulates the nonstationary of the time series. The VMD decomposes the DGA time series into subcomponents to reduce the nonstationary, where the boost marine predators algorithm (BMPA) optimizes the parameters <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${K}$ </tex-math></inline-formula> and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\alpha $ </tex-math></inline-formula> of the VMD to retain the useful information of the original signal. The ARIMA predicts the future result for each subcomponent, and the ARIMA parameter is optimized by the BMPA for optimal solutions of the Bayesian information criterion (BIC). The performance of the proposed approach is measured by predicting the health status of the diurnal new values of the DGA of a power transformer. The experiment result shows that the proposed model exhibits high efficacy in predicting the DGA time series and obtains the health status of the power transformer.
Read full abstract