Prediction Of The Gas Dissolved In Power Transformer Oil Using Multivariable Model And Combined Model

The power transformer is a major apparatus in a power system. It is of great importance to detect incipient failures in power transformers as early as possible in order to minimize system outage. Therefore, the prediction of gas dissolved in transformer oil is investigated in this paper. The contributions and conclusions are made as followings:The development of the gas dissolved in power transformer oil is analyzed. An overview and classification of conventional forecasting models are presented. The suitable range and applications of these models are also shown. Moreover, the trend of dissolved gas prediction is discussed. The great importance of using multivariable forecasting and combined forecasting is proposed.There are many factors influencing gas dissolved in transformer oil. Among these factors, there are not only the definite ones but also the uncertain ones, i.e. 'grey'. So it is feasible to apply grey model to predict gas dissolved in transformer oil. The discrete grey model (DGM (1, 1) model) is the precise form of the GM (1, 1) model, so it has better precise and stability. Therefore, it is suggested to replace the GM (1,1) model in gas-in-oil prediction.For precise and reliable fault detection, it is essential to consider simultaneously the changes in several gases dissolved in oil. Since these gases are from the same oil, their changes are interrelated. The multivariable grey model-MGM (1, n) and BP neural networks can describe the changes in each gas from the views of systems, indicating their coupling relationship, therefore it is highly suitable to predict the failures in power transformers.There are various trends in the development of the gases in transformer oil, so the prediction transformer failures should be performed by means of combined models. The optimally combined model, variable weight combined model based on BP neural networks and grey-time series combined forecasting model are used to predict the gas dissolved in transformer oil. The effectiveness and practicability of these models is verified by some examples.An efficient failure forecasting procedure for power transforms is developed using above mentioned prediction methods. The effectiveness and advantages of the proposed procedure is shown by the data of a 110 kV power transformer .