Investigation of Prediction with Grey Model for Different Combinations of Transformer Characteristics

Converter transformer is an important part of UHVDC transmission system, which undertakes the work of energy transfer and energy conversion. In addition, the prediction of the characteristics of dissolved gas in oil is significant for the performance evaluation of the transformer body. In this paper, the mechanism of the prediction approach is presented particularly, and the influence of different combinations of characteristics for the dissolved gas in oil of a converter transformer is investigated. From the results, it can be clearly seen that the prediction error declines obviously with increasing of the combination number. In addition, the prediction error of the scheme with high correlation coefficient is lower obviously than the scheme with low correlation coefficient. Particularly, the prediction error of the scheme with high correlation coefficient reaches the minimum value when the combination number is greater than or equal to six. This can be guaranteed the prediction accuracy of the transformer characteristics. This paper presents a technical support for the prediction of other similar characteristics.