Prediction of Power Transformer Oil Dissolved Gas Concentration Based on Modified Gray Model

This paper put forward the Equal Dimension Markov Grey Model based on gray model for prediction concentration of power transformer dissolved gases. First, original gas concentration was processed and transformed into equal interval data. Then, improved gray model with modified background value formula forecasted equal interval data and residual error. The prediction of power transformer oil dissolved gas concentration was composed of predictive gas concentration value and residual error. Markov process identified polarity between predictive gas concentration value and residual error. Finally, equal dimension constrained condition was introduced to optimize gray model. Experiments on prediction of power transformer oil dissolved gas concentrain demonstrates that equal dimension markov gray model in the paper is suitable for predicate concentration of power transformer oil dissolved gases, and its forcasting accuracy is higher than GM(1, 1, ρ) model.