A GA-based grey prediction model for predicting the gas-in-oil concentrations in oil-filled transformer

摘要

Dissolved-gas-analysis (DGA) techniques are widely used to diagnose oil-filled transformer insulation, but the conventional procedure acquiring the gas-in-oil concentrations is not timely. To make up the disadvantage, a new method based on a genetic algorithm and the grey theory to predict the gas-in-oil concentrations is proposed in This work. The grey model (GM(1,1)) has been improved and a new optimized grey model (GM(1,1,/spl beta/)) has been constructed. The genetic algorithm has been applied to search the optimal parameters of the GM(1,1,/spl beta/) model. The validity of the GA-based GM(1,1,/spl beta/) model was verified with two prediction examples.