Fault Prediction And Diagnosis Of Gas Chromatograph In Transformer Oil Based On SVM

As the ultimate element in the electric power system, the safe of operating of transformer is correlative with the safety of the power system, so the condition assessment and fault diagnosis of transformer is of great importance.On the bases of oil chromatographic analysis, this paper inreoduces the prediction model ,diagnosis model based on improved least square support vector machine by genetic algorithm, and example analyses established that the model is effective.And this paper introduces the data management and fault diagnosis expert system used SQL Server 2005 and Visual Studio 2005, which is simple and practical. The system runs steadily, and it can improve power transformer's state information management and maintenance efficiency.