New method for improving the reliability of dissolved gas analysis

The diagnostic methods and the dissolved gas analysing standards or regulations generally are based on experts practical and experimental knowledge while detecting and evaluating the faults. Dissolved Gas Analysis (DGA) is a widely used technique to estimate the condition of oil-immersed transformers. The measurement of the level and the change of combustible gases in the insulating oil is a trustworthy diagnostic tool which can be used as indicator of undesirable events occurring inside the transformer. There are standards available for this purpose the DGA interpretation should also be based on other information about the reliable particular transformer. This paper describes a realistic method for power transformers using readily available data. The method considers practical limitations on obtaining data and possible constraints on the parameters utilize IEC, IEEE, CIGRE criteria. The calculation considers no t only typical test results but also other parameters such as physical observations, tap changer and bushing condition, load history, maintenance work orders, age, trends of the transformer failures etc. The calculation includes condition ratings, weighting factors, and assigned scores for specific condition parameters by using fuzzy logic. A neural network using the DGA results is applied to achieve the initial conclusion firstly. Then, several fuzzy equations are established to realize the detailed diagnosis. As many sorts of data and relevant DGA information are selected in different fuzzy equations, which cause that the accuracy of the detailed diagnosis will be higher. Today, in modern electricity system it is very important for the equipments to have a modern expert system with suitable efficiency so the past record is also taken into consideration during the diagnosing. The DGA expert system improved aggregates 5 diagnostic methods and prepares the classification of t he equipment based on the measured state markers/features. The evaluation of the single diagnostic procedures is based on the fuzzy logic. As a result the diagnoses provided by the investigational results can be precised, so the lifetime and the cost effectiveness is increasable. This paper illustrate the advantages of diagnosing with the fuzzy logic compared to the procedures with normal evaluation for international results and Hungarian transformer examples.