What is it about?

Dissolved Gas Analysis is a technique by which incipient faults in power transformer are detected analysing the gases dissolved in the transformer oil. A novel Dissolved Gas Analysis technique where the advantages of both the ratio methods and graphical approach are utilised to deal with the problems of both the methods. A pentagon is developed based on the principles of Duval Pentagon with more flexible fault zones. Some regions where more than one faults have the possibility to coexist are separated and the features of Gas Ratio Combination are applied with certain modifications. Finally, a fuzzy logic approach is used for the fault prediction along with giving a measure of how reliable the prediction itself is.

Featured Image

Why is it important?

Firstly, this technique proposes a very effective amalgamation of graphical approach and ratio analysis for developing a more flexible fault prediction method. The problem with the existing graphical methods is its rigidly separated fault zones. Here the concept of overlapping fault zones is introduced with the development of two regions, 'Regions of Certainty' where only one type of fault exist and 'Regions of Uncertainty' where multiple fault types can exist. In the ROCs, prediction accuracy achieved is almost 100% where as it comparatively low in the ROUs. Prediction is aided by ratio analysis in the ROUs. Also, this method calculates how reliable the prediction itself is, which another unique feature of this technique.

Perspectives

1. The flexibility of the approach increases the overall accuracy greatly. 3. Separation of the ROCs and ROUs gives a more intelligent and effective way of discriminating zones with different prediction accuracy. 2. The information provided about the reliability of prediction provided the person using it a measure of how much trust can be put upon the predicted result.

Kingshuk Chatterjee
Indian Institute of Technology Kanpur

Read the Original

This page is a summary of: A Novel Prediction-reliability Based Graphical DGA Technique Using Multi-layer Perceptron Network &Gas Ratio Combination Algorithm, IET Science Measurement & Technology, April 2019, the Institution of Engineering and Technology (the IET),
DOI: 10.1049/iet-smt.2018.5397.
You can read the full text:

Read

Contributors

The following have contributed to this page