Measurement and evaluation of transformer oil properties using advanced multi attribute texture classification method

Abstract

Electrical power systems play an important role in both manufacturing and services in both industrial and commercial sectors. Any energy transmission and distribution system is a major asset transformer. The transformer performance depends on the quality of the insulated oil. The traditional methods were the cost and time consuming. The proposed task uses a fast and reliable multi- attribute image processing technique to overcome all existing problems. In the proposed image processing analysis, the transformer oil images inputs are processed using quick software based analysis that determines the key features of the transformer oil processing it based on the multi-attribute structure image classification technique. Due to the changes color in the transformer oil, the oil properties also deteriorating. These factors and the transformer oil color are very close to each other. In this proposed method is found between the two intermittent entropy pattern and linear regression techniques. The performance of the proposed multi attribute texture classification method is validated through simulation, the simulation is created using Matlab2017a software. The transformer's work and health dependence is by using this multi-attribute texture classification system to find the basic quality of the transformer oil.