ITEE Innovation Expo 2008 : Project Details

Expert system based transformer diagnosis based on chemical and dielectric measurements.

Transformers are a major asset in any power system around world. So maintaining them and performing condition monitoring tasks is essential in any power system..

The aim of this project is to investigate all different methods of assessing transformer insulation to perform condition monitoring. Once the techniques have been researched, to then design and build an expert system that implements all of these different techniques. The main dielectric measurement techniques that were used in this project were, Dissolved Gas Analysis (DGA), Polarisation and Depolarisation current (PDC), Return Voltage measurement (RVM) and Furan analysis.

Since both chemical and electrical measurements are taken into account the developed system is an extremely versatile application. The program was design to take as input all of the test results, then from there use the interpretation techniques to give out some results and a recommendation for further actions based on these.

To interpret DGA a modified version of IEC ratios was used. The method was modified in a way such that instead of giving only a fault or no fault condition it will give out a probability of fault which is much more useful for condition monitoring purposes. The other interpretation techniques were mimicked from existing systems and can give us an indication of moisture content as well as the degree of polymerization measurement for the insulation.

So the deliverable for the project is an expert system that can give an accurate description of transformer insulation, however can be used by non-experts.