Clustering Techniques for the Recognition of Partial discharge Sources

Clustering Techniques for the Recognition of Partial discharge Sources

Project description

 One of the main challenges with Partial discharge (PD) measurements is that they are affected by the measuring circuit and as a consequence the results of PD parameter computation can differ from one user to another or between different instruments. This is one of the main concerns particularly for unconventional PD measuring systems, which are out of the recommendations given by the Standard IEC60270 regarding standardize measuring circuits, measuring quantities, measuring instruments and calibrating methods.

The objectives of the research on partial discharge measurements in the High Voltage Laboratory of TU Delft includes:

• Design test set-ups with proper characteristics to control the effect of the measuring circuit on the measured response. Fundamentals of the partial discharge phenomena.
• Test out for robustness (noise and disturbance rejection) of new algorithms for the computation of PD parameters such as electric charge and energy.
• Development of the novel IQE clusters for the separation of combined PD sources.
• Design and construction of PD sensors with bandwidths suitable for each application.
• Development of a software tool for the research and teaching of partial discharge phenomena.