Pattern recognition in large datasets to improve high voltage electrical equipment reliability (FULLY FUNDED)

Supporting company: EA Technology

Project Outline:

High voltage electrical equipment is often monitored for “partial discharge”, which is an indication of the integrity of the electrical insulation. However, continuous monitoring generates very large, multi-GB datasets. Existing techniques to analyse these datasets do not make full use of this data. This project will identify efficient techniques for storing, accessing and analysing multi-gigabyte datasets; apply these techniques to determine patterns and trends within these datasets; correlate these patterns and trends with observed degradation of electrical insulation and develop analytical tools and techniques to apply these findings.

Benefit to/impact on industry:

Improvement in reliability of electrical systems. Reduced operating cost. Improvement in safety.

What novelty will the student base their PhD on?

Determining the correlation between identified PD patterns and actual degradation of insulation. Development of analytical tools and techniques to apply these findings.