Hampton and Perkel both work at the Georgia Tech National Electric Energy Testing Research and Applications Center (NEETRAC), where Hampton is a principal research engineer and Perkel is a senior research engineer. The title of their award-winning presentation is “Health Index Methods for Using Forensic Diagnostics to Manage Water-Treeing in MV Cables.” Their NEETRAC colleagues, Thomas Parker and Dean Williams, were also co-authors on this presentation. Parker is a senior research technologist, and Williams is a research coordinator II.

Electric utilities are often interested in the condition of underground power cables that have been in service for many years. Many of these cables are insulated with a polymeric material that deteriorates over time via a phenomenon called water treeing. Water trees “grow” within the insulation due to the presence of moisture, voltage, and some type of imperfection such as voids or contaminants.

When a sample of cable is removed from service, water trees can be detected viewing wafers of the cable insulation after they are dyed in a methylene blue solution. The size, density, location, and type of water trees observed provide some indication of the cable condition. However, until now that relationship was unclear. 

The work described in this paper showed how a more definitive relationship between water tree growth and cable condition could be assessed by analyzing water tree growth in hundreds of samples and processing the results using machine learning. Now, when utilities send aged underground cable samples for evaluation, NEETRAC has the unique ability to deploy the high-level data analytics developed in this work to help them make an informed decision on whether to replace the cable or leave it in service.