Algorithms for Guided Wave Detection of Corrosion in Ground Rods

Abstract

A key element in the distribution of electric power is the substations, where high voltage power is transformed to lower voltages for distribution to final users. A critical element of the substation is the network of rods driven into the ground that provide the electrical “ground” for the equipment and personnel. As electric substations age, their grounding rod network can corrode and the safety of personnel and equipment can be compromised. Maintaining the integrity of the ground rods is vital to the safe operation of substations, so it is critical to detect if ground rods have corroded to the point where they no longer provide adequate protection to the equipment and operating personnel. In this work, a signal clustering technique was applied to guided wave pulse-echo detection of corrosion damage in grounding rods. The clustering method was tested on 2.9 m long, 19 mm diameter steel rods with machined simulated corrosion pits. Wet clay was applied to the rods to simulate the acoustic loss due to the presence of soil. The corrosion resulting in greater than 50% cross-sectional loss is easily detectable. Defect detection using Fuzzy c-means for clustering produced low false positives and achieved a depth resolution of approximately 0.3 m. The known damage locations on the rods and the algorithm estimated damage locations were in good agreement, even with varying soil loss conditions. Acoustic pulse echo sensors with Fuzzy c-means clustering is a promising tool to survey ground rod corrosion.