Top-hat feature accumulation-based localization algorithm for buried conductive grid topologies

Abstract The substation grounding grid, a shallow-buried mesh conductive structure, is critical for electrical equipment safety. Accurate topological identification and localization are key for detecting faults like conductor breakage or corrosion, but traditional electromagnetic methods suffer from interference in complex substation environments, reducing accuracy. This paper proposes an optimized topological feature calculation method. It first derives weak-signal characteristics in the weak excitation zone via surface magnetic induction amplitude analysis. Then, it evaluates the top-hat algorithm (an image processing tool for highlighting data features) for extracting topological features in weak-signal zones, and investigates the correlation between detection probe width, top-hat feature values, and localization accuracy. This optimized algorithm integrates multi-dimensional feature information to enhance topological feature saliency, improve precision, and suppress noise. Simulations show the method effectively identifies grounding grid topologies in noisy environments, outperforming traditional approaches, with substation field measurements yielding an average error of only 0.0694 m. This provides a new technical pathway for grounding grid topological detection.