The Prediction of Grounding Grid Corrosion Rate Based on Vector Similarity and Fuzzy Closeness

Abstract

Aiming at the grounding grid corrosion prediction involves many factors, the relationship is complex, and there are uncertainty and ambiguity in gaining involved factors, an efiective grounding grid corrosion rate prediction model is proposed. Consider sample set of every corrosion factories as feature vectors and then standardized them; by calculating the similarity of feature vectors and corrosion rate vectors, the weight of each corrosion factor on corrosion rate was obtained; treat membership function based on relevant relationship and half-Gaussian distribution as the core to determine the membership; use difierent sample properties have difierent corrosion rate contributions to do classiflcation, and combine closeness principles to predict corrosion rate. The application results show that the model can efiectively predict the grounding grid corrosion rate of those ground web sites who have near distance or similar soil physical and chemical indicators.