Corrosion prediction of galvanized steel electrode in soil using deep learning‐based model

• Published in: Electrochemical science advances Vol. 2; no. 6
• Main Authors: Wu, Kongyong, Zhang, Guofeng, Dong, Manling, Zheng, Wei, Peng, Mingxiao, Lei, Bing
• Format: Journal Article
• Language: English
• Published: Aachen John Wiley & Sons, Inc 01.12.2022
• Subjects: Accuracy; Artificial intelligence; Barium; Chloride; Corrosion potential; corrosion rate prediction; Corrosion tests; Deep learning; Electrodes; Galvanized steel; grounding electrode; LSTM; Machine learning; Methods; Moisture content; Neural networks; Support vector machines
• ISSN: 2698-5977; 2698-5977
• DOI: 10.1002/elsa.202100133

Accurate prediction for the corrosion status of grounding electrodes is critical for the safe and stable operation of power systems. However, the corrosion rate of grounding electrodes changes dramatically with the soil environmental parameters, making it hard to be precisely predicted. To address this problem, a deep learning method was proposed to numerically predict the corrosion rate of a galvanized carbon steel grounding electrode in this paper. The long‐short term memory method is selected as the modeling algorithm, and also chosen as the hidden layer for the Recurrent Neural Network, while the soil environmental parameters are used as input features. The predicted results match well with the experimental data when evaluated using different soil parameters, such as soil moisture, chloride (Cl–) concentrations, and sulfate (SO42–) concentrations. The threshold corrosion rate related to each parameter is obtained to estimate the corrosion rate with more accuracy. We proposed a deep learning method to numerically predict the corrosion rate of a galvanized carbon steel grounding electrode in this paper. The long‐short term memory method is selected as the modeling algorithm, and also chosen as the hidden layer for the recurrent neural network, while the soil environmental parameters are used as input features. The predicted results match well with the experimental data when evaluated using different soil parameters, such as soil moisture, chloride (Cl–) concentrations, and sulfate (SO42– )concentrations. The threshold corrosion rate related to each parameter is obtained to estimate the corrosion rate with more accuracy.