IDBO-RBF Neural Network Method for Predicting the Surface Field Strength of Live Working on UHV DC Transmission Lines

• Published in: 2023 3rd Power System and Green Energy Conference (PSGEC) pp. 980 - 985
• Main Authors: Li, Jian, Zhang, Yadi, Chen, Gang, Zhou, Lin, Sun, Wencheng, Peng, Yuhui, Qiu, Jie, Zhang, Bide
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-08-2023
• Subjects: DBO algorithm; Knee; live work; Neurons; Personnel; Prediction algorithms; Predictive models; Radial basis function networks; RBF neural network; Safety; surface field strength
• DOI: 10.1109/PSGEC58411.2023.10255796

The surface field strength of the live work personnel is an important safety indicator for the live work of UHV DC transmission system. If the surface field strength is too big, it will cause tingling sensation of personnel, which will lead to safety accidents in serious cases. However, the presence of various meteorological conditions can cause changes in the distribution of space charges, ion diffusion rates on the UHV DC transmission system, resulting in changes in the surface field strength of the live working personnel. Prediction the surface field strength of the live working personnel can be utilized to assess the safety of live work on UHV DC transmission system in order to reduce the likelihood of accidents. A prediction model using Improve Dung Beetle Optimizer (IDBO) to optimize Radial Basis Function (RBF) networks is proposed for the prediction of body surface field strength. This model uses the meteorological conditions as input neurons and the field strengths of the feet, knees, hands, chest, and head of the live working personnel as output neurons. The IDBO algorithm is used to optimize the RBF network, including the centroids, widths, weights of the connection layers, and biases. It is shown by examples that the RBF neural network optimized using IDBO algorithm has higher prediction accuracy compared to the traditional RBF neural network.