Research on an icing-thickness calculation model for power transmission lines based on KPCA and GWO

• Propose a novel KPCA-GWO-SVM model to calculate the icing thickness of transmission power lines under different micro-meteorological environments. • Integrates KPCA to extract principal features in nonlinear data relationships, employs GWO for global optimization of the SVM hyperparameters. • Demonstrate the superior convergence speed and prediction accuracy of the proposed mode on real-world datasets. Addressing the limitations of poor accuracy and slow convergence in icing-thickness detection using conventional models on transmission lines, our paper presents a new calculation model based on Kernel Principal Component Analysis (KPCA) and Grey Wolf Optimizer (GWO). The model combines the advantages of KPCA in dealing with nonlinear relationships and GWO in global optimization and low computational complexity. KPCA is used to downscale micro-meteorological data. GWO is used for the purpose of efficiently optimizing SVM kernel parameters and penalty factors. GWO-SVM of inputting dimensional micro-meteorological data is used to calculate icing-thickness. In addition, this study analyzes the general principle and algorithm framework of KPCA-GWO-SVM model. Compared with the traditional computational models, the simulation results show that the convergence speed has increased up about to 75.0 %, 90.6 % respectively, the computational accuracy has improved 5.92 % at least.