新月形覆冰导线气动力特性预测及舞动响应研究OA

The icing of transmission conductors alters their aerodynamic characteristics,potentially leading to conductor galloping.Addressing the limitations of traditional wind tunnel tests—such as high costs and narrow variable coverage—as well as the low accuracy and long computation time of computational fluid dynamics(CFD)simulations,this study establishes a multi-parameter database based on existing wind tunnel data,incorporating factors such as ice thickness,conductor diameter,wind attack angle,and turbulence intensity.Using algorithms including multiple nonlinear regression(MNR),decision tree regression(DTR),and Gaussian process regression(GPR),high-precision prediction models for the aerodynamic coefficients of iced conductors under multi-parameter coupling effects are developed,overcoming the limitations of traditional single-variable polynomial fitting methods.Comparative analysis in MATLAB shows that the GPR model achieves the highest accuracy(R2>0.98)with slightly lower computational efficiency,the DTR model achieves R2>0.96 with optimal efficiency,and the MNR model achieves R2>0.908 but with a complex expression.Galloping simulations of iced conductors,based on aerodynamic coefficients predicted by the GPR and DTR models,reveal multi-modal coupled vibrations dominated by low-frequency oscillations.The amplitudes of vertical and horizontal galloping increase with ice thickness,while comparisons with wind tunnel data indicate that errors in galloping amplitude and frequency are both below 10%.Using the established aerodynamic coefficient prediction models and simulation procedures,galloping simulations of practical engineering cases effectively reproduce field-observed galloping amplitudes,frequencies,and vibration modes.This demonstrates the practical application value of the proposed methodology for predicting galloping in engineering contexts.