The fracture reason of steel wire cable is complex, and the corrosion and local bending effect of anchorage end of steel wire cable under tension are one of the main factors. Taking the steel wire of an arch bridge cable as the research object, the notch method was used to simulate the corrosion pits on the surface of the steel wire, and the tension and bending mechanical properties of the high strength notched steel wire were tested. The bending finite element model of the high strength steel wire was established by ANSYS WORKBENCH, and the tension and bending mechanical properties of the notched steel wire under different vertical loads and pretension were studied. The test and calculation results show that the test data are close to the finite element calculation results and the variation law is consistent. Under the same vertical load, the deformation of steel wire notch decreases with the increase of pretension; The stress at the bottom of the notch is the largest at 180˚ direction and the smallest at 90˚ direction of the vertical load.Under the same vertical load and pretension, the stress of spherical shape at the notch is the largest, followed by ellipsoid shape, and groove shape is the smallest, and there is a high stress zone at the edge of groove shape. When the pretension is applied, the initial stress increases with the increase of pretension, while the stress at the notch caused by bending decreases with the increase of pretension.

The equivalent circuit of traditional capacitive voltage transformers often faces the problem of complex data calculation and difficulty in grasping the internal nonlinear characteristics of transformers when constructing broadband models, resulting in poor power accuracy and stability. Therefore, with the help of electromagnetic transient physical models, the admittance sub model and nonlinear model are established by considering the frequency and saturation characteristics of the transformer. Based on the characteristics of capacitive voltage transformers, the model is processed in parallel to obtain a broadband coupling transformer model. The results showed that the error between the simulated current peak amplitude and voltage results of the model and the measured values was less than 2% and 1%, respectively. In the fault results, the harmonic error of the load voltage of the improved transformer was relatively small, far less than the error result of 2.73% of the traditional transformer. The proposed transformer model can better characterize its characteristics and has good transient response ability, providing reference tools and value for the operation and state detection of power systems.