The conventional stress-grading system of form-wound motor coils has been found to perform well under sinusoidal, power-frequency voltage. However, when the motor is fed by inverter-based drives employing fast switching, a significant impact on the electric stress-grading system occurs. Simulations of the stress-grading systems using finite element analysis indicate that, during the fast rise-time, a high electric stress can develop right at the slot exit of the stator if the conductivity of the conductive armor tape (CAT) is not high enough. Increasing the conductivity of the CAT alleviates the problem but increases the eddy current loss in the stator laminations. To address this problem, the conductivity of the area just beyond the slot exit needs to be increased by coating the surface of the CAT with a high-conductive material. The stress-grading at the end of this conductive paint section then is achieved by two coordinated layers, which help to distribute the electric field and heat that is produced

In this paper, the effectiveness of stress grading coatings (SGCs) on form wound stator coil groundwall insulation under fast rise time pulses is analyzed. The combined performance of the stress grading coating and the semiconductive slot coating is studied under fast rise time transient voltages. During the rapid rise of voltage, the high stress moves from the stress grading coating to the semiconductive coating at the end of the slot. According to the results, it is evident that the design of both the stress grading and semiconductive slot coatings must be done together for effective stress relief. The use of materials with varistor behavior in the stress grading coating, combined with a high conductivity semiconductive slot coating, is analyzed as a solution to the high stress that develops at the slot end of form wound stator coil groundwall insulation of inverter fed drive (IFD) motors.

In this work two stress grading options for 115 kV non-ceramic suspension insulators are analyzed by means of 2D and 3D simulations. Both options were optimized in order to compare the maximum reduction that can be obtained on the electric field at the surface of the insulator. In the first option the shape and permittivity value of the housing material next to the energized end were the optimized parameters. For the second option, the installation of a corona ring, the position and dimensions of the ring were optimized. Electric field simulations were performed with the finite element method (FEM) while for the optimization process different functions of the MATLAB optimization toolbox were used. The optimization was performed with 2D FEM simulations and the optimal designs were then used on 3D models, in order to verify that these designs remain as the best option. According to the results, it was found that the use of a corona ring in 115 kV lines produces the maximum reduction of electric field as long as it is installed in the optimal position. A modification of the housing profile next to the energized end will also produce a significant reduction on the maximum electric field on 115 kV non-ceramic insulators.

The paper discusses the material properties and designs of stress grading coatings for cable terminations that can help to reduce the adverse effects of stepped voltages present in voltage source converters. Simulations using transient and quasistatic finite element method, as well as experimental measurements of stress grading materials are presented in this study. The use of two stress grading layers in cable terminations is suggested as an alternative design to reduce the hot spot temperature in conductive stress grading coatings.

Fast front pulses from lightning and switching conditions produce transient overvoltages and dielectric stresses that can result in premature deterioration or failure of transformer windings. Predicting these conditions in new transformer designs requires the implementation of internal winding models. This paper analyzes the accuracy and performance of different transformer winding models available in the literature, by means of comparisons with time- and frequency-domain measurements on a 2000-turns disk-type winding prototype, as well as evaluation of the computer time needed by each model to perform the corresponding simulation of the same prototype. The results show that including the frequency dependence of the model parameters is a decisive factor to accurately reproduce the measurements. They also show that a lumped parameter model described in a turn-by-turn basis can provide accurate results and requires shorter simulation time than a distributed parameter model.