This study aims at proposing a criterion for transformer insulation risk (IR) due to lightning overvoltages (LOVs) generated by back-flashover (BF) in ultra-high voltage substations using the improved limiting parameter (LP) method. In the proposed method, besides avoiding heavy computational burden imposed by a large number of transient simulations, the Monte Carlo (MC) and LP methods are combined to deal with the uncertain nature of LOVs. Consequently, the calculations in the proposed method are based only on the expected stresses on the transformer. Moreover, any overvoltage stress has structurally destructive consequences depending on its amplitude and duration. Insulation behavior may be different after and before applying stress. Therefore, the present study exploits the non-self-restoring behavior of transformers against various insulation stresses to develop a voltage-time-dependent strength accumulation characteristic. The proposed MC-LP method selects the appropriate distribution of expected strokes to estimate IR of the transformer based on the overvoltage surge. Therefore, the specifications of the substation, transformer, and incoming lines are used to determine the transformer insulation coordination with minimum calculations. The simulations are performed based on the field and experimental data, and the results show 25 more insulation risk compared to previous methods. a significant improvement in IR estimation.
展开▼
