Research on fault diagnosis of grounding girds based on intelligence hybrid method

Grounding grid is one of the most important equipments in electric power system. Its performance is directly related to stable operation of power system, therefore, the grounding grid for condition monitoring and fault diagnosis is of great significance. In order to determine the resistance of each branch in the actual grounding grid, a practical grounding will eventually be translated into the intrinsic grounding grid model using network topology decomposition, and then this paper establishes augmented fault diagnosis equations of intrinsic grounding grid. In order to solve the nonlinear equation sets, this paper proposes the hybrid method based on particle swarm optimization and least squares iteration. At first, the approximate solution of grounding resistance is obtained by particle swarm optimization algorithm, then the approximate solutions are treat as the initial value, the optimal solution of fault diagnosis can be obtained by the iterative least square method. The hybrid method has significant advantages on speed and accuracy of calculation. At last, the fuzzy branch resistances are accessible by binary search scope. The results show that the actual resistance value can be gained by the hybrid method proposed in this paper, which can judge the corrosion of grounding grid effectively.