A validation of the boundary element method for grounding grid design and computation

Abstract

[Abstract] Several widespread intuitive techniques developed during the last two decades for substation grounding analysis, such as the Average Potential Method (APM), have been recently identified as particular cases of a more general Boundary Element formulation [1]. In this approach, problems encountered with the application of these methods [3] can be explained from a mathematically rigorous point of view, and innovative advanced and more eficient techniques can be derived [2]. Numerical results obtained with low and medium levels of discretization (equivalent resistance and leakage current density) seem to be reasonable. However, these solutions still have not been validated. Unrealistic results are obtained when domain discretization is increased, since no one procedure is yet available to eliminate the above mentioned problems. Hence, numerical convergence analyses are precluded. The obtention of highly accurate numerical results by means of standard techniques (FEM, Finite Differences) implies unapproachable computing requirements in practical cases. On the other side, neither practical error estimates have been derived, nor analytical solutions are known for practical cases, nor suficiently accurate experimental measurements have been reported up to this point. In this paper, we present a validation of the results obtained by the Boundary Element proposed formulation, including the classical methods. A highly accurate solution to a specially designed test problem is obtained by means of a 2D FEM model, using up to 80; 000 degrees of freedom. Results are compared with those carried out by Boundary Elements.