Protection scheme for a distribution system with distributed generation using neural networks

Abstract

Interconnection of distributed generation (DG) to a radial distribution system has some impacts on the coordination of protection devices. In this case, faults are fed by multi-sources; networks and DGs. Existence of multi-sources in fault condition is the main difficulty in the identification and location of faults and therefore distribution system protection. New protection methods should be introduced for isolation of faulted sections. This paper presents a novel approach that can overcome the above problems. A feed forward neural network is used to identify the fault location. The predicted results from the algorithm prove the effectiveness of the proposed procedure for identifying fault location and then protection of the system for a wide range of system conditions.