Soft fault identification in electrical network using time domaine reflectometry and adaptive neuro-fuzzy inference systeme

The main purpose of this work is to implement a new and accurate approach based on Time Domain Reflectometry (TDR) combined with Adaptive Neuro-Fuzzy Inference System (ANFIS) to solve the problem of soft faults detection and localization on complex wiring electric network. Firstly, the response of the transmission line is given by applying the Finite Difference Time Domain method (FDTD) on the transmission line equations. Then, the ANFIS method is used to solve the inverse problem which permits to detect and localize the soft fault. Finally, very acceptable results are obtained and many problems are solved at the same time as: defining the exact position and exact resistance values of the fault, defining the state of electrical network in the real time.