Overcurrent relay curve modeling using adaptive neuro fuzzy inference system

In this paper, modeling of overcurrent relay (OCR) curves using adaptive neuro fuzzy inference system (ANFIS) are proposed. The accurate models of OCR curve with inverse time relay characteristics have an important role for protection coordination of power system. Models of OCR curve are appropriate with IEC standard. This model implements of microcontroller AT mega 128 as digital relay and personal computer as facility to design of OCR curve. ANFIS is developed to OCR curve modeling with different types of membership function and each membership function is trained for 10 iterations. Input for training to OCR curve using the load current and current setting or I_L/I_S. Time to opening the circuit breaker or T_CB is used as output for training of OCR curve. ANFIS is developed using visual basic. The simulation results are compared with different types of membership function to obtain the optimal design of OCR curve. Moreover, the testing results are compared with OCR curve modeling to check validation and accuracy of the proposed model.