Improved Power Flow Initialization of Dynamic Studies with Non-linear Composite Load Model

Abstract

A power flow solution using a static constant MVA load model is a straightforward approach to initialize power system dynamic simulations. Using this model to represent induction motor load, however, may not be suitable due to a natural mismatch between initial bus scheduled powers and actual input powers that are computed afterward using their converged bus voltages. To avoid this mismatch occurrence, the Newton-Raphson power flow algorithm has been newly extended to incorporate non-linear characteristics of a composite load model in order to get an exact operating point of each individual induction motor and static load. The power flow solution of the applied 13-bus industrial power system reveals that the extended algorithm is capable of giving an exact active and reactive power input of the motors in correlation with their converged slips, internal bus voltages, and mechanical load torque characteristics and of voltage-dependent loads in relation to their exponent indices. Moreover, the computational efficiencies of the algorithm have been investigated using IEEE 14-bus and 57-bus networks. The results show that embedding induction motor and static exponent load models directly into the power flow mismatch function and Jacobian can improve computational efficiency since the power flow solution is well converged in a satisfactory quadratic manner.