A MEMETIC ALGORITHM APPLIED TO INDUCTION MACHINE PARAMETERS IDENTIFICATION BASED ON AN OUTPUT ERROR

Abstract

Based on an output error, several evolutionary methods have been applied to identify the parameters of an Induction Machine (IM). The main drawback of these methods is their premature convergence in many situations. To overcome this issue and achieve a more accurate solution, this paper proposes a Memetic Algorithm (MA), which combines a Genetic Algorithm (GA) and a local search method. This approach uses the Hooke-Jeeves (HJ) method for the local search as a mutation operator.GA has proven good ability in global search. The HJ method has a good ability to refine the local search and achieve the optimal accuracy solution. The proposed MA, which maintains a tradeoff between exploration and exploitation strategies, is applied to minimize the related objective function to obtain the electrical and mechanical machine parameters. The validation of the method is confirmed by an experiment carried out on an (0.4 kW) IM with parameters estimated using the measured data. Using the estimated parameters, the computed transient and steady-state currents agree well with the measured data.