Comparative Study Between Levenberg Marquardt And Genetic Algorithm For Parameter Optimization Of An Electrical System

Parameter optimization is an important problem in control field of power system. In this paper, a comparative study between Levenberg Marquardt (LM) and genetic algorithm (GA) for parameter optimization of nonlinear systems is presented. Some basic steps and technologies used for implementing these approaches, as well as its advantages and disadvantages are discussed. The main goal is to show the merits of genetic algorithm optimization and to determine its suitability in the area of numerical field analysis. Since the DC series motor is widely used in several industrial sectors, the algorithm developed is potentially useful in order to implement a robust closed-loop control. Accordingly, the application of these two approaches to the estimation of the field winding and the armature resistance and inductance armature of DC motor show that the genetic algorithm have a better dynamic performance then the traditional one (LM).