Journal of Applied Research in Electrical Engineering

Journal of Applied Research in Electrical Engineering

Performance Analysis of Dynamic Selections Based on Exponential Ranking Inside Genetic Algorithms for Induction Machine Parameters Estimation

Document Type : Research article

Authors
1 Electronics Department, Hassiba Benbouali University, 02000, Chlef, Algeria
2 Faculty of Technology, Laboratory of electrical systems and control LABSET, Blida 1University,9000, Blida, Algeria
3 Electronics Department, Morsli Abdallah University, 42000, Tipaza, Algeria
4 Automatics Department, M’hamed Bougara University, 35000, Boumerdes, Algeria
Abstract
In genetic algorithms (GAs), the imbalance between exploration and exploitation leads to premature convergence. To mitigate this drawback and enhance GA performance, the proposed dynamic selections incorporate the progressive control of selection pressure (SP) through exponential curve beams that dynamically adjust the exploration–exploitation balance. This paper analyzes the performance of the proposed dynamic selections based on exponential ranking generating equidistant curve beams relative to a horizontal straight line (GA-EH) and to a linear function (GA-LF). These algorithms are applied to parameter identification of an induction machine (IM). Both electrical and mechanical parameters are simultaneously estimated using only the IM’s starting current and the corresponding phase voltage measurements. The unknown parameters are identified by minimizing the quadratic error between the measured and calculated currents at each time instant using the output error method. The performance of the proposed dynamic selections is compared to static selections such as linear ranking (GAL) and nonlinear ranking (GANL). Evaluation is performed on simulated data and experimental data from an IM (1.5 kW). The simulation and experimental results highlight that GA-LF converges faster in terms of number of iterations. The accuracy of the estimated parameters is confirmed by the close match between the measured and computed currents in both transient and steady-state regimes.
Keywords
Subjects


Articles in Press, Accepted Manuscript
Available Online from 30 July 2026

  • Receive Date 24 November 2025
  • Revise Date 09 May 2026
  • Accept Date 30 July 2026