This paper aims to propose a novel global optimization algorithm for fast and precise parameter identification of the inverse Preisach hysteresis model.
An enhanced parallel Runge-Kutta (ERUN) algorithm is proposed to identify the nine-parameter inverse Preisach model. Integrates chaotic mapping, parallel processing and adaptive perturbation to strengthen global exploration and convergence robustness. The Preisach model used in this paper is established by analytically deriving the inverse Everett function from first-order reversal curves (FORCs) and validated against experimental hysteresis loops of 20SW1200 nonoriented (NO) silicon steel, B30P105 grain-oriented silicon steel and 50WW470 NO silicon steel.
The ERUN algorithm achieves a 1.83% error and a computation time of 40.6 s, outperforming the genetic algorithm, simulated annealing, particle swarm optimization and the original Runge-Kutta (RUN) optimization method. The average relative root mean square error analysis confirms that all simulated hysteresis loop errors remain below 10%, even at low magnetic flux densities.
The proposed ERUN algorithm extends the original RUN method by introducing chaotic mapping and parallel computing strategies, effectively alleviating the premature convergence problem. This work presents a progressive improvement to the global optimization algorithm for inverse Preisach hysteresis model parameter identification, extending the original RUN optimizer with chaotic mapping and parallel computing strategies to achieve faster convergence and higher accuracy, achieving the lowest parameter estimation error (1.83%) and the shortest computational time (40.6 s) among the benchmarked algorithms.
