Nonlinear marine predator algorithm for robust identification of fractional hammerstein nonlinear model under impulsive noise with application to heat exchanger system

Abstract

Identification of stiff nonlinear systems is considered as one of the challenging tasks and research community is providing promising solution for identification of these systems. Researchers have concluded that integration of fractional calculus provides better insight and understanding of complex systems by keeping the previous history. In this study, nonlinear marine predator optimization algorithm (NMPA) is used for the identification of fractional Hammerstein control autoregressive system (FHCAR) with Gaussian as well as impulsive noise. Further, a practical example of heat exchanger system modeled with FHCAR structure, is considered to analyze the knacks of NMPA in terms of convergence, robustness and stability. Grunwald-Letnikov's concept of fractional calculus derivative is used to transform standard Hammerstein control autoregressive system into FHCAR system. Mean square error-based fitness function is used to examine the performance of NMPA for identification of 4th order nonlinear FHCAR system for all three case studies i.e., FHCAR with Gaussian noise, FHCAR with impulsive noise and heat exchanger system identification. The performance of NMPA is observed in terms of fast convergence, accuracy, stability, robustness and accuracy in estimation of correct parameters of the system for multiple noise scenarios and the superiority is endorsed through comparison with the recent counterparts i.e., Gazelle optimization algorithm, Runge Kutta optimization method, Whale optimization algorithm, Harris Hawks optimization algorithm and African vulture optimization algorithm.