Load frequency control for multi-area power systems: A new type-2 fuzzy approach based on Levenberg-Marquardt algorithm

Ali Dokht Shakibjoo; Mohammad Moradzadeh; Seyed Zeinolabedin Moussavi; Ardashir Mohammadzadeh; Lieven Vandevelde

doi:10.1016/j.isatra.2021.03.044

Load frequency control for multi-area power systems: A new type-2 fuzzy approach based on Levenberg-Marquardt algorithm

ISA Trans. 2022 Feb:121:40-52. doi: 10.1016/j.isatra.2021.03.044. Epub 2021 Apr 15.

Authors

Ali Dokht Shakibjoo¹, Mohammad Moradzadeh², Seyed Zeinolabedin Moussavi³, Ardashir Mohammadzadeh⁴, Lieven Vandevelde⁵

Affiliations

¹ Department of Electrical Engineering, Shahid Rajaee Teacher Training University, Tehran, Iran. Electronic address: a.shakibjoo@sru.ac.ir.
² Department of Electrical Engineering, Shahid Rajaee Teacher Training University, Tehran, Iran; Department of Electromechanical, Systems and Metal Engineering, Ghent University, Belgium. Electronic address: m.moradzadeh@sru.ac.ir.
³ Department of Electrical Engineering, Shahid Rajaee Teacher Training University, Tehran, Iran. Electronic address: smoussavi@sru.ac.ir.
⁴ Department of electrical engineering, East Azerbaijan province, University of Bonab, Iran. Electronic address: a.mzadeh@ubonab.ac.ir.
⁵ Department of Electromechanical, Systems and Metal Engineering, Ghent University, Belgium; FlandersMake@UGent-Corelab EEDT-DC, Flanders Make, Belgium. Electronic address: lieven.vandevelde@ugent.be.

PMID: 33894972
DOI: 10.1016/j.isatra.2021.03.044

Abstract

In this study, a new fuzzy approach is proposed for load frequency control (LFC) of a multi-area power system. The main control system is constructed by use of interval type-2 fuzzy inference systems (IT2FIS) and fractional-order calculus. In designing the controller, there is no need for the system dynamics, therefore the system Jacobian is obtained by a multilayer perceptron neural network (MLP-NN). Uncertainties are modeled by IT2FIS, and for training fuzzy parameters, Levenberg-Marquardt algorithm (LMA) is used, which is faster and more robust than gradient descent algorithm (GDA). The system stability is studied by Matignon's stability method under time-varying disturbances. A comparison between the proposed controller with type-1 fuzzy controller on the New England 39-bus test system is also carried out. The simulations demonstrate the superiority of the designed controller.

Keywords: LFC; Levenberg–Marquardt algorithm; MLP-NN; Type-2 fuzzy.