A new predictive energy management system: Deep learned type-2 fuzzy system based on singular value decommission, By Ardashir Mohammadzadeh

Research

Title	A new predictive energy management system: Deep learned type-2 fuzzy system based on singular value decommission
Type	Article
Keywords	Machine learning Microgrids Deep learning Fuzzy systems Singular value decomposition
Researchers	Yan Cao، Ardashir Mohammadzadeh، Jafar Tavoosi، Saleh Mobayen، Rabia Safdar، Afef Fekih

Abstract

A new predictive frequency management system is designed for multi-area microgrids (MGs). The uncertainties are online modeled by a deep learned type-2 (T2) fuzzy-logic system (FLS) and singular-value decomposition (SVD) approach. The predictive controller is designed based on the SVD-T2FLS model. The robustness against perturbations is ensured by the adaptive learning laws and adaptive compensators. The learning rules are extracted from stability and robustness investigation. The effect of demand response (DR), multiple load changes, high dynamic perturbations and variation of wind speed and solar energies are studied. The simulations on two case study systems and comparisons with other fuzzy logic controllers (FLCs), demonstrate that the suggested approach is well effective and results in good regulation accuracy.