Theory and simulation of a neuro-fuzzy controller for 20MW turbo generator control
This document focuses on the development of a learning tools for the dynamic model of a steam turbine 20MW capacity type able to accommodate the different variations of the machine's operating parameters.
Considering all the advantages and disadvantages of the optimization methods that we have studied (genetic algorithm and neural networks), the neuro-fuzzy option was remarkable in view of the results obtained in the modeling of the systems. The ANFIS model (Adaptive Neuro-Fuzzy Inference System), known for its performance in the learning domain of complex non-linear systems, is chosen for a better approximation of functions for modeling.
The active power depending of the high pressure steam flow inlet turbine and the low pressure steam flow extraction as well as the voltage across the generator depending of the power factor and the active power shall be controlled from 2 ANFIS models mounted in parallel, their outputs acting respectively on the control of the HP (High Pressure) steam inlet valve opening for the variation of the HP steam flow inlet and the rotor excitation current for the variation of the voltage across the alternator. The dynamic model thus obtained from the adjustment of the parameters of the membership functions describes the behavior of the turbine to stabilize its operating point in the face of any disturbance
Auteur(s) : Babacar KEBE , Oumar Ba , Lamine Thiaw
Pages : Pages: 94 - 99
Année de publication : 2018
Revue : IEEE
N° de volume : 2018 7th International Conference on Renewable Energy Research and Applications (ICRERA)
Type : Article
Mise en ligne par : BA Oumar