Open AccessANN‐based adaptive current controller for on‐grid DG system to meet frequency deviation and transient load challenges with hardware implementation
Author(s) -
Patowary Moushumi,
Panda Gayadhar,
Naidu Bonu Ramesh,
Deka Bimal C.
Publication year - 2018
Publication title -
iet renewable power generation
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.005
H-Index - 76
eISSN - 1752-1424
pISSN - 1752-1416
DOI - 10.1049/iet-rpg.2017.0142
Subject(s) - control theory (sociology) , controller (irrigation) , transient (computer programming) , computer science , grid , filter (signal processing) , electric power system , stability (learning theory) , voltage , power (physics) , engineering , control (management) , mathematics , geometry , artificial intelligence , agronomy , biology , operating system , electrical engineering , physics , quantum mechanics , machine learning , computer vision
To maximise functional efficacy and reliability of distributed generations (DGs), this study leads to modelling, control, stability analysis, and hardware validation of a new adaptive current controller in the application of an on‐grid voltage source converter (VSC) system. For effective mitigation of power system hindrances without affecting the power quality (PQ), self‐tuning of weights associated with the proposed variable leaky adaptive step‐size‐least mean square (VLAS‐LMS) control algorithm based on artificial neural network (ANN) is updated in natural frame and felicitous shaping of VSC outputs are witnessed. The selection of a constant step‐size associated with proposed controller usually yields updating of same weights by all the sampling periods and hardly takes care of rate of convergence factor, which decides the stability of the controller. It can no longer avoid wavering of weights during grid disturbances, resulting in high‐filtering gains. Again, a constant leaky factor may lead to over‐ or under‐parameterisation of regularisation component. These disputes can be overcome in the proposed algorithm by the introduction of an adaptive step size along with a variable leaky factor. Furthermore, PQ is maintained as trade‐off by the inclusion of detuned LC filter. Experimental outcome ensures the validation and effectiveness of the proposed controller.