Title: Machine learning-based novel DSP controller for PV systems

Authors: Subramanya Bhat

Addresses: Department of Electronics and Communication Engineering, N.M.A.M. Institute of Technology, Karkala Taluk, Nitte 574 110, Karnataka, India

Abstract: As fossil fuels are getting depleted it is very much essential to harvest solar energy. The harvesting and conversion methods available in literature are simulation-based. A very few work has been reported using hardware circuitry. However, machine learning-based DSP controller for solar energy harvesting is not available in literature. In the proposed study, machine learning-based DSP controller is implemented. The genetic algorithm (GA)-based DSP controller has been designed for enhancing the efficiency of solar PV. In the proposed work, perturb and observe (P&O) technique and genetic algorithm (GA) have been considered to achieve maximum power point and precise control parameters of PID controller respectively. Single DSPTMS320F28377s has been used to implement both P&O and GA and it is revealed that the proposed DSP-based hardware model provides better speed, efficiency and reliability than the existing simulation-based controller. The proposed work will bring a paradigm shift in solar energy harvesting and control.

Keywords: converter; tuning; control algorithm.

DOI: 10.1504/IJAAC.2021.113343

International Journal of Automation and Control, 2021 Vol.15 No.2, pp.226 - 239

Received: 15 Aug 2019
Accepted: 20 Nov 2019
Published online: 01 Mar 2021 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article