Machine learning-based novel DSP controller for PV systems

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.