A hybrid approach for PV based grid tied intelligent controlled water pump system

• Published in: International journal of adaptive control and signal processing Vol. 38; no. 4; pp. 1281 - 1308
• Main Authors: Selvakumaran, S., Baskaran, K.
• Format: Journal Article
• Language: English
• Published: Hoboken, USA John Wiley & Sons, Inc 01-04-2024; Wiley Subscription Services, Inc
• Subjects: Algorithms; bidirectional power flow; Cuk integrated Landsman converter; Electric motors; Flow control; Hybrid systems; machine learning RBFNN; Neural networks; Photovoltaic cells; Power factor; Power flow; PV system; Radial basis function; Reluctance Synchronous Motor (RSM); Resonant frequencies; Switching; System reliability; Weather
• ISSN: 0890-6327; 1099-1115
• DOI: 10.1002/acs.3748

Summary
This article proposes a hybrid POA‐RBFNN approach for PV (photovoltaic) based grid tied intelligent controlled water pump system. The proposed method is the hybrid wrapper of Pelican Optimization Algorithm (POA) and Radial Basis Function Neural Network (RBFNN) and later it is termed as POA‐RBFNN method. On the basis of bidirectional power grids, this research presents bidirectional power flow control for PV systems. The water pump is driven through a three‐phase synchronous reluctance motor (RSM). A novel Cuk integrated Landsman converter can operate at low duty cycles to provide larger DC voltages with negligible switching losses is used. In addition, to enhance the pumping system reliability, solar panels and electric pumps can be used to their maximum potential. Despite fluctuating weather conditions, the system supplies nominal displacement all over the day and surplus energy to the grid. Furthermore, if one of the sources fails, the pump performance is unaffected. Switching losses are minimized since the Voltage Source Inverter (VSI) operates at a fundamental frequency. In all operational modes, this control delivers unique grid‐side power quality and unity power factor functioning. Here, the performance of the proposed hybrid system is implemented in the MATLAB and is compared with several existing methods.