Performance investigation on ANFIS and FFNN assisted direct and indirect PV-fed switched reluctance motor water pumping system

Alternative energy sources have impacted as the ideal choice for reducing the burden on the grid and producing clean energy for water pumping. As an alternative to conventional motors for pumping, switched reluctance motor (SRM) is promising, cost-effective and highly efficient. This paper presents artificial intelligence (AI) supported variable voltage angle control (VVAC) for indirect and direct photovoltaic (PV)-fed SRM-based water pumping systems (WPS). Proposed feedforward neural network (FFNN) and adaptive neuro-fuzzy inference system (ANFIS) models are simulated and analysed for their percentage average error and real-time execution speed in the dSPACE-1104 control board. ANFIS resulted in 1/10^th reduction of execution time and improved performance accuracy in contrast to FFNN counterparts. A 4kW, 8/6 SRM model, has been developed in MATLAB/Simulink environment to validate the proposed system for both direct and indirect schemes. Further, the performance comparison of peak current, RMS current, torque ripple, and efficiency for direct and indirect PV-fed systems are presented in this work. These comparative assertions reveal the feasibility of both the methods with good performance. Moreover, a reduction in overall system size and cost with the absence of an intermediate converter, the direct method provides a reduced torque ripple as compared to the indirect counterpart.