DESIGN OF ARTIFICIAL NEURAL NETWORK -BASED MAXIMUM POWER POINT TRACKING CONTROLLER FOR SOLAR-POWERED WATER PUMPING SYSTEM (CASE STUDY: DAWURO ZONE YALOW LALA KEBLE

Abstract

Energy has great importance for our life and economy. Designing a stand-alone Photovoltaic power and water pumping system is highly essential for supplying energy to rural communities. This study presents the design of an artificial neural network (ANN) based maximum PowerPoint tracking (MPPT) controller for a solar-powered water pumping system. Yalow Lala Keble is one of the villages most affected by getting basic needs such as water supply and reliable electricity. The government has made sufficient efforts towards rural electrification through the use of renewable energy, particularly PV, in stand-alone systems. This effort aims to solve the water supply problem for the isolated community’s people who transport water from the river using plastic cans, bottles, pots, and vessels. ANN based MPPT drives a DC-DC boost converter, which maximizes the power from the PV module while also giving the Brushless DC (BLDC) motor constant input. The speed of the BLDC motor was managed through a Pulse Width Modulation (PWM) control of the voltage source inverter using a DC Link voltage controller. The built-in encoder has been used to generate a PWM signal, which is then used to carry out the electronic commutation by hall signal sensing. The system's efficacy was evaluated in comparison to Perturb and Observe (P&O) and Incremental (INC) MPPT control methods. Based on the results of the study, a population of 5,200 with a daily water demand of 520 m3 at the minimum design flow rate has been set as the target size. Water needs to be pumped at a minimum head of 82.5 m at a rate of 12 L/s with a minimum power rating of 17 kW from the pump being considered for this site. The estimated total power demand from the solar PV array was 29 kW. The analysis leads to the selection of the 305-watt module, resulting in a total of 96 modules to be connected in 6 series and 16 parallel. The systems were sized appropriately before being modeled in MATLAB / Simulink