Improve the performance of renewable energy conversion and storage via ANN in a system of solar water heater with variable speed photovoltaic circulating pump

Summary A solar heater with a variable speed circulation pump is analyzed based on the manufacturer's data sheet of the PV generator, the DC pump as well as the solar collector under a specific climate condition via ANN approach. Direct normal irradiance, global horizontal irradiance and ambient temperature data of Tabuk city were used for system analysis. The detailed electrical characteristics of PV generator are performed based on a single diode model and transient thermal modeling of the storage tank is performed based on the Crank‐Nicolson numerical method which is developed based on thermodynamic energy balance. Hourly electrical performance output of PV generator is predicted considering maximum power point as MPPT is part of the PV pumping system which optimizes the power output of from the PV module. An empirical curve fitting correlation that determines the performance of the DC pump is developed from the performance curve of the pump which is obtained from the pump manufacturer. Similarly, an empirical correlation that relates flow rate and head loss at the solar collector is developed from the solar collector's manufacturer data sheet. Hydraulic features of the system such as hourly flow output and hourly head loss at the collector are determined based on a developed algorithm. Furthermore, hourly thermal characteristics are determined based on the Crank‐Nicolson technique. The power of solar cells was highly dependent on voltage and radiation intensity. To evaluate the sensitivity of the PV arrays, an artificial neural network (ANN) was used and it was found that the ANN with an R2 of .9998 had an error of less than 5% for more than 97% of the data points.