Neural network modelling for accurate prediction of thermal efficiency of a flat plate solar collector working with nanofluids

Abstract

In the present study, the performance of the flat plate solar collector with three different working fluids (pure water, Al2O3/water nanofluid and CuO/water nanofluid) is simulated using Artificial Neural Network (ANN). The solar radiation heat flux varied between 650 and 950 W/m2 and the flow rate varied from 1 to 4 L/min. The effect of radiation heat flux, mass flow rate, ambient and inlet temperature on the thermal efficiency was analysed. The predicted results of the three above-mentioned working fluids are compared and validated with those of the measurements. The output of ANN for all three working fluids was found to be reasonably capable of estimating the performance of the flat plate collector system with the deviation less than ±2%, while the trend of each working fluid is different against the mass flow rate. Experimental investigations are usually time-consuming and their equipment is expensive. Therefore, the advantages of the ANN model compared to the conventional testing methods are speed, simplicity and the capacity of the ANN to learn from limited experimental data.