Intelligent modelling (NARX) of evacuated tube heat pipe solar collectors

A nonlinear interface identification simulation for the temperature of the Evacuated Tube - Heat Pipe Solar Collector (ETHPSC) was included in this work. To collect input data (volumetric flow rate) and output data (outlet water temperature), experimental data were initially derived via experimental work. To estimate the collector temperature as a system output, the nonlinear self-regressive exogenous (NARX) model has then been proposed as one of the Neural Network techniques. A mean square error (MSE) was used to validate the simulation of this method. The best result was obtained when the number of hidden layers was equal to 8, the number of delays was equal to 2, with the recorded lowest (MSE) value was equal to 0.4060.