Optimization of the PVT performance with various orientations of jets and MFFNN-RSA prediction model for smart buildings

Abstract

The combined thermal and photovoltaic technology in PV/T systems is considered as a greatly promising technology for smart buildings. Thus, investigations for enhancing the PV/T performance are still proceeding. This research presents an investigation for novel configurations of cooling jets for the PVT system. The linear and circular distribution for the inlet jets considering regular and irregular positioning for all the jets as new cooling configurations are implemented. Moreover, the proposed geometrical configurations are optemized regarding the performance to identify the most suitable configuration that achieves the optimum efficiency and temperature. Furthermore, a novel hybrid ANN model is presented for predicting the performance of the PVT systems. This model combines the multi-feedforward neural network (MFFNN) with an optimization technique called reptile search algorithm (RSA). The proposed model can process the studied parameters to predict the PVT performance parameters (top surface temperature, temperature un-uniformity, outlet temperature, and efficiencies). The proposed MFFNN-RSA model minimized the mean square error to less than 0.4857×10-3. The maximum temperature decrease achieved by the presented configuration reached 60.62K compared to the uncooled case, while the minimum temperature un-uniformity reached 1K and 6K for 400 and 1000 W/m2, respectively. The increase of the ambient temperature found to decrease the temperature un-uniformity in all the cases. The irregular jet with the linear distribution was found to achieve the optimum performance of the overall, thermal, and electrical efficiencies of 63.5%, 49.6%, and 14.25%, respectively. Furthermore, the electricity production cost was reduced by 11.6%, and the yearly CO2 emissions were reduced by 215.3 kg/m2 compared to the normal PV system. The proposed irregular-line distribution of the jets is found to be the best configuration regarding the temperature of the PV model and the overall efficiency considering the pumping losses.