Abstract
The use of photovoltaic (PV/T) system that converts solar radiation to electricity and provide thermal needs concurrently stands as one of the most effective means of utilizing renewable energy. In recent times, machine-learning techniques have been extensively used in solar system applications due to their high accuracy in predicting the performances without necessarily going through physical modelling. The main objective of this study is to implement an intelligent algorithm adaptive neuro-fuzzy inference system (ANFIS) model to simulate and predict the thermal and electrical efficiencies of a water-based photovoltaic-thermal (PV/T) system. Thorough experimentation for the fabricated set-ups (conventional PV and a water-based PV/T system) was carried out. The real experimental results was validated using ANFIS model. Base on the results obtained it was confirmed that there was an excellent agreement between the predicted model outputs and the actual experimental data. However, the ANFIS model gave the highest prediction accuracy with the lowest error margin of 0.00021, 0.0089 and 0.0459 for MSE, RMSE and MAE with strong correlation (R2) of 0.9998. Base on the results obtained it was concluded that this intelligent algorithm is a reliable tool in predicting the PV/T system performances.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have