Machine learning predictive models for optimal design of photovoltaic/thermal collector with nanofluids based geothermal cooling

Abstract

This study aims to estimate the performance of photovoltaic/thermal (PV/T) collector using alumina water‐based nanofluid with geothermal cooling through machine learning (ML) approach. A mathematical model is developed for the first law of thermodynamic analysis of nanofluid in PV/T system integrated with geothermal cooling and is validated with experimental results. Further, a machine learning‐based approach has been employed to simulate the cooling performance of a nanofluid cooling based PV/T system. In the study, Multi‐layer perceptron (MLP) is proposed for estimating the thermal and electrical performance of PV/T system based on design parameters like nanofluid concentration, Reynolds number, and time. The same is then compared with other state‐of‐the‐art machine learning techniques and it is evaluated based on various quality metrics such as mean square error (MSE), root mean square error (RMSE), and R2 test. The designed network is compared with the other ML algorithms available in literature like linear regression (LR), support vector machine (SVM) and decision tree (DT). The proposed MLP network is provided a significant outcome with an average accuracy of 98% and predicted PV panel temperature of 32.1–36.5°C for 0–18 sequences. It was also observed that electrical efficiency of PV/T system improved from 10.51% to 10.66% for 0–18 sequences through MLP predictions.

CNN-LSTM to Predict and Investigate the Performance of a Thermal/Photovoltaic System Cooled by Nanofluid (Al2O3) in a Hot-Climate Location

“…They presented a highly accurate predictive model with R 2 ranges of 0.9997-0.9999 for the dataset used in their study. Recently, Jakhar et al developed a multi-layer perceptron (MLP) in order to match other researchers' targets in predicting the performance of a thermal/photovoltaic system with nanofluids-based geothermal cooling [20]. The average rate of determination (R 2 ) of their predictive model reached 98%, and the improved predicted cell temperature and electrical performance were 32.1 • C and 10.66%, respectively.…”