Performance Augmentation Study on a Solar Flat Plate Water Collector System with Modified Absorber Flow Design and its Performance Prediction Using the XGBoost Algorithm: A Machine Learning Approach

This study compares the thermal performance between a conventional and a newly configured solar flat plate water collector system. The principal goal of this newly configured system is to increase the heat transfer rate between the tube and sheet assembly by increasing the contact surface length. The comparative experiments between the conventional and newly configured system were conducted in Tiruchirappalli, India. The experimental outcomes indicated that the overall performance delivered by the modified system with increased contact surface length is 12.41% higher than the existing system. Further, this research attempts to predict the thermal performance delivered by the system using the XGBoost algorithm, a machine-learning technique. The XGBoost algorithm proposed in this study includes four features (three inputs and one output). Real-time data were used to measure the performance delivered by the XGBoost algorithm. The results recorded by this XGBoost algorithm are closer to real-time values. The accuracy of the proposed XGBoost algorithm in predicting the thermal power of a flat plate collector is 99.80%.