Parameter-Driven Efficiency Evaluation of Solar Thermal Collectors Using CFD–ML Integration

Abstract

This research presents a comprehensive computational framework integrating Computational Fluid Dynamics (CFD) with Adaptive Neuro-Fuzzy Inference System (ANFIS) for parameter-driven efficiency evaluation of parabolic trough solar thermal collectors. The methodology combines high-fidelity CFD simulations with advanced ANFIS machine learning techniques to predict thermal performance parameters across diverse operational conditions. A validated CFD model was developed using ANSYS Fluent to simulate the complex thermal and fluid dynamic behavior within parabolic trough collector systems, generating 920 numerical datasets encompassing varying flow rates, inlet temperatures, solar irradiation levels, and geometric configurations. The ANFIS model demonstrated exceptional predictive capabilities, achieving coefficient of determination (R²) values of 0.94 for thermal efficiency predictions. The integrated CFD-ANFIS framework reduces computational time by 82% compared to traditional CFD approaches while maintaining prediction accuracy within 4.2% error bounds. Key findings indicate that solar irradiation intensity and heat transfer fluid mass flow rate are the most influential parameters affecting collector efficiency, with optimal performance achieved at Reynolds numbers between 10000-15000. The developed framework enables rapid parametric optimization and real-time performance assessment for parabolic trough solar thermal collector systems, providing significant advantages for design engineers and system operators in concentrating solar power applications.