Experimental and machine learning prediction of a triangular absorber tube configuration in a parabolic trough collector using ZnO- based Nanofluids

Abstract

This study investigates the thermal and hydraulic performance of a novel parabolic trough collector (PTC) design incorporating a triangularly arranged absorber tube system and ZnO/DIW nanofluids at various volume concentrations ranging from 0.1 to 0.4%. Experimental analyses were conducted under real solar irradiation conditions to evaluate heat transfer coefficient, friction factor, and thermal efficiency across different Reynolds numbers. Among all tested conditions, a ZnO/DIW volume concentration of 0.4% exhibited the highest enhancement in thermal conductivity, heat transfer coefficient, friction factor, and thermal efficiency. In addition to the experimental investigation, machine learning (ML) models including Gaussian Process (GP), Random Forest (RF), Linear Regression (LR), and M5P were employed to predict key thermohydraulic parameters. Out of 320 data points, 200 were used for training and 120 for testing. The GP model consistently showed the highest predictive accuracy across all parameters, especially at 0.4%, followed by RF, while LR and M5P showed lesser performance. This integrated experimental-ML approach validates the feasibility of enhancing thermal efficiency in PTC systems through optimized nanoparticle loading and soft computing techniques. The findings support the use of ZnO-based nanofluids and triangular absorber configurations as an effective strategy to improve the thermal performance of parabolic trough collectors without relying on conventional vacuum insulation.