Integrated MCRT–CFD modeling and neural-network estimation of tracking error for a parabolic trough collector

Author : Belkacem Agagna, Ahmed Sayadi, Abdelhamid Bouhelal

Abstract : We present an integrated workflow that couples a Monte-Carlo ray tracing (MCRT) optical model with a Computational Fluid Dynamics (CFD) thermal model to quantify the impact of tracking errors on a parabolic trough collector (PTC), and a lightweight neural network (NN) to estimate tracking error from synthetic flux/LCR maps. The testbed is the MicroSol-R pilot plant composed of three 12-m PTCs with model inputs and operating conditions drawn from the facility. The optical model produces Local Concentration Ratio (LCR) curves that degrade systematically with tracking error (0–20 mrad), and ray-path visuals that explain power spillover at 10 mrad. CFD predictions of HTF outlet temperature and thermal efficiency agree well with MicroSol-R measurements under the same conditions. Trained on MCRT-generated data, the NN reproduces measured tracking error with milliradian-level residuals and a slight underestimation bias. The combined approach provides a practical tool for diagnosing mis-tracking and quantifying associated thermal losses in PTCs.

Keywords : Parabolic trough solar collector, Numerical study, Tracking error, Neural network, Performance analysis.

Conference Name : International Conference on Renewable Energy and Climate Change Impacts (ICRECCI - 26)

Conference Place : Istanbul, Turkey

Conference Date : 19th Jan 2026