Three-dimensional thermal modelling and heat transfer analysis in the heat collector element of parabolic-trough solar collectors

Highlights: • Fast and realistic unsteady formulation of a 3D thermal model in the HCE is applied. • FVM is used in thermal-thermohydraulic model and ray-tracing for optical modelling. • The inability of standard HTC to rightly predict absorber thermal fields is proved. • A correction factor for the standard HTC is successfully validated in the DISS loop. • A new volumetric approach is evaluated for heat transfer through the glass envelope. The prediction of thermal distributions around the heat collector element (HCE) is a key issue for the safety and efficiency in parabolic-trough solar collectors. Obtaining 3D temperature fields involving the non-uniform heat flux distribution (NUHFD) around the receiver becomes an essential matter for modelling and simulation tools. Several 3D numerical studies have been implemented using computational fluid dynamics (CFD) commercial software, but with high computational effort. As an alternative, 3D HCE models coupled to 1D heat transfer fluid (HTF) problem results in a much lower computational cost and accuracy enough. In the present work, a realistic 3D HCE − 1D HTF model under an unsteady formulation of the partial differential equations is implemented to properly calculate the receiver thermal distribution. The model is solved using the finite volume method, involving the NUFHD through a Monte Carlo ray-tracing method implemented in SolTrace. Its main novelty is to involve a correction factor (CF) in the standard heat transfer coefficient (HTC) correlations for uniform boundary conditions (BC), due to their inability to correctly predict the absorber thermal profiles. The suggested CF is based on the azimuthal local Nusselt reported in past studies for circumferentially-varying BCs, and on the absorber experimental data from the Direct Solar Steam (DISS) test facility. The model is validated in the AZTRAK platform and the superheated steam region of the DISS facility under steady-state conditions. The heat transfer variables mean deviations are lower than $2.4\%$ and $7.0\%$ for AZTRAK and DISS facilities, respectively. The involvement of a CF in the DISS facility improves the accuracy of absorber cross-section thermal gradients predictions, reducing the mean deviations from $22.2\%$ (without considers it) to $6.9\%$. Otherwise, the verifications against previously models in AZTRAK platform certify the necessity to correct the standard HTC, but the absence of absorber thermal profiles experimental data inhibits its validation.