Eulerian multiphase study of direct steam generation in parabolic trough with OpenFOAM: Fid-Bau Portal

Eulerian multiphase study of direct steam generation in parabolic trough with OpenFOAM

Achi, Alladdine / Demagh, Yassine

Direct steam generation (DSG) in parabolic trough solar collectors is a feasible option for economic improvement in solar thermal power generation. Three‐dimensional Eulerian two‐fluid simulations are performed under OpenFOAM to study the turbulent flow in the evaporation section of the parabolic trough receiver and investigate the phase change, and pressure drop of water as a heat transfer fluid. First, the model's validity has been tested by comparing the numerical results of a laboratory scale boiler with the available correlations and semi‐correlations of boiling flows from the literature. Simulations agreed well with Rouhani–Axelsson correlation for horizontal tubes, with a mean relative error of less than 7.1% for all studied cases. However, despite a mean relative error of less than 13.19% compared to the experimental data in the literature, the reported pressure drop factor remains valid; overprediction remains tolerable for most engineering applications. Second, the scaling effect on the mathematical model, from laboratory to commercial‐scale configuration, was tested with experimental data of the DISS test loop in Platforma Solar de Almeria, Spain. The Monte Carlo Ray Tracing method under the Tonatiuh package allowed for obtaining the nonuniform heat flux distribution. Due to the large size of the evaporation section in the DISS loop (eight collectors), each collector is considered independently in the simulations. Thus, simulations follow each other, taking the numerical results of each collector output as input data in the next collector and so on until the last. The numerical results showed an excellent agreement for the void fraction with 3.53% against the Rouhani–Axelsson correlation. Frictional pressure losses are within a 17.06% error of the Friedel correlation, in the range of previous work in the literature, and the heat loss is less than 4.69% error versus experimental correlation.