Direct absorption collectors (DAC) with nanofluid are among the most promising yet least studied in solar energy technology. There are numerous micro- and macroscopic factors that determine their efficiency. This complicates in situ optimization of DACs using physical prototypes. The present paper describes a multiphase CFD model of the collector, which was validated against two independent experimental datasets. The model was used for a multiparametric numerical analysis, where we altered concentration and size of the nanoparticles, as well as the geometry and inclination of the collector. The optimization resulted in up to 10% improvement in the collector's efficiency. Finally, we considered the process of thermomagnetic convection in the collector using a magnetic nanofluid. This resulted in a 30% increase in the collector performance. (C) 2018 Elsevier Ltd. All rights reserved.