CFD modeling and predicting the performance of direct absorption of nanofluids in trough collector

In this paper, two types of nanoparticles are used in a solar direct absorption parabolic trough collector (DAPTC). Nanosilica and multi-wall carbon nanotube (MWCNT) with two different volume fractions are used in ethylene glycol (EG) in the experimental section. A glass-glass absorber tube is utilized as the receiver, and the working nanofluid goes through the tube. The computational fluid dynamics (CFD) method is used for simulating the process, and it is validated by experimental data. The volume fractions of the nanoparticles in the base-fluid are 0.1-0.5% and 0.1-0.6% for nanosilica and MWCNT, respectively. The nanofluid including 0.6% MWCNT/EG has the highest outlet temperature which is equal to 346.1 K. The experimental results used to validate the CFD procedure. In the last step, a multi-layer perceptron neural network is trained using the data obtained from the CFD model. It has been shown that the outputs of the network have good agreement with the CFD results. Both the CFD and the ANN methods were compared, and the advantages and disadvantages of each technique in modelling the problem have been discussed.