Second-law analysis of nanofluid-based photovoltaic/thermal system modeling and forecasting model based on artificial neural network

The PVT solar collectors can produce the thermal energy and power in a same frame. The improvement of the PVT's efficiency leads to reducing the system size and capital costs. To this end, this paper studied the irreversibilities of the Al₂O_3[sbnd]Cu/water hybrid nanofluid (NF) in a PVT solar collector considering two single and double serpentine channels (SS and DS). The influences of Re and nanoparticle concentration (φ) on the thermal and frictional entropy generation rates (S˙_th and S˙_fr) were investigated and the thermal, electrical and overall exergy efficiencies (ψ_th, ψ_e, ψ_ov) of the PVT with SS and DS channels were compared and discussed. Based on the results, the DS channel exhibited S˙_fr of almost 75 % lower than SS channel due to lower nanofluid inlet velocities and velocity gradients. In addition, S˙_th for the DS channel is nearly 65 % lower and 26 % higher than that for the SS channel at Re numbers of 500 and 2000, respectively. Besides, the Re escalation from 500 to 2000 intensifies S˙_fr by almost 94 % at different φs in the SS and DS channels. The increase in φ from 0 % to 1 % escalates S˙_fr by almost 99.98 % times for two configuration regardless of the Re number. ψ_th of the DD channel is nearly 14.5 % and 12.77 % higher than that of the SS channel at Res of 500 and 2000, respectively. Besides, ψ_e of the PVT with the DS channel is 2.36 % higher than that with SS channel at Re=500 at four studied φs. Moreover, the maximum ψ_e for the PVT with the DS and SS were obtained as 22.29 % and 21.28 %, respectively, which are associated with Re=1500 and φ=0.25 %. Additionally, a predictive model was presented to determine the total entropy generation rate based in the Re and φ as the inputs.