A Novel Cooling System by Surface Corrugation and Nanofluid Utilization for the Performance Improvement of Photovoltaic Module Coupled with Thermoelectric Generator and Efficient Computations by Using Artificial Neural Network-Based Hybrid Scheme

For a photovoltaic module coupled with thermoelectric generator, a unique wavy cooling channel is proposed, and its performance is numerically assessed by using three-dimensional computations. The cooling channel uses nanofluid of alumina–water with various shaped nanoparticles (spherical, cylindrical and brick). Numerical simulations are performed for a range of parameters for the corrugation amplitude (\(0 \le \text {Amp} \le 0.1\)), wave frequency (\(2 \le \text {Nf} \le 16\)), nanoparticle loading quantity (\(0 \le \text {SVF} \le 0.03\)), and nanoparticle shape (spherical, brick, and cylindrical). We analyze the photovoltaic module’s average temperature and temperature uniformity for a variety of parameter variations. When nanofluid and greater channel corrugation amplitudes are utilized, the average panel surface temperature is decreased more. A wavy shape of the cooling channel at the maximum corrugation amplitude yields a cell temperature reduction of 1.88 \(^\text {o}\)C, while frequency has little impact on average cell temperature and its uniformity. The best-performing particles are those with cylindrical shapes, and the drop-in average photovoltaic temperature with solid volume fraction is essentially linear. As utilizing cylindrical-shaped particles, the average temperature of corrugated cooling channels decreases by around 1.9 \(^\text {o}\)C as compared to flat cooling channels with base fluid at the greatest solid volume fraction. As compared to un-cooled photovoltaic, cell temperature drops by around 43.2 \(^\text {o}\)C when employing thermoelectric generator. However, temperature drop value of 59.8 \(^\text {o}\)C can be obtained by using thermoelectric generator and nano-enhanced wavy cooling channel utilizing cylindrical-shaped nanoparticles. An hybrid computational strategy for the fully coupled system of photovoltaic with cooling system is provided, which reduces the computational time by a factor of 75.