Thermal management for conjugate heat transfer of curved solid conductive panel coupled with different cooling systems using non-Newtonian power law nanofluid applicable to photovoltaic panel systems

Abstract

Thermal performance features for a coupled conjugate thermo-fluid system with different cooling configurations (flat channel (F-C), grooved channel (G-C) and impinging jets (I-J)) are explored numerically by using non-Newtonian nanofluid. The numerical work is performed for different Reynolds numbers (100 <= Re <= 300), index of power law (0.8 <= n <= 1.2), height (0.1H <= b <= 0.6..) and number (2 <= N <= 9) of corrugation in the G-C system, number (2 <= N-j <= 9) and distance (5 omega <= sx <= 25 omega) between jets in the I-J flow system. Different volume fractions (0 <= phi <= 0.04) and particle sizes (20 nm <= dp <= 80 nm) of nanoparticles are used. When systems operating at the highest and lowest Re are compared, 9 K, 11 K and 8 K temperature reduction are achieved for F-C, G-C and I-J cooling systems. However, I-J flow system at higher Re is very effective on the thermal performance improvement when shear thickening fluid is used. For the G-C flow system, increasing the height and number of the corrugation waves resulted in improvement in the thermal performance. Up to 46% increment in the Nu number (average) and reduction of 6.5 K in the average surface temperature are achieved with varying the height of the corrugation while these values are 17% and 1.6 K when wave number is increased. The average Nu number rises by about 32% and temperature drops by about 6.5 K when jet number is varied from 3 to 11, while these values are obtained as 8% and 4 K for when distance between jets are varied from sx = 5 omega to sx = 25 omega. For F-C, G-C and I-J flow systems, average Nu rises by about 15.5% and 14.5% and 16.3% for shear thinning fluid while they become 16.6%, 9.94% and 12.8% for shear thickening fluid at the highest solid volume fraction. As the nanoparticle size is increasing, there is 6% and 7% reduction in the average Nu number. Thermal performance estimations are made with four inputs and four outputs system by using artificial neural networks.