Parametric optimization of solar air heaters with dimples on absorber plates using metaheuristic approaches

The performance optimization analysis of artificially roughened solar air heaters (SAH) becomes complex due to the inherent difficulties of the optimization analytical processes. Metaheuristic algorithms can be employed as optimization methodologies to handle complex engineering problems with minimal computational efforts and high accuracy. However, due to their stochastic nature, only some metaheuristic algorithms are suitable for solving a specific problem. Hence, there is a need to perform a comparative performance analysis of different metaheuristic algorithms in optimizing SAH. The current paper presents a comparative performance analysis of four metaheuristics algorithms, i.e., Firefly algorithm (FA), Particle swarm optimization (PSO), Differential evolution (DE), and Teaching learning-based optimization (TLBO), in optimizing the system and operating parameters of SAH. Two types of artificially roughened SAH absorber plate design patterns have been taken: dimpled plate in a zig-zag pattern and straight-line pattern. Both single and multi-objective optimization techniques are carried out while viewing the maximization of Nusselt number (Nu) and minimization of friction factor (Ff) as two objective functions in single-objective optimization. Results showed that the TLBO algorithm delivered better optimal solutions for single and multi-objective optimization than the other three metaheuristic algorithms. In the single-objective optimization, the maximum improvement in Nu and Ff is 16.81 % and 3.61 %, respectively. In the multi-objective optimization of a zig-zag pattern, the suitable parametric setting is Re = 20000, S/e = 30.83, L/e = 30.83, and d/D = 0.2515 for the optimal value of Nu and Ff as 204.7034 and 0.0134, respectively. For a straight-line pattern, the optimal values for Nu and Ff are 108.6901 and 0.0237, respectively at Re = 12000, p/e = 12, and e/D = 0.0379.