Multi-objective optimization and thermodynamic assessment of a solar unit with a novel tube shape equipped with a helical tape

Abstract

The performance of several heat transfer mechanisms, including solar energy sources, is being improved through optimization, which also effectively supports the most efficient feasible design. A numerical study has been done on a turbulator-enhanced flat plate solar collector (FPSC) while also introducing the utilization of a quad-lobed tube as a pioneering innovation in the system’s design. The incorporation of nanoparticles into the working fluid was implemented to enhance performance, with aluminum oxide particles selected for this purpose. Furthermore, assessing the second law and analyzing exergy generation was employed to discern irreversibility within the system. Optimization of both geometric and non-geometric elements has been accomplished through the integration of a multi-objective optimization (MOO) method with machine learning (ML). Random Forest (RF), LASSO Regression (LaR), and Support Vector Regression (SVR) are the machine learning models that were utilized in order to ascertain the correlations that exist between the system’s inputs and outputs. According to the findings, RF was found to be the most appropriate algorithm for capturing the impacts of parameter variations. This was demonstrated by the high coefficient of determination (R2) quantities that RF possessed, which were 0.95, 0.99, and 0.99 for friction factor (f), exergy loss (Xd), and second law effectiveness (ηII), respectively. However, the SVR and LaR algorithms showed significant deviations from the FPSC simulation results. The SVR algorithm achieved R2 values of 0.54 for f, 0.78 for Xd, and 0.84 for ηII, while the LaR algorithm achieved R2 values of 0.65 for f, 0.99 for Xd, and 0.98 for ηII. This research aims to find the optimal input parameters that minimize f and Xd while maximizing ηII. Therefore, to determine the Pareto optimal solutions the non-dominated sorting genetic algorithm (NSGA-II) has been utilized. The result was an illustration of the Pareto set, which stands for a collection of optimal solutions. Every solution in this collection achieved the desiredbalance, obtaining the best outcomes without compromising any objectives.