Soft computing analysis of thermohydraulic enhancement using twisted tapes in a flat-plate solar collector: Sensitivity analysis and multi-objective optimization

Flat-plate solar collectors are a cost-effective technology for water heating in residential buildings. Comprehensive sensitivity analysis and multi-objective optimization of five crucial factors in a flat-plate solar collector equipped with twisted tapes are investigated. The studied factors are the Nusselt number (Nu), friction factor (f), solar collector efficiency (eta), thermal performance factor (xi), and the temperature difference of the heat transfer fluid across the device (Delta T). The simulation of functions is carried out by four soft computing methods, such as the linear and cubic form of multivariate adaptive regression splines (MARS), group method of data handling (GMDH), and multivariate polynomial regression (MPR). The linear form of the MARS algorithm shows the highest accuracy in predicting Nu, f, and eta by R-2 = 0.99986, 0.9984, and 0.98821, respectively. The Reynolds number (Re) is from 4000 to 14,000, and the twisted tape ratio (Y) is between 0 and 15. The simulated annealing optimization is employed to find the global optima. According to the study, Nu, eta, xi, and Delta T reach their maximum when f is minimum, which occurs for Y = 5 and Re in the 13,200 to 13,400 range for different global solar radiation levels. The maximum temperature difference happens at noon, about 6.54 K at the Re = 5362.5 and Y = 5 with the collector surface area by 2 m(2). When the appropriate working conditions are met so that these factors are simultaneously optimal, Delta T, hence the efficiency, can be increased.