Improving the performance of a condensation water production system through support vector machine modeling and genetic algorithm optimization

Abstract

Abstract Water scarcity is recognized as a critical global concern and one viable solution involves extracting water from atmospheric humidity by leveraging subterranean coldness. This study meticulously evaluates the operational efficacy of a water production system by examining four pivotal factors: buried pipe length (TL), air flow rate (AFR), air temperature (AT), and air humidity (AH). A positive correlation between these variables and water vapor production is established, with AT exerting the most pronounced influence. Significantly, the analysis of variance reveals the main and interactive effects of the variables, except for TL × AFR, at a 5% significance level. To enhance understanding of the intricate interplay among these factors, a proficient least squares support vector machines model is devised, employing a radial basis function kernel. This model demonstrates an impressive 98% concurrence between projected and empirical data, with a minimal error of 0.66 mL and 5.99%. An in-depth sensitivity analysis underscores the differential impact of AT, AH, TL, and AFR on water vapor (WV) prediction. Optimal values of 3.98 m, 6.89 m3/h, 46.30 °C, and 86.62% for TL, AFR, AT, and AH, respectively, are obtained through subsequent optimization of independent variables using genetic algorithms, resulting in a notable water production of 23.61 mL.