A large-scale stochastic simulation-based thermodynamic optimization for the hybrid closed circuit cooling tower system with parallel computing

The emerging multi-mode cooling tower can cool down the circulating water by flexibly switching the operating modes according to varying weather conditions. Herein, a computational framework for addressing a large-scale stochastic simulation-optimization task is developed to obtain the optimal thermodynamic performance of the multi-mode cooling system. First, the numerical model is constructed using a well-validated evaporative cooler in the wet and wet-heating modes, as well as an air cooler in the dry mode. A well-suited experimental design is performed for generating an optimal set of samples by approximating the multivariate probability distributions of uncertain data. To reduce the computational burden, a customized parallel computing strategy is presented via parallelization of the task using the message-passing interface. Finally, an example illustrates that the time reduction is up to 93.5%, while the optimal exergy efficiency ratios are expected to be 37.0%, 17.3%, and 22.6% for the wet, dry, and wet-heating modes, respectively.