Performance Evaluation and Comparison of Commonly Used Optimization Algorithms for Natural Gas Liquefaction Processes

The mixed refrigerant (MR) process is the preferred choice for most LNG plants. However, due to the complexity of refrigerant components and parameters, and their interaction, the MR process must be optimized to take full advantage of its efficiency. The optimization of processes has received great attention. Nevertheless, there are limited studies that analyze and compare the performance of different algorithms for optimizing processes. Thus, this article presents a comparative analysis of the performance of different algorithms, utilizing the widely-used GA and BOX method, and a long-neglected method, the PSO algorithm to optimize high-energy-efficient SMR, DMR, C3MR, and MFC processes. For the given conditions, it was found that the PSO optimization had the best performance and the lowest specific energy consumption, with a maximum reduction of 20.92%. Furthermore, the exergy analysis showed that the lowest exergy loss after the PSO optimization, resulting in a maximum reduction of 31.13%. The study also discovered that the PSO method has better stability than the GA, whereas unit computational time is reversed. This study would show that the results of the PSO optimization could be used as an important indicator to evaluate the performance of algorithms under similar working conditions