Abstract
Liquefied natural gas (LNG) plants are energy intensive. One way to reduce their energy consumption is to apply optimization methods when designing such plants. In this paper, genetic algorithm (GA) from Matlab optimization toolbox was used to optimize a propane pre-cooled mixed refrigerant (C3-MR) LNG plant that was originally designed by Mortazavi et al.
[1]. GA was chosen because it can reach a global optimum with any problem. A computer model was developed for the LNG plant using HYSYS and verified with the model developed by Mortazavi et al., with good agreement.
The optimization problem has 22 variables and 24 constraints. In order to reduce the complexity of the problem, optimization was carried out in two stages. First, MCR cycle optimization and then Propane cycle optimization were conducted with respective constraints. New refrigerant mixtures were found, with savings in power consumption as high as 13.28%. Propane cycle optimization resulted in a savings of 17.16% in power consumption. The optimized C3-MR LNG plant model consumes 100.78 MW, whereas the baseline consumes 110.84 MW. The optimum composition of refrigerant mixture obtained was compared with two optimized compositions of refrigerant mixtures from the open literature. The resulting power consumption utilizing the literature-referenced mixtures is 6.98% and 13.6% more than this work’s optimum composition of refrigerant mixture.
C3-MR LNG plant optimization was conducted with four pinch temperatures (0.01, 1, 3 and 5 K) that represent different common heat exchangers in LNG applications (e.g., spiral-wound heat exchanger and plate fin heat exchanger). Power savings is increased significantly with a pinch temperature of 1 K as compared to 3 or 5 K, but with little improvement as compared to 0.01 K. This figure can have a significant impact on LNG plants selection.
[Display omitted]
►Verified computer model for a propane pre-cooled mixed refrigerant (C3-MR) LNG plant was developed using HYSYS software. The LNG plant model was optimized for minimum power consumption with the help of GA from Matlab software. ►Different optimum composition of refrigerant mixtures were found as a function of the cryogenic heat exchanger’s pinch temperature which represent different heat exchanger type. ►This work optimum composition of refrigerant mixtures was compared with two optimized compositions of refrigerant mixtures from the open literature.