Active constraint regions for a natural gas liquefaction process

Optimal operation of liquefaction processes is little studied in the open literature. In particular, the issue of how optimal operation changes with disturbances has received very little attention. This paper addresses optimal operation of a simple natural gas liquefaction process – the PRICO process. The focus is on how the active process constraints change with disturbances. It is shown that the feasible part of the disturbance space can be divided into five regions with different sets of active constraints. We also suggest control structures for the process, and find that as little as two control structures may be needed despite of the fact that there are five regions. It is suggested to use compressor speed to control the margin to surge at minimum, and to keep the turbine outlet stream at saturation at all times.

► We address optimal operation of a natural gas liquefaction process, the PRICO process.
► We study how the active constraints for optimal operation vary with disturbances.
► The considered disturbances are feed rate and ambient temperature.
► We find that there are five sets of active constraints inside the feasible region.
► We suggest control structures based on the self-optimizing control approach.