Dynamic modeling and offset-free predictive control of LNG tank

초록

Due to tightened regulations on sulfur oxide (SOx) and nitrogen oxide (NOx) emissions, especially, in the ocean industry, liquefied natural gas (LNG) is becoming more important as an eco-friendly fuel source because of its extremely low SOx and NOx emissions. However, since LNG is usually stored at below its boiling temperature (-160 degrees C) and slightly above the atmospheric pressure, the invasion of natural heat causing evaporation of LNG is inevitable even with proper insulation. This evaporated LNG called boil-off gas (BOG) accumulates inside the tank, and without proper venting, the pressure inside the tank will build up and eventually reach the maximum allowable working pressure (MAWP) of the storage tank. The violation of this threshold will lead to serious consequences. Therefore, safe control of storage tank pressure is critical to its safe operation. In this paper, we propose to use a model predictive control (MPC) system to regulate the LNG tank pressure. Firstly, a dynamic model of the LNG storage tank is developed using the film evaporation theory. Secondly, a reduced-order model is developed using the N4SID algorithm. Finally, the offset-free MPC system scheme is applied to control the tank pressure. For demonstrating the performance of the offset-free MPC system in unnatural circumstances, a fire around the tank is applied. The results show that the offset-free MPC system is able to regulate the pressure even when a fire accident occurs despite the plant-model mismatch in such conditions.

키워드