Predictive control using mirror plant and application to LNG regasification plants

We describe the development of Mirror Plant, a system in which a virtual plant through constant connection with a real plant is able to reproduce its behavior at any time. Mirror Plant can support plant operation by not only visualizing difficult to measure values but also predicting their behavior, and is already being used in commercial LNG regasification plants. In Japan, the combination of increasing disturbances to calorific value adjustment control systems caused by diversification of raw materials plus strictly regulated gas calorific values presents many difficulties to the operation of LNG regasification plants. In this paper, we introduce a novel robust control method for calorific value adjustment based on the visualization and prediction capabilities of Mirror Plant. Simulation studies show that the proposed system provides a much higher degree of control over conventional schemes.