Predictive Simulation Applied To Refinery Hydrogen Networks For Operators' Decision Support

Hydrogen networks are essential in modern oil refineries for they supply the main reactant for hydrodesulphurisation processes. These networks are usually integrated in a complex fashion, and sensitive to changes in any of their process units. In addition, plant measurements and prediction tools available, tend to be insufficient to make educated decisions when unexpected changes realize, especially on-line. This study presents an embedded dynamic estimator within a simulation framework, which uses plant data to estimate states and parameters of a hydrogen network and predicts its future behavior. A representative process network of three consumers and two sources is the case study utilised to demonstrate the usefulness in their decision-making process of operators. For this purpose, a brief what-if-analysis is discussed. In the last section, future challenges and directions of this predictive simulation tool are highlighted. (C) 2019, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.