(539a) Reactive Scheduling for Ethylene Cracking Furnace System

Thermal cracking is crucial for ethylene plants because the major product yields of an ethylene plant are mainly determined in this operation. The cracking furnace system follows a certain schedule to process all the coming feeds. In practice, because of the uncertainties hidden in the feedstock market and/or during the shipment, the lead time and the incoming quantity may not follow the expectation very precisely. Thus, it is better for the furnace scheduling conducted in a reactive way, such that various feeds from new shipments and existing inventories can be smartly allocated to different furnaces for processing to obtain the maximum production profit per unit time. The main challenges for this reactive scheduling study include: i) thermal pyrolysis process is a performance decay semi-continuous process, and thus a furnace has to be shut down after some processing time for decoking; ii) non-simultaneous decoking policy usually has to be ensured for mitigating the productivity upsets from the cracking furnace system; iii) the cracking product of ethane has to be recycled for a secondary cracking and usually needs a specific furnace for such purpose; iv) reactive scheduling has to seamlessly connect the previous schedule and the being determined schedule for taking care of different feeds from new shipments and existing inventories. Facing these challenges, this paper developed a novel reactive scheduling strategy, which can dynamically generate a reschedule based on the new coming feeds and the leftover feeds from the current schedule. A new MINLP model is also developed to maximize the profit per time under various stringent constraints. The efficacy of the study and its significant economic impact are demonstrated by case studies.