Thermal cracking furnace system is the most critical section of an ethylene plant, where multiple furnaces in parallel are used to break various hydrocarbon feedstock (e.g. naphtha, ethane, and propane) into smaller-molecular hydrocarbon products such as ethylene, propylene and methane. Based on the provided inputs including product yield coefficients with respect to multiple types of feedstock, furnace performance constraints, product sale prices, and feedstock/manufacturing costs, as well as designated operational constraints, an optimal decoking schedule for the entire ethylene cracking furnace system has to be determined in order to achieve the maximized economic performance. Due to its performance-decaying operational characteristic, the cracking furnace system has to experience periodic hot shutdowns, which consequently leads to decreased throughput for the downstream section and unwanted upsets. Apart from the hot shutdown, a longer outage is required to conduct major maintenance, which is called as cold maintenance. In order for a cold maintenance to occur, the project team will provide the estimated starting/ending times to the furnace operation team, according to which a feasible schedule will be conceived. Also, in practice, because the feed supply conditions change constantly, it is necessary to implement a reactive scheduling strategy which can smartly reschedule the furnace operations with respect to any new feed delivery.
The previously published works are either focused on simulation and control aspects, or simply scheduling the furnaces for maximum average daily profits without incorporating some realistic operational constraints. In the proposed scheduling model, most of the major scheduling issues concerning an ethylene cracking furnace system have been addressed, such as multiple feedstock, multiple furnaces, hot shutdowns for decoking operations, cold shutdowns for mechanical maintenance, decaying cracking performance, non-simultaneous shutdowns, split cracking and recycled C2/C3 cracking. Specially, to mitigate the throughput upsets caused by furnace shutdowns to the downstream section, the developed scheduling model considers different measures commonly applied in industry, such as additional furnace capacity makeup. Furthermore, the scheduling model is constructed in a reactive and iterative way to seamlessly make transitions from current schedule to reschedule under any delivery of new feeds and on a rolling forward scheduling horizon. The efficacy of the developed scheduling model is demonstrated by some case studies.
See more of this Group/Topical: Computing and Systems Technology Division