Computational fluid dynamics‐based steam cracking furnace optimization using feedstock flow distribution

Nonuniform temperature fields in steam cracking furnaces caused by geometry factors such as burner positions, shadow effects, and asymmetry of the reactor coil layout are detrimental for product yields and run lengths. The techniques of adjusting burner firing (zone firing) and feedstock mass flow rate (pass balancing) have been practiced industrially to mitigate these effects but could only reduce the nonuniformities between the so‐called modules (a group of many coils). An extension of the pass balancing methodology is presented to further minimize the intra‐module nonuniformities, that is, variation between the coils within a module, in floor fired furnaces. Coupled furnace‐reactor computational fluid dynamics‐based simulations of an industrial ultraselective conversion (USC) furnace were performed to evaluate four different feedstock flow distribution schemes, realizing equal values for coil outlet temperature, propene/ethene mass ratio, maximum coking rate and maximum tube metal temperature (TMT), respectively, over all the reactor coils. It is shown that feedstock flow distribution creates a larger operating window and extends the run length. Out of the four cases, the coking rate as criterion leads to the highest yearly production capacity for ethene and propene. Uniform maximum coking rates boost the annual production capacity of the USC furnace with a nameplate ethene capacity of 130 10³ metric tons per year with 1000 metric tons for ethene and 730 metric tons for propene. For industrial application, achieving uniform maximum TMT is more practical due to its measurability by advanced laser‐based techniques. Most steam cracking furnaces can be retrofitted by optimizing the dimensions of venturi nozzles that regulate the feedstock flow to the coils. © 2017 American Institute of Chemical Engineers AIChE J, 63: 3199–3213, 2017