No abstract

A novel catalytic coating that converts coke to carbon oxides through a reaction with steam has been developed. Several coating formulations were tested in a jet-stirred reactor setup, and the best performing formulation was further evaluated in a pilot plant setup. Application of the coating during steam cracking of ethane at industrially relevant conditions resulted in a reduction of the asymptotic coking rate by 76%. The coating activity remained constant over several coking/decoking cycles. Coupled furnace-reactor run length simulations of an industrial ethane cracking unit were performed and resulted in an increase of the run length by a factor of 6. However, the simulated CO2 yield is higher than the design value of a typical caustic tower.

The use of one-dimensional reactor models to simulate industrial steam cracking reactors has been one of the main limiting factors for the development of new reactor designs and the evaluation of existing three-dimensional (3-D) reactor configurations. Therefore, a 3-D computational fluid dynamics approach is proposed in which the detailed free-radical chemistry is for the first time accounted for. As a demonstration case, the application of longitudinally and helicoidally finned tubes as steam cracking reactors was investigated under industrially relevant conditions. After experimental validation of the modeling approach, a comprehensive parametric study allowed to identify optimal values of the fin parameters, that is, fin height, number of fins, and helix angle to maximize heat transfer. Reactive simulations of an industrial Millisecond propane cracker were performed for four distinct finned reactors using a reaction network of 26 species and 203 elementary reactions. The start-of-run tube metal skin temperatures could be reduced by up to 50 K compared to conventionally applied tubular reactors when applying optimal fin parameters. Implementation of a validated coking model for light feedstocks shows that coking rates are reduced up to 50%. However, the increased friction and inner surface area lead to pressure drops higher by a factor from 1.22 to 1.66 causing minor but significant shifts in light olefin selectivity. For the optimized helicoidally finned reactor the ethene selectivity dropped, whereas propene and 1,3-butadiene selectivity increased with a similar amount. The presented methodology can be applied in a straightforward way to other 3-D reactor designs and can be extended to more complex feedstocks such as naphtha.

No abstract

No abstract