(583n) Combining CFD and Statistical Methods to Optimize the Oxidative Coupling of Methane in a Fixed Bed Reactor | AIChE

(583n) Combining CFD and Statistical Methods to Optimize the Oxidative Coupling of Methane in a Fixed Bed Reactor


Eppinger, T. - Presenter, Technische Universität Berlin
Wehinger, G. D., Technische Universität Berlin
Kraume, M., Technical University Berlin

In recent years, the use of computational fluid dynamics (CFD) has made large progress due to increasing CPU power and decreasing costs for electronic devices. Therefore, CFD can be used to calculate and design chemical reactors. But still, CFD simulations of spatially resolved geometries in combination with chemical reactions are time and resource consuming. Furthermore, a parameter study to optimize a reactor involves many simulations. Consequentially, a method for the efficient use of computational experiments is beneficial.

Hydrocarbons show significant importance in the field of industrial organic chemistry. Until now, they are mainly produced by using crude oil. But industry is seeking for alternatives. The oxidative coupling of methane (OCM) represents one direct path of converting methane to higher hydrocarbons. This process is characterized by its exothermic homogeneous heterogeneous behavior. Until now, no industrial application of this promising process have been realized.

In the following contribution a time and resource saving way is discussed for optimizing the OCM with the help of 3D CFD simulations in terms of C2 yield. A single-pass fixed bed reactor is simulated by varying temperature, feed composition, gas flow rate, nitrogen gas dilution ratio, catalyst mass and inert material mass. CFD simulations are realized by a pseudo-homogeneous reactor model based on a 10-step OCM kinetics. In this model the structure of the catalyst is not considered explicitly. Therefore, effective thermal conductivity and diffusion coefficients are used. All simulations are carried out with the commercial tool STAR-CCM+ by CD-adapco.

The optimization is carried out in consecutive steps. First, a parameter screening identifies the significant process variables. Second, with design of experiment (DoE) a small number of simulations are carried out. Third, with these results a statistic based meta model describes the behavior in a wide parameter range. Finally, this model identifies the maximum yield by a response surface and is validated by a final CFD simulation.

It is demonstrated that the design of experiment with the consecutive optimization represents an efficient practical method in combination with CFD. The time saving while using DoE is highly promising and can be expanded in this field of research.