(258i) Computationally Efficient CFD Simulations With Detailed Free-Radical Mechanisms

Chemical engineers are continuously trying to increase process efficiency to increase process throughput and profitability while lowering environmental cost. One such example is the replacement of conventional tubular reactors by more advanced three-dimensional reactor configurations in steam cracking units such as the patented Swirl Flow Tube of Technip or the MERT technology of Kubota. By means of improved heat transfer, the reactor metal temperatures and the corresponding coke formation on the reactor inner wall decreases, which enables longer process operation and/or higher throughput [1]. To assess the effect of the geometry of these three-dimensional reactor configurations on product yields and coke formation, a three-dimensional reactor model is necessary. Furthermore, to accurately describe the chemical kinetics over a wide range of process conditions and feedstocks, a detailed microkinetic model is required. Only these microkinetic models capture the essential chemistry and allow prediction of the effect of reactor geometry on product yields.

The incorporation of these detailed kinetic models within complex computational fluid dynamics (CFD) simulations requires extremely long CPU times [2]. The reason is twofold. First, every species in the reaction network adds an extra continuity equation that needs to be integrated. Second, the large differences in the species time scales due to the distinct reactivity of molecules and radicals makes the resulting system inherently stiff. To alleviate the computational burden while retaining the accuracy of detailed kinetic models, different methodologies have been developed and evaluated [3-7]. In chemical reaction engineering the quasi-steady-state approximation (QSSA) is often applied [8, 9]. The applicability of QSSA on steam cracking kinetics has been validated previously using a 1D plug flow reactor model [9]. It allows to obtain the radical species’ concentrations from a set of algebraic equations, greatly reducing  stiffness and computational time.

In the present contribution, QSSA is applied in three-dimensional simulations of steam cracking reactors. The implemented reaction mechanisms contain over 50 species and a few hundred reactions. The Navier-Stokes equations, turbulence equations, energy equation and molecular species’ continuity equations are solved using the commercial FLUENT/13.0 software. The radical concentrations are determined from algebraic equations by application of QSSA outside the main FLUENT solver. This reduces both the number of continuity equations and stiffness, yielding a speed-up of up to a factor 5. First, the approach is validated by comparison to reference simulations, i.e. without QSSA, for reaction networks of different sizes and complexity. Next industrial steam cracking reactors with a three-dimensional configuration are simulated  and compared with conventional tubular reactors in terms of  product yields and coking tendency. The advantages and disadvantages of using three-dimensional reactor configurations instead of conventional tubular reactors are discussed.

ACKNOWLEDGMENTS

CMS acknowledges financial support from a doctoral fellowship from the Fund for Scientific Research Flanders (FWO). The authors also acknowledge the financial support from the Long Term Structural Methusalem Funding by the Flemish Government – grant number BOF09/01M00409. The computational work was carried out using the STEVIN Supercomputer Infrastructure at Ghent University, funded by Ghent University, the Flemish Supercomputer Center (VSC), the Hercules Foundation and the Flemish Government – department EWI.

REFERENCES

[1] J.J. De Saegher, T. Detemmerman, G.J. Froment, Three dimensional simulation of high severity internally finned cracking coils for olefins production, Revue de l'Institut Francais du Petrole, 51 (1996) 245-260.

[2] W.J. Pitz, C.J. Mueller, Recent progress in the development of diesel surrogate fuels, Progress in Energy and Combustion Science, 37 (2011) 330-350.

[3] K.Y. He, I.P. Androulakis, M.G. Ierapetritou, Multi-element Flux Analysis for the Incorporation of Detailed Kinetic Mechanisms in Reactive Simulations, Energy & Fuels, 24 (2010) 309-317.

[4] K.Y. He, M.G. Ierapetritou, I.P. Androulakis, Integration of On-The-Fly Kinetic Reduction with Multidimensional CFD, Aiche J., 56 (2010) 1305-1314.

[5] L. Liang, J.G. Stevens, J.T. Farrell, A dynamic adaptive chemistry scheme for reactive flow computations, Proceedings of the Combustion Institute, 32 (2009) 527-534.

[6] Y. Shi, L. Liang, H.W. Ge, R.D. Reitz, Acceleration of the chemistry solver for modeling DI engine combustion using dynamic adaptive chemistry (DAC) schemes, Combustion Theory and Modelling, 14 (2010) 69-89.

[7] S.B. Pope, Computationally efficient implementation of combustion chemistry using in situ adaptive tabulation, Combustion Theory and Modelling, 1 (1997) 41-63.

[8] G.M. Come, Radical reaction-mechanisms - Mathematical-theory, Journal of Physical Chemistry, 81 (1977) 2560-2563.

[9] K.M. Van Geem, M.F. Reyniers, G.B. Marin, J. Song, W.H. Green, D.M. Matheu, Automatic reaction network generation using RMG for steam cracking of n-hexane, Aiche J., 52 (2006) 718-730.