Turbulent combustion models can broadly be classified as topology free and topology based [1, 2]. Topology free models do not make any assumptions on the underlying flame structure or combustion mode and are consequently more broadly applicable to a variety of reacting flows. Probability density function (PDF) methods fall in the topology free category and have been shown to accurately capture flames with strong turbulence chemistry interactions. PDF methods are known to be more computationally intensive than simpler topology based approaches such as steady laminar flamelet models [4], where magnitude of the actual difference is strongly dependent on the chemical mechanism.

Liang et al. [3] have recently proposed a PPAC methodology to mitigate the cost of using particle PDF methods while maintaining the accuracy. This methodology includes an offline preprocessing stage, during which reduced models are generated for a representative database of compositions. At runtime, these models are then selectively utilized to integrate the particles. Additionally, the particles are only required to carry a reduced representation. A key assumption of the method is that compositions representative of those encountered at runtime can be generated a-priori with minimal computational expense. In the first part of this talk, I will briefly explain the PPAC methodology, followed by a critical examination of the aforementioned assumption by comparing compositions from a conventional reduced order model to compositions from a well known DNS database. An improved algorithm which will be the focus of future work, will also be briefly introduced.

The second part of the talk will detail recent efforts to combine PPAC with existing and complementary dimension reduction and storage retrieval techniques. Specifically, the integration of PPAC with rate-constrained chemical equilibrium (RCCE), in-situ adaptive tabulation (ISAT), and RCCE-ISAT will be discussed. Finally, we show that the combined PPAC-RCCE-ISAT methodology outperforms the standalone PPAC by reducing the number of variables that need to be retained at runtime and avoiding redundant integration during the reaction fractional step.

References:

[1] H. Wu, Y.C. See, Q. Wang, and M. Ihme. Combustion and Flame, 162(11):4208-4230, 2015.

[2] S.B. Pope. Proceedings of the Combustion Institute, 34(1):1-31, 2013.

[3] Y. Liang, S.B. Pope, and P. Pepiot. Combustion and Flame, 162(9):3236-3253, 2015.

[4] V. Hiremath, S.R. Lantz, H. Wang, and S.B. Pope. Combustion and Flame, 159(10):3096-3109, 2012.