Cerfacs Enter the world of high performance ...

Large Eddy Simulation of Constant Volume Combustion: numerical modeling for turbulent propagating flames in non homogeneous mixtures

Classical gas turbine thermodynamic cycle has undergone no major changes over the last decades and the most important efficiency improvements have been obtained reducing thermal losses and raising the overall pressure ratio and peak temperature. Despite the efforts in research and development aiming at enhancing especially combustion chambers performances, current technologies may fall short of complying the increasingly stringent environmental constraints. Consequently, a technological breakthrough is essential to shape the future of thermal engines. Pressure Gain Combustion (PGC) emerges as one of the most promising solutions, introducing new thermodynamic cycles where, unlike the Brayton cycle, pressure increases across the combustion process. This can lead to a lower entropy raise, benefiting the overall cycle efficiency.
Several PGC concepts are currently studied by the combustion community, ranging from deflagration, such as constant volume combustion (CVC), to detonation, including Rotating Detonation Combustion (RDC) and Pulse Detonation Engine (PDE). Large Eddy Simulation (LES) has assumed an increasingly significant role in combustion science thanks to its high capability in capturing reacting flows. However, with the increasing complexity of combustion systems, advanced physical models are crucial to ensure predictive simulations.
In this work, constant volume combustion technology is assessed and the main numerical challenges posed by these combustion systems are scrutinized. Ignition, high pressure combustion, dilution, flame-turbulence interaction, flame-stretch effects, heat fluxes are just part of the physics that CVC systems encompass and their interplay leads to complex physical phenomena that have to be modeled. The numerical models developed in this work are primarily scrutinized in simple test cases and then applied in complete 3D LES framework to compute the constant volume combustion chamber CV2, operated at Pprime laboratory (Poitiers, France).
First, novel boundary conditions, based on NSCBC formalism, are derived from nozzle theory to mimic intake and exhaust valve effects. With this strategy no moving part is introduced in the LES and the flow properties are imposed both at the inlet and the outlet of these valvescontrolled systems.
Second, a two-step chemistry for propane/air mixtures is derived for multiple pressure, temperature and composition of fresh gases. The chemical kinetics is optimized for different concentration of dilutants, composed by burnt productssuch as carbon dioxide and water vapor.
Like piston engines, constant volume chambers operate cyclically and each combustion event
is affected by the residual burnt gases coming from the previous cycles. For this reason, a numerical model to detail the local composition of diluted flammable mixtures is proposed to provide all the fresh gas information required by the kinetics and the combustion model. Based on a generalization of the classical Thickened Flame (TF) model, a new combustion model, the Stretched-Thickened Flame (S-TF) model, is developed to overcome the TF model limitations in predicting stretch effects on the laminar flame burning velocity. This is crucial to well capture transient events of propagating flames, which are fundamental in CVCs. Eventually, the ignition modeling is assessed and the Energy Deposition model is coupled with the S-TF model by tracking the kernel size in time.
The models developed in this thesis are then applied to the CV2 chamber, highlighting their positive impact in capturing the unsteady physics involved in such systems.

Jury