Embedded Files
ARC-M1 Combustor
The Army Research Combustor-M1 (ARC-M1) was designed to replicate a single swirl-cup of a helocopter-sized gas turbine engine. It is equipped with four-sided optical access, enabling a wide array of laser and optical diagnostics, which can be used to conduct high-speed flame imaging, air velocity measurements, and fuel spray characterization.
Air is supplied to the ARC-M1 through two primary means: the main air, which flows through the dome from the air plenum, and the dilution air, which is split into six jets, three on each of the top and bottom of the combustor.
Low NOx Gas Turbine Simulation
Experimental rig designed at Purdue University
Experimental rig designed at Purdue University
The experimental apparatus is designed to produce longitudinal mode instabilities with a system frequency of approximately 400 Hz. The LDI rig is composed of wave plenum/chamber with different lengths and operates at around 1 MPa of mean chamber pressure. The heated air up to 800 K is provided through a choked inlet section and is swirled using a helical swirler assembly. Fuel (kerosene) is injected from the centrally located injector just downstream of the swirler assembly.
Linear Eddy Model (LEM) for sub-grid scale turbulent scalar mixing
Linear Eddy Model (LEM) for sub-grid scale turbulent scalar mixing
In the LEM approach, the scalar transport is modeled using a hybrid Eulerian-Lagrangian method whereby, the scalar fields evolve on another grid that is embedded inside the LES cell and in which the reaction-diffusion processes are resolved at a much finer resolution. It suffices to note here that with this approach, the LES grid quality does not directly impact the scalar mixing and combustion process since the subgrid resolution is much finer.
The LEM field is resolved in each LES cell for scalar mixing process. The LEM domain length is equal to the local LES filter length scale, and the number of LEM cells are determined based on the high shear layer region and chosen to resolve all the relevant length scales within the subgrid. In this approach, it is assumed that the smallest eddy is typically the local Kolmogorov scale. The number of LEM cells is fixed in all LES cells and the choice is based on the need to resolve the features in the regions of highest subgrid turbulent fluctuation. This region typically occurs in the high shear layer regions resulting in a rough estimate of the Kolmogorv length scale as 60µm. A LEM resolution of 18 cells is used in every LES cell, and typically 9 LEM cells are used to resolve a subgrid eddy. Hence, for the regions of high turbulence this results in a resolution of 0.5η.
Unsteady Flow Features
Unsteady Flow Features
Since the flow is highly swirling and with a swirl number above the critical value a vortex breakdown bubble (VBB) is formed in the sudden expansion region due to the adverse pressure gradient. This bubble is prominently seen in the time-averaged visualization as shown in the figure below. Flow is strongly recirculating in this region and temperature is quite high. Typically, it is made up of hot burned products and their recirculation provides an aerodynamic flame holder for the flame that is anchored ahead of this VBB.
Chamber Acoustics
Chamber Acoustics
The pressure signature in the combustor has been recorded at various locations for acoustic analysis. We focus on the data at the locations where the experimental study also measured the signal. The typical pressure signal and power spectral density for pressure fluctuation at combustor dump plane are shown below.
Single point LDI combustor
Single point LDI combustor
9-point LDI combustor
9-point LDI combustor
The LEM solver was successfully tested over multiple combustor configurations.
High-fidelity simulations with the advanced TCI model were achieved.
Highly scalable code performance was demonstrated.
More details can be found in my paper: https://arc.aiaa.org/doi/abs/10.2514/6.2014-0131
Page updated
Google Sites
Report abuse