A computational fluid dynamic module for the simulation of small (1-10 m) reactive solid and molten volcanic particles in gas turbine engines has been proposed. Volcanic ash experimental data is used as a basis for the material properties. A first principles based particle framework has been described that will account for the various physical phenomena in the system. A dual approach has been considered utilizing both Lagrangian and Eulerian methods. Lagrangian particles are more natural but become prohibitively expensive for fine particles (such as the sizes for this project). Eulerian particle transport is physically appropriate for fine particles but becomes inaccurate for large particle sizes. A dual approach using both Eulerian and Lagrangian methods allows for optimal computational cost at maximum accuracy. Methods of particle transport, agglomeration, and deposition have been described for both approaches. Relevant gas turbine engine simulations will be used for verification and validation of the module. The developed software and module will be provided to the Navy for improved design and development of ash and sand mitigation designs for gas turbine engines.
Benefit: The benefits of our proposed approach, are the following: 1) Provides grid and discretization-order- independent LES framework (EFLES), 2) When RANS is used in one-way coupling mode instead of LES, compute steady state mean flow only once, turbulent fluctuations are then obtained through a fast synthetic turbulence computation added to the mean flow, 3) Accounts for essential particle dynamic processes: inter-particle breakup and agglomeration validated through DNS, phase change, wall sticking/bouncing, and wall surface thickening/shape-change due to particle deposition, 4) Provides substantial cost reduction through dynamic space dependent particle binning while producing comparable accuracy to full range binning, 5) Performs efficient wall surface thickening/shape-change computation employing level-set technique: no need for remeshing, low frequency operation (not required every time-step), and 6) Coupled dual Eulerian-Lagrangian particulate dynamic approach, allowing entrainment of particles between Eulerian and Lagrangian domains, hence, efficient simulations of dens small particulates which would be much more expensive if it is done through full Lagrangian computations. The focus, however, is initially on full Lagrangian particulates coupled with EFLES. The fast methods: dual Eulerian-Lagrangian particle module coupled with RANS synthetic-turbulence will be completed and validated at later stage of Phase II. Further, we propose a flexible implementation approach. The developed methods will be incorporated as independent physics modules that can be plugged and run on different versions the same software package or third party CFD codes as hooks or user-subroutines by adjusting the API and/or interface data-structure. There will be two primary avenues of commercialization from the successful completion of this project: a reactive glass particle library and engineering services and consulting. The goal of the project is to develop a modular reactive glass particle library that can integrate into a variety of computational tools that customers are working with. The tool will be able to simulate a variety of particles including, but not limited to, salt, sand, ash, and volcanic glass. Engineering services will be provided in the form of consulting and simulation. We can either implement our developed library into the customers code or use our code to model our customers problem and provide analysis for them. Mitigation design and analyses are expected. For the DoD, there are approximately 8000 fielded gas turbine engines. Approximately 1/3 of these are affected from sand ingestion problems. An example is a dust storm affecting helicopter flight. A $30,000 rotor that is supposed to last for 6,000 hours instead lasts only 400 hours due to sand effects. Estimating an engine at around $1,000,000 (a small GE T-700 is approximately $700,000), the DoD currently has $8 billion of engines in use. Sand ingestion therefore currently affects roughly $3 billion of hardware. To match the same hardware life as without sand, replacements which cost half a million per engine are required. Current design efforts utilize a build and try approach which can be costly. A fast and accurate CFD solver can be used for designs to further improve current development efforts of sand mitigation to improve engine life and durability. The primary commercial customers would include General Electric, Pratt & Whitney, Rolls-Royce, and Honeywell Aerospace. All of these companies provide engines to commercial entities (Boeing, Airbus) as well as government bodies (US, UK, etc.). As the outcome of reactive particle ingestion is engine replacement, the desire for effective mitigation is paramount to obtain a competitive advantage. Once one of these companies successfully solves the particulate problem with regards to sand and glass, other companies will want to follow suit.
Keywords: Large Eddy Simulation, hybrid Eulerian Lagrangian, Gas Turbine, Reynolds Averaged Navier-Stokes, glass particles, Volcanic ash, sand dust, compressible multiphase flow
