An Improved Mesh-Conversion Methodology for Particulate Deposition Simulations in Gas Turbines

Dynamic deposition of fine particulate matter inside gas turbine engines leads to changes in geometrical morphology and degradation of aero-thermal performance. Due to physical and optical limitations in existing measurement methods, comprehensive experimental and operational data on dynamic deposition are generally not available. It is thereby vital to develop a reliable and robust numerical simulation method to observe the changes in aero-thermal performance inside turbines susceptible to deposition. To date, the most widely used simulation strategy is a coupled particle deposition-dynamic mesh morphing approach. However, movement of the mesh nodes on domain boundaries leads to a rapid deterioration in mesh quality, making this technique not possible in complex geometries or for dense deposition. A promising alternative to simulate deposition growth is a mesh-conversion approach, in which the fluid domain cells accommodating deposition are converted to solid domain cells to mimic the solid deposition while the mesh topology is retained without modifying the mesh, allowing for continuous simulations in complex geometry and dense deposition cases. Due to a couple of inherent defects, however, the mesh-conversion technique has not been widely applied in gas turbine deposition. This study thereby attempts to improve the technique by integrating added source terms, near-wall mesh refinement, and a wall roughness model, with the expectation to overcome the shortcomings. The reliability of the improved mesh-conversion method is validated using data from a canonical jet impingement deposition experiment in the literature. Following that, deposition simulations in an actual externally and internally cooled turbine vane are performed to verify its feasibility. The validation and verification demonstrate that the improved mesh-conversion method generates accurate deposition effects comparable to the conventional dynamic mesh technique, and it is superior in deposition simulations in complex geometries with excessive deposition thickness, showing its potential in gas turbine deposition simulations.