Solution-Based Mesh Adaption Criteria Development for Accelerating Flame Tracking Simulations

Abstract

Abstract Accurate flame tracking plays a vital role in predicting the combustion characteristics of a system. This is even more critical for systems that evolve over time. Predicting relight performance of an aero combustor, predicting flame propagation due to gas leakage from a storage tank or during the thermal runaway of batteries, are some examples of such dynamic systems. Predicting accurate flame position also plays an important role in deriving the correct pollutant formation rate from a combustion system. The challenge with flame tracking through a 3D computational fluid dynamics (CFD) simulation comes from the requirement to have a good resolution of gradients along the flame front. This requirement can push the overall mesh count of any industrial cases to a very large value (several million-mesh count). Further, the global drive towards using hydrogen or hydrogen blended fuels for different combustion applications pushes the limits on having even finer cells since hydrogen is a fast-burning fuel and has a much thinner flame front compared to hydrocarbons. Solution-based mesh adaption approaches have been widely studied and tested by different research groups to generate the required finer meshes in the critical regions on the fly while keeping the overall mesh count to a manageable level. However, these approaches are typically applicable for a set of problems, and therefore, there is a need for a generic approach suitable for a broader range of problems. This work explores various parameters and specific weightage factors to predict correct flame-tracking outcomes for different types of flames. The selections of flow quantities (flow-variables, their gradients, curvatures) are performed using simple flames and flow configurations. The functions based on selected flow-quantities derived from these studies are then tested to predict the results for the more complex set of published flames like the Engine Combustion Network (ECN) spray flame and Knowledge for Ignition, Acoustics and Instabilities (KIAI) five-burner configuration (liquid and gas fuel). Derived adaption criteria are found to predict the correct flame tracking behavior in terms of transient evolution of flame front, flame propagation, and ignition timing of burners. The parameters used for the study are identified keeping genericity as the key point, and thus making sure that the derived adaption functions can be applied across different types of fuel blends, combustion systems (gaseous or liquid fuel-based systems) and combustion models, for example species transport or mixture fraction-based models.