Numerical exploration and modelling of a novel environmentally friendly combustion technique: droplet-laden MILD combustion

A small reduction in NOx emission per kilo-watt of generated power will have a significant reduction in environmental impact of combustion used for power generation. The MILD (Moderate or Intense Low-Oxygen Dilution) combustion technique offers an opportunity to drastically reduce emissions while improving thermal efficiency of furnaces and boil-ers. In gas turbines, though overall fuel-air mixture is fuel-lean and MILD combustion is not directly applicable, fuel-rich regions in the primary zone of the combustor exhibit localised MILD regimes, particularly for liquid fuel operation How-ever, the physical and chemical intricacies of this novel technique are not well understood and thus identifying key con-trol parameters for using this technique for power generation and industrial processes over wide range of conditions is challenging. This project aims to provide a step change in physical understanding and modelling of this combustion technique and to identify key control parameters. The aim is to investigate MILD combustion of high calorific value gaseous and liquid fuels for practical application using Direct Numerical Simulations (DNS) and Large Eddy Simula-tions (LES), with high-fidelity mathematical description for physical and chemical processes involved. The droplets of liquid fuel spray will be tracked using the Lagrangian approach while the gas phase is treated using the Eulerian ap-proach for the simulations.
The effects of droplet diameter, equivalence ratio (both for gaseous and liquid fuels), extent of dilution by combustion products, volatility (by considering different fuels), turbulence intensity and its length scale on the burning rate, flame structure (in terms of chemical reaction pathways analysis and flame and flow topologies) and pollutants formation will be analysed based on a judicious parametric analysis based on three-dimensional detailed chemistry DNS data. In this project, the fundamental physical understanding extracted from DNS data will be utilised to develop high-fidelity models for engineering Computational Fluid Dynamics (CFD)-based simulations to identify key control parameters using LES after validating these models against the available experimental results. This project will provide (1) a ro-bust modelling framework for MILD combustion technique, which would be a cost-effective reliable tool for designing energy-efficient and clean gas turbines and industrial furnaces and (2) the key control parameters identified can help to design retro-fit "greener" combustion systems.