Transported PDF and MMC modelling of local extinction in turbulent piloted NH3/H2/N2-air jet flames

Abstract

Ammonia is a potential alternative fuel for decarbonising hard-to-abate sectors. Practical utilisation is hindered by unfavourable combustion properties that include slow chemical kinetics, low flame speeds and high nitrogen oxide emissions. These challenges are further exacerbated by local extinction in turbulent flames driven by turbulence–chemistry interactions. This study uses the joint-scalar transported probability density function (JPDF) and Multiple Mapping Conditioning (MMC) frameworks, both of which inherently provide a closed chemical source term treatment, to investigate such interactions in two turbulent ammonia–hydrogen–nitrogen–air flames exhibiting local extinction. The flames have been experimentally characterised and correspond to 59.2% (Flame D) and 88.9% (Flame F) of the blow-off velocity. The performance of JPDF methods, featuring Modified Curl’s (JPDF-MC) and Euclidean Minimum Spanning Tree (JPDF-EMST) closures for transport in scalar space, is evaluated alongside the MMC-based MMC-MC and MMC-IEM models for predicting local extinction. All four models provide generally good predictions for Flame D, but show noticeable differences for Flame F, particularly where local extinction is extensive. The JPDF-EMST closure predicts the least amount of local extinction, followed by MMC-IEM, with JPDF-MC and MMC-MC providing closer agreement with experimental data. The presence of NH₃ containing fluid in fuel lean regions for Flame F is related to local extinction events with computed results found to be sensitive to very minor changes ($\simeq 1\text{\%}$) in the fuel jet exit velocity. The MMC-MC formulation improves predictions of temperature PDFs in fuel-rich regions and OH PDFs in fuel-lean regions due to the enforced localness of transport in scalar space.