Reduced chemistry for numerical combustion of NH3/H2 fuel blend

Ammonia is increasingly recognized worldwide as a promising carrier for hydrogen and energy. One effective strategy is blending ammonia with hydrogen to achieve combustion characteristics comparable to those of natural gas, including stable flame anchoring, controlled flame length, and sufficient heat release for energy production. The design and optimization of ammonia combustion systems rely heavily on computational fluid dynamics (CFD). Accurate CFD simulations of furnaces and gas turbines require chemical kinetic models that strike a balance between simplicity and fidelity, minimizing computational complexity (typically less than 20 species to be transported with the flow) while effectively capturing the essential thermochemistry of hydrogen-enriched ammonia combustion. This study begins with a detailed ammonia/air combustion mechanism and combines various canonical problems to derive two reduced chemical schemes. The methodology employs automated analyses to identify the most influential chemical species and elementary reactions. Five key reactive scenarios representative of premixed and non-premixed burners with eventual dilution by burnt gases are explored: auto-ignition, chemistry interacting with turbulent micro-mixing, freely propagating laminar premixed flames, strained counterflow diffusion flames, and mixing layers. This comprehensive approach facilitates the development of reduced kinetic models specifically tailored to ammonia/hydrogen–air combustion under a set of given operating conditions.

Novelty and Significance Statement

Novel reduced chemical schemes are proposed from a reference detailed mechanism for simulating ammonia–hydrogen-enriched combustion. To ensure their applicability in turbulent flame simulations, the reduction methodology incorporates a specific combination of various canonical test cases, including turbulent micro-mixing, for validation and performance assessment. These schemes achieve a significant reduction in stiffness and the number of degrees of freedom to be solved. To demonstrate their feasibility in computational fluid dynamics, a reactive mixing layer is simulated, and the results obtained with the reduced schemes are compared against those from the reference detailed mechanism.