Experimental investigation on ammonia combustion behavior in a spark-ignition engine by means of laminar and turbulent expanding flames

Ammonia combustion appears as a meaningful way to retrieve stored amounts of excess variable renewable energy, and the spark-ignition (SI) engine has been proposed as a practical conversion system. In this study, the combustion characteristics of ammonia blends in SI engine conditions were elucidated. Laminar and turbulent flame experiments were carried out in a constant-volume vessel at engine-relevant conditions of 445 K and 0.54 MPa to assess the combustion behavior of ammonia/hydrogen/air, ammonia/methane/air and methane/hydrogen/air mixtures observed in an all-metal single-cylinder SI engine. Results demonstrated that the respective accelerating or decelerating effects of hydrogen or methane enrichment observed in the SI engine could not be sufficiently explained by the measured laminar burning velocities of the mixtures. Since the latter are very low, the studied combustion regimes are at the boundary between the thin and broken reaction zones regimes, and thus strongly influenced by flame-turbulence interactions. The quantification of the flame response to turbulence revealed much higher effects for ammonia blends, than for methane-based fuels. The opposite effects of ammonia enrichment with hydrogen or methane were observed on the turbulent burning velocity during the turbulent flame experiments and correlated to the thermochemical properties of the reactants and the flame sensitivity to stretch. The latter may explain an unexpected bending effect on the turbulent-to-laminar velocity ratio when increasing the hydrogen fraction in the ammonia/hydrogen blend. A very good correlation of the turbulent velocity was found with the Karlovitz and Damkoehler numbers, indicating that ammonia combustion in SI engines may be described following the usual turbulent combustion models. This encourages further investigations on ammonia combustion for the optimization of practical systems, by means of dedicated experiments and numerical simulations