The debate over synthetic fuels is intense especially in sectors with a high energy demand like maritime [1, 2]. Hydrogen production from renewable sources is growing, but immediate measures for decarbonization are needed [3, 4]. In this context, the project MethMag was funded, and a gas engine for methane combustion with an innovative cooling concept and a purged prechamber (PC) spark plug was virtually developed [5, 6]. Validation with data from the test bench demonstrates that the simulations accurately represent the operating conditions [7, 8].
This combustion process is adapted for ammonia, which is being considered as a climate-friendly fuel of the future, particularly in maritime transportation [4, 9]. This fuel faces significant combustion challenges and is therefore mostly considered in complex, bivalent systems [10]. In particular, the prechamber is examined regarding the ignitability of ammonia. The overarching objective is to eliminate the necessity for a secondary fuel system, thereby reducing system complexity and associated costs.
The transition to ammonia highlights the need for further adjustments. The geometry of the PC cap significantly affects turbulence and mixture formation in the prechamber [11]. While swirl caps generate high turbulence, the mixture formation is inadequate. Tumble caps, on the other hand, provide advantages in mixture formation by achieving an earlier increase in turbulence, even though the maximum turbulence is lower. For ammonia combustion, PC wall conditioning is not essential, given the inherently low combustion temperatures. However, conditioning can improve cold-start behavior by accelerating PC combustion and offering greater flexibility in ignition timing [12]. Direct injection into the prechamber enhances fuel mixing and reduces sensitivity to ignition timing adjustments. This leads to higher efficiency and better combustion characteristics, particularly at lean air-fuel ratios [13, 14]. Operating with a lean ammonia-air mixture is challenging but offers benefits for non-selective catalytic reduction (non-SCR) of nitrogen oxides. Simulations show that operation with λ = 1.2 and λ = 1.4 is feasible, although efficiency decreases at leaner mixtures [15].