Combustion engines operating on a hydrogen-argon power cycle (H-APC) offer potential for superior thermal efficiency with true zero exhaust emissions. The high specific heat ratio of argon allows extrapolation of the theoretical efficiency of the Otto cycle to almost 90%. However, this potential is significantly constrained by challenges in combustion control, excessive thermal loading, and system integration, particularly regarding argon recovery. This study investigates these trade-offs, within the context of real-world engine-based peaking power plants. An experimentally validated 1D-simulation model of a prototype Wärtsilä 20 DF engine serves as reference for analysis of a retrofit incorporating a closed-loop argon cycle, with dedicated H₂ and O2 injectors, a water condenser and water separator. Engine performance is evaluated at reference operating point of 75% load, considering pre-ignition, peak pressure and exhaust temperature constraints, condenser limitations, and impurity accumulation. Argon emerges as the best monoatomic gas for H-APC. Helium, the second-best candidate, offers superior thermal conductivity and specific heat, but its low density and molecular weight reduce power output. A 90% argon and 10% oxygen mixture offers the optimal trade-off between power output, efficiency, and durability. A compression ratio of 11.90:1 ensures stable combustion within design constraints, while stoichiometric operation and condenser inlet pressure of 3.23 bar enhances performance, achieving the best indicated gross efficiency of 59.10%. This is over 10 percentage points better than the reference engine at 75% load. Nevertheless, practical implementation is limited by pumping losses in a packaging-optimized argon-path layout, reducing extractable efficiency to 56.70%. Furthermore, just 2% impurities in fuel/oxidizer stream causes progressive efficiency decline, falling below the reference threshold after approximately 10 minutes of operation. This highlights the necessity of a membrane-based separator and system volume optimization. The findings establish a validated computational framework for optimizing closed-loop hydrogen combustion and provide valuable insights for progressing demonstrator development.