Work

A system-level simulation framework for developing next-generation marine gas engines

Year: 2026

Type: article

Abstract: Reactivity-controlled compression ignition (RCCI) is a proven highly efficient and fuel-flexible combustion concept, yet its industrialisation is constrained by a plethora of operating parameters with strong nonlinear interactions. These characteristics challenge development of control algorithms, hindering efficient engine calibration. Existing model-based approaches struggle to resolve these nonlinearities with sufficient speed and fidelity, relying either on heuristic submodels or computationally heavy computational fluid dynamics simulations. The present study addresses this by employing an advanced performance-oriented combustion model, featuring detailed chemical kinetics, semi-predictive fuel-stratification and in-cylinder mixing grounded on the turbulence-energy cascade. Validation against a 200 mm-bore, single-cylinder research engine reveals combustion phasing predicted within ±3 °CA, and NO X and CH₄ emissions below 35% error, representing best-in-class predictive performance. The mesoscale model is subsequently coupled with a validated 1D air-path model, enabling efficient multi-cylinder simulations with turbocharger and air-path thermal management. For the first time, this comprehensive simulation framework is applied to model-based development of a new engine, based on the Wärtsilä 20-series turbo-diesel platform. The retrofit transmutes towards gas–diesel RCCI operation, assuming minor hardware modifications. The optimal compression ratio using the stock turbocharger is identified as 14:1, yielding a peak indicated efficiency of 49.4% at 75% load. Load-range-averaged NO X is half IMO's Tier II limit, while CH₄ emissions reach 2.9 g/kWh, except at low loads. Air-management optimisation further elevates indicated efficiency to 51.2%, with NO X reduced to 0.16 g/kWh, below the automotive Euro VI limit. The above explorations cover 1880 simulation runs, each taking under 20 min until convergence.

Source: Applied Thermal Engineering

DOI: 10.1016/j.applthermaleng.2026.131209