In the maritime transport sector, renewable fuels are increasingly being used. This development contributes to the reduction of CO₂ and pollutant emissions and thus promotes the sustainable transformation of shipping towards a more climate-friendly transport system. Supporting this transition requires engine technologies that are compatible with the new fuels, as well as precise and traceable measurement systems to demonstrate progress and validate performance.
The project "MaritimeMET" focuses on the development of improved measurement systems to enable a more qualitative assessment of these new fuels. Under the leadership of the Physikalisch-Technische Bundesanstalt (PTB), the seventeen project partners are conducting research not only on new measurement systems but also on optimized simulation models. These models are used to describe, in detail, the emission behaviour and combustion process of engines operating with alternative fuels.
To meet the new emission targets set by the International Maritime Organization (IMO) by 2030, the shipping industry must adopt low- or zero-emission fuels such as renewable methanol, dimethyl ether, and ammonia. This requires reliable and traceable emissions measurements. However, existing measurement and calibration methods are reaching their limits—particularly in terms of accuracy, operational conditions, and the availability of suitable standards.
To support both research and industry, new precise sensor solutions and measurement data are needed for machine learning models. These advancements aim to reduce uncertainties and enable cost-efficient monitoring concepts.
The aim of the project is to sustainably improve emissions monitoring through traceable measurements and machine learning. To this end, the Institute of Combustion Engines and Fuel Cells is developing a numerical prediction model for a dual-fuel engine operating on methanol and diesel. The model will be calibrated using measurement data obtained from an engine test bench. Computationally intensive components of the model will be simplified through the use of machine learning. This simulation model will enable the estimation of parameters that are difficult to measure directly or allow for the replacement of expensive sensors.