Project researcher Saana Hautala presented a conference paper “Towards a digital twin of a mid-speed marine engine: from detailed 1D engine model to real-time implementation on a target platform” at the 9th International Congress on Combustion Engines earlier this week. The paper is part of the work done in work package 3 (Novel combustion and advanced aftertreatment) of the Clean Propulsion Technologies project. The process started with an example model that Wärtsilä had previously provided in the INTENS project. The paper is a result of teamwork, which shows in the list of co-authors: Maciej Mikulski, Emma Söderäng, Xiaoguo Storm, and Seppo Niemi.
“It has made the process easier when you are surrounded by intelligent people who are enthusiastic about their work! During the work, we have received valuable support from Wärtsilä as well“ says Hautala.
Saana Hautala has worked for four years at the University of Vaasa. She has been working with 1D engine modeling as well as model validation and optimization in the latest research projects at the university. She works as a primary resource for the Task 3.5 (Engine-aftertreatment model coupling) in the Clean Propulsion Technologies project.
This study explores development of a control-oriented digital twin of a Wärtsilä 4L20 marine engine located in the VEBIC engine laboratory at the University of Vaasa. Starting from a detailed one-dimensional model (GT-Suite), it explores reduction strategies towards a fast-running engine model (FRM), balancing the calculation speed and accuracy trade-off. Finally, the FRM is tested for real-time implementation on a target platform, which is seen as a critical step towards digital twin implementation in the laboratory.
“During the study, it was concluded that the FRM reduces the level of complexity while maintaining a good level of predictivity. Real-time application is achieved by applying explicit solver optimized for speed and the accuracy loss of the FRM compared to the detailed engine model is minor” says Hautala.
The outcome of the real-time implementation on a target platform was that a more powerful CPU for the RT machine is needed to avoid errors in high-frequency signals. However, mean value control outputs were already calculated correctly.
System complexity is challenging for the development of marine mid-speed engines when striving to meet increasingly stringent emission targets. Digital twins (real-time capable, physics-based models) offer fast and predictive solutions for engine development. Thus, proof of concept for digital twin applications in mid-speed marine engines can provide a solution for this issue.