Â鶹ÉçÇø

SLGREEN - Green Transition of the Blue Denmark through Digitalization, Decarbonization, and Safety

Launched in May 2024, SLGREEN is a three-year cross-sector innovation project under the ShippingLab platform. Supported by Innovation Fund Denmark, the Danish Maritime Fund, and the Lauritzen Foundation, the project brings together more than 20 partners from industry, academia, and public institutions to accelerate the green transition of the Danish maritime sector.

SLGREEN focuses on three interconnected pillars: digitalisation, decarbonisation, and safety. By advancing digital tools, data-driven operations, and autonomous technologies, the project aims to enable smarter and more efficient ship operations. At the same time, it supports the adoption of low- and zero-emission fuels through practical, scalable solutions. Underpinning both areas is the development of robust safety concepts and validation methods to ensure that innovation goes hand in hand with risk management and regulatory acceptance. Through hands-on testing, cross-sector collaboration, and shared innovation, SLGREEN aims to provide the foundations for a more sustainable, digital, and safe future for the Blue Denmark.

WP3 – Digital twin for condition monitoring of engines

The section of Engineering Operations Management (EOM) Â鶹ÉçÇø is in charge of the third work package (WP3) of the SLGreen project, collaborating with two industrial partners including Everllence and Navigator Gas. Meanwhile, this project is collaborating with Amazon Web Service (AWS) which partially supports computational resources, cloud and general Machine learning operations (MLOps) for digital model development. The research team so far consists of Jie Cai (Senior researcher), Niels Gorm Maly Rytter (Senior researcher), Aurelian-lonut Dodis (Phd student) and several Master’ students. The WP3 focuses on digital twins development for condition monitoring and predictive maintenance modelling of marine engines using high-frequency operational data, aiming for better performance and longer engine life. This project is also aiming to contribute to the delivery of digital models, IT applications and improved work practices which can enable DK shipping companies and partnering equipment vendor(s) to be on the forefront on data driven condition monitoring, predictive maintenance and optimized spare part logistics. Additionally, the purpose is to assess how digital solutions and improved work practices can improve uptime of equipment, reduce the likelihood of breakdowns, avoid costly off-hire periods, reduce schedule recovery expenses as well as improve spare parts procurement and logistics. Eventually the project is expected to have an impact on fuel efficiency and emissions from vessels.

Marine diesel engines work as the heart of a vessel. In the current research, comprehensive data related to engines are collected for ongoing vessels from the shipping company, which includes real time engine operational data, engine incident reports, maintenance records, scrape down reports, scavenge inspection reports, ship off-hire reasons and vessel operational data. Typical data science methods have been utilized accounting for data qualities, data preparation and data modeling to produce scientifically robust insights on engine performance, root causes for failures and component wear under different operational settings and conditions. Frameworks for digitalization and PdM of marine combustion engines are developing which will facilitate the implementation of PdM of engines in practice. Besides, investigation focuses on digital modeling based on the latest Artificial Intelligence (AI), Machine Learning (ML) algorithms, and MLOps tools to develop prediction models for engine anomalies. Created knowledge will be essential for the industry if to meet future international legal and market requirements for safety, fuel efficiency and emissions of ships. The project will contribute to bringing the DK maritime industry to the forefront on the area of digitalization and sustainable ship operations, improving marine systems uptime, reducing risk for breakdowns, shortening off-hire periods and lowering expenses for vessel schedule recovery, spare parts procurement and logistics.

Fig: MAN 2 stroke marine diesel engine (Isometric view)

In the shipping industry, there are other big challenges and obstacles for PdM and condition monitoring of engines and typical systems on board. Among others, it includes the time-consuming training process based on ML for individual marine system, existing skills gap between marine engineers and data scientists, low model generalization ability due to complex operational statuses in the sea during ship maneuvering, the integration of all critical systems on board for condition monitoring and PdM, and data privacy concerning. For instance, when it comes to different marine systems and different critical components, similar data science processes for PdM model development are needed which are repetitive and expensive. The developed individual model is relatively difficult to implement due to limited generalization ability, even with the same systems but across different vessels. The skills gap has further increased the difficulties of such model development, interpretation and implementation, for instance, the lack of domain knowledge for data scientists and the lack of data knowledge for crew, operators and marine engineers. Additionally, if zooming out to an overall picture, operational data is crucial to the PdMs. However, shipping companies are, in general, not willing to share data with each other due to commercial reasons and market competition. Data privacy and data security due to such data silo will inevitably become constraints for PdM using AI to gain customer trust and achieve a large-scale implementation. Therefore, in order to realize the full implementation and integration of PdM and condition monitoring into the current shipping industry, our current project is still the initial but a critical step to figure out the above-mentioned challenges. Future research can be easily extended through the investigation of PdM modelling methods with high generalization ability for typical marine systems across vessels, and the development of rapid PdM tools to alleviate previously described skills gaps in shipping industry. In addition, accounting for data privacy and data security, how the PdM, tools and digital solutions using AI to gain customer trust and achieve a large-scale implementation across companies need to be studied, which will break down such a data silo. Overall, PdM has the potential to modernize shipping operations by using real time data, AI, and analytics to predict failures before they occur. Addressing these challenges, it will help to realize a more sustainable and environmentally friendly ship operations, and accelerate the green transition of shipping.