Design and Evaluation of a Ship and Data Model for Digital Twin Applications

Abstract

The digital twin concept has emerged as a promising technology in various industries, including maritime applications. This thesis is focused on the design and evaluation of a ship and data model for a digital twin application. The primary thesis objective revolves around the detailed design of a functional remote-controlled ship model equipped with various sensors to gather real operational data which later will be utilized in creating a decision-support framework with a focus on hybrid energy systems. Following this, to effectively manage and analyze the generated data, the development of appropriate technology becomes imperative. In pursuit of this objective, a data modeling approach capable of proficiently storing, organizing, and retrieving data was proposed. Additionally, a data preprocessing algorithm was developed and tested on the collected data, aiming to enhance the data quality.

To achieve these objectives, an electro-mechanical system integrated with multiple sensors able to capture real operational data was designed for a scaled-down ship model. Although the use of a scaled model posed challenges in acquiring suitable components, it proved cost-effective, low-risk, and straightforward to design. The collected operational data effectively represented a single operation scenario, demonstrating the potential to expand to multiple scenarios for more extensive data analysis.

The research also emphasized that no one-size-fits-all data pre-processing technique exists, and finding the most suitable algorithms requires experimentation and comparative analysis. Various data pre-processing algorithms were explored, such as filtering outliers, noise reduction, data synchronization, and imputing missing data. These explorations aimed to elevate the quality of the data.

The outcomes of this research provide substantial insights into the utilization of scaled-down models to advance the development of digital twins, while also underscoring the various challenges associated with this implementation. This approach not only minimizes risks but also highlights the potential to establish a cost-effective testbed for decision-making support platforms. Additionally, the results underscore the significance of utilizing a variety of data pre-processing techniques to proficiently manage distinct data types and scenarios.

Ultimately, this study contributes to the advancement of digital twin applications in the maritime domain, opening new opportunities for harnessing the advantages of scaled-down models in creating and testing digital replicas. The data gathered from such models can significantly contribute to providing crucial decision support for ship operations with minimal risk compared to full-scale implementations.