Knowledge Management Strategy

Knowledge Management Strategy

Pengcheng Ni, Jussi Kantola

Exploring the factors influencing knowledge management strategy in the European shipbuilding industry: A pilot study.

AI generated illustration of a ship and a brain​​​

Knowledge is an important asset in shipbuilding. This paper aims to explore the key factors of knowledge management (KM) in European Shipbuilding to improve its strategic performance. This pilot study can form an important reference for knowledge management strategy implement in European shipbuilding. Data collected from project partners are used to test seven hypotheses by means of linear regression analysis, to identify the key factors of knowledge management strategy in European shipbuilding. This paper provides ideas for shipyards to consider when conducting knowledge management as part of the Smart European Shipbuilding (SEUS) project. A full article is available at doi: https://doi.org/10.62477/jkmp.v25i6.595

Generative Algorithms

Generative Algorithms

Diego De León, Herbert Koelman
 

Generative Algorithms in Early Ship Design: An Exploration of Hull Subdivision Generation 

An Scenarion for Data Driven Early Ship Design​​​ This paper explores the potential for a data-driven tool to aid in the early ship design process, through the generation of subdivisions for the general layout via a proof-of-concept prototype which leverages a GAN to create plausible layout alternatives. The software implementation integrates a BSP tree structure for parametrisation, and a CAD geometry implementation. To work within the intrinsic limitations of generative algorithms, the decision-making is made by a naval architect, targeting facilitating the evaluation of multiple concepts and broadening the design possibilities. The paper describes the functioning of the proof-of-concept prototype, considerations on its creation and applicability. The full article is available at doi: https://doi.org/10.5281/zenodo.17534156

AI Naval Architecture

AI Naval Architecture

Karolina Bierkowska, Henrique Gaspar, Thomasz Hinz

Navigating AI in Naval Architecture: A Comparative Effectiveness Study of Machine Learning Models for Ship Stability 

AI generated image - ship in rough sea​​​

This study comparatively analyses diverse AI/ML models on an established dataset of hull variants and Second-Generation Intact Stability Criteria metrics, a time-consuming task in the early stages of ship design. This selection encompasses diverse AI techniques, each recognised for its unique strengths, featuring Artificial Neural Networks, Decision Trees, Probabilistic Models, and Large Language Models. In this article, we focus primarily on one of the failure modes: excessive acceleration. This research guides naval architects in selecting suitable emergent AI tools to enhance design space exploration, ultimately contributing to filter the current AI "hype" into useful NA practices in the industry. A full article is available at doi: https://doi.org/10.5281/zenodo.17493840

AR for digitalisation training

AR for digitalisation training

Zeynep Tacgin, Miia Martinsuo

An augmented reality solution for digitalisation training in shipbuilding: Systematic review and application development

AI generated image with a ship in the rough waters​Manufacturing and construction industries, including shipbuilding, are developing various digital solutions to increase efficiency in their processes, and digitalisation requires new kinds of capabilities from workers at all levels. Research on the use of augmented reality (AR) in various industries suggests that AR systems could be incorporated into training on digitalisation and Industry 4.0 solutions also in the shipbuilding industry to improve worker performance and productivity. This literature review of 29 publications synthesises the AR studies in the shipbuilding industry, with an attempt to identify goals, AR tools, and development needs in applying AR for training in shipbuilding. The analysis takes into account recently developed AR applications, the software and hardware components used for application development, the scope of developed AR apps, the utilised AR types, and the target audience of these applications by discussing their implementation. A full article is available at doi: 10.21125/inted.2025.0574

Interoperability

Interoperability

Gökce Yilmaz, Miia Martinsuo, Henrique Gaspar, Janica Altea Bronson

Interoperability in Project-Based Industries: Learnings and Challenges

AI generated image - ship in rough sea​​​

The paper reviews the ongoing issue of interoperability at an organizational level and its importance in the context of Industry 5.0, focusing on the building and maritime industries. The paper explores the potential for cross-industry learning between the two industries, highlighting data domains that can improve project flow while analyzing lessons learned and challenges in the adoption of modern tools for data sharing and management. Findings suggest that establishing clear legal frameworks, adaptable work processes, and cultural dynamics are necessary for improving interoperability. A full article is available at doi: https://doi.org/10.1016/j.ifacol.2025.09.240

Retrevial-agumented generation

Retrevial-agumented generation

Maryam Teimouri, Jenna Kanerva, Filip Ginter

A Deep Dive into Multi-Head Attention and Multi-Aspect Embedding

​​​

Multi-vector embedding models play an increasingly important role in retrieval- augmented generation, yet their internal behaviour lacks comprehensive analysis. We conduct a systematic, head-level study of the 32-head Semantic Feature Representation (SFR) encoder with the FineWeb corpus containing 10 billion tokens. For a set of 4,000 web documents, we pair head-specific embeddings with GPT-4o topic annotations and analyse the results using t-SNE visualisations, heat maps, and a 32-way logistic probe. The study offers practical guidance on where to extract embeddings, which heads may be pruned, and how to aggregate them to support more transparent and controllable retrieval pipelines. A full article is available at doi: https://doi.org/10.26615/978-954-452-098-4-146

Ship design - tracing changes

Ship design - tracing changes

Jisang Ha, Janica Altea Bronson, Henrique Gaspar

Managing Design Changes in Shipbuilding: Proposing a Real-Time Simulation Dashboard 

AI generated image with a ship in the rough waters​​A design simulation model with a unified data format is demonstrated in this work, focusing on: i) how design changes affect other objects and attributes; ii) what interfaces are required for this; and iii) the advantages of the implemented model through a real-time ship 3D model. Tracking design changes and revisions throughout the ship design and production process is known to be a challenge, especially when it involves multiple stakeholders working on 2D, 3D, and real-world objects. In this study, we utilized the NTNU experimental vessel “Gunnerus” as the target vessel. As a result, we were able to implement a model to verify the ship design that changes according to the design version of each design object, especially the compartments. Furthermore, we provided both version-based and time-based views of the target ship’s design to assist designers and producers. In addition, we proposed an initial structure for representing changes over versions of each design object. A full article is available at doi: https://doi.org/10.5281/zenodo.17306358

A RAG-based LLM Approach

A RAG-based LLM Approach

Janica Altea Bronson, Maryam Teimouri, Henrique Gaspar, Icaro Fonseca, Karolina Bierkoszka, Filip Ginter, Herbert Koelman

A RAG-based LLM Approach for Data Validation and Harmonization in Ship Design 

AI generated image with a ship in the rough waters​​​Validating ship design data across systems is challenging due to fragmented information from multiple sources, file types, and formats – from 2D drawings, 3D models, and specifications, often found in unstructured text files. While unified 3D models aim to serve as a single source of truth, ensuring accuracy and consistency across all representations remains a complex task. This paper presents a retrieval-augmented generation (RAG) solution for extracting and comparing design parameters from diverse files and formats. The approach aims to detect inconsistencies between documents and versions,helping designers maintain data integrity and reduce manual effort throughout the ship design process. A full article is available at doi: https://doi.org/10.5281/zenodo.17493949