AI Foundation Model for Ship Shape Design

Generative AI foundation model development for ship shape design in collaboration with KRISO

Project Overview

Conceptual diagram of the AI-based Ship Shape Design Support System (NARNIA LABS × KRISO)

This project develops an AI-based ship shape design support system in collaboration with the Korea Research Institute of Ships and Ocean Engineering (KRISO). As a co-research institution, Narnia Labs is developing a generative AI foundation model dedicated to the creation and exploration of ship geometries. The initiative aims to redefine the traditionally experience-driven ship design workflow by leveraging large-scale generative models and design optimization — capturing the know-how of experienced designers as a system so that hull and propulsor shapes can be derived in a much shorter time, and helping strengthen the competitiveness of small and medium-sized Korean shipbuilders.

My Role

Project Manager, Narnia Labs

  • Lead overall project management: scope definition, milestone planning, deliverables, and coordination with the KRISO consortium
  • Drive hands-on technical work on the generative AI foundation model for 3D ship shape design
  • Design methodologies for shape generation, exploration, and optimization tailored to naval architecture constraints
  • Bridge AI research and industrial ship design requirements together with academic and industry partners

Project Details

Consortium

  • Lead: Korea Research Institute of Ships and Ocean Engineering (KRISO)
  • Co-research (Industry): Narnia Labs
  • Co-research (Academia): Chungnam National University, Sungshin Women’s University, Kyung Hee University, University of Ulsan, Korea Maritime and Ocean University

Scale

  • Duration: June 2026 - December 2030
  • Total R&D budget: approx. 16.2 billion KRW

Goals

  • Develop an AI foundation model capable of generating and exploring diverse ship shapes
  • Reduce ship design time by more than 50% compared to conventional workflows
  • Move beyond designer-experience-dependent, iteration-heavy ship shape design

Background

Ship shape design has traditionally relied heavily on the accumulated experience of designers and repetitive analysis cycles, making the process both time-consuming and costly. Narnia Labs has previously applied AI-driven design to shipbuilding projects with major Korean shipbuilders including Samsung Heavy Industries and Hyundai Heavy Industries, and this project extends those efforts toward a general-purpose AI foundation model for ship shape design.

Impact

Establishing a foundation model for ship shape design is expected to:

  • Dramatically shorten the concept-to-detail design cycle
  • Enable systematic exploration of a much larger design space than manual approaches
  • Provide a reusable AI backbone that can be transferred across different vessel types and design requirements
  • Strengthen the AI-driven design capability of the Korean shipbuilding industry

Project Duration: June 2026 - December 2030 Partner: KRISO (Korea Research Institute of Ships and Ocean Engineering) Organization: Narnia Labs

Source: E2 News 기사