Projects

Maritime Data Analysis

Post-processing of emission and fuel consumption (includes PoC application)

Visit Alfa Laval

Data-Driven Decision Tools in Commodity Shipping

How can a user-facing frontend application be designed and developed to support one of CM Navigator’s core algorithms?

Visit Copenhagen Merchants Group

Change Detection in Dynamic Shallow Areas

Investigate dynamic shallow-water areas (e.g., fjord outlets) using satellite data and remote-sensing techniques and identify and quantify changes over time.

Visit Geodatastyrelsen

Deep Learning

Deep Learning, either in:

- Feature detection (Rocks in Greenland, coastline delineation, etc.)

- Satelite-derived bathymetry

- Automated cleaning of CSB data

Visit Geodatastyrelsen

Foundation Models for Forecasting Water Time Series

Visit DHI

Reservoir detection using deep learning approaches

Visit DHI

Refinement of flow accumulation raster using information from water surface frequency and digital elevation models

Visit DHI

Estimation of reservoir volume beneath recorded DEM using ML or geostatistical approaches

Visit DHI

Koopman Autoencoders as surrogates for MIKE 21 coastal models with integration of observed data

Visit DHI

Neural Operators as Surrogates for Coastal Ocean Modeling with MIKE 21 HD

Visit DHI

GNN-Autoencoders as Surrogates for Coastal Ocean Modeling with MIKE 21 HD

Visit DHI

Self-Supervised Deep Learning in Remote Sensing for Wetland Classification based on Satellite Time Series

Visit DHI

LIDAR-corrected Digital Terrain Model for Hydrological Applications using Deep Learning

Visit DHI

A Hull Inspector Platform

Development of Hulinspector platform to digitize data from diving companies and onboard crews conducting GoPro/ROV hull inspections

Visit Hafnia

Simulation of liner shipping network execution

This thesis will focus on developing a simulation model that represents the execution of a liner shipping network design under realistic operational conditions. The simulation will incorporate key elements such as vessel schedules, port calls, transshipment flows, and potential disruptions. The goal is to create a robust framework that can replicate the dynamic behavior of the network and evaluate its performance under varying scenarios, while tackling the scalibility challenge of the problem.

Visit Maersk

Heuristics for Terminal friendly Liner Shipping Network Designs

Container terminals are critical for global shipping operations, as they determine where and when vessels can dock. When network designs ignore terminal constraints, it often leads to inefficiencies and execution challenges later on. This project focuses on developing algorithms to bridge the gap between design and execution by making shipping network designs more aligned with terminal realities. Because the problem is highly complex, the approach will rely on heuristic method.

Visit Maersk

Inspection Prediction

Development of a risk indication for successful inspection based on the vessels condition, people onboard, and location.

Visit Navigator Gas

Technical Fault-finding Assistant

A Fault-finding assistant that helps identify problems with mechanical and electrical equipment onboard

Visit Navigator Gas

A Smart Model for Crew Planning

Planning crew member relief is as complex as it is changing.

It is currently done in Excel.

Visit Navigator Gas

Biofouling Pressure Model

Hull fouling causes speedloss and increase fuel consumption, posing a significant challenge to vessel efficiency.

A biofouling pressure model can help improve this.

Visit Vessel Performance Solutions

Projects

Projects

Maritime Data Analysis

Data-Driven Decision Tools in Commodity Shipping

Change Detection in Dynamic Shallow Areas

Deep Learning

Foundation Models for Forecasting Water Time Series

Reservoir detection using deep learning approaches

Refinement of flow accumulation raster using information from water surface frequency and digital elevation models

Estimation of reservoir volume beneath recorded DEM using ML or geostatistical approaches

Downscaling AI weather models to local areas

Sensor-agnostic coastline delineation using foundational models

Time Series Forecasting for Water Applications Using N-HiTS

Super-Resolution for Groundwater Models Using Multi-Resolution MIKE SHE Grids

Multi-Resolution Modeling for Flexible Mesh Coastal-Ocean Models

Koopman Autoencoders as surrogates for MIKE 21 coastal models with integration of observed data

Neural Operators as Surrogates for Coastal Ocean Modeling with MIKE 21 HD

GNN-Autoencoders as Surrogates for Coastal Ocean Modeling with MIKE 21 HD

Self-Supervised Deep Learning in Remote Sensing for Wetland Classification based on Satellite Time Series

LIDAR-corrected Digital Terrain Model for Hydrological Applications using Deep Learning

A Hull Inspector Platform

Simulation of liner shipping network execution

Heuristics for Terminal friendly Liner Shipping Network Designs

Inspection Prediction

Technical Fault-finding Assistant

A Smart Model for Crew Planning

Biofouling Pressure Model