Graduation Internship: Offshore Weather Simulations

How you can make your mark

At Boskalis, logistics simulations are increasingly used to support tenders and projects in offshore energy and dredging. These simulations help engineers evaluate work methods, fleet configurations and the impact of weather downtime on project performance. Weather conditions and offshore workability for vessels are key drivers of uncertainty and project risk.

The current workability assessment approach relies on hindcast analysis, using more than 40 years of hourly metocean data to statistically evaluate weather downtime and operational windows. While this provides robust insight into long-term weather uncertainty, it generally considers a predefined operational strategy and does not explicitly capture how vessel crews adapt to changing conditions.

Capturing such adaptive behavior requires modelling decision points and branching pathways within a Discrete Event Simulation (DES). In addition, DES models are increasingly used not only to evaluate project performance under uncertainty but also to optimize operational strategies, vessel deployment, and work sequences. Both objectives require many simulation runs: adaptation introduces multiple possible project pathways for each weather scenario, while optimization requires repeated evaluation of alternative decision configurations and operational plans.

As a result, model complexity and computational requirements increase substantially. Rather than evaluating a single project trajectory for each weather scenario, the simulation must assess numerous possible pathways across thousands of weather realizations derived from the hindcast dataset, often repeatedly as part of an optimization procedure. This computational burden can become a limiting factor when applying DES to large-scale weather workability analyses.

To address the computational burden associated with weather-adaptive DES models, various model reduction and approximation techniques can be explored. Potential approaches include clustering weather conditions into representative weather states, modelling weather evolution using Markov-chain-based methods, and developing surrogate models that emulate the behavior of computationally intensive simulations.

The challenge lies in striking an appropriate balance between computational efficiency and the fidelity required to accurately represent weather-driven offshore operations.

The goal of this research is therefore twofold:

1. To investigate which modelling techniques can significantly accelerate probabilistic DES for weather workability applications.
2. To determine whether the additional modelling detail gained from explicitly representing operational decision-making under weather uncertainty actually leads to more accurate and valuable project insights compared to current workability analyses.

You will work within the AI Department and collaborate closely with simulation engineers, R&D engineers, metocean engineers, and offshore energy planners to develop, evaluate, and validate next-generation weather workability simulations for Boskalis projects.

During this project you will:

• Perform a literature review on various model reduction and approximation techniques, uncertainty quantification and workability assessments.
• Analyze existing Boskalis weather workability and logistical simulation models.
• Identify operational decisions and weather-driven branching mechanisms suitable for modelling in DES.
• Design and implement the selected technique approaches in Python.
• Compare accuracy, robustness and computational performance across different modeling approaches.
• Develop demos for practical engineering use cases.
• Present findings to engineers, data scientists and business stakeholders.

Your qualities

You are a Master’s student in Applied Mathematics, Computational Engineering, Operations Research, Marine Engineering, Transport and Logistics, or a related quantitative discipline.

You have:

• Strong Python programming skills
• Affinity with simulation, optimization, data analysis and uncertainty modeling
• Knowledge of Gaussian Processes, Bayesian methods, surrogate modelling, or machine learning
• Experience with computing libraries such as SimPy, Numpy and SciPy
• Interest in offshore operations and maritime engineering
• Strong analytical and problem-solving skills
• Good communication and report-writing skills

We offer

• A challenging MSc thesis project with direct impact on real-world engineering projects.
• Supervision from AI engineers, simulation experts and domain specialists.
• Access to industrial-scale simulation models and datasets.
• Opportunity to work on cutting-edge applications of AI in maritime engineering.
• A collaborative environment combining data science, optimization and operational decision support.
• Possibility to continue the work through an internship or future employment opportunities.