About
Sander van den Broek
Dr.ir. · Founder · Computational mechanics engineer
Computational mechanics engineer specialising in the stochastic and nonlinear analysis of thin-walled and composite structures, surrogate modelling, and expert engineering-mechanics data for AI.
Background
Structural mechanics, uncertainty and machine learning
- Based in
- The Netherlands
- Education
- Dr.-Ing., Leibniz Universität Hannover
MSc Mechanical Engineering, TU Delft - Current roles
- R&D Engineer, Netherlands Aerospace Centre (NLR)
Honorary Senior Research Associate, University of Bristol - ORCID
- 0000-0002-6186-6057
Sander van den Broek works where structural analysis, uncertainty and machine learning meet. He founded Van den Broek Computational Mechanics to offer this expertise to engineering teams and AI developers.
He studied mechanical engineering at Delft University of Technology, with an MSc thesis on aeroelastic analysis of wind turbines. His doctorate at the Institute of Structural Analysis, Leibniz Universität Hannover, supervised by Prof. Raimund Rolfes, turned the random variations that normally weaken structures into a design tool. Sensitivity patterns from random-field analyses guide changes in thickness, stiffness or fibre paths that raise buckling loads, improve post-buckling behaviour and extend fatigue life. The research was funded by the EU Marie Skłodowska-Curie network FULLCOMP and the DFG Cluster of Excellence SE²A.
At the Bristol Composites Institute, University of Bristol, he developed higher-order finite elements for wind turbine blades (with the Offshore Renewable Energy Catapult) and Gaussian-process surrogates for landing-gear design loads (with Airbus). He is an Honorary Senior Research Associate at Bristol and an R&D engineer in computational mechanics of structures at the Netherlands Aerospace Centre (NLR).
He also creates engineering-mechanics data for training and evaluating language models, and wrote abq2ccx, an open-source Abaqus-to-CalculiX converter.
Career
Experience & education
-
Current
Founder
Van den Broek Computational Mechanics
Computational mechanics consulting, AI for simulation and engineering-mechanics data for language models.
-
Since 2026
R&D Engineer
Netherlands Aerospace Centre (NLR), Amsterdam
Computational mechanics, mainly of structures.
-
Current
Honorary Senior Research Associate
University of Bristol
-
2021 – 2026
Postdoctoral researcher
Bristol Composites Institute, University of Bristol
Nonlinear analysis of wind turbine blades with variable-kinematics finite elements (with the Offshore Renewable Energy Catapult); surrogate modelling of landing-gear design loads (with Airbus).
-
2016 – 2023
Doctoral researcher · Dr.-Ing.
Institute of Structural Analysis, Leibniz Universität Hannover
Thesis: Tailoring structures using stochastic variations of structural parameters (2023). EU Marie Skłodowska-Curie network FULLCOMP and DFG Cluster of Excellence SE²A; research visit to the University of Bristol.
-
2010 – 2013
MSc Mechanical Engineering
Delft University of Technology
Thesis: Towards high-fidelity aeroelastic analysis of wind turbines: coupling and verification.
Expertise
Methods & domains
- Nonlinear finite element analysis
- Buckling & post-buckling
- Thin-walled & composite structures
- Fibre steering
- Random fields
- Monte Carlo methods
- Uncertainty quantification
- Sobol sensitivity analysis
- Gaussian processes
- Surrogate modelling
- Robust design
- Additive manufacturing
- Fatigue
- Higher-order finite elements
- Aircraft loads
- Wind turbine blades
- LLM training data
Co-authors include engineers and researchers from Airbus, DLR (German Aerospace Center), the Offshore Renewable Energy Catapult and the National Composites Centre, as well as the universities of Bristol, Hannover and Limerick.
abq2ccx
Converts Abaqus input decks for the open-source CalculiX solver. It expands mesh-generation cards, flattens part/instance assemblies, translates keywords, elements and materials, and reports anything CalculiX cannot run instead of dropping it. Dependency-free Python 3, validated against NAFEMS benchmarks and 1,022 test decks.
Next step
Let’s talk about your project
A structure to analyse, a surrogate to build or a dataset to create: a short call is the best place to start.