Services
Engineering simulation you can rely on
Analysis, method development and AI-accelerated simulation for structures where stability, variability and weight matter, scoped to the decision you need to make.
01 / Structural analysis
Finite element analysis of demanding structures
Linear and nonlinear analysis of thin-walled, composite and 3D-printed structures, from a quick verification check to a full nonlinear study.
Built on a doctorate in structural analysis and research on panels, shells, wind turbine blades and landing gear.
-
Buckling & post-buckling
Linear eigenvalue and nonlinear path-following analyses of plates, panels and shells, including mode interaction and imperfection sensitivity.
-
Composite structures
Laminates and fibre-steered, variable-stiffness designs: stiffness tailoring, stability and the effect of fibre misalignment.
-
Additive manufacturing
How process variability in printed parts affects stiffness, strength and fatigue, and how to design around it.
-
Fatigue-driven redesign
Stress-life assessment and local design changes that extend fatigue life, checked against test data where available.
-
High-fidelity modelling
Higher-order and variable-kinematics finite elements that capture 3D stress fields with fewer degrees of freedom.
-
Dynamics & loads
Structural dynamics and early-stage design loads, such as landing-gear impact and load spectra.
-
Independent review
A second opinion on existing models: assumptions, boundary conditions, mesh convergence and the interpretation of results.
-
Verification & benchmarking
Checks against analytical solutions and recognised benchmarks, such as NAFEMS, before results are trusted.
02 / Uncertainty & robust design
Design for the structure you will actually build
Real structures are never ideal. Quantify how variation in geometry, material and loading affects performance, then use it to make designs better and less sensitive.
Demonstrated in peer-reviewed work: up to +29% linear buckling load from stiffness redistribution, and roughly 4–5× fatigue life in tested 3D-printed specimens.
-
Random-field modelling
Spatially correlated variation of thickness, stiffness, fibre angle or geometry on any finite element mesh, including curved and complex shapes.
-
Stochastic FEA & Monte Carlo
Distributions of buckling loads, stresses or fatigue life instead of a single deterministic number.
-
Sensitivity analysis
Correlation maps and Sobol indices that show which parameters and regions drive the response.
-
Robust design & tailoring
Redistribute thickness, stiffness or fibre paths to raise performance while reducing sensitivity to imperfections.
-
Reliability & knock-down factors
The probability of meeting a design criterion, and evidence-based alternatives to blanket knock-down factors.
-
Optimisation under uncertainty
Design optimisation that accounts for scatter in parameters, rather than stacking conservative assumptions.
03 / AI for computational mechanics
Machine learning where it genuinely helps
Surrogate models pay off when they are built on sound simulation data and report their own uncertainty. Use them to explore a design space in seconds instead of days.
Recent work: Gaussian-process surrogates with principal component analysis for early-stage landing-gear design loads, developed with Airbus engineers.
-
Surrogate models
Gaussian-process and neural-network surrogates trained on finite element or dynamic simulations, for rapid prediction of loads, responses and fields.
-
Dimensionality reduction
Principal component analysis of time histories and fields, making high-dimensional outputs learnable from a modest number of simulations.
-
Uncertainty quantification at speed
Global sensitivity analysis and uncertainty propagation that would be unaffordable with full simulations.
-
Design-space exploration
Rapid trade studies and optimisation in early design, with uncertainty bands that show where the surrogate can be trusted.
-
Simulation data pipelines
Automated design of experiments, batch simulation and curated datasets with full provenance, ready for training ML models.
-
AI in engineering workflows
Advice on where AI assistants can be trusted in a simulation workflow, and how to check their output.
04 / Software & automation
Tools that make analysis repeatable
Custom tools around finite element solvers cut manual work and errors. Open-source pipelines remove licence bottlenecks for large studies.
Open source: abq2ccx converts Abaqus input decks to CalculiX and is validated against NAFEMS benchmarks and a corpus of 1,022 decks.
- Abaqus
- CalculiX
- DIANA FEA
- Python
- MATLAB
-
Analysis automation
Python tools for pre- and post-processing, parametric studies and reporting.
-
Solver migration
Moving Abaqus models to the open-source CalculiX solver, with every translated or unsupported keyword reported rather than silently dropped.
-
Method implementation
Methods from the literature, such as random fields, reduced-order models and surrogates, built into your toolchain.
-
Verification test suites
Benchmark-based tests that keep in-house simulation code trustworthy as it evolves.
05 / Working together
Engagements that fit the question
Based in the Netherlands, working remotely with clients worldwide. On-site visits by arrangement. NDAs are welcome, and your models and data stay yours.
Fixed-scope study
A defined question, a fixed price and a clear deliverable, such as a buckling assessment, a sensitivity study or a surrogate model.
Flexible expert support
Day-rate support alongside your team: method development, troubleshooting nonlinear analyses, or extra capacity when deadlines are tight.
Independent review
A second opinion on an existing model, report or simulation process before you act on it.
-
01
Introductory call
A short, no-obligation conversation to understand the problem and the decision behind it.
-
02
Proposal
Scope, deliverables, timeline and price in writing, as a fixed price or day rate.
-
03
Delivery
Regular check-ins and early sight of intermediate results, so there are no surprises.
-
04
Handover
A clear report with the models and scripts, so you can reproduce the work.
Next step
Tell me about your structure
Share what you are analysing and what you need to decide. I will reply with a few questions and a proposed next step.