Simulation & Design Solutions
Structural, thermal, CFD, modal, impact and multi-physics analysis with optimisation.
Read more
Conceptual design, CAD, FEA, thermal and fluid analysis, optimisation and validation.
An engineering thesis has to be right twice: the physics has to be modelled correctly, and the modelling has to be shown to be correct. Most of the questions in a viva are about the second of those.
The seven areas on the poster describe the arc of an engineering research project: conceptual design and analysis, CAD modelling and simulation, structural analysis using finite elements, thermal and fluid flow analysis, optimisation and parametric study, experimental design and validation, and technical reporting and documentation.
Every one of those stages produces something an examiner can question. Why this geometry simplification? Why this element type? Why this turbulence model? What was the mesh independence result? How does the simulation compare with the experiment, and where it disagrees, why? A model that gives a plausible answer without these checks is not yet a result.
We work in ANSYS, SolidWorks, Autodesk Inventor, MATLAB and COMSOL, and we document the work so that it can be reproduced. That means the CAD file, the mesh settings, the boundary conditions, the solver setup and the convergence history — not just a colourful contour plot.
Six service areas from the poster, and the analyses inside each.
Verification asks whether you solved the equations correctly. Validation asks whether you solved the correct equations. They are different questions and both have to be answered. Verification is largely a numerical matter: mesh independence, time-step sensitivity, residual convergence, conservation checks. Validation is physical: comparison with an experiment, an analytical solution, or trusted published data.
A mesh independence study is not optional and takes an afternoon. Run the case at three or four mesh densities, plot the quantity of interest against element count, and show that it has stopped changing. Present that plot in the thesis. It answers a question the examiner would otherwise ask, and its absence suggests the result may be a meshing artefact.
Where validation data does not exist, say so and treat the result as a comparative study rather than an absolute prediction. Comparative conclusions — this configuration is better than that one under the same assumptions — are defensible even without validation. Absolute claims are not.
Geometry, materials, loads, boundary conditions and the quantity you actually need. Written down before anything is modelled.
CAD prepared and simplified with the simplifications justified, then meshed with quality metrics recorded.
Material models, contacts, boundary conditions and solver settings, all documented so the run can be reproduced.
Mesh independence, convergence and sensitivity checks, presented as plots you can put straight into the thesis.
Comparison against experiment, analytical solution or published data, with discrepancies discussed rather than hidden.
Results, figures, tables and a written interpretation, plus every project file so you can re-run and extend the work.
The most common problem we see with inherited simulation work is that nobody can say how it was set up. A results folder full of images with no record of the mesh, the material data or the boundary conditions is not reusable, and when a reviewer asks a question about it there is no answer.
Every project we deliver comes with a setup record: geometry version, simplifications made, mesh statistics, material properties with their source, boundary conditions, solver and convergence criteria, and the software version. It takes an extra hour and it is the difference between work you can build on and work you have to repeat.
You also receive the native project files. If you later need a parametric sweep or a different load case, you or another engineer can open the model and run it, rather than starting from the CAD again.
| What you receive |
|---|
| CAD model and manufacturing drawings |
| Meshed model with quality statistics |
| Documented boundary conditions and solver setup |
| Mesh independence and convergence plots |
| Result contours, plots and tables at publication resolution |
| Validation comparison against experiment or literature |
| Written technical report with interpretation |
| All native project files for future re-use |
Yes. A dimensioned sketch, a photograph with a scale, or a technical drawing is enough to build a parametric CAD model. We confirm the geometry with you before any analysis, because an error there propagates through everything after it.
You should not take it on trust, and neither should your examiner. That is why every project includes a mesh independence study, convergence evidence and, where data exists, a validation comparison. If none of these can be provided we tell you that the result should be presented as comparative rather than absolute.
We help design the experiment, choose instrumentation and sample points, plan the measurement uncertainty analysis and interpret the results against the simulation. We do not run experiments in your laboratory on your behalf.
Whichever your department has a licence for, wherever possible, so that you can open and continue the work. If you have no licence we will tell you which parts can be done in free or student editions and where that limits the analysis.
Use a published property set for the nearest standard grade, state the source explicitly, and run a sensitivity check on the property that matters most for your result. Reporting that a ten per cent change in modulus moves your answer by two per cent is far stronger than quoting a single number with no provenance.
Structural, thermal, CFD, modal, impact and multi-physics analysis with optimisation.
Read more
Steady-state, transient, conjugate heat transfer, thermal stress and cooling design.
Read more
Theoretical modelling, experimental data interpretation, computation and validation.
Read moreSend your topic, your dataset or one draft chapter. We will tell you honestly what it needs — before you pay anything. The first consultation is free.