Mechanical Engineering Project Reports: Australia Guide
How Australian mechanical engineering projects are assessed, covering FEA and CFD validation, AS standards, Engineers Australia competencies and reporting uncertainty properly.
Australian mechanical engineering programmes are accredited against Engineers Australia's competency standard, and that shapes how final-year projects are marked. Simulation results alone score poorly. What earns marks is a validated model, a stated uncertainty, and a design decision that follows from both. The single most common failure is a beautiful contour plot that nobody checked against anything.
Simulation Without Validation Is Decoration
Finite element and computational fluid dynamics packages will produce a colourful result for almost any input, including a wrong one. An examiner's first question is how you know the result is right. If the report has no answer, the analysis carries little weight regardless of how much work went into it.
Validation means comparing your model against something independent: an analytical solution for a simplified case, published experimental data, or your own test rig. Verification is different and also required, meaning evidence that you solved the equations correctly, principally through mesh independence.
The Mesh Independence Study
This is the cheapest marks in the report and it is routinely missing. Run the model at three or more mesh densities, plot the quantity of interest against element count, and show where it stops changing meaningfully.
| Mesh | Elements | Peak von Mises stress | Change from previous |
|---|---|---|---|
| Coarse | 12,000 | 186 MPa | reference |
| Medium | 48,000 | 203 MPa | 9.1 per cent |
| Fine | 190,000 | 209 MPa | 3.0 per cent |
| Very fine | 760,000 | 210 MPa | 0.5 per cent |
The figures above are an illustrative pattern, not measured data from a specific model. Report your own numbers and then state the conclusion: which mesh you adopted and why. Note also that peak stress at a sharp re-entrant corner will never converge, because the elastic solution is singular there. Recognising that, and reporting a stress at a defined distance or using a fillet, demonstrates real understanding.
Boundary Conditions Decide the Answer
More student models are wrong because of boundary conditions than because of meshing. A fully fixed face where the real joint is a bolted connection will stiffen the model and understate deflection, sometimes by a large margin.
| Real condition | Common student choice | Effect on the result |
|---|---|---|
| Bolted flange | Fully fixed | Understates deflection, overstates natural frequency |
| Weld along one edge | Fixed on the whole face | Misses local stress at the weld toe |
| Bearing support | Pinned in all directions | Suppresses axial growth, creates false thermal stress |
| Distributed pressure | Point load | Creates an artificial stress singularity |
| Contact between parts | Bonded contact | Prevents separation and sliding that occurs in reality |
State every constraint and load with a one-line justification. That table alone often distinguishes a credible model from a plausible-looking one.
Safety Factors and Material Data
Two questions recur in vivas. Where did your material properties come from, and what factor of safety did you apply and why?
Material data taken from a supplier datasheet is acceptable if cited. Data taken from a textbook table is acceptable if the grade matches. Data typed from memory is not. For polymers and composites in particular, properties are strongly dependent on temperature, moisture and orientation, so state the conditions the figures apply to.
On safety factors, name the basis rather than the number alone. A factor chosen because a standard requires it is defensible. A factor of 2 chosen because it feels comfortable is not, and an examiner will ask.
CFD-Specific Checks
- y-plus in range for the turbulence model and wall treatment you selected.
- Residual convergence reported, and monitored on a physical quantity rather than residuals alone.
- Domain independence, showing the boundaries are far enough away not to influence the solution.
- Turbulence model justified, not left at the solver default.
- Mass and energy balance checked across the domain.
Australian Standards and Accreditation
| Reference | Relevance |
|---|---|
| AS 1100 | Technical drawing conventions |
| AS 4100 | Steel structures, common in mechanical support design |
| AS 1210 | Pressure vessels |
| AS/NZS 1170 | Structural design actions |
| AS 4024 | Safety of machinery |
Engineers Australia accredits programmes against Stage 1 competencies, which include knowledge of engineering fundamentals, application of established methods, and, importantly, ethical conduct and professional accountability. A project report that maps its own work to two or three of those competencies in the conclusion is unusual and reads as professionally aware.
Structuring the Report
| Section | Job | Rough share |
|---|---|---|
| Abstract | Problem, method, key result, implication | Under 5 per cent |
| Introduction and aims | Why the problem matters, what you will answer | 10 per cent |
| Literature and background | What is already known, and the gap | 15 per cent |
| Methodology | Model, rig, instrumentation, procedure | 20 per cent |
| Results | Data with uncertainty, no interpretation | 20 per cent |
| Discussion | What it means, against the literature | 20 per cent |
| Conclusions and further work | Answers to the aims, stated plainly | 10 per cent |
The commonest structural error is interpretation leaking into results. If a sentence needs a citation, it belongs in the discussion.
Reporting Uncertainty Properly
An experimental result without an uncertainty is incomplete. Combine instrument accuracy, repeatability and any calibration uncertainty, and state the confidence level.
- Type A uncertainty comes from statistical analysis of repeated measurements.
- Type B uncertainty comes from other information, typically manufacturer specifications.
- Combined standard uncertainty follows from the root sum of squares of the components.
- Expanded uncertainty multiplies by a coverage factor, commonly k equals 2 for roughly 95 per cent.
Then compare simulation against experiment with the uncertainty band shown. A model within the experimental uncertainty is validated. A model outside it is not, and saying so honestly scores better than quietly presenting both and letting the reader assume agreement.
Common Mistakes and Fixes
| Mistake | Fix |
|---|---|
| Simulation with no mesh independence study | Three densities minimum, tabulated, with a stated conclusion |
| No validation against anything independent | Analytical case, published data or your own rig |
| Results reported without uncertainty | Combine Type A and Type B, state the coverage factor |
| Contour plots with no scale or units | Legend, units and the quantity named on every figure |
| Peak stress quoted at a singularity | Fillet the geometry or report at a defined distance |
| Boundary conditions not stated | List every constraint and load, with its justification |
| No link from analysis to a design decision | Say what changed in the design because of the result |
Suggested Images, Free to Reuse
Hosted on Wikimedia Commons and free to reuse with attribution. Confirm the licence on the file page before publishing.
| Subject | Source | Licence |
|---|---|---|
| Finite element mesh example | Wikimedia Commons | CC BY 4.0 |
| Meshing in finite element analysis | Wikimedia Commons | CC BY-SA 4.0 |
| Pre-engineered steel building frame | Wikimedia Commons | CC BY-SA 3.0 |
Where Ethical Support Fits
Discussing whether your boundary conditions are physically sensible, or whether your validation case is appropriate, is exactly what supervision is for. Technical staff and simulation specialists advise on this routinely, and a PhD expert writer in mechanical engineering can review the same reasoning.
What stays yours is the model, the measurements and the interpretation. In a project viva you will be asked why you constrained a face the way you did, and that answer has to be yours.
Frequently Asked Questions
How many mesh densities are enough?
Three at minimum, four is better. What matters is showing the trend flattening and stating which mesh you adopted.
What if my simulation disagrees with my experiment?
Report it and investigate. Disagreement with a stated explanation scores better than agreement with no evidence. Check boundary conditions, material properties and instrument calibration first.
Do I need both FEA and experimental work?
Not always, but a project with neither validation nor experiment struggles. Where a rig is impossible, validate against published data and say why experiment was not feasible.
Which referencing style do Australian engineering schools use?
IEEE and Harvard are both common, and some schools use APA. Check your unit outline, since it varies by institution and even by unit.
How do I cite an Australian Standard?
Give the full designation, year and title, and the clause where you rely on a specific requirement. Standards are usually accessed through your library subscription.
Should I include CAD models in the report?
Include key views and a general arrangement in the body, with full drawing sets in an appendix. The body should carry only what the argument needs.
Your Next Step Today
Before refining your model further, write one sentence stating what you will validate it against. If you cannot name an analytical case, a dataset or a measurement, the simulation cannot yet support a conclusion, and that is the gap to close before adding any more elements.
Trusted Sources
- Engineers Australia, Stage 1 competency standards and accreditation. Accessed 12 August 2026.
- Standards Australia. Accessed 12 August 2026.
- Institution of Mechanical Engineers. Accessed 12 August 2026.
- ASME Codes and Standards. Accessed 12 August 2026.
- Wikimedia Commons, freely reusable images. Accessed 12 August 2026.
Standards and unit requirements change. Where this guide and your unit outline or the current standard differ, follow those.
Related services
Related guides
- Civil Engineering Design Reports: A UK Student Guide
- Electrical Engineering Capstone Reports: A US Guide
- IEEE Referencing for Engineering and Computer Science
All academic writing guides or browse our academic writing services.