Additive-manufacturing parameter optimisation
A design-of-experiments workbench for modelling and optimising additive-manufacturing parameters against strength, surface, accuracy, time, and material objectives.
Project definition
Problem statement
Layer height, speed, temperature, orientation, infill, material, and machine condition interact to affect part quality, time, and material use. Changing one factor at a time can miss interactions and produce conclusions that do not generalise.
The engineering problem is to design efficient experiments, build valid response models, optimise competing objectives, and confirm selected settings with held-out or physical tests.
Project objectives
- Define a bounded process, material, specimen, factors, levels, responses, and measurement methods.
- Generate a factorial, response-surface, Taguchi, or selected space-filling experiment design.
- Model parameter effects and interactions with diagnostics and uncertainty.
- Find Pareto-optimal settings for selected quality, time, and material objectives.
- Verify recommended settings through held-out or confirmation experiments.
System design
System modules
Experiment designer
Defines factors, ranges, constraints, replicates, randomisation, blocks, and the selected design method.
Measurement manager
Records specimen, machine, process, strength, roughness, dimensional, time, mass, and failure data.
Response modeller
Fits selected statistical or machine-learning models and displays residual and validation checks.
Optimisation engine
Searches feasible settings and constructs Pareto tradeoffs across weighted or constrained objectives.
Confirmation dashboard
Compares predicted and measured confirmation results with uncertainty and experiment traceability.
Methodology
System workflow
- 01Define study
The student fixes the process, material, specimen, factors, responses, and equipment limits.
- 02Create design
The system produces a randomised experiment table with repeats and constraints.
- 03Enter results
Measured outcomes and failed builds are recorded with specimen and process traceability.
- 04Model and optimise
Validated response models generate feasible settings and Pareto choices.
- 05Confirm
Selected settings are tested separately and prediction error is added to the final comparison.
Demonstration scenario
An FDM study varies layer height, print speed, nozzle temperature, and infill for strength, dimensional error, time, and mass. The workbench fits response models, displays significant interactions, creates a Pareto front, and compares one selected setting with a separate confirmation build.
Engineering
Technical architecture
- Web application
- Nuxt and Vue for study setup, experiment tables, effect plots, Pareto views, and confirmation results.
- Experiment API
- FastAPI for designs, observations, models, optimisation jobs, diagnostics, and exports.
- Data layer
- PostgreSQL for studies, factors, runs, specimens, measurements, models, candidates, and confirmations.
- Analysis engine
- Python, Pandas, statistical tools, and scikit-learn for design generation, response fitting, validation, and optimisation.
- Reproducibility
- Randomisation seeds, measurement units, machine and material records, versioned datasets, and saved model settings.
Testing
Evaluation
Evaluation measures
- Design balance, factor coverage, randomisation, and repeatability
- Cross-validated response error and residual diagnostics
- Accuracy of predicted parameter effects and interactions on prepared data
- Pareto feasibility and objective improvement against baseline settings
- Prediction error on held-out or physical confirmation runs
- Sensitivity to measurement noise, model choice, and factor-range changes
System boundaries
- Conclusions apply only to the documented process, machine, material batch, specimen, and parameter ranges.
- Prepared data can demonstrate the software, but a full engineering study requires controlled physical experiments.
- Model optimisation cannot recommend settings outside tested or explicitly bounded ranges with confidence.
- Machine safety, material handling, and test procedures remain governed by laboratory rules and qualified supervision.
Included
- 01Experiment design and measurement-entry interface
- 02Response modelling, optimisation, and confirmation modules
- 03Parameter effects, Pareto tradeoffs, and uncertainty dashboards
- 04Prepared or experimentally collected build data and results
- 05Complete source code in a private GitHub repository
- 06Complete project documentation with synopsis, abstract, methodology, specimen and process diagrams, experiment results, screenshots, and conclusion
- 07Setup and usage guide
Project record
No buyer information is collected on this page.
- Permanent project ID
- GP-ME-0NND5GR
- Catalogued
- 21 Aug 2026
- Completed
- Pending
- Verified
- Pending
- Demonstration
- Added when ready
Handover
After purchase
- 01Payment is confirmed
The project is marked unavailable and cannot be purchased again.
- 02Repository access is granted
The buyer's submitted GitHub account receives access to the private repository.
- 03The purchase record is delivered
The certification sheet is prepared from the reviewed buyer details and sent privately by email.