Define factors, levels and responses precisely enough to design a valid experiment.
Quality & Productivity
Design of Experiments for Optimising Process Parameters and Yield
A practical course in design of experiments that teaches engineers to plan factorial and response surface studies, analyse results statistically and optimise process yield.
🗓️Duration5 training days
📚Content4 modules · 8 sessions
🎓On completionAccredited attendance certificate
About the programme
Course Overview
Changing one process parameter at a time to chase better yield is slow, expensive and frequently misleading once parameters begin interacting with each other. This course teaches design of experiments as a structured alternative: defining factors, levels and responses before a single trial is run, choosing between full factorial, fractional factorial and response surface designs, and reading the resulting main effects, interaction plots and analysis of variance output correctly. Sessions work through a realistic process optimisation scenario, from screening a long list of candidate factors down to the few that matter, through fitting a response surface to locate an optimum, to confirming that optimum holds up under normal production conditions. Participants also cover the practical side experiments are usually let down by: randomising and blocking runs to avoid hidden bias, sizing an experiment against the resources actually available, and communicating results to people who were not in the room when the study was designed. The course closes with a look at robust design ideas for keeping a process on target despite noise factors that cannot be controlled in normal production.
Expected Learning Outcomes
Choose between full factorial, fractional factorial and response surface designs for a given problem.
Use a screening design to reduce a long list of candidate factors to the vital few.
Interpret main effects plots, interaction plots and analysis of variance output from an experiment.
Apply randomisation and blocking to protect an experiment from hidden sources of bias.
Fit a response surface model to locate optimum process parameter settings for yield.
Confirm an experimentally derived optimum with verification runs before changing standard settings.
Who Should Attend
Process and manufacturing engineers seeking to improve yield without prolonged trial and error.
Expert improvement practitioners preparing the analyse and improve phases of a project.
R&D and formulation scientists optimising a product or process with multiple interacting variables.
Quality engineers supporting process validation studies that require statistical evidence.
Production supervisors who approve parameter changes and need to judge experimental evidence.
Continuous improvement leads introducing structured experimentation in place of ad hoc trials.
Course Modules
Select any module to see its sessions and points.
01Foundations of Planned Experimentation
2 sessions · 8 points
Session 1Factors, Levels, Responses and Objectives
- Translate a vague improvement goal into a measurable response variable suitable for experimentation.
- Select factors and set realistic levels that reflect the true operating range of the process.
- Distinguish controllable factors from noise factors that cannot be fixed during normal production.
- Write a clear experimental objective that states what decision the results are meant to support.
Session 2Why One-Factor-at-a-Time Testing Fails
- Demonstrate how one-factor-at-a-time testing misses interactions between two or more process parameters.
- Compare the number of runs and information gained from one-factor-at-a-time testing against a factorial design.
- Identify a past process trial that reached the wrong conclusion because of an unrecognised interaction.
- Explain why designed experiments generally need fewer total runs to reach a reliable conclusion.
02Factorial and Screening Designs
2 sessions · 8 points
Session 1Full and Fractional Factorial Designs
- Construct a full factorial design for a small number of factors and calculate the required run count.
- Build a fractional factorial design when the full factorial is too large to run economically.
- Explain confounding and aliasing in fractional designs and their consequence for interpretation.
- Choose a design resolution appropriate to how much interaction detail the project actually needs.
Session 2Screening Designs for Narrowing Down Factors
- Apply a Plackett-Burman or similar screening design to test many factors in relatively few runs.
- Rank factors by effect size to separate the vital few from the trivial many before further study.
- Decide which factors to carry forward into a more detailed optimisation experiment.
- Recognise when screening results suggest a factor was set at the wrong range rather than being unimportant.
03Analysing Experimental Results
2 sessions · 8 points
Session 1Reading Effects, Interactions and Variance
- Calculate and plot main effects to see how each factor individually shifts the response.
- Read an interaction plot to judge whether one factor's effect depends on the setting of another.
- Interpret an analysis of variance table to judge which effects are statistically significant.
- Check residual plots to confirm the model's assumptions before trusting its conclusions.
Session 2Response Surface Methodology and Optimisation
- Fit a response surface model once a promising region of factor settings has been identified.
- Use contour and three-dimensional surface plots to visualise how yield changes across factor combinations.
- Locate an optimum setting using the fitted response surface rather than by further guesswork.
- Apply desirability functions when optimising more than one response, such as yield and a quality characteristic.
04Practical Execution and Robustness
2 sessions · 8 points
Session 1Running a Credible Experiment on the Shop Floor
- Randomise the run order of an experiment to prevent time-related drift from disguising itself as a factor effect.
- Use blocking to remove the effect of a known nuisance variable such as a raw material batch change.
- Plan an experiment's resourcing, including material, machine time and operator availability, before scheduling it.
- Brief operators and shift teams so an experiment is executed exactly as designed rather than adapted informally.
Session 2Confirming Results and Building in Robustness
- Run confirmation trials at the predicted optimum before updating standard operating parameters.
- Apply basic robust design thinking to choose settings that keep yield stable despite noise factors.
- Update control plans and operator instructions once an experimentally derived optimum is confirmed.
- Archive experimental designs and results so future projects can build on rather than repeat the work.
What the participant receives
4 course modules
A structured syllabus
8 training sessions
across 5 days
32 detailed points
Applied, detailed content
Accredited attendance certificate
On completing the programme
Complete your registration
We will contact you within one business day to confirm.
Ready to start?
Reserve your seat and start building the skill.