Download the course curriculum
This course focuses on designing and analyzing mixture experiments that enable you to predict and optimize system characteristics, which are critical to product and process design and quality. Products and processes that involve mixtures of liquids, solids, or gases are pervasive and impact the sustainability of your organization. Mixture systems involve inherent constraints and complex kinetics, thereby making their behavior difficult to characterize and predict.
This course introduces an innovative and cost-effective approach to designing and analyzing mixture experiments. You will design experiments that are much smaller and more cost-effective than traditional classical designs. You will apply a modern machine learning method, Self-Validating Ensemble Modeling (SVEM), to build accurate and stable predictive models of system behavior, using small data sets. A variety of design situations and numerous case studies will show you how to apply the methodology and demonstrate how well it works in practice.
Who Should Attend
Engineers, scientists, and researchers who work directly with mixtures, mixture designs, and mixture processes in various industries, including chemical, specialty chemical, pharmaceutical, biotechnology, semiconductor, solar, consumer products, and automotive
24 hours (3 days) of instruction
- On-site classroom
- Virtual live
- Virtual on-demand (coming soon)
No prior, formal training in design of experiments is assumed. Some familiarity with basic statistics is desirable. Options for gaining basic knowledge of statistics are available for participants without this background.
You and your team will have the methodology and tools to construct cost-effective mixture experiments and fit accurate and stable models, thereby enabling you to leverage these models.
Day 1 (8 hours)
About the Modern Mixture Experimentation Course
Factorial versus Mixture Designs
Mixture Spaces, Components, Blending
Modeling Mixture Experiments
Case Study: Etch Rate
Types of Mixture Designs
Case Study: Yarn
Case Study: Etch Rate (Space-Filling Design)
Exercise: Construct a Space-Filling Design
Modern Machine Learning
Day 2 (8 hours)
Modern Machine Learning Example
Self-Validating Ensemble Modeling
Self-Validating Ensemble Modeling (SVEM) Add-In*
Saving Profiler Scripts
Exercise: Conduct a SVEM Analysis
Bounds on Components
Case Study: Flare (Bounds)
Case Study: Flare (Linear Constraints)
Classical vs Modern Mixture Experimentation
Example: Waste Glass
Responses, Factors, and Constraints Tables
Day 3 (8 hours)
Mixture of Mixtures Experiments
Case Study: Photoresist Blending
Case Study: Harvey Wallbanger
Exercise: Superfood Drink
Optimization and Multiple Responses
Case Study: Detergent
Case Study: High Alloy Steel
Mixture-Process Factor Experiments
Case Study: Fly Ash
Case Study: Fabric Finish
Additional session covering Functional Responses (4 hours)
Modeling Functional Responses
Case Study: Fly Ash Curves (Fit Curve)
Functional Responses and Functional Data Explorer (FDE)**
Case Study: Fly Ash Curves (FDE)**
Other Course Add-on Options
- Additional half-day session covering functional responses
- Virtual office hours with instructors for questions, feedback, and discussion
- Consulting services for project work and specific requirements
Request More Information about This Course
Please get in touch with us if you have questions or would like more information on how to register for this course. Enquire about our corporate discount for group registrations from the same organization.
* SVEM is run using JMP (standard) software, for which a trial license is available.
** Specialized topic exercises use JMP Pro software, for which an evaluation license is available.