These will be added as speakers submit their talks. Check back to see what's new!
Register now!
ImproClean – Intelligent and resource-efficient monitoring system of process quality of disinfection and cleaning along the healthcare supply chain
by Oliver Thunich, STATCON GmbH
During the government funded research project ImproClean, Statcon performed multiple DoEs with the goal of predicting hygiene parameters along the healthcare supply chain. DoE strategies used for investigating laundry and surface disinfection processes such as optimal designs and design augmentation and their implementation with the easy to use features of Statease 360 are presented. Additionally, the application of the models generated in a data streaming and prediction platform, utilizing data from innovative sensor systems provided by the project partners will be discussed.
DOE: Developing Optimal Espressos
by Andrew Macpherson, Prism Training & Consultancy
Making a cup of coffee is simple, isn’t it? Well, yes… unless you’re a DoE trainer searching for a suitable topic to present at a glamorous international statistics conference!
When viewed through this Prism, brewing the perfect espresso transforms into a complex multi-stage process, allowing us to replicate many of the difficulties experienced by proper scientists.
This highly caffeinated presentation will offer practical solutions to real-world challenges faced by scientists, engineers and formulators from every industry. We will demonstrate DoE techniques essential to constructing practicable designs, generating reliable data, and reaching otherwise impossible levels of process understanding!
Definitive screening designs in the creation of robotic coating application for vehicle refinishes business
by Niels Dekker, AkzoNobel
In the vehicle refinishes business it is crucial that the color of the repair paint matches the vehicle that is up for repair in the bodyshop. AkzoNobel has several processes for the bodyshops to ensure this is achieved independent of the car that enters the repair process. One of the aspects of this process is the ability to internally apply coatings in a fast and efficient way using robots. Before this can be done, the robot application needs to match the manual application as executed in the bodyshop. In this talk, I will discuss and show how definitive screening designs add value to this process.
Addressing a multivariate and multilevel challenge
by Erik Coppens, ONDRAF/NIRAS
In this presentation, we will explore the successful application of Design of Experiments (DoE) in optimizing the development of a conditioning process for liquid radioactive waste. By leveraging Generalized Least Squares (GLS), we systematically analyzed the effects of key production settings while accounting for batch-to-batch variability in the liquid waste and other raw materials.
This data-driven approach enabled us to identify optimal process conditions that ensure compliance with product specifications while unlocking a potential cost reduction—potentially saving multiple millions. This work demonstrates the power of DoE in addressing not only multivariate problems but also multi-level challenges involving multiple sources of variability.
The one that got away – Experiments avoided through DOE
by Morten Bormann Nielsen, Danish Technological Institute
One of the central value-drivers in Design of Experiments is avoiding unnecessary work, but the process of how this is done in practice is rarely paraded in front of an audience. In this talk I will share examples where DOE was used to find the minimum (but still adequate) amount of work required for the given task and show what my decision process looks like. As a counterexample, we will also dive into a horror story where DOE was not used to plan a large experiment to see what the poor souls involved missed out on.
Using Stat-Ease 360® for simulation and optimization of first-principle and simulation models (metamodeling)
by Frank Westad, Idletechs AS & the Norwegian University of Science and Technology
DoE is, as we know, the best strategy for planning and performing experiments for process optimization and product development. After acquiring the empirical data, ANOVA is typically applied for analysis and visualization. Nevertheless, DoE has also a huge potential also as a tool for understanding and optimizing first-principle (physical, mechanistic) models and simulation systems, so-called metamodeling. Some of the objectives of metamodeling are the following:
- Perturb the parameters in the physical/simulation model
- Estimate the sensitivity of the parameters (including interaction and square terms)
- Establish an empirical model between input and output
- Find the optimal settings given constraints
- Predict output while circumventing the physical model
Firstly, the concept of metamodeling will be described, followed by considerations regarding what kind of experimental designs that might be appropriate in various situations. Results from selected cases studies involving metamodelling will be presented.
State of Stat-Ease Software: A Look Back & The Path Forward
by Martin Bezener, Ph.D., President & CTO, Stat-Ease
Full abstract coming soon!
Diving into Detail on New Stat-Ease Features
by Hank Anderson, Vice President of Software Development, Stat-Ease
Full abstract coming soon!