This talk provides a briefing on design of experiments (DOE) for rapid and effective screening, characterization and optimization of factors affecting reliability. It covers a broad range of design of experiment (DOE) tools, including factorial design, Weibull regression, split plots for hard-to-change factors and response surface methods (RSM) for minimizing failure rates. Anyone tasked with improving reliability will benefit by learning how to make the most from every experiment via these multifactor testing methods.
This webinar upgrades robustness considerations previously discussed for characterization via factorial design to the optimization phase of process or product development. It lays out response surface methods (RSM) that identify regions within the design space that not only meet all response specifications, but also minimizes variation caused by wandering factor settings and extraneous noise factors. Attend this webinar to learn advanced RSM tools such as propagation of error (POE) to discover the robust high plateaus of performance.
The goal of robustness studies is to demonstrate that our processes will be successful upon implementation in the field when they are exposed to anticipated noise factors. Join this webinar to learn about the assumptions and underlying concepts that need to be understood when setting out to conduct a robustness study, and what kind of design you need to fit each situation.
Building up from the Mixture DOE Crash Course, this webinar explains how formulators can create experiment designs that combine mixture components with process factors, include categorical factors, and deal with hard-to-change variables.
This presentation provides practical aspects for combining mixture, process and categorical variables into one optimal experiment-design using innovative "KCV" models. Learn about structuring combined designs with a case study on making delightful chocolate chip cookies.
DOE is often presented as a “one shot” approach. It may be more efficient to divide the experiment into smaller pieces, thus expending resources in a more adaptive manner. This sequential approach becomes especially suitable when beginning with very little information about the process, for example, when scaling up a new product. It allows for better definition of the design space, adaption to unexpected results, estimation of variability, reduction in waste, and validation of the results.
By removing known sources of variation such as day-by-day, blocking reduces error in statistical analysis, thus providing higher power for estimating effects and greater precision for fitting response surfaces. The technique, however, encounters limitations in some settings. This quick, 30-minute webinar will discuss the basics of blocking and demonstrate fixed vs. random block effects using the powerful, yet intuitive Stat-Ease® 360 software.
In many cases, experimental data is the result of a deterministic simulation rather than a lab experiment. These may be referred to as computer experiments. In other cases, physical experiments may produce low or zero-error response measurements. Such situations need special experimental designs and data analysis tools. See how Stat-Ease 360 fills this need with via space-filling designs and Gaussian process models.
In this talk, Mark Anderson details cost-saving mixture-process methods invented by statisticians Kowalski, Cornell and Vining (KCV) and implemented by Stat-Ease. The KCV tools streamline combined designs by focusing on the interactions—the hidden gold remaining buried by traditional experimentation. Via a real-world example, Mark will present experiment-design and modeling methods that make combined mixture-process studies practical for chemists.
In this advanced-level webinar, Stat-Ease Consultant Pat Whitcomb discusses robust design, propagation of error, and tolerance analysis. Propagation of error (POE) accounts for variation transmitted from deviations in factor levels. It finds the flats—high plateaus or broad valleys of response, whichever direction one wants to go—maximum or minimum; respectively. Tolerance analysis drills down to the variation of individual units, thus facilitating improvement of process capability.
Pat Whitcomb details the cost-saving mixture-process models developed by Scott Kowalski, John Cornell and Geoff Vining (KCV). Design-Expert® software drops this modeling tool right into the user's hands. See how it reduces the number of model terms and thereby reduces the number of runs required to estimate the complex relationship between mixture and process variables.