After a Designed Experiment …

You can build a model to predict future process results!

Contour Plot from STATISTICA visualizing corrosion = f(Mn, Mg)Following a designed experiment, a model may be made to visualise and predict a response as a function of the factors varied. At the right is a graph predicting corrosion variation as a function of Mg and Mn. Here, this is an interaction effect – the effect of Mn depends on the level of Mg. A Designed Experiment allows the measurement of effects of factors and their interactions. Often prediction of process behaviour is not intuitively obvious, due to the presence of interactions!

You can simultaneously optimize multiple responses!

Visualization of Models predicting Popcorn Taste and Unpopped Kernels, DesignExpert
Usually when a designed experiment is run, measurements are taken of as many responses as possible. Sometimes improving one response results in making another response not as desirable. Design of Experiments software such as Design Expert, JMP, Minitab, and STATISTICA allow the modeling of several responses, setting desirable targets for each response – and then predicting the best combination of experimental factors to simultaneously optimize each response!

Visualization of models to predict popcorn taste and unpopped kernels, JMP
The diagram to the right shows the responses of taste and unpopped kernels in microwave popcorn, varying the factors time (x axis) and power (y axis). The best (highest) taste ratings occur with low power and low time (bottom left corner). The best (lowest) unpopped kernel response is with high time and high power (top right corner). It seems to be difficult to obtain best values of both of these responses! The white region shows a region where taste is between 60-70, and unpopped kernels is < 1.5 oz. This is just about the best we can do to optimize both responses!

Download pdf’s visualizing models from microwave popcorn designed experiment: Design-Expert, JMP.

You can reduce the number of factors for future investigation!

Another outcome of Designed Experiments is the screening of the number of important factors from the “trivial many” to the “vital few” factors which are critical to efficient process running and product quality. Once the critical factors have been identified, these may be monitored using SPC charts, which allow the operator to distinguish between “common cause” and “special cause” variation.

Learn how to set up, run, analyse and present results from designed experiments, at our intensive hands on Design of Experiments Workshop.