Question

Briefly describe how a Design of Experiments methodology can be used when testing a new pump design that has three variables (number of impeller blades, channel depth, outlet diameter). Why would a factorial design of experiments be better than optimizing each variable independently?

Answer 1

Design of Experiments (DOE) involves testing the design (in this case, the pump) at various levels of the variables that impact the design performance (in this case, number of blades, channel depth and outer dia).

To use design of experiments methodology, we need to decide the minimum and maximum values of the design variables and also their discrete levels. We need to find the dimensionless parameters which affect the system performance and ensure that the DOE testing covers the wide range of values of the dimensionless parameters that lie in the operating zone.

Factorial design will be better than optimizing each variable independently because it allows for simultaneous consideration of many variables at different levels and the interactions between those variables, using a smaller number of observations than conventional procedures.