Design of experiment (DOE) approaches

Design of experiment (DOE) approaches for parameter studies and sensitivity analyses

Analysis of a cause-and-effect relationship using Design of Experiment approaches (DOI)
© Fraunhofer IWS Dresden
Analysis of a cause-and-effect relationship using Design of Experiment approaches (DOI)

Laser materials processing is characterized by a high number of variables which influence the processing results. On the one hand, these are user-selected parameters such as laser power and processing speed and, on the other hand, many disturbance variables causing processing faults and/or deviations from the expected results. In addition, unknown relationships and often non-linear interactions between user-selected parameters and disturbance variables cause statistical variations of interesting responses which must be considered by a proper process analysis. The experimental effort of conventional one-factor-at-a-time methodology becomes very high due to the necessary replications of trials to determine data scattering and standard deviation. In addition, there are limitations with respect to interactions between different variables and it is often difficult to find the true dependencies between control parameters and quantities of interest.

A promising alternative is offered by Design-of-Experiment (DOE) approaches including 2-level factorial designs and multi-level response-surface methods (RSM). These methods allow for a detailed and statistically validated evaluation of results, either achieved experimentally or numerically, with a defined scope of experiments. 

We support our customers by defining appropriate experimental designs to reveal cause-and- effect relationships using of statistical software. This approach is applicable both for experimental tests and numerical investigations. Often, the most significant results are achieved if both methodologies are simultaneously performed.