Friday, 06 June 2014, 11:00 a.m.
Sala Grande Palazzina B via alla Cascata 56/C
The method of Design of Experiments: general concepts and example of application
IUT Le Mans University Institute of Technology,
Université du Maine, Le Mans, France
The method of Design of Experiments (DoE) is a statistical approach of experimentation where all the setting parameters are varied simultaneously. This statistical technique enables to get a maximum amount of reliable information in a minimum number of experiments.
The basic concepts of the method will be presented in a first part particularly i) the concept of experimental domain, which must be defined by the user when implementing the DoE and before making any experiments and ii) the essential concept of mathematical modeling which enables the experimental domain to be investigated while avoiding too many experiments to be performed.
The method will be illustrated in the second part with an example of process improvement : 12 parameters were investigated by using a screening DoE including a reasonable number of 20 experimental runs. We will show that the method offers simple but efficient modeling tools to analyze and interpret the measured data.