In order to optimally setup a machine, the study of the tool life according to the material and the machining parameters is critical. The expression of the tool life could be estimated using an appropriate experimental campaign, which should have a limited size in order to reduce the experimental costs. This approach becomes of primary importance when the production is not in series where the costs can be spread over a large number of pieces.
The aim of this paper is to propose a new methodology that stops the experimental campaign as soon as the expected gain in carrying on the experimentation does not justify the marginal cost of experimentation.
To prove our idea, a simple problem from the well-known turning cutting condition optimization is used and the optimization technique Response Surface Methodology is selected.