ENBIS: European Network for Business and Industrial Statistics

Process improvement using designed experiments: a case study in an aerospace manufacturing company

23 September 2009, 10:45 – 11:05

Abstract

Submitted by

Roberto Marrone

Authors

B. Palumbo (1), G. De Chiara (2), R. Marrone (2)

Affiliation

(1)Dept of Aerospace Engineering, Univ. of Naples Federico II, Naples-Italy; (2)AVIO S.p.A., Manufacturing Technologies Dept, Pomigliano, Naples-Italy

Abstract

This paper shows the strategic role that a systematic and sequential approach to industrial experimentation plays in technological process improvement. It enables to integrate engineering and statistical knowledge and allows to put into action a virtuous cycle of sequential learning consistent with Six Sigma approach.

A case study concerning the pre-brazing tack-welding process of an aero-engine honeycomb structured component is proposed. A first pre-experimental phase (i.e. the phase that precedes the actual experiments) lays the experimental foundations of the successive screening and optimization phases: an half fraction of a five-factor factorial (with three replicates) and a three-factor central composite design with three centre points and alpha fixed at 1 (with two replicates) are adopted, respectively. Five response variables, each of which related to a specific type of defect, are taken in consideration. A multicriterion optimization approach based on an overall desirability function is adopted in the optimization phase. Finally, some SPC tools are proposed for monitoring process in production context.

The proposed case study is a positive example of synergic collaboration and partnership between academic statisticians and industrial practitioners; it has been developed in AVIO industry, an international aerospace company at the leading edge of propulsion technology.

Return to programme