Title: Accelerating product launches of high-volume applications with data mining

Authors: Sebastian Körner; Elisabeth Ladstätter; Klaus Drechsler

Addresses: BENTELER SGL Composite Technology GmbH, Fischerstraße 8, Ried im Innkreis 4910, Austria ' Department of Mechanical Engineering, Technical University of Munich, Boltzmannstr. 15, Garching 85748, Germany ' Department of Mechanical Engineering, Technical University of Munich, Boltzmannstr. 15, Garching 85748, Germany

Abstract: Launches of new products or manufacturing processes are challenging due to requirements regarding quality and capability. This article observes a production process for GFRP leaf springs based on the VARTM technology. The production of structural automotive applications in high volumes requires automated production lines. A multiplicity of sensors alongside that production processes acquires a huge quantity of data about the current production conditions. The fields of statistics and data mining offer new possibilities to analyse these data. An approach for the exploration of an optimised set of parameters is developed. Thereby the main influencing factors for particular defects are determined. Predictive models to forecast the probability of possible defects are established. These findings confirm that data mining encourages and accelerates the launch of new products and processes at an early stage of production. The time to market and the manufacturing costs for fibre-reinforced plastics can be reduced.

Keywords: automated serial production; data mining; GFRP; predictive analytics; process optimisation; product launch; quality improvement.

DOI: 10.1504/IJAUTOC.2016.084326

International Journal of Automotive Composites, 2016 Vol.2 No.3/4, pp.189 - 208

Received: 07 Sep 2016
Accepted: 04 Feb 2017

Published online: 25 May 2017 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article