First-class risk management from second-use data sources: How intelligent data processing could make risk management more efficient and affordable for SMEs

In modern manufacturing, large data sets from different sources are permanently generated along the production chain. This data are supposed to be used to optimize products and production chains. However, in most cases, process participants only focus on acquiring data and leave it then to decision makers for interpretation. While there is nothing wrong with that on principle, comparable results might be achieved in a more efficient and productive manner by preprocessing already collected data and analyzing the outcome of legacy decisions. This requires connecting information from different sources within the product life cycle (horizontally: product stages, vertically: stakeholder - decision maker - implementer) and enriching models in an iterative way (comp. software engineering, consumer marketing). The goal of this paper is to describe how to use existing data in the industry in order to reduce the translation of "idle" information into failures in decision making. Often, critical characteristics have been discovered already, in a legacy product, a former product version or an abandoned project, and recognizing them in new product structures can be facilitated with IT-based quality management. In addition, the paper shows how intelligent selection and recombination of second-use data sources can help to detect and treat risks proactively, before adverse effects occur. Examples from our current research in risk identification for MedTech and during injection molding processes shall illustrate the fields of opportunities.