Analyse von Maschinendaten zur Entscheidungsunterstützung bei der Produktverbesserung durch die Anwendung eines Feedback Assistenz Systems

Technological progress leads to an increasing number of generated and managed data. This trend can also be observed in the context of the product life cycle. During the use phase data are collected from the product automatically, with increasing tendency, by means of on-site, embedded or mounted sensors. In addition, the operator or service staffs insert substantial data into information systems. The accumulated data are used for specific purposes, such as billing of services, and afterwards are archived in the repositories. Knowledge can be generated from the collected data with the intention to support development and especially improvement of the product.
A successful product development leads to high-quality goods and customer satisfaction which ultimately influences on the company’s success in the market with large scale sales of goods. Hence, the product development is consistently a subject of research. Currently in the product life cycle, the subsequent phases, especially the use phase is not considered. The aim of this work is to realize a knowledge transfer where applicable knowledge from the product use data is fed back into product development. The focus is on industrial goods, where the feedback data are structured in multiple data sources, and also manufacturers and customers are in close contacts.
For the management and analysis of data, a feedback assistance system (FAS), is designed and developed. Here the data from different sources are transferred into a unified database. On this data layer, knowledge-based methods can be applied i.e. from the field of data-mining. These methods should assist product developers in the improvement of existing product generations. The captured, extensive amounts of data are, therefore, condensed, and patterns are detected. As a result, the FAS provide the product developer with decision-relevant and intuitively under-standable knowledge. In the scope of product improvement, important is to determine the areas in which the feedback data can be used successfully. So for the FAS, a combination of methods for three main applications will be integrated (1) verification of customer requirements, (2) fault diagnosis, and (3) evaluation of improvement potential. Cost and time indicators are deployed for the verification of requirements which consequently reveal the non-achieved objectives of the product improvement process. So the diagnosis should be applied to detect weak points and failure causes as improvement potential such as reduction of the error rate. To eliminate the existing weaknesses and deficiencies, the product developers should evaluate a variety of alternatives according to the objectives of the product development on the basis of the feedback data. For this purpose, a method from the multi-criteria decision theory is implemented.