Herausforderungen im RQM mit KI-Lösungen meistern

Kevin Wrasse

Contact Software GmbH

Abstract

Artificial intelligence (AI) has the potential to increase efficiency in Product Lifecycle Management (PLM) processes by automating manual process steps. Through Natural Language Processing (NLP)-based comparison of requirements, similar requirements can be identified via a data pool, avoiding unnecessary duplicate processing or reusing underlying information and models. This will increase the efficiency of product development.
The author presents an approach and its implementation in a python application to identify the most similar requirements out of a large number in an automated and highly accelerated way. It implements the common natural language processing pipeline for text similarity tasks like pre-processing, optimization of the mathematical word representations and the calculation of a similarity metric. All processing steps are implemented with open source libraries. We present the application itself and the result of experiments with several trained neural network models, pre-processing steps and text similarity-approaches and discuss the best approach for automated comparison of text requirements. Complementary AI solutions like translations and quality assurance support can enhance the benefit when integrated as services in the RQM tool. The author will demonstrate an integration of this new applications into the PLM technology CONTACT Elements, in order to validate applicability in proven, real PLM software environments. This integration illustrates the potentials of processing data objects that contain natural language more precisely and less time consuming in automotive-relevant requirements management use cases.

Biografie

Kevin Wrasse studierte Informationstechnik im Maschinenwesen an der TU Berlin mit dem Schwerpunkt Konstruktion und Fertigung.  Er war von 2013 bis 2017 Wissenschaftlicher Mitarbeiter am Fraunhofer Institut für Produktionsanlagen und Konstruktionstechnik (IPK) und arbeitete anschließend für zwei Jahre als Systems Engineer bei MELAG Medizintechnik. Seit  2019 ist Herr Wrasse Projektingenieur PLM/Industrie4.0 im Produktmanagement bei CONTACT Software.