Concept for a tailor-made visual inline inspection system for fibre layup processes depending on material and process characteristics

Fibre layup processes like Automated Fibre Placement (AFP) are state of the art technologies for the production of large scale lightweight components like wing covers or fuselages. These composite components are produced layer by layer using narrow material strips. During this material layup it is possible to generate defects e.g. wrinkles, twists, gaps, overlaps, etc. with probably a huge influence on the mechanical properties of the produced component. Recognizing these defects, inline quality assurance (QA) systems are part of ongoing research and industrial development. These systems are mostly inflexible in the use of different sensors and algorithms. One single system is used for every production case. For different materials and various processes, this system performs differently and possibly non-deterministic. In order to compensate this issue it is necessary to understand the relation between optical material properties and the used sensors and algorithms. The approach considered in this paper is based on analysing optical characteristics of fibre materials and link them to the characteristics and properties of various feasible sensors and algorithms. This proposal has the potential to increase the performance of an inline QA system significantly. Using this optical properties data it will be possible to build a tailor made inline inspection system. This system is specifically designed by knowledge and had a well-known error range. During this work, the following scientific questions have to be answered: - Which dependencies in optical characteristics had to be analysed to configure a tailor-made inline QA system for a specific material? - How do we describe layup defects by its optical characteristics? - How robust performs an inline QA system for different materials? Firstly, it is planned to implement the general system concept, including multiple sensors and algorithms. Following the optical material characterisation and the linking to the corresponding sensor-algorithm behaviour will be considered.