Hand lay-up of complex geometries-prediction, capture and feedback

This paper presents a process improvement framework built on previous research activities at the University of Bristol. The work focusses on hand lay-up and seeks to reduce variability, improve productivity and increase manufacturability of future designs. The framework is based on a double-loop learning model which incorporates prediction, capture and feedback. The predictive method employed uses a kinematic drape model as part of an expert system. The expert is needed to translate the model outputs into a more realistic set of drape instructions. The lay-up is captured by video analysis and quality data captured using an on-line tool. This data is then fed back to the user to facilitate decision making.