AutoPack - Automatic packaging of pipe ant tube installations based on optimization and machine learning

Purpose and goal

AutoPack´s purpose is to streamline the development process by connecting models from different disciplines in digital computational chains, and to automate engineering work based on tools such as design automation, multidisciplinary optimization and machine learning. The application for the project is hose routing in engine compartments, which is a time-consuming and labour-intensive process. The goal is to reduce the amount of engineering hours for packaging work by 50% and to achieve the corresponding cost savings, as well as a 25% reduction in lead time.

Expected results and effects

The overarching effect of the AutoPack project is to strengthen the Swedish automotive industry´s international competitiveness through a more efficient and data-driven product development process. The project has resulted in a number of computer tools and algorithms integrated into the AutoPack framework. The framework has been evaluated by engineers at Volvo Cars, and the results show a potential in reducing engineering working by 5-8 times. Furthermore, the automatically generated solutions are comparable to the manual solutions in terms of quality.

Planned approach and implementation

The AutoPack project is led by Linköping University and comprises the partners Fraunhofer-Chalmers Center for Industrial Mathematics, Volvo Car Corporation, and Intellium Engineering. The project is divided into four work packages (WPs) where models and methods for design automation are developed in WP1. Tools and algorithms for creating optimal hose routings are established within WP2. Machine learning algorithms are developed in AP3 together with an integrated AutoPack framework which has finally been implemented and validated at VCC in WP4.