Robotic disassembly technology as a key enabler of autonomous remanufacturing

点击翻译按钮获取中文摘要

Remanufacturing is "the process of returning a used product to at least OEM original performance specification from the customers' perspective and giving the resultant product warranty that is at least equal to that of a newly manufactured equivalent". Remanufacturing can be more sustainable than manufacturing de novo "because it can be profitable and less harmful to the environment ...". Remanufacturing is a sizable industry. For example, in the USA, there are more than 73,000 companies engaged in remanufacturing. They employ 350,000 people and have turnovers totalling $53 billion. A key step in remanufacturing is disassembly of the returned product to be remanufactured. As it is complex, disassembly tends to be manually executed and is labour intensive. We propose to develop robotic technology allowing disassembly to be carried out with minimal human intervention or in a collaborative fashion by man and machine. We aim to facilitate the cost-effective robotisation of this critical step in remanufacturing to unlock the potential of remanufacturing and make it feasible for many more companies to adopt, thus helping to expand the UK's £2.35 Billion remanufacturing industry. Our research will start with a detailed investigation of disassembly processes aimed at fundamentally understanding them. Such a fundamental understanding does not currently exist but is necessary to support the development of robust disassembly strategies and systems that can autonomously handle variability in the product. We will study basic common tasks such as unscrewing, removal of pins from holes with small clearances, separation of press-fit components, extraction of elastic parts (e.g. O-rings and circlips) and breaking up of 'permanently' assembled components. We will analyse those generic disassembly tasks for feedback information that can be obtained while a robot is performing them. We will employ different types of sensors to provide feedback appropriate to a given task. In addition to visual sensing, we will focus on using contact forces and moments as a means to gauge the state of the disassembly operation. To counteract uncertainties, such feedback will be helpful in guiding the robot and avoiding damage to the components being taken apart. We will apply the acquired basic process knowledge methodically to create models, scheduling algorithms and learning tools to enable autonomous or semi-autonomous disassembly by robotic systems. We will develop strategies for planning and implementing multi-robot operation when the disassembly task is too complex for one machine. We will devise techniques for effective collaboration between humans and robots in cases where the work is too difficult for people or for machines on their own. We will validate these plans, strategies and techniques experimentally and will give public demonstrations of collaborative robotic disassembly using real products as examples. Our multi-disciplinary project team, with experience in robotic assembly, intelligent systems, CAD/CAM and process modelling, will be supported by three industrial partners (Caterpillar, Meritor and MG Motor). These user companies will supply case studies for evaluating the research results. Two technology translators (the Manufacturing Technology Centre and the High Speed Sustainable Manufacturing Institute) will contribute to converting laboratory-based technology into solutions ready for deployment on an industrial scale.