Automatic Control Sequence Generation for a Hybrid Disassembly System

Abstract

Disassembly plays a key role in a life-cycle-based economy since it enables the recovery of resources in product and material cycles. In order to set up future disassembly plants, various problems have to be solved in advance. This includes varying quantities as well as a wide range of product types and states. A partly automated disassembly system could adapt to a large variety of products and different degrees of devaluation. Differences between a given disassembly plan and the actual situation at the shop floor require a highly flexible control system. High flexibility can be achieved by rapid rearrangement of transport processes and disassembly operations based on the actual system conditions in relation to requested disassembly tasks. This requires an adaptive control based on modules. Additionally, the ability to exchange control modules, variables, and parameters online will be required. Consequently, at the device level, robots have to be adaptive to new requirements in cooperation with the cell controller. Modularized program blocks and online exchanges are also required for the robot controller.Note to Practitioners-In this paper, a concept for the generation of flexible control sequences for a partly automated disassembly system has been presented, which can cope with the disassembly object variety, different product conditions, as well as product amounts. Washing machines were chosen as an example for the product group domestic appliances. The approach is directed at the planning level and cell-control level. The concept for the generation of disassembly control sequences consists of a function unit for the disassembly control system. The procedure, which will be implemented in the module, considers the availability of disassembly tools and the technological feasibility of disassembly processes and tools. A modular control sequence will be generated by the programmed function unit. The developed control concept can also be applied for very small lot-sized manufacturing and assembly tasks.

Modelling and condition-based control of a flexible and hybrid disassembly system with manual and autonomous workstations using reinforcement learning

“…Regarding disassembly systems, almost all control approaches generally exclude defective parts and failing processes because of the otherwise significantly increased complexity (Kim et al 2007a). On the other hand, it is undisputed that in any reverse logistics system, where discarded products are collected and disassembled, the aforementioned variance in the condition of products brings enormous uncertainty into the disassembly process and can, thus, lead to an even increased chance of errors (Altekin and Akkan 2012).…”

“…There is a consensus in research that hybrid production systems consisting of manual and automated workstations offer the greatest potential for a productive and economically reasonable industrial application (Kim et al 2007a). Besides a high number of variants and low quantities, the main reason is the high fluctuation of product conditions.…”

Disassembly 4.0: A Review on Using Robotics in Disassembly Tasks as a Way of Automation

“…Moreover, 14 papers (8 %) involve an automation approach without necessary a priori knowledge. These papers deal with learning models for robotic controllers [46] or agent-based configurations, mostly equipped with cameras and sensors to identify fasteners and components autonomously [47][48][49][50][51][52][53] and more abstract concepts of potentially autonomous systems and approaches [54][55][56][57][58].…”

Simulation-Supported Verification of Methods for Controlling Disassembly Lines

“…As part of the planning of a hybrid disassembly system, Kim, Harms and Seliger [13] developed reactive control measures. Their application, however, also involves a robotic automated disassembly system, but not the manual systems considered in this paper.…”