In a future circular economy, return flows and disassembly processes will play a dominant role to recover valuable components and materials for reuse. In order to contribute to the rationalisation of disassembly operations, this work presents a method to generate machine-readable disassembly plans for both human and robot resources. Previous research work on (dis)assembly operations in which products are modelled as graphs has provided methods for optimal sequencing of the disconnect actions. But disassembly also entails part and component manipulation tasks for which part orientation and clustering are relevant. Therefore, this paper presents a method based on a topological description of a product in the form of a hypergraph, to capture additional relevant properties. To further develop this model into a practical tool for circular businesses, which rely on efficient disassembly operations, an exploration is conducted on how to store disassembly product models in a suitable graph database. This needs a disassembly ontology to be mapped onto a graph schema, considering scalability to multiple products and product types. As a case study, a computer mouse product model was created and loaded into a graph database, from which a disassembly sequence was successfully retrieved. Additionally, scripting was developed to provide visualizations and metrics of the product to illustrate its disassemblability (based on existing python modules, with good user interaction possibilities). While further work is needed to enhance the user-friendliness of the tools and data flow, it provides a base for a robust product model repository for disassembly operations, suitable for generating disassembly plans and instructions.
Fraanje et al. (Thu,) studied this question.