Disassembly is a decisive step in the remanufacturing process of End-of-Life (EoL) products. As an emerging semi-automatic disassembly paradigm, human–robot collaborative disassembly (HRCD) offers multiple disassembly methods to enhance flexibility and efficiency. However, HRCD increases the complexity of planning and determining the optimal disassembly sequence and scheme. Currently, the optimisation process of heuristic methods is difficult to interpret, and the results cannot be guaranteed as globally optimal. Consequently, this paper introduces a general ontology model for HRCD, along with a rule-based reasoning method, to automatically generate the optimal disassembly sequence and scheme. Firstly, the HRCD ontology model establishes the disassembly-related information for EoL products in a standardised approach. Then, customised disassembly-related rules are proposed to regulate the precedence constraints and optional disassembly methods for each disassembly task of EoL products. The optimal disassembly sequence and scheme are automatically generated by combining supportive rules with the ontology model. Lastly, the human–robot collaborative disassembly planning of a gearbox is presented as a case study to validate the feasibility of the proposed methods. Our method generates an optimal disassembly scheme compared with other heuristic algorithms, achieving the shortest process time of 308 units and the fewest number of disassembly direction change of 3 times. Additionally, the reasoning procedure can be easily tracked and modified. The proposed method is both universal and easily reproducible, allowing it to be extended to support the entire remanufacturing process.

中文翻译：

---

智能再制造中基于本体和规则的人机协同拆卸规划方法

拆卸是报废 (EoL) 产品再制造过程中的决定性步骤。作为一种新兴的半自动拆卸范式，人机协作拆卸（HRCD）提供了多种拆卸方法来提高灵活性和效率。然而，HRCD 增加了规划和确定最佳拆卸顺序和方案的复杂性。目前，启发式方法的优化过程难以解释，并且不能保证结果是全局最优的。因此，本文引入了 HRCD 的通用本体模型以及基于规则的推理方法，以自动生成最佳拆卸序列和方案。首先，HRCD本体模型以标准化的方式建立了EoL产品的拆卸相关信息。然后，提出定制的反汇编相关规则来规范EoL产品的每个反汇编任务的优先约束和可选的反汇编方法。通过支持规则与本体模型的结合，自动生成最优的拆卸顺序和方案。最后，以变速箱人机协同拆卸规划为例，验证所提方法的可行性。与其他启发式算法相比，我们的方法生成了最优的拆卸方案，实现了最短的处理时间308个单元和最少的拆卸方向改变次数3次。此外，推理过程可以很容易地跟踪和修改。所提出的方法既通用又易于重现，使其可以扩展以支持整个再制造过程。