Research on Knowledge-based Assembly Process Planning Method Based on Representation Learning Reasoning

For multi-variety and small-batch assembly parts, the assembly process updates and iterations are very frequent. The assembly process planning method based on manual experience cannot effectively utilize the large amount of hidden assembly process knowledge generated during product iteration. To solve this problem, this paper proposes an effective assembly process planning method based on knowledge graph. Fristly, the entire assembly is constructed from each layer of assembly components to reduce the computational complexity of assembly process planning. Then, assembly process planning starts from the last layer of assembly components. The entire assembly process planning is obtained by concatenating the results of each layer’s assembly process planning from bottom to top. Assembly operations are determined based on the features of the base component and neighboring assembly components. If there is no corresponding assembly operation for the assembly features, the assembly operation can be inferred through representation learning. After that, arrange and combine all possible assembly solutions to form a feasible solution space. Inferring the optimal assembly process sequence through rules. Combining the optimal assembly sequence with specific assembly operations to complete assembly process planning. Finally, Macro and micro adjustments are made to the assembly process planning of different products in the same series by comparing the differences between current product and classic product. In this paper, the upper door of elevator is used as an example to verify the proposed method. The results show that the proposed method can effectively realize assembly process planning.