In the aviation field, the reuse rate of process design knowledge for parts with long cycles and complex processes was low, and there was a lack of effective knowledge recommendation and inheritance. The current methods for process knowledge generalization mainly focused on using fuzzy knowledge and reasoning algorithms to overcome the dependence of process methods on the environment. These methods did not significantly help in addressing the incompleteness and inconsistency of formalized expressions. This paper was based on the standardized expression of process data, the knowledge graph construction and heuristic search for process recommendations were carried out, resulting in a higher knowledge reuse rate, which provided a way of thinking for the high-reusability, high-efficiency, and knowledge-driven process planning of parts with long cycles and complex processes.
因为所有工序流程都是由“供料模块”始、到“入库模块”终,所以路径的起点和终点确定,故任务转换成了条件最短路问题:即在有向图中寻找一条起点和终点确定的最短路,经过特定的若干必经点。考虑到每次必经点都会变化,但是有向图不会变,所以选择用启发式搜索中的A-star算法来解决,首先采用广度优先搜索(breadth first search,BFS)算法或其他单源最短路算法计算出每个点到起点的距离和到终点的距离,然后将经过必经点个数加入A-star的估值函数计算,从而推出最短路径。A-star搜索算法是一种在人工智能领域广泛使用的搜索方法,该算法对评价函数进行了优化,利用启发式函数对即将遍历的状态节点进行评估,然后选择函数值较小的状态节点进行扩展,直到找到目标节点或遍历完所有状态节点。A-star算法的核心是第x个节点的代价函数,其计算式如下:
LIUYueyan. Research on Feature-oriented Decision-making Method for Machining Spacecraft Section Parts [D]. Harbin: Harbin Institute of Technology, 2023.
[5]
李征. 基于知识图谱的工艺路线推荐方法研究[D]. 北京:北京交通大学,2021.
[6]
LIZheng. Research on Process Route Recommendation Method Based on Knowledge Graph[D]. Beijing: Beijing Jiaotong University, 2021.
PENGShixin, XIAOBiao, ZHAOZhengcai, et al. Construction and Application of Knowledge Graph for Processing Technology of Complex Thin-walled Aerospace Parts[J]. Journal of Mechanical & Electrical Engineering, 2024, 41(4): 709-719.
[9]
杨圣斌. 炼油工艺设计集成系统标准化研究[J]. 山东化工,2024,53(9):252-255.
[10]
YANGShengbin. Research on Standardization of Refining Process Design Integration System[J]. Shandong Chemical Industry, 2024, 53(9): 252- 255.
ZHANGChaochao, YELei, CAOZhiyong, et al. Development of WEB-based CAPP System for Aerospace Ring Forgings[J]. Journal of Netshape Forming Engineering, 2023, 15(5): 164-175.
WEIHaikun, XUSixin, SONGWenzhong. Generalization Theory and Generalization Methods For Neural Networks[J]. Acta Automatica Sinica,2001,27(6):806-815.
[15]
ZHAOLong, PENGXi, CHENYuxiao, et al. Knowledge as Priors: Cross-modal Knowledge Generalization for Datasets Without Superior Knowledge[C]∥IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR 2020).Seattle, 2020:6527-6536.
[16]
MAChenyu, JIAJinfang, HUANGJianqiang, et al. DisRot: Boosting the Generalization Capability of Few-shot Learning via Knowledge Distillation and Self-supervised Learning[J]. Machine Vision and Applications,2024,35(3):1-13.
SHANGYaqin, TENGWei, CAOYang, et al. Research on Design of Process Recommendation System Based on LLM[J]. Papermaking Equipment and Materials, 2024, 53(5): 28-30.
[19]
伍缘杰. 激光选区熔化工艺数据库与工艺推荐研究[D]. 武汉:华中科技大学,2022.
[20]
WUYuanjie. Research on Laser Selective Melting Process Database and Process Recommendation[D]. Wuhan: Huazhong University of Science and Technology, 2022.
[21]
曹勇. 基于数据挖掘的工艺知识发现与重用研究[D]. 济南:山东大学,2019.
[22]
CAOYong. Research on Process Knowledge Discovery and Reuse Based on Data Mining[D]. Jinan: Shandong University, 2019.
JIGuiyang, WANGPeiyan, YUZhuo. Research on Knowledge Injection Method of Large Language Model for Process Specification Texts[J]. Journal of Frontiers of Computer Science and Technology, 2024, 18(9): 2361-2369.