摘要
由于直升机组件数量多且组件之间约束关系复杂,传统的装配序列规划方法存在指数爆炸和易陷入局部最优解的问题。为了应对装配序列求解时间随组件数量指数增长和多约束关系矩阵运算复杂的难点,引入知识图谱(KG)建立直观的语义装配信息模型,并提出了一种基于KG的直升机部件装配序列规划方法。首先,基于三维模型解析和知识推理技术实现结构信息和约束关系等关键装配信息的提取。其次,基于本体构建KG形式的装配信息模型。最后,利用带反馈的图规划算法从KG中求解直升机部件的装配序列。KG提供的优先关系减少了算法搜索空间,在优先关系的约束下,带反馈的图规划算法以装配方向改变次数最小和装配工具变化次数最少为目标,逐步规划序列并反馈规划结果以避免重复搜索。直升机的中机身中段部件被作为实验对象以验证该方法的有效性,对比启发式算法能获得更高适应度值的装配序列且求解时间更短。
Due to the large number of helicopter components and the complex constraint relationships between them,traditional assembly sequence planning methods encounter the problems of exponential explosion and local optimal solution.To address the difficulties posed by the exponential increase of running time with the number of components and complex calculations of multi-constraint relationship matrix,knowledge graphs(KG)were introduced to establish intuitive semantic assembly information models.A KG-based method for planning the assembly sequence of helicopter components was proposed.Firstly,key assembly information such as structure information and constraint relationships was extracted based on 3D model analysis and knowledge reasoning.Secondly,an assembly information model in the form of KG was constructed based on ontology.Finally,a graph planning algorithm with feedback was employed to determine the assembly sequence of helicopter components from the KG.The priority relationships provided in the KG reduced the search space of the algorithm.Under the constraint of these priority relationships,the graph planning algorithm with feedback aimed to minimize the number of assembly direction changes and the number of assembly tool changes.It gradually planned the sequence,and fed back the planning results to avoid repeated searches.The middle piece of the helicopter mid-fuselage was used as an experimental object to verify the effectiveness of the proposed method.The proposed method achieved higher fitness for the assembly sequence and shorter solution times compared to heuristic algorithms.
作者
蒋明杰
张炜才
荣浩明
张俊起
黄少华
JIANG Mingjie;ZHANG Weicai;RONG Haoming;ZHANG Junqi;HUANG Shaohua(College of Mechanical and Electrical Engineering,Nanjing University of Aeronautics and Astronautics,Nanjing Jiangsu 210016 China;AVIC Changhe Aircraft Industrial(Group)Co.,Ltd.,Jingdezhen Jiangxi 333002,China)
出处
《图学学报》
CSCD
北大核心
2024年第5期1062-1070,共9页
Journal of Graphics
关键词
知识图谱
直升机部件
装配序列规划
语义信息模型
图算法
三维模型解析
knowledge graph
helicopter component
assembly sequence planning
semantic information model
graph algorithm
3D model analysis
