基于多属性交叉聚类的实木定制柜门零件族划分

Furniture part family division of customized wooded cabinet doors using multi-attribute cross clustering

定制家具生产制造需平衡大批量和多品种之间的关系,合理的成组方式直接影响生产效率。实木零件分类成组受多种指标影响,针对其复杂性,选取莫干山整木工厂的40余款典型实木造型柜门进行试验研究。柜门拆解后零部件共458个,其中实木零部件314个。选取影响生产制造分类的4个主要属性——尺寸、工艺、榫槽结构和弯曲程度进行多属性交叉聚类,步骤包括:利用层次分析对属性指标进行重要性排序,各单属性分类(分类方法包括尺寸-层次聚类、工艺-模糊聚类、榫槽结构-连接方式、弯曲程度-弯曲半径和弧度半径),单属性分类交叉综合得零件族划分结果。层次分析法得出属性排序为:尺寸、工艺、榫槽结构、弯曲程度。单属性分类为尺寸4类、工艺5类、榫槽结构8类、弯曲程度4类,交叉综合后零件族共17类。结果表明:多属性交叉聚类方法适用于实木定制柜门的零件族划分,且此类方法具有良好的分组性能和可调整性,可广泛应用于实木家具的零件成组加工;4个属性的选取和各分类方法符合定制柜门家具企业的实际生产,零件族中的单个属性可选择,顺序可调换,可根据企业的需求进行灵活调整,满足不同工段和设备等的需求;属性选取数量和单属性分类数直接影响零件族的划分,属性数量越多,单属性分类数越多,零件族数也就越多越细致,故在组织成组加工生产制造时需根据加工节拍合理控制零件族数。

Customized furniture production and manufacturing need to balance the relationship between large quantities and many varieties,and the reasonable group mode directly affects the production efficiency.The classification of solid wood parts into groups is affected by the variety of indicators.By considering its complexity,more than 40 typical solid wood shaped cabinet doors in the Moganshan Whole Wood Factory are selected for this experimental research.After dismantling the cabinet doors,there are 458 parts in total,among which 314 are solid wood parts.Four main attributes affecting manufacturing classification,i.e.,size,process,tenon and groove structure and bending degree,were selected for the calculation of multi-attribute cross clustering.The steps included:the hierarchical analysis was used to rank the importance of attribute indexes,and each single attribute was classified.The classification methods included size-hierarchical clustering,process-fuzzy clustering,tenon and groove structure-connection mode,bending degree-bending radius and radian radius.The result of part family division was obtained by the cross synthesis of single attribute classification.The analytic hierarchy process showed that the attributes were sorted as follows:size,process,tenon and groove structure and bending degree.The single attribute classification included four kinds of dimension,five kinds of process,eight kinds of tenon and groove structure,four kinds of bending degree,and a total of 17 part families after the cross synthesis.The results showed that the multi-attribute cross clustering method is suitable for the part family division of the solid wood customized cabinet doors,and this method has good grouping performance and adjustment,and can be widely used in the solid wood furniture parts group processing.By selection of four attributes and the classification method in accordance with customized cupboard door furniture enterprises in the actual production,a single attribute of part family can be chosen and the order can be flexibly exchanged according to the needs of the enterprises to meet the requirements of different sections and equipment.The number of selected attributes and the number of single attribute classes directly affect the division of part families.The more attributes,the more single attribute classes,the more and more detailed the number of part families.Therefore,the number of part families should be controlled reasonably according to the processing beat when organizing into groups for processing and manufacturing.