基于深度信念网络的非标刀具设计知识挖掘与重用研究

doi:10.11996/JG.j.2095-302X.2026020411

摘要/Abstract

摘要：

非标刀具设计中刀具特征与零件加工特征之间具有强耦合的关联关系，是一种典型隐性设计知识，具有数据多模态性和多维不确定性，导致难以捕获和重用，因此提出了基于深度信念网络(DBN)的刀具设计知识挖掘与重用方法。首先，面向加工特征和刀具特征所具有的二维图像和属性文本2种模态数据，设计了一种双通道DBN实现了特征的提取与融合。其次，设计了一种面向关联关系挖掘的DBN，实现加工特征与刀具特征之间隐含关系的获取。最后，通过关联规则推理和改进Rake算法对已有刀具案例进行评价和实现重用。以非标专用内孔槽刀设计过程为例，通过重用结果与实际结果在结构信息和属性信息方面的对比，验证了方法的有效性。

关键词: 非标刀具, 深度信念网络, 隐性知识挖掘, 多模态融合, 关联关系

Abstract:

In the design of non-standard tools, a strongly coupled correlation between tool features and part-machining features is identified as a typical type of implicit design knowledge. It exhibits data multimodality and multi-dimensional uncertainty, leading to difficulties in capture and reuse. Therefore, a method for tool design knowledge mining and reuse based on a Deep Belief Network (DBN) was proposed. First, targeting the two-modal data of 2D images and attribute texts associated with machining features and tool features, a dual-channel DBN was designed to perform feature extraction and fusion. Second, a DBN oriented to correlation mining was designed to obtain implicit relationships between machining features and tool features. Finally, existing tool cases were evaluated and reused through association-rule reasoning and an improved Rake algorithm. Taking the design process of a non-standard special inner-hole-groove tool as an example, the effectiveness of the method was verified by comparing the reuse results with the actual results in terms of structural and attribute information.

Key words: non-standard tool, deep belief network, implicit knowledge mining, multimodal fusion, cross-modal correlation

中图分类号:

TG71
TP183

王明微, 赵建骅, 孙志宏, 睢鹏, 路晓君. 基于深度信念网络的非标刀具设计知识挖掘与重用研究[J]. 图学学报, 2026, 47(2): 411-422.

WANG Mingwei, ZHAO Jianhua, SUN Zhihong, SUI Peng, LU Xiaojun. A study on knowledge mining and reuse for non-standard tool design based on deep belief network[J]. Journal of Graphics, 2026, 47(2): 411-422.

图/表 17

图1 基于深度学习的非标刀具设计知识挖掘与重用过程

Fig. 1 Non-standard tool design knowledge mining and reuse process based on deep learning

图2 刀具特征提取与融合网络的结构

Fig. 2 The structure of tool feature extraction and fusion network

图3 加工特征提取与融合网络的结构

Fig. 3 The structure of processing feature extraction and fusion network

图4 面向刀具特征-加工特征关联关系挖掘的DBN模型

Fig. 4 DBN model for tool feature-machining feature association mining

图5 刀具设计知识重用过程

Fig. 5 Tool design knowledge reuse process

图6 部分刀具结构特征像素图

Fig. 6 Part of the tool structure feature pixel map

图7 部分刀具对应的属性信息

Fig. 7 Attribute information corresponding to some tools

图8 部分加工特征的像素图

Fig. 8 Pixel map of partial processing features

图9 部分加工特征对应的属性信息

Fig. 9 Attribute information corresponding to some processing features

图10 加工特征的提取与融合网络的重构误差((a) 结构特征提取网络；(b) 属性特征提取网络；(c) 特征融合网络)

Fig. 10 Extraction of machining features and reconstruction error of fusion network ((a) Structural feature extraction network; (b) Attribute feature extraction network; (c) Feature fusion network)

图11 刀具各局部特征融合网络最后一层RBM重构误差曲线((a) 刀具刀片连接部分特征融合网络; (b) 刀具刀片与刀尖部分特征融合网络; (c) 刀具刀头构造部分特征融合网络; (d) 刀具前刀面结构部分特征融合网络; (e) 刀具刀杆截面部分特征融合网络)

Fig. 11 Reconstruction error curves of the last layer RBM in local feature fusion networks of the cutting tool ((a) Feature fusion network of tool insert connection part; (b) Feature fusion network of tool insert and tip part; (c) Feature fusion network of tool head structure part; (d) Feature fusion network of tool rake face structure part; (e) Feature fusion network of tool shank cross-section part)

图12 关联关系挖掘网络重构误差曲线图

Fig. 12 Association relationship mining network reconstruction error curve diagram

图13 内孔槽加工特征

Fig. 13 Machining characteristics of inner hole groove

表1 各局部特征相似案例的相似度值

Table 1 The similarity value of each local feature similar case

案例名	案例编号	刀片连接	刀片与刀尖	刀头构造	前刀面结构	刀杆截面	刀具整体
x-T0855	1	0.743	0.664	0.765	─	0.679	2.084 2
x-T1012	2	0.717	─	─	─	0.701	1.059 1
x-ZD0700	3	─	─	─	0.726	─	0.538 3
x-T0857	4	0.682	─	0.692	─	0.668	1.517 2
x-VL8010	5	─	0.823	0.728	0.735	─	1.631 3
x-S2269	6	─	0.759	─	─	─	0.528 8
x-T0762	7	─	─	─	0.641	─	0.484 2

图14 刀片连接方式重用结果((a) 实际设计结果；(b) 推荐实例)

Fig. 14 Blade connection reuse results ((a) Actual design results; (b) Recommended examples)

图15 刀片与刀尖结构重用结果((a) 实际设计结果；(b) 推荐实例)

Fig. 15 The reuse results of blade and tool tip structure ((a) Actual design results; (b) Recommended examples)

表2 推荐案例的物理参数

Table 2 Physical parameters of recommended cases

实例	材料	加工精度	是否涂层	前角/后角/主偏角/刃倾角	刀片连接		刀片与刀尖
实例	材料	加工精度	是否涂层	前角/后角/主偏角/刃倾角	宽	刀片宽	刀尖半径	刀尖角度
0	42CrMo	IT8	是	7/7/20.5/3.0	18.0	9.50	1.75	90
1	42CrMo	IT8	是	7/7/45.0/3.0	17.1	7.98	1.50	90
2	42CrMo	IT7	是	─	12.0	─	─	─
3	42CrMo	IT8	否	─	─
4	42CrMo	IT8	是	10/10/30.0/3.5	20.8
5	42CrMo	IT7	是	7/7/20.0/3.0	─	9.00	1.75	90
6	42CrMo	IT8	是	─	─	9.60	0.50	90
7	42CrMo	IT8	否	─	─	─	─	─

参考文献 22

[1]	丁明娜, 刘献礼, 岳彩旭, 等. 面向智能制造过程的刀具设计、制备与管控技术[J]. 机械工程学报, 2023, 59(19): 429-459. DOI
	DING M N, LIU X L, YUE C X, et al. Design, manufacturing, management and control technology of cutting tools for intelligent manufacturing process[J]. Journal of Mechanical Engineering, 2023, 59(19): 429-459 (in Chinese). DOI
[2]	张效栋, 房丰洲, 程颖. 自由曲面超精密车削加工路径优化设计[J]. 天津大学学报, 2009, 42(3): 278-282.
	ZHANG X D, FANG F Z, CHENG Y. Optimized design of cutting path for freeform surface in ultra-precision turning[J]. Journal of Tianjin University, 2009, 42(3): 278-282 (in Chinese).
[3]	FENG X M, MENG C, LIAN W, et al. Optimization analysis of microtexture size parameters of tool surface for turning GH4169[J]. Ferroelectrics, 2023, 607(1): 186-207. DOI URL
[4]	丁伟, 李春旭, 刘柄宏. 螺旋面方程建立及铣削时刀具刃型设计[J]. 沈阳工业大学学报, 2020, 42(2): 185-190. DOI
	DING W, LI C X, LIU B H. Establishment of helicoid equation and design of blade profile for cutter milling[J]. Journal of Shenyang University of Technology, 2020, 42(2): 185-190 (in Chinese). DOI
[5]	李明星, 岳彩旭, 刘献礼, 等. 钛合金框架类零件铣削用立铣刀设计与应用研究进展[J]. 机械工程学报, 2023, 59(23): 283-309. DOI
	LI M X, YUE C X, LIU X L, et al. Research on design and application of Endmill for milling titanium alloy frame parts[J]. Journal of Mechanical Engineering, 2023, 59(23): 283-309 (in Chinese). DOI
[6]	JIANG T Y, ZHOU J T, ZHAO J H, et al. A multi-dimensional cognitive framework for cognitive manufacturing based on OAR model[J]. Journal of Manufacturing Systems, 2022, 65: 469-485. DOI URL
[7]	JIANG T Y, ZHOU J T, CAO Y, et al. A connecting rod assembly deformation cognition method based on quality characteristics probability network[J]. Advanced Engineering Informatics, 2024, 62: 102580. DOI URL
[8]	LI E M, ZHOU J T, YANG C S, et al. Part machining deformation prediction based on spatial-temporal correlation learning of geometry and cutting loads[J]. Journal of Manufacturing Processes, 2023, 92: 397-411. DOI URL
[9]	ZHOU J T, LI X Q, WANG M W, et al. Thinking process rules extraction for manufacturing process design[J]. Advances in Manufacturing, 2017, 5(4): 321-334. DOI URL
[10]	ZHOU J T, ZHAO H, WANG M W, et al. Tool selection method based on transfer learning for CNC machines[J]. Mechanical Sciences, 2018, 9(1): 123-146. DOI URL
[11]	WU X Y, FENG Y X, LOU S H, et al. A cognitive computing approach for product configuration design based on associative rule mining and deep reinforcement learning[EB/OL]. [2025-07-20].https://doi.org/10.1080/00207543.2024.2428426.
[12]	WAN Y W, LIU Y, CHEN Z Y, et al. Making knowledge graphs work for smart manufacturing: Research topics, applications and prospects[J]. Journal of Manufacturing Systems, 2024, 76: 103-132. DOI URL
[13]	冯毅雄, 钟锐锐, 张志峰, 等. 设计知识驱动的不规则多边形排样算法及应用[J]. 计算机集成制造系统, 2023, 29(2): 593-603.
	FENG Y X, ZHONG R R, ZHANG Z F, et al. Irregular polygons nesting algorithm driven by design knowledge and application[J]. Computer Integrated Manufacturing Systems, 2023, 29(2): 593-603 (in Chinese).
[14]	柯庆镝, 李自生, 王玲, 等. 设计信息特征驱动下绿色设计知识挖掘方法[J]. 机械工程学报, 2025, 61(9): 424-436.
	KE Q D, LI Z S, WANG L, et al. A knowledge mining method for green design knowledge driven by design information characteristics[J]. Journal of Mechanical Engineering, 2025, 61(9): 424-436 (in Chinese). DOI URL
[15]	张一鸣, 刘金锋, 陈亚杰, 等. 零件加工隐性工艺知识获取方法研究[J]. 图学学报, 2024, 45(2): 399-408. DOI
	ZHANG Y M, LIU J F, CHEN Y J, et al. Tacit process knowledge acquisition methods for the parts machining[J]. Journal of Graphics, 2024, 45(2): 399-408 (in Chinese). DOI
[16]	FAN M C, BI C X, LIU X L, et al. Study on parametric design and cutting performance of double-arc milling cutter[J]. The International Journal of Advanced Manufacturing Technology, 2024, 134(9): 4907-4921. DOI
[17]	SUN Y, SONG Q F, LIANG F M. Consistent sparse deep learning: theory and computation[J]. Journal of the American Statistical Association, 2022, 117(540): 1981-1995. DOI PMID
[18]	KALE A P, WAHUL R M, PATANGE A D, et al. Development of deep belief network for tool faults recognition[J]. Sensors, 2023, 23(4): 1872. DOI URL
[19]	何俊, 张彩庆, 李小珍, 等. 面向深度学习的多模态融合技术研究综述[J]. 计算机工程, 2020, 46(5): 1-11. DOI
	HE J, ZHANG C Q, LI X Z, et al. Survey of research on multimodal fusion technology for deep learning[J]. Computer Engineering, 2020, 46(5): 1-11 (in Chinese). DOI
[20]	ARCHANA R, JEEVARAJ P S E. Deep learning models for digital image processing: a review[J]. Artificial Intelligence Review, 2024, 57(1): 11. DOI
[21]	ZHAO F, ZHANG C, GENG B. Deep multimodal data fusion[J]. ACM Computing Surveys, 2024, 56(9): 1-36.
[22]	DONG Y N, LIU Q W, DU B, et al. Weighted feature fusion of convolutional neural network and graph attention network for hyperspectral image classification[J]. IEEE Transactions on Image Processing, 2022, 31: 1559-1572. DOI URL