基于数字孪生和Transformer-LSTM-XGBoost的液压机器人故障诊断方法

李亚洁; 吴瑞龙; 李炜

doi:10.13700/j.bh.1001-5965.2025.0815

基于数字孪生和Transformer-LSTM-XGBoost的液压机器人故障诊断方法

doi: 10.13700/j.bh.1001-5965.2025.0815

李亚洁^{1, 2},
吴瑞龙^{1, 2},
李炜^{1, 2, ,}

1.
兰州理工大学自动化与电气工程学院，兰州 730050
2.
兰州理工大学有色冶金智能机器人创新中心，兰州 730050

基金项目:

国家自然科学基金(62163022)；甘肃省产业支撑计划项目(2024CYZC-18)；甘肃省科技重大专项计划(23ZDGA007)；甘肃省联合科研基金(24JRRA827)；兰州市科技计划项目(2024-2-5)

详细信息

通讯作者:
E-mail：liwei@lut.edu.cn

中图分类号: TH165⁺.3；TP181；V245.1
计量
- 文章访问数: 303
- HTML全文浏览量: 109
- PDF下载量: 13
- 被引次数: 0
出版历程
- 收稿日期: 2025-11-25
- 录用日期: 2026-02-05
- 网络出版日期: 2026-03-03
- 整期出版日期: 2026-06-30

Fault diagnosis method for hydraulic robots based on digital twin and Transformer-LSTM-XGBoost

LI Yajie^{1, 2},
WU Ruilong^{1, 2},
LI Wei^{1, 2
, ,}

1.
School of Automation and Electrical Engineering，Lanzhou University of Technology，Lanzhou 730050，China
2.
Innovation Center of Intelligent Robotics for Nonferrous Metallurgy，Lanzhou University of Technology，Lanzhou 730050

Funds:

National Natural Science Foundation of China (62163022); Gansu Provincial Industry Support Program (2024CYZC-18); Gansu Provincial Major Science and Technology Project (23ZDGA007);Gansu Provincial Joint Research Fund (24JRRA827);Lanzhou Science and Technology Plan Project (2024-2-5)

More Information

Corresponding author: E-mail：liwei@lut.edu.cn

摘要

摘要:
随着液压机器人的应用范围日益广泛，其健康诊断与维护需求也愈发迫切。通过将数字孪生与深度学习技术相结合，提出一种基于Transformer、长短期记忆网络(LSTM)与极端梯度提升(XGBoost)算法相融合的液压机器人智能故障诊断方法。提出基于数字孪生技术的液压机器人故障诊断架构与方案；建立涵盖三维模型与属性模型且具备虚实同步能力的液压机器人数字孪生体，并研究了提升孪生模型精度的校准方法。针对常见的液压系统泄漏、阀门卡滞、阻尼孔堵塞、过滤器堵塞4种典型故障，在所建数字孪生体上进行故障机理分析与演化过程模拟，构建覆盖正常工况及多类故障工况的高质量数据集，用于支撑数据驱动建模与智能诊断。提出融合Transformer与LSTM的深度特征提取模型，以捕获多维特征间的全局依赖关系与时序动态特征，并引入XGBoost分类器实现多故障模式识别。实验验证结果表明：所提方法在多种液压系统故障诊断中可获得96.6%的诊断准确率，在不同噪声扰动条件下仍保持稳定的诊断性能，所提方法具有良好的鲁棒性与泛化能力，可为液压机器人故障诊断与维护提供有效的技术支撑。
- 数字孪生 /
- 液压机器人 /
- 故障诊断 /
- 深度学习 /
- 机器学习
Abstract:
Hydraulic robots are increasingly used in industrial systems, which raises the need for reliable health diagnosis and maintenance under complex operating conditions. An intelligent fault diagnosis method was developed by integrating digital twin technology with deep learning and combining a Transformer, a long short-term memory networks (LSTM), and extreme gradient boosting (XGBoost). A diagnosis architecture based on digital twins was built, and cyber-physical synchronization was made possible by a digital twin with an attribute model and a three-dimensional model. Calibration improved twin accuracy and virtual-physical consistency. Fault mechanism analysis and fault evolution simulation were conducted for four typical hydraulic system faults: leakage, valve sticking, damping orifice blockage, and filter blockage. The simulations generated a dataset covering normal and fault conditions for data-driven modeling and diagnosis. A Transformer-LSTM feature extractor captured global dependencies and temporal dynamics in multi-dimensional time-series data, and XGBoost performed multi-fault classification. The experimental verification results show that the proposed method demonstrate steady performance under various noise disturbances and 96.6% diagnostic accuracy across many problems, suggesting high resilience and generalization and supporting hydraulic robot fault detection and maintenance in industrial applications.
- digital twin /
- hydraulic robot /
- fault diagnosis /
- deep learning /
- machine learning

HTML全文

图 1 液压机器人数字孪生五维架构

Figure 1. Five-dimensional digital twin architecture of hydraulic robot

下载: 全尺寸图片幻灯片

图 2 液压机器人故障诊断方案

Figure 2. Fault diagnosis scheme for hydraulic robot

下载: 全尺寸图片幻灯片

图 3 液压机器人实体图

Figure 3. Physical hydraulic robot

下载: 全尺寸图片幻灯片

图 4 液压机器人建模流程

Figure 4. Hydraulic robot modeling process

下载: 全尺寸图片幻灯片

图 5 液压机器人SolidWorks模型

Figure 5. SolidWorks model of hydraulic robot

下载: 全尺寸图片幻灯片

图 6 3ds Max模型轻量化处理

Figure 6. Lightweight processing of 3ds Max model

下载: 全尺寸图片幻灯片

图 7 CoppeliaSim机器人树状图

Figure 7. CoppeliaSim robot tree diagram

下载: 全尺寸图片幻灯片

图 8 机器人的连杆坐标系

Figure 8. Link coordinate system of robot

下载: 全尺寸图片幻灯片

图 9 液压驱动系统的模型

1.液压油箱；2.油箱温度计；3.风冷式冷却器；4.油泵；5.过滤器；6，8，18.溢流阀；7.流量计；9，17.换向阀；10.压力表；11，15.摆动缸；12，13，14，16.直线缸；19.止回阀。

Figure 9. Model of hydraulic drive system

下载: 全尺寸图片幻灯片

图 10 数字孪生模型校准流程

Figure 10. Digital twin model calibration process

下载: 全尺寸图片幻灯片

图 11 机器人故障形成分析

Figure 11. Analysis of robot fault formation

下载: 全尺寸图片幻灯片

图 12 Transformer-LSTM-XGBoost的液压机器人故障诊断方法框架

Figure 12. Transformer-LSTM-XGBoost-based framework for hydraulic robot fault diagnosis method

下载: 全尺寸图片幻灯片

图 13 不同故障类型下压力信号的变化趋势

Figure 13. Pressure signal variations under different fault types

下载: 全尺寸图片幻灯片

图 14 关节角度对比

Figure 14. Comparison of joint angles

下载: 全尺寸图片幻灯片

图 15 关节角度误差对比

Figure 15. Comparison of joint angle errors

下载: 全尺寸图片幻灯片

图 16 补偿后的关节误差对比

Figure 16. Comparison of joint errors after compensation

下载: 全尺寸图片幻灯片

图 17 性能指标对比

Figure 17. Performance metrics comparison

下载: 全尺寸图片幻灯片

图 18 不同模型的混淆矩阵

Figure 18. Confusion matrices of different models

下载: 全尺寸图片幻灯片

表 1 液压机器人的基本参数

Table 1. Basic parameters of hydraulic robot

参数	数值
质量/kg	40
水平伸展范围/mm	1330
向上伸展范围/mm	1640
向下伸展范围/mm	293
抓举质量/kg	30
肩部转向范围/(°)	180
肩部俯仰范围/(°)	110
肘部俯仰范围/(°)	100
腕部俯仰范围/(°)	100
腕部回转范围/(°)	310
钳口张开范围/mm	100

下载: 导出CSV

表 2 液压机器人改进D-H参数表

Table 2. Modified D-H parameters of hydraulic robot

连杆$ i $	$ {\alpha }_{i-1} $/(°)	$ {{a}}_{i-1} $/mm	$ {d}_{i} $/mm	$ {\theta }_{i} $/(°)
1	0	0	286.923	$ {\theta }_{1} $
2	−90	0	0	$ {\theta }_{2} $
3	0	646.901	0	$ {\theta }_{3} $
4	0	244.866	0	$ {\theta }_{4} $
5	−90	0	461.049	$ {\theta }_{5} $
注：$ {\alpha }_{i-1} $表示绕着$ {x}_{i-1} $轴，从$ {{\textit{z}}}_{i-1} $旋转到$ {{\textit{z}}}_{i} $的角度。

下载: 导出CSV

表 3 数字孪生模型精度评价结果

Table 3. Accuracy evaluation results of digital twin model

MAE/(°)	RMSE/(°)	BIAS/(°)	R²/%
0.26	0.32	−0.16	98.3

下载: 导出CSV

表 4 补偿后的数字孪生模型精度评价结果

Table 4. Accuracy evaluation results of compensated digital twin model

MAE/(°)	RMSE/(°)	BIAS/(°)	R²/%
0.059	0.127	0.07	99.9

下载: 导出CSV

表 5 样本分布

Table 5. Sample distribution

故障类别	类别标签	样本数/条
滤油器堵塞	1	392
阻尼孔堵塞	2	392
阀门卡滞	3	392
系统泄漏	4	392
正常工况	5	779

下载: 导出CSV

表 6 多模型性能指标对比

Table 6. Multi-Model performance comparison chart

模型	准确率/%	精确率/%	召回率/%	F₁值/%
SVM^[34]	52.8	60.8	49.6	45.3
CNN^[30]	81.7	83.6	78.6	80.0
BiLSTM^[31]	68.0	77.7	63.4	66.2
GRU^[35]	62.1	70.6	56.7	58.1
Transformer^[32]	60.9	75.6	56.4	56.9
LSTM^[31]	66.4	76.3	62.1	62.9
XGBoost^[33]	94.4	93.3	94.4	93.6
Transformer-LSTM	72.3	78.4	69.2	68.1
LSTM-XGBoost	96.6	95.5	96.3	96.1
Transformer-XGBoost	95.3	95.3	95.1	95.1
Transformer-LSTM-RF^[36]	96.2	94.8	96.2	95.1
Transformer-LSTM-XGBoost	96.6	96.0	96.3	96.1

下载: 导出CSV

表 7 不同噪声强度下的模型性能对比

Table 7. Performance comparison under different noise intensities

噪声强度	准确率/%	精确率/%	召回率/%	F₁值/%
无额外噪声	96.6	96.0	96.3	96.1
低强度噪声	95.5	95.1	95.3	95.2
中强度噪声	93.1	93.7	94.0	93.8
高强度噪声	89.8	91.0	89.5	88.8

下载: 导出CSV

参考文献(36)

[1]	DI LALLO A, YU S Y, SLIGHTAM J E, et al. Untethered fluidic engine for high-force soft wearable robots[J]. Advanced Intelligent Systems, 2024, 6(11): 2400171.
[2]	朱明君, 张鹏飞, 黄德民, 等. 液压系统的故障诊断与健康管理研究综述[J]. 润滑油, 2023, 38(3): 7-12. ZHU M J, ZHANG P F, HUANG D M, et al. A research review of fault diagnosis and health management of hydraulic systems[J]. Lubricating Oil, 2023, 38(3): 7-12(in Chinese).
[3]	陈斌, 王占林, 裘丽华. 机载液压系统的主要发展趋势[J]. 航空学报, 1998, 19(1): 2-7. CHEN B, WANG Z L, QIU L H. Main developmental trend of aircraft hydraulic systems[J]. Acta Aeronautica et Astronautica Sinica, 1998, 19(1): 2-7(in Chinese).
[4]	REN X, GUO Q, LI T S, et al. Fault detection and tolerant control with a variable-bandwidth extended state observer for electro-hydraulic servo systems with lumped disturbance and measurement noise[J]. ISA Transactions, 2025, 162: 287-300.
[5]	DAI J Y, TANG J, HUANG S Z, et al. Signal-based intelligent hydraulic fault diagnosis methods: review and prospects[J]. Chinese Journal of Mechanical Engineering, 2019, 32: 75.
[6]	PENG Z K, CHU F L. Application of the wavelet transform in machine condition monitoring and fault diagnostics: a review with bibliography[J]. Mechanical Systems and Signal Processing, 2004, 18(2): 199-221.
[7]	周汝胜, 焦宗夏, 王少萍. 液压系统故障诊断技术的研究现状与发展趋势[J]. 机械工程学报, 2006, 42(9): 6-14. ZHOU R S, JIAO Z X, WANG S P. Current research and developing trends on fault diagnosis of hydraulic systems[J]. Chinese Journal of Mechanical Engineering, 2006, 42(9): 6-14(in Chinese).
[8]	夏婷, 曹戈, 朱加升, 等. 工程机械故障诊断技术研究综述及展望[J]. 工程机械与维修, 2024(4): 6-9. XIA T, CAO G, ZHU J S, et al. Review and prospect of fault diagnosis technology for construction machinery[J]. Construction Machinery & Maintenance, 2024(4): 6-9(in Chinese).
[9]	PENG P, WANG J G. NOSCNN: a robust method for fault diagnosis of RV reducer[J]. Measurement, 2019, 138: 652-658.
[10]	姜少飞, 邬天骥, 彭翔, 等. 基于XGBoost特征提取的数据驱动故障诊断方法[J]. 中国机械工程, 2020, 31(10): 1232-1239. JIANG S F, WU T J, PENG X, et al. Data driven fault diagnosis method based on XGBoost feature extraction[J]. China Mechanical Engineering, 2020, 31(10): 1232-1239(in Chinese).
[11]	郑彪, 高丙朋, 程静. 基于DNN-CapsNet的液压泵故障程度诊断方法[J]. 液压与气动, 2023, 47(8): 41-49. ZHENG B, GAO B P, CHENG J. Fault degree diagnosis of hydraulic pump based on DNN-CapsNet[J]. Chinese Hydraulics & Pneumatics, 2023, 47(8): 41-49(in Chinese).
[12]	SUN J C, DING H, LI N, et al. Intelligent fault diagnosis of hydraulic system based on multiscale one-dimensional convolutional neural networks with multiattention mechanism[J]. Sensors, 2024, 24(22): 7267.
[13]	LIU Y J, LI W H, LIN S Y, et al. Hydraulic system fault diagnosis of the chain jacks based on multi-source data fusion[J]. Measurement, 2023, 217: 113116.
[14]	袁亮. 基于数字孪生的机械臂抓取作业仿真与监控系统[D]. 上海: 东华大学, 2023: 5-15. YUAN L. Simulation and monitoring system for mechanical arm grasping operations based on digital twins[D]. Shanghai: Donghua University, 2023: 5-15(in Chinese).
[15]	TAO F, ANWER N, LIU A, et al. Digital twin towards smart manufacturing and industry 4.0[J]. Journal of Manufacturing Systems, 2021, 58: 1-2.
[16]	蔡宝, 康亮, 顾鸿良. 工业机器人数字孪生系统研究[J]. 自动化与仪表, 2025, 40(8): 62-66. CAI B, KANG L, GU H L. Research on digital twin of industrial robots[J]. Automation & Instrumentation, 2025, 40(8): 62-66(in Chinese).
[17]	丁珈, 周炜, 张永, 等. 数字孪生驱动的装备全寿命周期智能运维技术进展与展望[J]. 系统工程理论与实践, 2025, 45(6): 1828-1845. DING J, ZHOU W, ZHANG Y, et al. Digital twin-driven full lifecycle intelligent equipment maintenance: progress and prospects[J]. Systems Engineering—Theory & Practice, 2025, 45(6): 1828-1845(in Chinese).
[18]	LIU X R, DAVID I. AI simulation by digital twins: systematic survey, reference framework, and mapping to a standardized architecture[J]. Software and Systems Modeling, 2025: 1-26.
[19]	李亚洁, 刘强, 李炜. 基于离心泵数字孪生流场云图的叶轮故障诊断方法与应用[J]. 北京航空航天大学学报, 2025, 51(1): 193-201. LI Y J, LIU Q, LI W. Impeller fault diagnosis method and application based on digital twin flow field contour of centrifugal pump[J]. Journal of Beijing University of Aeronautics and Astronautics, 2025, 51(1): 193-201(in Chinese).
[20]	李志祥, 唐宏宾, 王永胜, 等. 基于数字孪生的液压机械臂运动仿真与状态监控系统设计[J]. 计算机集成制造系统, 2025, 31(12): 4382-4391. LI Z X, TANG H B, WANG Y S, et al. Design of motion simulation and condition monitoring system for hydraulic manipulator based on digital twin[J]. Computer Integrated Manufacturing Systems, 2025, 31(12): 4382-4391(in Chinese).
[21]	简正豪, 陈卓贤, 刘云鸿, 等. 基于数字孪生的液压回路虚拟实验系统[J]. 实验技术与管理, 2024, 41(9): 111-117. JIAN Z H, CHEN Z X, LIU Y H, et al. Digital twin-based virtual reality experiment system for a hydraulic circuit[J]. Experimental Technology and Management, 2024, 41(9): 111-117(in Chinese).
[22]	赵大伟, 张太华, 卢尧, 等. 基于数字孪生体的液压系统知识模型构建及应用[J]. 现代制造工程, 2022(6): 17-24. ZHAO D W, ZHANG T H, LU Y, et al. Construction and application of hydraulic system knowledge model based on digital twin[J]. Modern Manufacturing Engineering, 2022(6): 17-24(in Chinese).
[23]	SUN J S, YANG Y S, SHI L, et al. Research on fault diagnosis method of the hydraulic system based on digital twin[J]. Journal of Physics: Conference Series, 2024, 2770(1): 012022.
[24]	WANG L T, LIU Y C, YIN H, et al. Fault diagnosis and predictive maintenance for hydraulic system based on digital twin model[J]. AIP Advances, 2022, 12(6): 065213.
[25]	LIU H T, ZHANG J C, CHENG Y, et al. Digital twin shop-floor: a complex system-oriented construction method and operation mechanism[J]. Procedia CIRP, 2023, 119: 52-57.
[26]	ZHOU D X, XIE M Z, XUAN P C, et al. A teaching method for the theory and application of robot kinematics based on MATLAB and V-REP[J]. Computer Applications in Engineering Education, 2020, 28(2): 239-253.
[27]	宋增来, 陈振华. N自由度机器人仿真的实现[J]. 计算机工程与设计, 2006, 27(24): 4813-4816. SONG Z L, CHEN Z H. Simulation of N-dof robot[J]. Computer Engineering and Design, 2006, 27(24): 4813-4816(in Chinese).
[28]	XU B, SHEN J, LIU S H, et al. Research and development of electro-hydraulic control valves oriented to industry 4.0: a review[J]. Chinese Journal of Mechanical Engineering, 2020, 33: 29.
[29]	赵洪利, 杨佳强. 基于融合卷积Transformer的航空发动机故障诊断[J]. 北京航空航天大学学报, 2025, 51(4): 1117-1126. ZHAO H L, YANG J Q. Aero-engine fault diagnosis based on fusion convolutional Transformer[J]. Journal of Beijing University of Aeronautics and Astronautics, 2025, 51(4): 1117-1126(in Chinese).
[30]	ZHAO J F, MAO X, CHEN L J. Speech emotion recognition using deep 1D & 2D CNN LSTM networks[J]. Biomedical Signal Processing and Control, 2019, 47: 312-323.
[31]	HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735-1780.
[32]	HUSSAN U, WANG H Z, PENG J C, et al. Transformer-based renewable energy forecasting: a comprehensive review[J]. Renewable and Sustainable Energy Reviews, 2026, 226: 116356.
[33]	LEI Y F, JIANG W L, JIANG A Q, et al. Fault diagnosis method for hydraulic directional valves integrating PCA and XGBoost[J]. Processes, 2019, 7(9): 589.
[34]	肖健华, 樊可清, 吴今培, 等. 应用于故障诊断的SVM理论研究[J]. 振动、测试与诊断, 2001, 21(4): 258-262. XIAO J H, FAN K Q, WU J P, et al. A study on SVM for fault diagnosis[J]. Journal of Vibration, Measurement & Diagnosis, 2001, 21(4): 258-262(in Chinese).
[35]	ENCALADA-DÁVILA Á, MOYÓN L, TUTIVÉN C, et al. Early fault detection in the main bearing of wind turbines based on gated recurrent unit (GRU) neural networks and SCADA data[J]. IEEE/ASME Transactions on Mechatronics, 2022, 27(6): 5583-5593.
[36]	MU H S, ZHAI X D, YIN D B, et al. A method of remaining useful life prediction of multi-source signals aero-engine based on RF-Transformer-LSTM[C]//Proceedings of the IEEE International Conference on Systems, Man, and Cybernetics. Piscataway: IEEE Press, 2022: 2502-2507.