基于多传感器数据关联的综合传动装置服役状态辨识方法

doi:10.12382/bgxb.2022.0664

摘要/Abstract

摘要：

针对综合传动装置在恶劣环境下传感器数据跳变、可靠性差,导致依赖单一传感器信息带来的状态判断虚警率高、服役状态难以准确辨识的问题,提出一种基于多传感器数据关联的综合传动装置服役状态辨识方法。通过时间窗均值关联网络,充分考虑关联传感器数据在各时间段的关联程度,可以有效表征复杂工况下传感器数据的关联关系。在时间窗关联度计算方法的基础上进一步构造了误差反向传播(Back Propagation,BP)数据映射模型,完成对关键传感器数据的映射。采用变分模态分解和样本熵(Variational Mode Decomposition-Sample Entropy,VMD-SE)方法对数据进行预处理;利用所提方法计算各传感器数据间的关联性,选取出相关性高的数据;将相关性高的数据输入构造出的BP数据映射模型进行映射。油压数据的案例验证结果表明,时间窗关联度计算方法能准确地衡量传感器数据间的关联性,BP数据映射模型输出的数据能够良好地表征关键传感器数据,二者结合能够有效提升服役状态判断的准确性。

关键词: 综合传动, 状态辨识, 数据关联, 映射模型

Abstract:

The sensor data jumps and the poor reliability of comprehensive transmission in harsh environment lead to a high false alarm rate in the state judgment brought by relying on a single sensor information, and make it difficult to accurately identify the service state. A service status identification method of comprehensive transmission based on multi-sensor data association is proposed. The proposed method fully considers the correlation degree of associated sensor data in each time period through the time window mean correlation network, and can effectively characterize the correlation relationship of sensor data under complex working conditions. Based on the calculation method of time window correlation degree, a back propagation (BP) data mapping model is constructed to complete the mapping of key sensor data. First, the variational mode decomposition-sample entropy (VMD-SE) method is used to preprocess the data; the correlation between the sensor data is calculated by the proposed method, and the data with high correlation is selected; finally, the data with high correlation is input into the constructed BP data mapping models to map the key sensor data. The case verification of oil pressure data is carried out. The results show that the time window correlation calculation method can accurately measure the correlation between sensor data, and the data output by the BP data mapping model can well characterize the key sensor data. The combination of the two can effectively improve the accuracy of service status judgment.

Key words: comprehensive transmission, status identification, data association, mapping model

中图分类号:

U463.22

徐保荣, 张金乐, 万丽, 吴昊阳, 王立勇. 基于多传感器数据关联的综合传动装置服役状态辨识方法[J]. 兵工学报, 2024, 45(2): 574-583.

XU Baorong, ZHANG Jinle, WAN Li, WU Haoyang, WANG Liyong. A Service Status Identification Method of Comprehensive Transmission Based on Multi-sensor Data Association[J]. Acta Armamentarii, 2024, 45(2): 574-583.

图/表 11

图1 关联数据映射模型

Fig.1 Linked data mapping model

图2 映射方法原理

Fig.2 Principle of mapping method

图3 多传感器数据映射的流程

Fig.3 Flowchart of multi-sensor data mapping

图4 综合传动装置测点示意图

Fig.4 Measuring points of comprehensive transmission

图5 综合传动装置部分测试信号

Fig.5 Test signal of comprehensive transmission

图6 样本数据时间窗均值关联网络

Fig.6 Sample data time window mean correlation network

表1 不同关联度传感器映射结果

Table 1 Mapping results of sensors with different correlation degrees

关联度	传感器数量	无干扰MSE	有干扰MSE
≥0.8	2	0.0090	0.0442
≥0.6	8	0.0065	0.0440
≥0.4	10	0.0069	0.0448
≥0.2	14	0.0124	0.0687

图7 操纵油压关联数据模型测试集输出结果

Fig.7 Manipulate the output results of oil pressure linked data model test set

表2 不同关联度传感器映射结果

Table 2 Mapping results of sensors with different correlation degrees

关联度	传感器数量	无干扰MSE	有干扰MSE
≥0.8	3	0.0042	0.0046
≥0.6	8	0.0034	0.0049
≥0.4	12	0.0039	0.0053
≥0.2	15	0.0125	0.0178

图8 润滑油压关联数据模型测试集输出结果

Fig.8 Test set output results of lubricating oil pressure correlation data model

图9 两种油压关联数据模型测试集输出结果

Fig.9 Test set output results of two oil pressure linked data models

参考文献 20

[1]	肖宇, 李英顺, 戴喜生, 等. 基于迁移学习的综合传动装置健康预测方法[J]. 兵工自动化, 2021, 40(10): 1-5, 9.
	XIAO Y, Li Y S, DAI X S, et al. Health prediction method of comprehensive transmission based on transfer learning[J]. Ordnance Industry Automation, 2021, 40(10): 1-5, 9. (in Chinese)
[2]	邹天刚, 闫清东, 盖江涛, 等. 基于扁平电机的轻混式履带车辆综合传动方案[J]. 兵工学报, 2021, 42(10): 2233-2241. doi: 10.3969/j.issn.1000-1093.2021.10.019
	ZOU T G, YAN Q D, GAI J T, et al. The scheme of lightweight Integrated mixing transmission based on flat motor for tracked vehicle[J]. Acta Armamentarii, 2021, 42(10): 2233-2241. (in Chinese) doi: 10.3969/j.issn.1000-1093.2021.10.019
[3]	闫书法, 马彪, 郑长松. 基于油液光谱分析的综合传动视情维护研究[J]. 光谱学与光谱分析, 2019, 39(11) : 3470-3474.
	YAN S F, MA B, ZHENG C S. Optimized method for shift valve used in integrated transmission[J]. Spectroscopy and Spectral Analysis, 2019, 39(11): 3470-3474. (in Chinese)
[4]	陈钢, 陈希祥. 综合传动装置状态监测与故障诊断系统研究[J]. 科学与信息化, 2021(8): 52.
	CHEN G, CHEN X X. The study on state monitor and failure diagnosis of integrated transmission system[J]. Technology and informationi, 2021(8): 52. (in Chinese)
[5]	郑长松, 冯毓庆, 汪宇, 等. 综合传动装置转速和温度传感器FMECA分析[J]. 车辆与动力技术, 2021(1): 55-59, 64.
	ZHENG C S, FENG Y Q, WANG Y, et al. FMECA analysis of speed and temperature sensor in the comprehensive transmission[J]. Vehicle & Power Technology, 2021(1): 55-59, 64. (in Chinese)
[6]	周锐, 蒋觉义, 乔丽. 综合传动装置一体化模型评价与验证研究[J]. 计算机测量与控制, 2020, 28(4): 246-250, 270.
	ZHOU R, JIANG J Y, QIAO L. Research on iIntegrated model evaluation and verication of integrated transmission[J]. Computer Measurement & Control, 2020, 28(4): 246-250, 270. (in Chinese)
[7]	闫书法, 马彪, 郑长松. 基于竞争失效的综合传动剩余寿命预测[J]. 汽车工程, 2019, 41(4): 426-431, 461.
	YAN S F, MA B, ZHENG C S. Remaining useful life prediction of power-shift steering transmission based on competing failures[J]. Automotive Engineering, 2019, 41(4): 426-431, 461. (in Chinese)
[8]	吴斌, 曲鹏, 甘辉. 基于虚拟余度的发动机伺服传感器故障检测[J]. 兵工自动化, 2021, 40(2): 25-28.
	WU B, QU P, GAN H. Fault detection method for engine servo sensor based on virtual redundancy[J]. Ordnance Industry Automation, 2021, 40(2) :25-28. (in Chinese)
[9]	闫静静, 王峥. 路灯节能调控系统矢量传感器误差修正方法[J]. 计算机仿真, 2021, 38(1): 256-260.
	YAN J J, WANG Z. Error correction method of vector sensor in street lamp energy saving control system[J]. Computer Simulation, 2021, 38(1): 256-260. (in Chinese)
[10]	吴耀春, 赵荣珍, 靳伍银, 等. 利用DCNN融合多传感器特征的故障诊断方法[J]. 振动、测试与诊断, 2021, 41(2) :362-369, 416.
	WU Y C, ZHAO R Z, JIN W Y, et al. Mechanical fault diagnosis method based on multi-sensor signal feature fusion using deep cConvolutional neural network[J]. Journal of Vibration, Measurement & Diagnosis, 2021, 41(2): 362-369, 416. (in Chinese)
[11]	隗寒冰, 白林. 基于多源异构信息融合的智能汽车目标检测算法[J]. 重庆交通大学学报(自然科学版), 2021, 40(8): 140-149.
	HUAI H B, BAI L. Intelligent vehicle target detection algorithm based on multi-source heterogeneous information fusion[J]. Journal of Chong Qing Jiao Tong University(Natural Science), 2021, 40(8): 140-149. (in Chinese)
[12]	李佳蔚, 崔涛, 刘宇航, 等. 基于排气温度动态模型的在线观测器研究[J]. 兵工学报, 2019, 40(8): 1562-1571. doi: 10.3969/j.issn.1000-1093.2019.08.002
	LI J W, CUI T, LIU Y H, et al. Construction for online observer based on exhaust gas temperature dynamic model[J]. Acta Armamentarii, 2019, 40(8): 1562-1571. (in Chinese) doi: 10.3969/j.issn.1000-1093.2019.08.002
[13]	路飞, 刘硕, 田国会. 智能机器人服务环境下传感数据映射及服务组合方法研究[J]. 机器人, 2019, 41(1): 30-39. doi: 10.13973/j.cnki.robot.18055
	LU F, LIU S, TIAN G H. Methods for sensor data mapping and automatic servicecomposition in intelligent robot service environment[J]. Robot, 2019, 41(1): 30-39. (in Chinese)
[14]	乔成林, 单甘霖, 段修生, 等. 多平台主动与被动传感器协同跟踪的长时调度方法[J]. 兵工学报, 2019, 40(1): 115-123. doi: 10.3969/j.issn.1000-1093.2019.01.014
	QIAO C L, SHAN G L, DUAN X S, et al. Non-myopic scheduling algorithm of multi-platform active /passive sensors for collaboration tracking[J]. Acta Armamentarii, 2019, 40(1): 115-123. (in Chinese)
[15]	王峻峰, 张玉帆, 邵瑶琪, 等. 面向生产性能数字孪生的仿真数据映射研究[J]. 系统仿真学报, 2021, 33(10): 2470-2477. doi: 10.16182/j.issn1004731x.joss.20-0506
	WANG J F, ZHANG Y F, SHAO Y Q, et al. Research on simulation data mapping for production performance digital twin[J]. Journal of System Simulation, 2021, 33(10): 2470-2477. (in Chinese) doi: 10.16182/j.issn1004731x.joss.20-0506
[16]	曾广迅, 龚光红, 李妮. 基于语义匹配的作战体系仿真想定生成方法[J]. 系统工程与电子技术, 2021, 43(8): 2154-2162. doi: 10.12305/j.issn.1001-506X.2021.08.17
	ZENG G X, GONG G H, LI N. Combat system of systems simulation scenario generation approach based on semantic matching[J]. Systems Engineering and Electronics, 2021, 43(8): 2154-2162. (in Chinese)
[17]	文必洋, 汤文成, 田应伟. 基于BP神经网络的高频雷达有效浪高场反演[J]. 华中科技大学学报(自然科学版), 2021, 49(4): 114-119.
	WEN B Y, TANG W C, TIAN Y W. Significant wave height field inversion of high frequency radar based on BP neural network[J]. Journal of Hua Zhong Universityof Science and Technology (Natural Science Edition), 2021, 49(4): 114-119. (in Chinese)
[18]	SHEN Y F, ZHENG W, YIN W J, et al. Feature extraction algorithm using a correlation coefficient combined with the VMD and its application to the GPS and GRACE[J]. IEEE Access, 2021(9): 17507-17519.
[19]	RAYI V K, MISHRA S P, NAIK J, et al. Adaptive VMD based optimized deep learning mixed kernel ELM autoencoder for single and multi-step wind power forecasting[J]. Energy, 2021(244): 122585.
[20]	JIA S, MA B, GUO W, et al. A sample entropy based prognostics method for lithium-ion batteries using relevance vector machine[J]. Journal of Manufacturing Systems, 2021(6): 773-781.