Depth Diagnosis of Spring Mechanical Faults of High Voltage Circuit Breakers Considering Wavelet Packet-Gray Level Co-occurrence Matrix Method
-
摘要: 弹簧操动机构作为高压断路器(High voltage circuit breakers, HVCBs)分合闸操作的储能单元,其可靠性对电力系统的安全运行具有重要意义。本文以六氟化硫高压断路器的弹簧操动机构为研究对象,分析分合闸弹簧的动作机理,对弹簧进行不同程度的故障设置。介绍了振动、声音传感器设备及采集参数,针对小波包时频分析法的缺点,提出一种基于小波包-灰度共生矩阵(Gray level co-occurrence matrix, GLCM)的特征提取方法。从诊断速度和诊断准确度两方面对比了支持向量机(Support vector machine, SVM)、决策树(Decision tree, DT)、朴素贝叶斯、K近邻(K nearest neighbors, KNN)4种诊断模型。实验结果表明,在模拟实际应用场景中,选用K近邻算法对分合闸弹簧故障进行深度诊断能够准确判断故障类型及故障程度,对高压断路器安全可靠运行具有实际应用价值。Abstract: As an energy storage unit for the opening and closing operations of high-voltage circuit breakers (HVCBs), the reliability of the spring operating mechanism is of great significance to the safe operation of the power system. In this paper, the spring operating mechanism of the SF6 HVCB is the research object, the action mechanism of the opening and closing spring is analyzed, and the mechanical failure of the spring is simulated. The vibration and sound sensor equipment and acquisition parameters are introduced. Aiming at the shortcomings of wavelet packet time-frequency analysis, a feature extraction method based on wavelet packet-gray level co-occurrence matrix (GLCM) is proposed. Then, the four diagnostic models of support vector machine (SVM), decision tree (DT), naive Bayes, and K nearest neighbors (KNN) were compared in terms of diagnosis speed and diagnosis accuracy. The experimental results demonstrate that in the simulation actual application scenario, the KNN algorithm is selected to perform an in-depth diagnosis of the opening and closing spring faults, which can accurately determine the type and degree of the fault, and has practical application value for the safe and reliable operation of high-voltage circuit breakers.
-
Key words:
- HVCBs /
- spring failure /
- sound-vibration signal /
- GLCM /
- KNN algorithm
-
表 1 弹簧疲劳故障模拟
Table 1. Simulation of spring fatigue failure
故障
名称原始
压缩量/mm故障
压缩量/mm故障程度 合闸弹簧
疲劳故障57 47 1级合闸弹簧疲劳 57 37 2级合闸弹簧疲劳 57 27 3级合闸弹簧疲劳 分闸弹簧
疲劳故障53 48 1级分闸弹簧疲劳 53 43 2级分闸弹簧疲劳 53 38 3级分闸弹簧疲劳 表 2 振动传感器技术参数
Table 2. Technical parameters of vibration sensors
参数 数值 电荷灵敏度 0.351 pC/ms2 最大冲击加速度 100 000 m/s2 频率响应 12 kHz 谐振频率 45 kHz 表 3 声音传感器技术参数
Table 3. Technical parameters of sound sensor
参数 数值 标称灵敏度 50 mV/Pa 频率范围 10~20 000 Hz 动态响应 17~140 dB 声级计频率特性 1级自由场 表 4 4种诊断模型用时及诊断准确度对比
Table 4. Comparison of time consumption and diagnostic accuracy among four diagnostic models
模型名称 训练时长/s 诊断时长/s 综合时长/s 诊断准确度/% SVM 3.14 2.66 5.80 96.6 DT 2.88 0.22 3.12 85.0 KNN 1.44 0.62 2.06 98.3 Bayes 6.82 1.46 8.28 93.3 表 5 分合闸弹簧疲劳故障种类及故障程度的准确率
Table 5. Accuracy of types and degrees of fatigue faults in opening and closing springs
故障类型名称 故障程度名称 故障程度准确率/% 合闸弹簧疲劳 合闸弹簧疲劳1级 100.0 合闸弹簧疲劳2级 87.3 合闸弹簧疲劳3级 80.2 分闸弹簧疲劳 分闸弹簧疲劳1级 72.5 分闸弹簧疲劳2级 83.3 分闸弹簧疲劳3级 67.0 -
[1] 段传宗, 鄢志平, 鄢志辉. 高压断路器故障检测与诊断技术[M]. 北京: 中国电力出版社, 2014: 24-28.DUAN C Z, YAN Z P, YAN Z H. Fault detection and diagnosis technology of high voltage circuit breaker[M]. Beijing: China Electric Power Press, 2014: 24-28. [2] HEISING C R, JANSSEN A L J, LANZ W, et al. Summary of CIGRE 13.06 working group world wide reliability data and maintenance cost data on high voltage circuit breakers above 63KV[C]//Proceedings of 1994 IEEE Industry Applications Society Annual Meeting. Denver: IEEE, 1994: 2226-2234. [3] 刘永超. 高压真空断路器振动特征提取及故障诊断方法研究[D]. 焦作: 河南理工大学, 2017.LIU Y C. Research on vibration characteristic extraction and fault diagnosis of high voltage vacuum circuit breaker[D]. Jiaozuo: Henan Polytechnic University, 2017. (in Chinese) [4] 刘明亮. 高压断路器机械振动信号特征提取及故障诊断研究[D]. 哈尔滨: 东北林业大学, 2017.LIU M L. Research of mechanical vibration signal feature extraction and fault diagnosis of high voltage circuit breaker[D]. Harbin: Northeast Forestry University, 2017. (in Chinese) [5] 赵国栋. 高压断路器在线监测与智能故障诊断方法研究[D]. 南京: 东南大学, 2017.ZHAO G D. Research on on-line monitoring and intelligent fault diagnosis method for HVCB[D]. Nanjing: Southeast University, 2017. (in Chinese) [6] 孙韬. 基于加权证据理论的高压断路器机械故障智能诊断技术[D]. 南京: 东南大学, 2017.SUN T. Intelligent mechanical fault diagnosis of high voltage circuit breaker based on weighted evidence theory[D]. Nanjing: Southeast University, 2017. (in Chinese) [7] 黄建. 特征评估高压断路器机械故障诊断方法的研究[J]. 高压电器, 2015, 51(12): 89-95.HUANG J. Research on machinery fault diagnosis of high voltage circuit breaker based on feature evaluation[J]. High Voltage Apparatus, 2015, 51(12): 89-95. (in Chinese) [8] CIABATTONI L, FERRACUTI F, FREDDI A, et al. Statistical spectral analysis for fault diagnosis of rotating machines[J]. IEEE Transactions on Industrial Electronics, 2018, 65(5): 4301-4310. doi: 10.1109/TIE.2017.2762623 [9] 袁建虎, 韩涛, 唐建, 等. 基于小波时频图和CNN的滚动轴承智能故障诊断方法[J]. 机械设计与研究, 2017, 33(2): 93-97.YUAN J H, HAN T, TANG J, et al. An approach to intelligent fault diagnosis of rolling bearing using wavelet time-frequency representations and CNN[J]. Machine Design and Research, 2017, 33(2): 93-97. (in Chinese) [10] 须颖, 李昊东, 安冬. 基于GLCM-SDAE的滚动轴承故障诊断方法[J]. 沈阳建筑大学学报(自然科学版), 2020, 36(4): 720-728.XU Y, LI H D, AN D. Fault diagnosis of rolling bearing based on GLCM-SDAE[J]. Journal of Shenyang Jianzhu University (Natural Science), 2020, 36(4): 720-728. (in Chinese) [11] 尚洁. 基于LTP灰度共生矩阵和SVM的织物疵点检测分类算法研究[D]. 成都: 成都理工大学, 2020.SHANG J. Research on fabric defect detection and classification algorithm based on LTP gray-level co-occurrence matrix and SVM[D]. Chengdu: Chengdu University of Technology, 2020. (in Chinese) [12] PALIWAL M, KUMAR U A. Neural networks and statistical techniques: a review of applications[J]. Expert Systems with Applications, 2009, 36(1): 2-17. doi: 10.1016/j.eswa.2007.10.005 [13] ZHANG J F, LIU M L, WANG K Q, et al. Mechanical fault diagnosis for HV circuit breakers based on ensemble empirical mode decomposition energy entropy and support vector machine[J]. Mathematical Problems in Engineering, 2015, 2015: 101757. [14] HUANG N T, CHEN H J, ZHANG S X, et al. Mechanical fault diagnosis of high voltage circuit breakers based on wavelet time-frequency entropy and one-class support vector machine[J]. Entropy, 2015, 18(1): 7. doi: 10.3390/e18010007 [15] 王英英, 罗毅, 涂光瑜. 基于粗糙集与决策树的配电网故障诊断方法[J]. 高电压技术, 2008, 34(4): 794-798.WANG Y Y, LUO Y, TU G Y. Fault diagnosis method for distribution networks based on the rough sets and decision tree theory[J]. High Voltage Engineering, 2008, 34(4): 794-798. (in Chinese) [16] 顾礼斌, 李勇刚. 基于模糊 k近邻的变电站主接线类型自动识别方法[J]. 广东电力, 2018, 31(2): 130-135. doi: 10.3969/j.issn.1007-290X.2018.002.021GU L B, LI Y G. Automatic identification method for substation main wiring modes based on fuzzy k-nearest neighbor algorithm[J]. Guangdong Electric Power, 2018, 31(2): 130-135. (in Chinese) doi: 10.3969/j.issn.1007-290X.2018.002.021 [17] 宁可, 孙同晶, 赵浩强. 基于属性关联的朴素贝叶斯分类算法[J]. 计算机工程, 2018, 44(6): 18-23. doi: 10.3969/j.issn.1000-3428.2018.06.004NING K, SUN T J, ZHAO H Q. Naive Bayesian classification algorithm based on attribute association[J]. Computer Engineering, 2018, 44(6): 18-23. (in Chinese) doi: 10.3969/j.issn.1000-3428.2018.06.004 [18] 张伟, 王志海, 原继东, 等. 一种局部属性加权朴素贝叶斯分类算法[J]. 北京交通大学学报, 2018, 42(2): 14-21.ZHANG W, WANG Z H, YUAN J D, et al. A locally attribute weighted naive Bayes classifier[J]. Journal of Beijing Jiaotong University, 2018, 42(2): 14-21. (in Chinese)