基于DSP的自确认传感器故障诊断算法研究及优化实现
|
摘要
传感器是信息获取的源头,是自动化测试与控制系统的重要组成部分。如果传感器发生了故障,输出测量值已经完全偏离了实际值,而使用时又无法判断真实的情况,有可能会造成严重的事故。因此,对传感器故障进行检测和诊断就显得非常重要,于是人们开始研究一种不仅能输出测量值,同时能够对自身的工作状态进行在线评估的新式传感器——自确认传感器。而自确认传感器的一个重要组成部分就是故障诊断单元。
本文研究了用于自确认压力传感器的故障诊断算法,包括故障特征提取算法和故障分类算法。故障特征提取算法针对压力传感器的六种故障类型,首先对传感器信号进行小波包分解,对分解得到的各个节点削减后再进行重构,以重构的各个频带内信号的能量及削减比构造特征向量。本文对不同小波函数提取故障特征的效果进行评估,并选择了效果较好的小波函数。同时,对小波包分解的层数进行了选择。得到故障信号的特征之后,采用支持向量机进行分类。本文根据所研究的故障类型的特点,设计了多分类支持向量机的拓扑结构,对支持向量机的模型进行了选择。
在此基础上,本文论述了在TMS320C6713 DSP平台上实现故障诊断算法的过程,并且设计了DSP的软件。调试工作完成之后,编写了启动引导程序以及Flash烧写程序,实现了系统上电后从外接Flash引导。为了满足实时性要求,对代码进行了优化,实现了对故障的实时诊断。
最后,通过实验来验证算法实现的正确性以及故障诊断的效果,通过算法运行时间来检验代码优化的效果。实验结果表明,算法运算速度满足要求,对于已知的六种故障类型能有效的进行诊断。
Sensor is the source of information acquisition, the important component of the automatic test and control system. If the sensor fault occurs, for example, the output value has completely deviated from the actual values, but is unable to identify the real condition, there may be a serious accident. Therefore, the sensor fault detection and diagnosis is very important. Then, a kind of new-type sensor -- self-validating sensor is begun to study. It can not merely export the measurement value, but also can carry on the online assessment to its own working state at the same time. And, an important component of self-validating sensor is fault diagnosis unit.
This paper has studied fault diagnosis algorithm used in self-validating pressure sensor, including the fault feature extraction algorithm and the fault classification algorithm. According to six pressure sensors fault patterns, first, the fault feature extraction algorithm makes wavelet packet decomposition to sensor signal. Second, the various nodes are reconstructed with some cutting algorithm. After that, the energy of each reconstructed signal, and the average cutting ratio of all nodes are calculated, which are regarded as the feature vector. This paper assesses different effects of fault feature extraction of wavelet function, and chooses the wavelet function with better effect. Meanwhile, this paper chooses the level of wavelet packet decomposition. After feature extraction, the support vector machines (SVM) is used for fault classification. According to the characteristic of the fault type studied in this paper, topological structure of multi-classification SVM has been designed. Meanwhile, chose the model of SVM.
Then, fault diagnosis algorithm is realized based on TMS320C6713. The software of DSP is also designed. After debugging, bootloader and Flash burning program is composed. The system has achieved load from external Flash. To meet the real-time requirements, the code has been optimized. The system also achieved the real-time diagnosis of every sampling point.
Finally, the effectiveness of fault diagnosis and whether the algorithm is correct is verified through experiments. And through the running time of algorithm to verify the effectiveness of code optimization. The experimental results showed that the proposed approach can be applied to the sensor fault diagnosis effectively, and meet the demand of the computing speed.
本文研究了用于自确认压力传感器的故障诊断算法,包括故障特征提取算法和故障分类算法。故障特征提取算法针对压力传感器的六种故障类型,首先对传感器信号进行小波包分解,对分解得到的各个节点削减后再进行重构,以重构的各个频带内信号的能量及削减比构造特征向量。本文对不同小波函数提取故障特征的效果进行评估,并选择了效果较好的小波函数。同时,对小波包分解的层数进行了选择。得到故障信号的特征之后,采用支持向量机进行分类。本文根据所研究的故障类型的特点,设计了多分类支持向量机的拓扑结构,对支持向量机的模型进行了选择。
在此基础上,本文论述了在TMS320C6713 DSP平台上实现故障诊断算法的过程,并且设计了DSP的软件。调试工作完成之后,编写了启动引导程序以及Flash烧写程序,实现了系统上电后从外接Flash引导。为了满足实时性要求,对代码进行了优化,实现了对故障的实时诊断。
最后,通过实验来验证算法实现的正确性以及故障诊断的效果,通过算法运行时间来检验代码优化的效果。实验结果表明,算法运算速度满足要求,对于已知的六种故障类型能有效的进行诊断。
Sensor is the source of information acquisition, the important component of the automatic test and control system. If the sensor fault occurs, for example, the output value has completely deviated from the actual values, but is unable to identify the real condition, there may be a serious accident. Therefore, the sensor fault detection and diagnosis is very important. Then, a kind of new-type sensor -- self-validating sensor is begun to study. It can not merely export the measurement value, but also can carry on the online assessment to its own working state at the same time. And, an important component of self-validating sensor is fault diagnosis unit.
This paper has studied fault diagnosis algorithm used in self-validating pressure sensor, including the fault feature extraction algorithm and the fault classification algorithm. According to six pressure sensors fault patterns, first, the fault feature extraction algorithm makes wavelet packet decomposition to sensor signal. Second, the various nodes are reconstructed with some cutting algorithm. After that, the energy of each reconstructed signal, and the average cutting ratio of all nodes are calculated, which are regarded as the feature vector. This paper assesses different effects of fault feature extraction of wavelet function, and chooses the wavelet function with better effect. Meanwhile, this paper chooses the level of wavelet packet decomposition. After feature extraction, the support vector machines (SVM) is used for fault classification. According to the characteristic of the fault type studied in this paper, topological structure of multi-classification SVM has been designed. Meanwhile, chose the model of SVM.
Then, fault diagnosis algorithm is realized based on TMS320C6713. The software of DSP is also designed. After debugging, bootloader and Flash burning program is composed. The system has achieved load from external Flash. To meet the real-time requirements, the code has been optimized. The system also achieved the real-time diagnosis of every sampling point.
Finally, the effectiveness of fault diagnosis and whether the algorithm is correct is verified through experiments. And through the running time of algorithm to verify the effectiveness of code optimization. The experimental results showed that the proposed approach can be applied to the sensor fault diagnosis effectively, and meet the demand of the computing speed.
引文
1 M P Henry, D W Clarke. The Self Validating Sensor: Rationale , Definitions and Examples. Control Engineering Practice. 1993, l(4):585~610
2冯志刚，王祁，徐涛等．自确认传感器技术研究．电子器件．2006，29(3)：848~854
3阎军.自确认传感器在过程控制中的应用.仪器仪表与分析监测.2001,(1):16~17
4 J C-Y Yang, D W Clarke. Control using self- validating sensors. Transactions of the Institute of Measurement and Control. 1996, 18(1): 15~23
5 D W Clarke. Sensor, actuator and loop validation. IEE Colloquium (Digest). 1999, 142:1~10
6 D W Clarke. Sensor, actuator and plant validation. IEE Colloquium (Digest). 1999, 160:1-8
7 Na. Seung. You, Lee. Heyoung, Kee, Chang - Doo. Self-Validating Sensors wit h Application to a Flexible Link Control System/ / Proceedings of SPIE - The International Society for Optical Engineering. 2001, 4563:108~115
8周东华，孙优贤．控制系统的故障检测与诊断技术．北京，清华大学出版社．1994：4~11．
9石志勇，李国章，付建平等．基于平均空间的故障检测与诊断．信息与控制．1999，28(6)：459~465
10陈金水，孙优贤．一种用于控制系统故障检测的U10观侧器设计方法．浙江大学学报．1996，30(1)：49～54
11王志毅，谷波，黎远光．小波变换应用于空调制冷机组故障先兆预测．暖通空调．2004，34(10)：117~120
12丁晖，刘君华，申忠如．在线小波变换技术在气体传感器漂移故障检测中的应用．西安交通大学学报．2002，36(12)：1219~1221
13 Ramesh Pwrla, Mukhopadhyay.s, Sensor fault detection and isolation usingartificial neural networks. 2004 IEEE Region 10 Conference. 2004, (4):676~679
14胡双俊，关起强，严桂．基于小波变换和支持向量机的电力电子故障诊断．煤矿机械．2008，29(4)：204~206
15唐浩，屈梁生．基于支持向量机的发动机故障诊断．西安交通大学学报．2007，41(9)：1124～1126
16宇德介，陈森峰，程军圣等．一种基于经验模式分解与支持向量机的转子故障诊断方法．中国电机工程学报．2006，26(1 6)：162～167
17 Zhang Jian Qiu, Yan Yong. A Wavelet-based Approach to Abrupt Fault Detection and Diagnosis of Sensor. IEEE Transactions on Instrumentation and Measurement. 2001, 50(5): 1389~1396
18徐红．小波包分解算法研究及其在机械故障诊断中的应用．仪器仪表学报．2005，26(8)：786～787
19韩朝晖．利用小波包分析提取信号分量．通信与信息技术会议论文集．2006：582~589
20 Jiang Wanlu, Zhang Shuqing. Wavelet transform method for fault diagnosis of hydraulic pump. Jixie Gongcheng Xuebao. 2001, 37(6):34~37
21杨福生．小波变换的工程分析与应用．北京，科学出版社．2001
22秦前清，杨宗凯．实用小波分析．西安，西安电子科技大学出版社．1998
23 Kon Max Wong, Jiangfeng Wu, T. N. Davidson and Qu(Gary)Jin. Wavelet Packet Division Multiplexing and Wavelet Packet Design Under Timing Error Effects, IEEE Trans. On Signal Processing. 1997, 45(12):2877~2889
24程正兴．小波分析算法与应用．西安，西安交大出版社．1998
25邓乃扬，田英杰．数据挖掘中的新方法——支持向量机．北京，科学出版社．2004
26彭璐．支持向量机分类算法研究与应用．湖南大学硕士学位论文．2007
27李方慧，王飞，何佩琨．TMS320C6000系列DSPs原理与应用．电子工业出版社．2003
28吴多，程德福．基于DSP的快速小波分解和重构．吉林大学学报．2004，12(3)：214～218
29徐波，王玉兰．基于矩阵形式的小波快速算法及DSP实现．科学技术与工程．2007，7(2)：195～198
30吕新华，武斌，攸阳等．小波变换Mallat算法实现中的边界延拓研究．天津理工大学学报．2006，22(2)：14～17
31 C J C Burges. A tutorial on support vector machines for pattern recognition. Data Mining and Knowledge Discovery. 1998, 2(2): 121-167
32 J C Platt. Fast training of support vector machines using Sequential MinimalOptimization. In: Advances in Kernel Method-Support Vector Learning. Cambridge. 1999:185~208
33 Texas Instruments. TMS320C6000 CPU and Instruction Set Reference Guide. USA, Texas Instruments Incorporated. 2000
34 Texas Instruments. TMS320C6000 Optimizing Compiler User's Guide. USA, Texas Instruments Incorporated. 2001
35 Texas Instruments. TMS320C6000 Programmer's Guide. USA, Texas Instruments Incorporated. 2001
36阳明晔，张志勇．基于TMS320C6000系列DSP的C代码优化方法研究．微处理机．2004，(2)：59～64
37刘朝晖，郑玉墙．用C语言进行DSP软件设计的优化考虑．空军雷达学院学报．200l，l5(2)：49～52
38 Q Zhao, J C Principe. Support vector machines for SAR automatic target recognition. IEEE Transactions on Aerospace and Electronic Systems. 2001, 37(2): 643~654
39赵晖，荣莉莉，李晓．一种设计层次支持向量机多类分类器的新方法．计算机应用研究．2006，(6)：34～37
40 S Keerthi, E G Gilbert. Convergence of a generalized SMO algorithm for SVM classifier design. Machine Learning. 2002, 46(1):351~360
41 Hsu Chihwei, Lin Chihjen. A comparison of methods for multi-class support vector machines. IEEE Transactions on Neural Networks. 2002, 13(2):415~425
42马笑天，黄席樾，柴毅．基于SVM的二叉树多类分类算法及其在故障诊断中的应用．控制与决策．2003，18(3)：272～276
43宋辛科．基于支持向量机和决策树的多值分类器．计算机工程．2005，31(14)：174～175
44 O Chapelle, V Vapnik. Choosing multiple parameters for support vector machine. Machine Learning. 2002, 46(1~3): 131~159
45 C Gold, P Sollichb. Model selection for support vector machine classification. Neurocomputing. 2003, 55(2):221~249
46陶珺，王鹏．基于TMS320C6000：DSP的嵌入式系统中引导方法的研究．计算机与数字工程．2005，(7)：27～3l
47李鹏，郑喜凤，丁铁夫．TMS320C6000系列DSPs外接FLASH引导方式的实现．长春理工大学学报2004，27(4)：52～54
48唐冰. TMS320C5410烧写Flash实现并行自举引导.单片机及嵌入式系统应用. 2003, (1):26~30
2冯志刚，王祁，徐涛等．自确认传感器技术研究．电子器件．2006，29(3)：848~854
3阎军.自确认传感器在过程控制中的应用.仪器仪表与分析监测.2001,(1):16~17
4 J C-Y Yang, D W Clarke. Control using self- validating sensors. Transactions of the Institute of Measurement and Control. 1996, 18(1): 15~23
5 D W Clarke. Sensor, actuator and loop validation. IEE Colloquium (Digest). 1999, 142:1~10
6 D W Clarke. Sensor, actuator and plant validation. IEE Colloquium (Digest). 1999, 160:1-8
7 Na. Seung. You, Lee. Heyoung, Kee, Chang - Doo. Self-Validating Sensors wit h Application to a Flexible Link Control System/ / Proceedings of SPIE - The International Society for Optical Engineering. 2001, 4563:108~115
8周东华，孙优贤．控制系统的故障检测与诊断技术．北京，清华大学出版社．1994：4~11．
9石志勇，李国章，付建平等．基于平均空间的故障检测与诊断．信息与控制．1999，28(6)：459~465
10陈金水，孙优贤．一种用于控制系统故障检测的U10观侧器设计方法．浙江大学学报．1996，30(1)：49～54
11王志毅，谷波，黎远光．小波变换应用于空调制冷机组故障先兆预测．暖通空调．2004，34(10)：117~120
12丁晖，刘君华，申忠如．在线小波变换技术在气体传感器漂移故障检测中的应用．西安交通大学学报．2002，36(12)：1219~1221
13 Ramesh Pwrla, Mukhopadhyay.s, Sensor fault detection and isolation usingartificial neural networks. 2004 IEEE Region 10 Conference. 2004, (4):676~679
14胡双俊，关起强，严桂．基于小波变换和支持向量机的电力电子故障诊断．煤矿机械．2008，29(4)：204~206
15唐浩，屈梁生．基于支持向量机的发动机故障诊断．西安交通大学学报．2007，41(9)：1124～1126
16宇德介，陈森峰，程军圣等．一种基于经验模式分解与支持向量机的转子故障诊断方法．中国电机工程学报．2006，26(1 6)：162～167
17 Zhang Jian Qiu, Yan Yong. A Wavelet-based Approach to Abrupt Fault Detection and Diagnosis of Sensor. IEEE Transactions on Instrumentation and Measurement. 2001, 50(5): 1389~1396
18徐红．小波包分解算法研究及其在机械故障诊断中的应用．仪器仪表学报．2005，26(8)：786～787
19韩朝晖．利用小波包分析提取信号分量．通信与信息技术会议论文集．2006：582~589
20 Jiang Wanlu, Zhang Shuqing. Wavelet transform method for fault diagnosis of hydraulic pump. Jixie Gongcheng Xuebao. 2001, 37(6):34~37
21杨福生．小波变换的工程分析与应用．北京，科学出版社．2001
22秦前清，杨宗凯．实用小波分析．西安，西安电子科技大学出版社．1998
23 Kon Max Wong, Jiangfeng Wu, T. N. Davidson and Qu(Gary)Jin. Wavelet Packet Division Multiplexing and Wavelet Packet Design Under Timing Error Effects, IEEE Trans. On Signal Processing. 1997, 45(12):2877~2889
24程正兴．小波分析算法与应用．西安，西安交大出版社．1998
25邓乃扬，田英杰．数据挖掘中的新方法——支持向量机．北京，科学出版社．2004
26彭璐．支持向量机分类算法研究与应用．湖南大学硕士学位论文．2007
27李方慧，王飞，何佩琨．TMS320C6000系列DSPs原理与应用．电子工业出版社．2003
28吴多，程德福．基于DSP的快速小波分解和重构．吉林大学学报．2004，12(3)：214～218
29徐波，王玉兰．基于矩阵形式的小波快速算法及DSP实现．科学技术与工程．2007，7(2)：195～198
30吕新华，武斌，攸阳等．小波变换Mallat算法实现中的边界延拓研究．天津理工大学学报．2006，22(2)：14～17
31 C J C Burges. A tutorial on support vector machines for pattern recognition. Data Mining and Knowledge Discovery. 1998, 2(2): 121-167
32 J C Platt. Fast training of support vector machines using Sequential MinimalOptimization. In: Advances in Kernel Method-Support Vector Learning. Cambridge. 1999:185~208
33 Texas Instruments. TMS320C6000 CPU and Instruction Set Reference Guide. USA, Texas Instruments Incorporated. 2000
34 Texas Instruments. TMS320C6000 Optimizing Compiler User's Guide. USA, Texas Instruments Incorporated. 2001
35 Texas Instruments. TMS320C6000 Programmer's Guide. USA, Texas Instruments Incorporated. 2001
36阳明晔，张志勇．基于TMS320C6000系列DSP的C代码优化方法研究．微处理机．2004，(2)：59～64
37刘朝晖，郑玉墙．用C语言进行DSP软件设计的优化考虑．空军雷达学院学报．200l，l5(2)：49～52
38 Q Zhao, J C Principe. Support vector machines for SAR automatic target recognition. IEEE Transactions on Aerospace and Electronic Systems. 2001, 37(2): 643~654
39赵晖，荣莉莉，李晓．一种设计层次支持向量机多类分类器的新方法．计算机应用研究．2006，(6)：34～37
40 S Keerthi, E G Gilbert. Convergence of a generalized SMO algorithm for SVM classifier design. Machine Learning. 2002, 46(1):351~360
41 Hsu Chihwei, Lin Chihjen. A comparison of methods for multi-class support vector machines. IEEE Transactions on Neural Networks. 2002, 13(2):415~425
42马笑天，黄席樾，柴毅．基于SVM的二叉树多类分类算法及其在故障诊断中的应用．控制与决策．2003，18(3)：272～276
43宋辛科．基于支持向量机和决策树的多值分类器．计算机工程．2005，31(14)：174～175
44 O Chapelle, V Vapnik. Choosing multiple parameters for support vector machine. Machine Learning. 2002, 46(1~3): 131~159
45 C Gold, P Sollichb. Model selection for support vector machine classification. Neurocomputing. 2003, 55(2):221~249
46陶珺，王鹏．基于TMS320C6000：DSP的嵌入式系统中引导方法的研究．计算机与数字工程．2005，(7)：27～3l
47李鹏，郑喜凤，丁铁夫．TMS320C6000系列DSPs外接FLASH引导方式的实现．长春理工大学学报2004，27(4)：52～54
48唐冰. TMS320C5410烧写Flash实现并行自举引导.单片机及嵌入式系统应用. 2003, (1):26~30