SIA OpenIR  > 工业信息学研究室  > 工业控制系统研究室
机电设备智能诊断方法研究及应用
Alternative TitleIntelligent Faults Diagnosis Technology for Electromechanical Equipment and Its Applications
郭前进1,2
Department工业控制系统研究室
Thesis Advisor于海斌
ClassificationTP277
Keyword状态维护 故障诊断 独立成分分析 混合粒子群算法 小波神经网络
Call NumberTP277/G95/2008
Pages170页
Degree Discipline机械电子工程
Degree Name博士
2008-01-26
Degree Grantor中国科学院沈阳自动化研究所
Place of Conferral沈阳
Abstract稳定、可靠的诊断系统能对机电设备早期故障及性能劣化提供适时监测,从而避免突发故障,为设备维护提供有力保障。目前,用来对机电设备进行监测诊断的方法很多,但都存在各自的不足。基于此,本论文对机电系统故障检测与诊断技术进行了方法及应用研究,主要工作有: 1.针对非周期、非平稳故障特征信号提取所面临的困难,本文构建了基于WVD-PCA的故障特征提取与诊断模型。模型克服了频谱分析方法缺乏局部分析能力,不能反映振动信号时域特征的问题,利用Wigner-Ville分布建立旋转机械状态时频谱图,在时域和频域内同时对非平稳信号进行分析;然后利用主成分分析方法的降维功能,通过构造特征矢量矩阵,将原始特征值投影到相互正交的矢量上,减少原始时频数据的冗余成分,获取可反映原始数据的主分量,达到了降维的效果。实验表明,模型可以对旋转类机械故障进行有效的诊断。 2.针对常规的信号处理方法对非平稳、非Gauss、非线性振动信号缺乏有效分析处理的问题,本文构建了基于CT-ICA的故障特征提取与诊断模型。模型首先对振动信号进行线调频小波时频分析,将故障信号映射到时-频空间,滤去与故障特征无关的干扰信号,然后利用独立成分分析从多维统计数据中寻找潜在故障因子或成分,运用SOM神经网络进行故障的定性和定量分析。实验表明,诊断模型对非平稳、非Gauss、非线性振动信号故障特征提取与分类具有良好的处理能力,改善了常规时频分析方法对非平稳故障信号的分析能力,提高了故障的早期诊断率。 3.提出了基于HGDPSO的小波神经网络的非线性智能诊断模型。模型运用改进的PSO算法与梯度下降算法相结合的混合粒子群方法来对小波神经网络结构及参数进行优化,该方法能在全局范围内进行鲁棒搜索,有效地防止寻优过程收敛于局部最优解,而且计算简单快速,因而应用更加方便。通过实验分析表明,基于混合粒子群算法的小波神经网络在故障诊断应用中具有训练性能优越、收敛速度快及诊断精确度高等优势,具有良好的应用前景。 4.提出了一种基于自组织补偿模糊小波神经网络的非确定性智能诊断系统。该系统使用补偿模糊运算,并将小波神经网络和模糊逻辑相结合,充分利用了相互间的优点,对处理故障诊断这类复杂、非线性及非确定性问题具有强大的功能。模糊小波神经网络不仅能自适应地调整输入输出模糊隶属函数,也能借助于补偿逻辑算法动态优化相应的模糊推理。由于补偿模糊小波神经网络引入了补偿模糊小波神经元,能使网络从初始定义的模糊规则开始训练,使网络容错率更高,系统更稳定。 5.分析了异步电机的主要故障模式及其故障机理,并建立了转子断条及匝间短路故障的暂态仿真模型,同时对转子断条故障及匝间短路故障进行了实验研究。通过对大量仿真及实验数据的详尽分析,系统地总结了转子断条及匝间短路故障特征。在此基础上,提出了基于小波神经网络与广义谐波小波包滤波技术的异步电机转子断条及匝间短路多故障检测方法;在进一步分析的基础上,提出了基于扩展Park矢量变换与模糊小波神经网络的转子断条及匝间短路故障检测方法。实验结果表明该方法是切实可行的。 6.本文最后就CBM开放系统的概念和结构进行研究,开发了基于CBM的故障维护应用平台。针对CBM体系结构和标准,充分利用现代分布式网络化技术,以Web Service技术为基础,完成CBM框架定义的组件模块设计,实现基于Web Service的分布式故障维护应用平台的开发工作。
Other AbstractA reliable and robust diagnostic/prognostic system is very useful for health condition monitoring of electromechanical machinery. Such a system can provide early warnings of malfunction or possible damage to avoid sudden failures, and to allow effective repair and maintenance. Currently, there are many methods that can be used to monitor the health condition of electromechanical machinery, but each has its advantages and disadvantages. This thesis presents a comprehensive investigation to assess the sensitivity and robustness of well-accepted fault detection techniques. The main works and contributions of this dissertation are summarized as follows: 1. In this study, the auto-regressive model based pseudo-Wigner-Ville distribution for an integrated time-frequency signature extraction of the machine vibration is designed. The method offers the advantage of good localization of the vibration signal energy in the time-frequency domain. Principal component analysis (PCA) is used for the redundancy reduction and feature extraction in the time-frequency domain, and the self-organizing map (SOM) was employed to identify the faults of the rotating machinery. Experimental results show that the proposed method is very effective. 2. Chapter 3 has proposed a kind of fault diagnosis method based on the combination of the adaptive Gaussian chirplet distribution and independent component analysis. The adaptive Gaussian chirplet distribution has been applied to electromechanical machinery fault diagnosis system due to their advantages in the representation of signals in both the time-frequency domains. This feature of the time-frequency analysis meets the requirements for analysing vibration signals that are non-stationary signals. Based on the features extracted from the time-frequency moments using ICA method, the machinery fault diagnoses were to be classified through the SOM network. 3. A HGDPSO-based wavelet neural network approach is developed for diagnosing incipient faults in electromechanical machinery. In this fault diagnosis system, wavelet neural network techniques are used in combination with a new evolutionary learning algorithm. This new evolutionary learning algorithm is based on a hybrid of the constriction factor approach for particle swarm optimization (PSO) technique and the gradient descent (GD) technique, and is thus called HGDPSO. The HGDPSO is developed in such a way that a constriction factor approach for particle swarm optimization (CFA for PSO) is applied as a based level search, which can give a good direction to the optimal global region, and a local search gradient descent algorithm (GD) is used as a fine tuning to determine the optimal solution at the final. The effectiveness of the HGDPSO based WNN is demonstrated through the classification of the fault signals in rotating machinery. 4. A new self-constructing fuzzy wavelet neural networks (SCFWNN) using compensatory fuzzy operators are proposed for intelligent fault diagnosis in Chapter 5. Hybrid intelligent systems, fuzzy wavelet neural networks possess the advantages of both wavelet networks and fuzzy rule-based systems and are particularly powerful in handling complex, non-linear and imprecise problems such as fault diagnosis. Two phases of learning, structure and parameter learning are used for constructing the neurofuzzy network. Two phases of learning algorithm are applied to automatically construct the SCFWNN. In the first phase the structure learning algorithm is used to find proper fuzzy partitions in the input space and create fuzzy logic rules. In the second phase all parameters are tuned using a supervised learning scheme. The hybrid intelligent system with adaptive fuzzy reasoning is more effective and adaptive than the conventional neural fuzzy system with non-adaptive fuzzy reasoning. The results of simulation show that this SCFWNN method has the advantage of faster learning rate and higher diagnosing precision. 5. In Chapter 6, the main fault modes such as rotor bar subsequent-breaking fault and stator winding inter-turn short circuit fault of motors and their fault mechanisms is analyzed. And the transient simulation of these fault modes has been completed successfully. Also, the corresponding experiments have been carried out. Based on multiple modulation zoom spectrum analysis, wavelet neural network and harmonic wavelet packets transform filter, this paper presents a novel detection scheme for these fault modes. Based on the square of the Park’s Vector Modulus and fuzzy wavelet neural network, this paper presents the other detection scheme for these fault modes. Results of simulation and experiments demonstrate that the detection scheme is valid and feasible. 6. To develop Open System Architecture for Condition-Based Maintenance (CBM), the architecture development had focused on the definition of distributed software architecture for CBM. The distributed software architecture was first defined for CBM, then the designation and application of CBM platform based on this open architecture were introduced, and finally an application of CBM platform was presented as a case. This study provided an approach to facilitate the integration and interchangeability between a variety of hardware and software components. The potential benefits of the Open System Architecture included the improved ease of upgrading for system components, more rapid technical development, and reduced cost to end-users and so on.
Language中文
Contribution Rank1
Document Type学位论文
Identifierhttp://ir.sia.cn/handle/173321/240
Collection工业信息学研究室_工业控制系统研究室
Affiliation1.中国科学院沈阳自动化研究所
2.中国科学院研究生院
Recommended Citation
GB/T 7714
郭前进. 机电设备智能诊断方法研究及应用[D]. 沈阳. 中国科学院沈阳自动化研究所,2008.
Files in This Item:
File Name/Size DocType Version Access License
10001_20041801470073(10825KB) 开放获取--Application Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[郭前进]'s Articles
Baidu academic
Similar articles in Baidu academic
[郭前进]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[郭前进]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.