【作者】 耿俊豹；

【导师】 黄树红；

【作者基本信息】 华中科技大学 ， 热能工程， 2007， 博士

【摘要】 本文以国家自然科学基金资助项目“定量评价旋转机械振动状态的融合信息熵方法研究”和海军青年科研基金项目“旋转机械振动状态的融合信息熵方法研究”为课题背景,以舰船动力装置为研究对象,对舰船动力装置技术状态综合评估技术进行了研究,提出将信息融合技术有机地嵌入到舰船动力装置技术状态综合评估中,构建了基于信息融合的舰船动力装置技术状态评估模型,并在理论和应用两个方面进行了研究。在综述设备状态监测和故障诊断技术,船舶技术状态评估技术、信息融合技术的基础上,指出了船舶技术状态评估存在的问题,提出将信息融合技术应用于船舶技术状态评估的思路。在阐述技术状态管理的概念和发展的基础上,论述了本文所研究的技术状态评估的含义。创建了基于信息融合的舰船动力装置技术状态评估模型,将信息融合技术有机地嵌入到舰船动力装置技术状态评估的相应环节。根据舰船动力装置的特点,建立了舰船动力装置技术状态评估的部分顶层指标体系。提出采用粗糙集理论的知识约简来优化技术状态评估中的特征指标。以QGF-1汽轮鼓风机为对象,利用粗糙集理论建立了QGF-1汽轮鼓风机技术状态评估的特征指标、决策规则。考虑到综合评估中证据的不确定性,提出了采用Dempster-Shafter(D-S)证据理论来对各证据进行组合,以便达到降低不确定性的目的。为此,首先描述了D-S证据理论的国内外发展情况,其次研究了D-S证据理论的基本方法,不足之处,以及某些改进措施,总结了3种基于证据理论的决策方法,最后,建立了基于D-S证据理论的舰船动力装置技术状态评估方法。考虑到动力装置的许多设备是旋转机械类,本文研究了基于信息熵的旋转机械故障诊断方法。首先利用试验建立4种信息熵特征(奇异谱熵、功率谱熵、小波空间状态特征谱熵和小波能谱熵)在旋转机械典型故障下的期望值,即获得基于信息熵的旋转机械故障诊断标准特征向量。在此基础上,根据越相似模式间的距离越短的思路,建立了基于信息熵贴近度的旋转机械故障诊断。考虑到D-S证据理论的功能,建立了基于信息熵贴近度和D-S证据理论的旋转机械故障诊断。通过实例计算描述了两种方法的适用性。从工程应用的角度上,提出一种基于任务剖面的复杂系统技术状态综合评估方法。将粗糙集理论、D-S证据理论、设备群和模糊理论等有机地嵌入到综合评估方法中。先由任务剖面和设备群概念确定需要评估的设备和设备群;其次采用粗糙集理论来优化单设备技术状态评估指标,然后采用基于D-S证据理论的单设备技术状态评估或基于指标分级的单设备技术状态评估来获得单设备的技术状态;通过各设备技术状态及在设备群中的权重综合评估获得设备群的技术状态;最后根据设备群的技术状态,采用D-S证据理论或基于设备群权重的综合评估方法,获得整个系统的技术状态。文中提出了基于专家模糊打分法的权重模型,使得评估易于操作,评估结果符合工程实际情况。更多还原

【Abstract】 This paper is supported by the National Natural Science Foundation of P.R. China“Research on Fusion Information Entropy of Quantitative Evaluation of Vibration Condition of Rotating Machinery”and Naval Youth Science of P.R. China“Research on Fusion Information Entropy of Vibration Condition of Rotating Machinery”. Taking plant power of ships as the object, technical condition evaluation of plant power of ships based on information fusion is constructed, which is studied on the theory and application.The general situations on the technique of device condition monitoring and fault diagnosis, the technique of technical condition evaluation for ships, and the technique of information fusion are summarized. The question in technical condition evaluation for ships is pointed out. It is brought out that the technical condition is evaluated by applying the information fusion.The concept and development of technical condition management are expatiated. The signification of technical condition evaluation is discussed. The model on the technical state evaluation for ships based on the information fusion is designed. Some top index systems are founded to be used in the technical state evaluation for ships.It is put forward that the character indexes for technical condition evaluation are optimized by the reduction function of rough set theory. Taking the QGF-1 steamer-fan as object, the character indexes and decision-making rules for technical condition evaluation of steamer-fan are made by rough set theory.In view of the uncertainty in the integrated evaluation, it is brought forwarded that various evidences are combinated by Dempster-Shafter (D-S) evidence theory, to reduce the uncertainty and enhance the nicety. Therefore, the development of D-S evidence theory is described. Secondly, the basic means, lack, and some mend measure of D-S evidence theory are studied and the decision-making means based on evidence theory is set forth. Lastly, the technical condition evaluation based on D-S evidence theory is made.Because many devices of plant power are rotating machinery, the fault of the rotating machinery based on information entropy is studied. Firstly, the various information entropy features are set up, which show the vibration energy of the fault of the rotating machinery such as singular spectrum entropy, power spectrum entropy, wavelet space state feature entropy, wavelet power spectrum entropy. By a lot of experiment on the rotor, the expectation of the value of the various entropy based on the different typical faults of the rotating machinery is set up, which can be the standard feature vector for fault diagnosis based on the information entropy. According to the thought that the more short of the distance, the more similar among the models, a method based on close degree to information entropy is set up. Considering the function of D-S evidence theory, the method of rotating machinery fault diagnosis based on close degree of information entropy and evidence theory is set up. The applicability of the various measures is depicted by the instances given.A kind of integrated technical condition evaluation method based on mission sections for complex systems is presented to deal with the engineering application, in which some theories are applied, such as rough set theory, D-S evidence theory, device cluster and fuzzy theory. Devices and device clusters needed are confirmed by the mission sections and the concept of device cluster in advance. Indexes of simple device are optimized by rough set theory. The technical condition can be gained by the evaluation method based on index classification or by the evaluation method based on D-S evidence theory. The technical condition of device cluster can be gained by the evaluation method based the technical condition and the weigh of simple devices. According to the technical condition of device clusters, the technical condition of the complex system is gained by the evalution method base D-S evidence theory or the weigh of device clusters.The weigh model based on expert fuzzy mark assures that the evaluation result accords with the actual instance.更多还原