【作者】 黄大荣；

【导师】 黄席樾；

【作者基本信息】 重庆大学 ， 控制理论与控制工程， 2006， 博士

【摘要】 近年来,因关键设备故障而引起的灾难性事故时有发生,这些严重的、或灾难性事故的不断发生,迫使各国政府、社会以及科研人员高度重视对复杂系统故障预测和诊断方面的研究。然而传统的故障诊断技术大量采用预防维护理念,不管设备是否发生故障,都对其进行定期检修。各类工业企业也投入了大量的资金用于工业系统的维护并使其能够延长正常运行的寿命。由于缺乏对系统运行准确的状态判断和健康分析,出于安全的考虑,对系统设备进行了大量不必要的维修,大大提高了运行成本。可随着现代工业及科学技术的迅速发展,工业设备日趋大型化、连续化、高速化和自动化,功能越来越多,结构也越来越复杂,传统的故障诊断技术已经不能适应复杂设备系统实际运行的需要。正是基于这个原因,复杂系统的故障预测理论及应用研究近些年了引起了科研人员的极大关注,并取得了一些较好的结果。与此同时,励磁装置系统作为大型电力发电厂的关键设备,其运行性能好坏直接影响着电厂发电能力的稳定,因而为复杂系统故障预测理论的应用提供了良好的试验平台。基于这个目标,本论文围绕复杂系统的故障预测理论及其应用这一关键问题,以国家自然科学基金项目“基于相似性原理和免疫应答理论重构故障诊断系统”、重庆市教委科学研究项目“基于非线性理论的复杂系统故障预测理论及应用研究”、重庆市科委自然科学基金项目“复杂系统的网络化智能故障预测理论及应用研究”以及企业横向项目“励磁装置在现故障预测和故障诊断系统”为背景,对复杂系统的运行机理以及结构进行了详尽分析,就其故障预测的理论进行了深入的研究与讨论,并同时将理论研究工作嵌入到电力设备系统的励磁装置的故障诊断预测中进行检验。本文的主要研究成果包括:1、在参考粗糙集理论的基础上,针对复杂系统故障影响因子的特点,提出了一种数据清洗模型,并利用专家系统知识库设计了一类基于粗糙集理论的远程监测数据简化系统,给出相应的算法规则,解决了励磁设备系统信息数据简化的问题。2、结合卡尔曼滤波和方差控制约束理论提出了一种复杂系统故障预测的理论模型。先对复杂系统的结构进行分析,利用随机逼近理论对非线性系统结构进行线性逼近解决这类非线性系统存在逼近误差的问题。在此基础上,建立重构系统的基本建模方法,推导出该重构系统的基本特征,解决了逼近系统存在内生扰动力的问题。然后,建立了复杂系统故障预测的卡尔曼滤波预测模型以及故障监控阈值的确立规则。最后设计了一种复杂系统故障预测集成系统,并利用励磁系统更多还原

【Abstract】 In recent years, the causable disaster trouble usually happened because the key equipments break down. These serious fault or disaster trouble compelled that all countries government, society and researches have a high attention for the study of the fault prediction and fault diagnosis in the complicated system. However, the traditional fault diagnosis technique usually adopted prevention maintenance principle and it’s carried on periodical check whether the equipments break down or not. All kind of industry enterprise also throws in a great deal of funds to use for the maintenance of the industry system and make it prolong the life of the normal running. Because of lack of circulate the accurate state judgment and the health analysis to the system, it should carry on a great deal of otiose maintain to the system equipment for safety consideration. And yet, with the quick development of modern industry and sciences technical, the industry equipments is gradually large-scale to run, continuously to turn, highly speed turns and automates. The function is more and more; the structure is also more and more complicated. So the traditional fault diagnosis method and technology already can’t adapt the running demand of the complicated equipments system. For this reason, the fault prediction theories and application studies of the complicated system has been mostly attended in recent years and obtained some good results. Simultaneously, owing to Excitation systems are the key equipment of the power plants, its performance should direct work on the stability of the dynamoelectric capacity of the power plant. Thus, this has provided a well testing plat for the fault prediction theories of the complicated system.By reason of that, the key problem of this thesis discussed is the fault prediction theories of the complicated system and its application in excitation systems, with the background of the national natural science fund item" Reconfiguration of Fault Diagnosis System Based on Immune Response Theory and Comparability Principle", chong qing municipal education commission item“Complicated System Fault prediction theory and its application based on non-linear theory”, Natural science fund item of Science and Technology department of Chongqing“Network Intelligence prediction theory and Application of complicted system”and Enterprise horizontal item“On-line fault diagnosis and prediction system of excitness system”,it is carried on the detailed analysis to the movement mechanism and structure of the complicated更多还原