大型风电机组故障诊断与状态综合评价方法研究

发布时间：2018-12-27 08:21

【摘要】：随着世界能源的紧缺,风能已经在各种场合扮演着重要的角色,风力发电技术在全世界范围内已广泛使用。然而,由于地理位置偏僻,维修技术复杂等各方面原因,使得风力发电过程中容易出现故障,且维修难度和运行维护成本较高,因此风力发电系统运行状态评估等工作就显得尤为重要。风力发电机组系统庞大,对其进行故障诊断和维修工程难度较大,很难靠单一的技术实现对设备准确、完整的诊断。综上,针对风力发电的故障诊断与评估需要一套合理的评价体系来解决这些难题。因此,本文的主要研究内容如下:(1)对当前风力发电机组的基本结构、工作原理进行了阐述,针对风电机组的运行特性,分析了机组的故障机理,并对机组故障多发部位及故障诊断常用方法进行了研究,提出了针对不同信号源的故障诊断方法。最后基于风电机组SCADA系统的实测运行数据,设计了两种故障诊断与状态综合评价方法。(2)运用熵权模糊综合评判方法,根据某风场SCADA系统选取的机组正常运行和故障数据,对机组建立综合评判模型。依照所建模型,对机组的运行状态进行划分,本文将机组运行状态划分为“优、良、中、差”四个等级来评估风电机组的健康状况,当计算得出有运行状态为“差”的子系统时,将其视为故障状态,需要立即停机检查,避免故障严重化,造成更大损失。(3)当故障发生时,对机组的各个区域和部件需进行全面的诊断。在此基础上,建立风电机组故障树模型。通过故障树模型生成了风电机组故障诊断知识库。介绍了三种故障推理机方式,并采用正反向混合推理设计分级存储方式下的故障诊断与状态评估推理过程流程图,实现了将复杂风机系统简单化,模块化。(4)本文最后通过分析机组齿轮箱系统和发电机系统正常和故障状态时对机组输出功率影响,结合故障树模型,发现即使细小的故障也能影响其正常运行。因此运用以上方法,当机组某个部件出现异常现象时,系统能够及时有效的发现故障原因并能尽快处理故障,为风电机组的故障诊断及处理提供一定参考。
[Abstract]:With the shortage of energy in the world, wind energy has played an important role in various situations. Wind power generation technology has been widely used in the world. However, due to the remote geographical location and complex maintenance technology, the wind power generation process is prone to failure, and the maintenance difficulty and operation maintenance costs are high. Therefore, wind power system operating state evaluation and other work is particularly important. The wind turbine system is huge, it is difficult to diagnose and maintain the wind turbine system, it is difficult to realize the accurate and complete diagnosis of the equipment by a single technology. In summary, it is necessary to solve these problems by a set of reasonable evaluation system for fault diagnosis and evaluation of wind power generation. Therefore, the main contents of this paper are as follows: (1) the basic structure and working principle of the wind turbine are expounded, and the fault mechanism of the wind turbine is analyzed according to the operating characteristics of the wind turbine. The common methods of fault diagnosis are studied, and the fault diagnosis methods for different signal sources are put forward. Finally, based on the measured operation data of wind turbine SCADA system, two kinds of fault diagnosis and state comprehensive evaluation methods are designed. (2) based on the normal operation and fault data of a certain wind field SCADA system, the entropy weight fuzzy comprehensive evaluation method is used. A comprehensive evaluation model is established for the unit. According to the established model, the operating state of the unit is divided into four grades: "excellent, good, medium and poor" to evaluate the health of wind turbine. When it is calculated that there is a subsystem with a "bad" running state, it should be regarded as a fault state, and it needs to be checked immediately to avoid the serious failure and cause more losses. (3) when the fault occurs, it is necessary to prevent the failure from becoming more serious. (3) when the fault occurs, All areas and components of the unit need to be fully diagnosed. On this basis, the fault tree model of wind turbine is established. The fault diagnosis knowledge base of wind turbine is generated by fault tree model. In this paper, three kinds of fault inference machine are introduced, and the flow chart of fault diagnosis and state evaluation reasoning is designed by using forward and backward hybrid reasoning, which simplifies the complicated fan system. (4) at last, by analyzing the influence of generator and gearbox system on the output power of the unit, and combining with the fault tree model, it is found that even small faults can affect the normal operation of the unit. Therefore, with the above method, the system can find the fault cause and deal with the fault as soon as possible when a unit has abnormal phenomenon, which provides a certain reference for the fault diagnosis and treatment of wind turbine.
【学位授予单位】：华北电力大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM315

【参考文献】