基于振动信号的堆内构件故障诊断研究

The reactor internals fault diagnosis based on vibration signals

反应堆堆内构件在高温、高压和高辐射的环境条件下运行,长期受冷却剂的高速冲击,一旦发生事故会影响压水堆机组的正常运转,甚至发生更高级别的核事故.以反应堆压力容器振动信号为研究对象,实现对异常振动行为的故障诊断十分必要.利用ANSYS Workbench堆芯吊篮和压力容器进行实体建模,得到吊篮梁式振动模态和壳式振动模态.通过瞬态分析模块,利用白噪声激励模拟正常工况下流质振动的情况,叠加冲击力的条件下模拟LOCA状态下的受力情况,得到正常和异常状态下压力容器上垂直方向加速度时程响应数据.在LabVIEW虚拟仪器平台上,利用分形关联维的方法对处理过的振动信号进行故障诊断,通过比较正常状态及故障状态时的关联维数,可以快速有效地进行故障模式的识别.

The reactor internals often work under high temperature, high pressure and strong radiation, as well as long-term high-speed impact by coolant during the reactor operation. Once in an accident condition, the internals will affect the normal operation of a nuclear power plant （ NPP ）, and even it is necessary to analyze the reactor pressure vessel vibration signals, thus cause higher level nuclear accidents. So detect and diagnose the abnormal vibra- tion behavior of the reactor internals. Using ANSYS Workbench reactor barrel and pressure vessel to build solid model to get beam vibration mode and shell vibration mode of reactor barrel in this paper, and then white noise ex- citation is used to simulate liquid vibration in the normal work condition, and to simulate the forces in LOCA state in superimposed impact condition. In this way, under normal and abnormal state the acceleration time response data of the pressure vessel in vertical direction can be obtained. The processed vibration signals are used to diagnose fault by fractal correlation dimension method on LabVIEW virtual apparatus platform. By comparing the correlation dimension of normal state and fault state, the fault modes of the system can be identified fast and effectively.