Study On Modeling And Evaluation Of Vibration State Of Hydropower Unit Based On Linear Regression

Hydropower plants undertake the responsibilities for frequency modulation, peak modulation, phase modulation and emergency standby in the grid. Its operation security directly relates to the grid security and stability. In the event of accident, it will bring significant economic losses and serious social influences. Hydropower generating unit is one of the core equipments of hydropower plants, and its operation stability is of great importance to the operation security of hydropower plants. The vibration state of hydropower unit can directly reflect the running status of the unit. The vibration level is one of the main indicators characterizing the operation stability of hydropower unit. Fault Symptoms of the unit tend to reflect on the vibration state of the unit. When the vibration of the unit exceed certain limits, it will affect the security and stability of the unit seriously. Therefore carrying out condition monitoring and evaluation of vibration state according to the data analysis of the vibration of the unit, is one of significant approaches to ensure the security and stability of the unit, and it’s the foundation of on-line fault diagnosis and condition based maintenance for the unit.Similar turbines refer to the turbines of which water flow movement in the flow components remain similar, which satisfy the geometric similarity, movement similarity and dynamic similarity conditions. Because of the similarity of water flow in the flow components, the acting force on the runner blade and turbine shaft is proportional. Correspondingly the ratio of dynamic component of stress to static stress on the surface of runner blade which caused by acting force of flow and rotational torque on the runner blades as well as the ratio of effective value of vibration displacement to static deflection are constants. An indirect confirmation of this proposition is a fact that for similar turbines occurs the constancy of relative turbine power loss taking into account that the power loss which goes to maintenance of turbine vibration is characterized with relationship of dynamic and static stress components. So for similar turbines, relative power consumption caused by the vibration which result from hydraulic factors is also constant. In view of this, a nondimensional parameter V*which characterizes the vibration performance of similar turbines was put forward. Then a regression model of vibration of hydropower unit is established based on V*.2regression models were carried out by using the linear regression theory in this paper taking oscillation of turbine guide bearings of Francis machines for example. The significance tests for the2regression models were performed, and after comprehensive comparison and analysis a regression model between the vibration parameter V*which characterizes the vibration performance of similar turbines and unit speed n11was finally established. The range of vibration was divided into A, B, C, D4areas according to different confidence levels to establish evaluation standards for evaluation of oscillation of turbine guide bearings of Francis machines. Then, regression models of other types of vibration of Francis turbine, including oscillation of generator guide NDE bearing(non-drive end)and generator guide DE bearing(drive end), vibration of upper and lower frame of generator, vibration of turbine head cover, were established in the same way. The ranges of vibration were divided into several areas according to different confidence levels to evaluate vibration of Francis turbines.