Non-contact On-line Detection And Diagnosis Of Abnormal Vibrations Of High-speed Rotating Blades

The rotor blade is a critical component in the gas turbine. High-speed rotatingblades under high temperature and pressure conditions, it is subjected to centrifugalforces, aerodynamic cycle alternating load and dynamic loads, so it is easy to generatevibration. The vast majority of blade damage accident is caused due to various abnormalvibration. Therefore, the online detection of the abnormal vibration of high-speedrotating blade has important significance. To this end, the cracks in the blade vibrationfault failure are regarded as an example, and the simplified blade-disc is regarded as theresearch object, the non-contact on-line detection and diagnosis of the abnormalvibrations of the high-speed rotating blades are carried out, the main work includes:(1) In order to analyze the impact of the crack on the blade vibration behavior,using the Euler-Bernoulli beam theory to establish the dynamic model of the crackedblades, derive its dynamic frequency expression, and a finite element model simulation.The results show that the crack lead to the blade dynamic frequency decreases, theamplitude increases, and closer to the root of the blade, the greater the impact.(2) High-speed rotating blade vibration displacement expression based on threeblades tip timing sensor is derived. Shannon sampling theorem signal reconstructionmethod for subsampling characteristics and smooth vibration signal preprocessingmethods are studied, and it provides data for the subsequent blade cracks featureextraction and fault diagnosis.(3) Blade vibration frequency extraction method based on the remainder theoremunder sampling signal, and the blade modal frequency extraction method based on thestate-space model under reconstructed signal. The blade crack fault diagnosis methodbased on the modal localization is proposed, and it can effectively identify differentcrack state.(4) A high-speed rotating blades experimental station is built, and a non-contactblade vibration signal on-line detection system is built.With the use of the artificiallyinjected blade crack, the experimental verification is completed.The results show that the research method of this paper can identify abnormalvibration of high-speed rotating blades, and detect the different degrees of crack failure.It can be expected for the online monitoring of the state of the gas turbine rotor blades.