Research On Key Technology Of Crack Detection In Rotating Blades By Non-contact On-line Blade Tip-timing Method

Rotated blades are key components in aero-engine,which operate in extreme conditions and usually change operating conditions.These can lead to blade vibration and cause fatigue damage.Therefore,blade vibration measurement during operation is of great importance from the perspective of safety and reliability.Blade Tip-Timing(BTT)is considered a promising technique for monitoring blade vibration non-contact and on-line for aero-engine.However,due to the restriction of sensor structure and operation requirements,the measured tip vibration signals are all under-sampled signals.But multiple-frequency vibration of the cracked blade are caused by nonlinear of the crack.Therefore,it is a key problem of crack detection in rotating blade by BTT method that how to use under-sampled BTT signals to obtain unknown multi-frequency blade vibration.In this paper,a dynamic model of blades with a breathing crack is established,which is used to analyze the crack on vibration characteristics of the cracked blade.Then,crack-sensitive vibration characteristics are extracted with unique features of BTT method.Simultaneously,compressed sensing is proposed to solve blind reconstruction of multi-frequency blade vibration from under-sampled BTT signals.Finally,vibration measurement and crack detection in rotating blades are achieved by non-contact on-line BTT method.The main works are shown as follows:(1)A single blade dynamic model with a breathing crack is established based on crack mechanism,which is used to analyze the effect of the crack on blade vibration characteristics.Then,nonlinear superharmonic characteristics of the cracked blade are extracted for under-sampled multi-frequency vibration reconstruction and crack detection in rotating blades.Simultaneously,a problem is discovered that the nonlinear multi-frequency of blade vibration displacement are insensitive to micro-crack.So power flow analysis method is proposed to improve the multi-frequency vibration of the cracked blade by micro-crack,which is validated by vibration test of the cracked blade on the vibrating table.This phenomenon can provide theoretical guidance for micro-crack detection in rotating blades.(2)A coupled multi-blade dynamic model with a breathing crack is established with mistuning of the lacing coupling blade-disk.Vibration characteristics of coupled blades are studied by mistuning and coupling strength,and then vibrational effect of mistuning blades are analyzed by the crack.Resonant frequencies,amplitudes and its variation factors of mistuning blades are extracted as crack features,which lays the foundation for crack detection in rotating blades by BTT method.(3)Aiming at the under-sampled characteristics of BTT measured signals,the compressed sensing model of BTT measurement is established.Then,the number of BTT sensors in actual measurement is determined by the constraints of solving the model.In addition,the condition number theory is used to optimize layout of BTT sensors.Finally,performances of multi-frequency blade vibrations reconstruction are studied by different sparse reconstruction algorithms.So the problem is solved by compressed sensing that unknown multi-frequency blade vibrations are reconstructed from under-sampled BTT signals.(4)Blade-tip vibrations of cracked blades model with mistuning are sampled by equivalent BTT method.Then,under-sampled signals are reconstructed by the compressed sensing.Simulation study of crack detection in rotating blades are achieved that crack characteristics of second harmonic response,resonance frequency and amplitude variation are extracted from reconstruction signals.Finally,the BTT on-line vibration measurement experiment platform is supported by aerodynamic excitation and magnetic excitation,which are used to verified multi-frequency blade vibrations blind reconstruction from BTT under-sampled signals solved by the compressed sensing theory.Using pre-cracking method,combined with compressed sensing and crack characteristics to achieve crack non-contact on-line detection in rotating blades by BTT method.