Under-sampled Vibration Signal Reconstruction For Blade Tip-timing At Non-constant Rotation Speeds

Blades are one kind of key rotating components in aero-engines,which have direct impacts on operation safety of aircrafts.At present,blade tip-timing(BTT)method has become an advanced technology for online blade vibration monitoring.Since BTT vibration signals are typically under-sampled,it must be reconstructed in order to obtain vibration characteristics of blades.Existing BTT vibration signal reconstruction methods are mainly based on the assumption of constant rotation speeds.In practice,however,the speed of rotating blades often fluctuates or changes,which leads to considerable reconstruction errors for BTT vibration signals and brings difficulties to identification of blade vibration parameters.Therefore,it's of great theoretical and practical value to study the reconstruction algorithm for BTT vibration signals at non-constant rotation speeds.In this paper,the types of non-constant speeds are divided into speed fluctuations and variable speed.At the same time,two BTT sensor configurations of uniform and non-uniform placements are considered.Thus four distinct signal reconstruction algorithms for BTT vibration signals are proposed.Main works of this dissertation are summarized as follows:(1)BTT sampling models under rotation speed fluctuations and variable speeds are built,respectively.Then basic principles of BTT vibration monitoring at non-constant rotation speeds are analyzed.(2)The paper uses the nonlinear time transformation method to approach inaccurate reconstruction problem caused by speed fluctuations.As for uniform and non-uniform distributions of BTT sensors,two algorithms for under-sampled BTT vibration signal reconstruction are then proposed.Numerical simulations are also carried out to validate the feasibility of two proposed methods.The results show that the proposed methods can effectively reduce reconstruction errors due to speed fluctuations.(3)In order to solve the problem that variable rotation speeds make BTT signal difficult to be reconstructed,the order analysis method was introduced.As for uniform and non-uniform distributions of BTT sensors,under-sampled BTT vibration signal reconstruction algorithms are derived in angular domain,respectively.Numerical simulations and experiments are done and the results show that proposed methods can effectively reconstruct BTT vibration signals under variable speed condition.In conclusion,compared with existing algorithms,proposed algorithms can better reconstruct BTT vibration signals at non-constant rotation speeds and identify blade vibration characteristics more accurately.