基于EDFT的非均匀欠采样叶尖定时信号分析

针对叶尖定时信号因严重的非均匀采样和欠采样导致的谱分析难题,提出了基于扩展离散傅里叶变换(EDFT)的分析方法.不同于传统的傅里叶变换,EDFT是以傅里叶积分变换为目标,在扩展的频率范围内通过构造和优化变换基函数替代传统FFT变换中的指数基来实现非均匀和欠采样信号的谱分析,在分析中利用原始数据迭代和近似拟合保证了分析精...

Analysis of nonuniform under-sampling blade tip timing signal based on EDFT

Aiming at the problem of blade tip timing signal’s spectral analysis being difficult due to serious nonuniform sampling and under-sampling,an analysis method based on the extended discrete Fourier transformation(EDFT)was proposed.It was shown that being different from the traditional Fourier transformation,EDFT takes Fourier integral transformation as the goal and uses the optimized transform basis functions to replace exponential basis functions in the traditional FFT transform within an extended frequency range,and realize nonuniform sampling and under-sampling signals’spectral analysis;in analysis,the original data iteration and approximate fitting are used to ensure analysis accuracy;this method breaks through the limitation of Nyquist sampling theorem,expands the range of analysis frequency,and the number of spectral lines is no longer limited by the number of sampling points,so the frequency resolution is improved.In order to verify the feasibility and reliability of the proposed method,the data generated by the sampling mathematical model of the blade tip timing system and the test data of the test bench were analyzed,respectively.The results showed that using either simulated data or test data,when the number of sensors is more than or equal to 3,the proposed method can accurately analyze and obtain spectral signals of blade vibration,so the proposed method can effectively solve the problem of spectral analysis of nonuniform under-sampling blade tip timing signal,and has good anti-interference performance.