Unit 44 Vibrationdata Sine Filtering Introduction Vibrationdata We

Unit 44 Vibrationdata Sine Filtering

Introduction Vibrationdata • We have covered time-domain, sine curve-fitting in previous units for signal identification • This unit will use this method for a new purpose, to separate sine and random components of a signal • The curve-fitting is done using random number generation and trial-anderror, with some convergence built-in • A measured sine vibration typically has some slight variation in amplitude and frequency, thus require a few sinusoids to match the data • The following method works best for “short “ time segments • Could be used to remove 60 Hz noise from measured data

Exercise 1 • Generate white noise • Add sine component • Then separate sine and random components Vibrationdata

Vibrationdata Save as: white

Displacement Vibrationdata

Displacement Vibrationdata Save as: combined

Sine-on-Random Signal Vibrationdata

Sine-on-Random Histogram Vibrationdata

Displacement Time History > Filters, Various > Remove Sine Tones Vibrationdata

Comparison Vibrationdata Results Case Amplitude fn(Hz) Phase(rad) 1 1. 0169 99. 9989 0. 0471 Standard Deviations original: 1. 232 synthesis: 0. 7191 residual: 1

Comparison, Close-up Vibrationdata

Original minus Sine Synthesis Vibrationdata

Exercise 2 Bombardier Q 400 Turboprop Acoustics Vibrationdata • The PW 150 A engine/propeller rotation rate during takeoff and climb is 1020 RPM, but is throttled back at cruise altitude to 850 RPM, or 14. 17 Hz • There are six blades on each engine, so the blade passing frequency is 85 Hz

Displacement Read in file: Q 400 Vibrationdata

Vibrationdata

Spectral Peaks Vibrationdata Next step: Separate blade passing frequencies from background noise

Displacement Vibrationdata

Results for 85 Hz Component Vibrationdata Results Case Amplitude fn(Hz) 1 0. 4765 85. 0661 2 0. 1591 85. 1877 3 0. 1311 85. 0024 4 0. 0601 84. 8849 5 0. 0541 85. 3313 6 0. 0328 85. 5232 7 0. 0269 85. 1293 8 0. 0264 85. 8288 9 0. 0249 84. 7275 10 0. 0188 84. 4948 • Phase(rad) 4. 5830 1. 7794 1. 1945 0. 9296 4. 5783 5. 4033 4. 7223 3. 8821 4. 0180 0. 4494 Diminishing returns effect as additional cases are included for a given frequency

Comparison Vibrationdata Difference is due to background noise and some spectral components unrelated to blade passing frequency

Original minus Synthesis Vibrationdata

Residual FFT • • Vibrationdata Peak at 354. 2 Hz is unrelated to blade passing frequency Could repeat exercise with more aggressive sine removal