(3) Vibration method
Next, we will consider the vibration method and the type of sensor.
For light and small objects, non-contact sensors are advantageous. Gap sensors and, more recently, laser Doppler vibrometers are commonly used, but when suspended in mid-air, the object's vibration can sometimes prevent accurate measurement. Also, for heavy and large objects, impacts with impulse hammers may not provide sufficient excitation due to the instantaneous energy, resulting in insufficient excitation force. In such cases, a vibration exciter is used. If a vibration exciter is not available on-site, an emergency measure is to randomly strike with an impulse hammer to generate energy while measuring, but in any case, it is important to measure the data without affecting the vibration characteristics of the object. Knowledge of techniques such as averaging to improve the signal-to-noise ratio is also necessary. If the sensor and excitation method are appropriate, good data can generally be obtained.
(4) Coherence function
We measure the coherence function to determine whether the data is of good quality.
In my personal opinion, a coherence function of 0.9 or higher is desirable. If the coherence function is small, check whether sufficient excitation was performed using power spectra, etc. If the coherence function is still small, it often indicates the presence of nonlinear elements, and the entire measurement system should be reviewed to eliminate them. For example, if there is play, secure the fixing firmly. If there is friction, reduce the excitation. The difficult part is when the material properties of materials such as rubber are nonlinear, but in this case, we adjust the excitation force to stabilize it.
As described above, we strive to perform measurements under conditions that offer the highest possible degree of linearity.
(5) FFT analyzer
Let's consider measurement techniques for FFT analyzers. Ideally, we should be able to measure a frequency response function (transfer function) with clear peaks and valleys. If this isn't the case, we should reconsider the FFT window selection. For response measurements using an impulse hammer, a rectangular window is used. If vibration damping is long, an indexed window may be used for Accelerometer signals. In many cases, unsatisfactory data obtained from impulse response measurements was due to using a Hanning window.
Furthermore, averaging is performed, and care is taken to avoid double hammering with the impulse hammer. To confirm these points, the measurement proceeds while monitoring the time-domain waveform and power spectrum to ensure that the correct waveform is being captured. After the measurement is complete, the display is switched to the frequency response function and coherence function for analysis.
(Excerpt from the email newsletter issued on June 17, 2004)
