Vibration After an Impact

Use the measured hammer force to locate the impact. Follow the accelerometer response, compare sections of the decay and use a periodogram to identify prominent frequency bands.

Impact test and unit conversion

Saullo G. P. Castro, Delft University of Technology, Impact hammer test of an Aluminium 6082 T6 plate, 260 × 246 × 1.5 mm, free boundary conditions (2023), CC BY 4.0. The plate was suspended on soft cotton threads to approximate free edges. Impact point 13 is at (0.130, 0.123) m and the single-axis accelerometer is fixed at point 5 at (0.260, 0) m in the author’s XY coordinate system. The selected file is Waveforms Point13.tdms, group Configuration0_Impact00. Zenodo record and protocol · Original TDMS file

force_n
Hammer load cell channel cDAQ1Mod1_ai0, in N.
accel_m_s2
Synchronized channel cDAQ1Mod1_ai1, converted from g to m/s² with the author’s factor of 9.81.
time_s
Derived from the common wf_increment of 0.0001171875 s. The sample rate is 8,533.333… Hz. All 32,768 samples are retained. Time zero is the maximum absolute force, at zero-based index 9. The preceding segment contains 9 samples, about 1.05 ms. The file records 8 pretrigger samples relative to its own trigger, which differs from the selected force maximum.
Measured hammer force around time zero
Plot of the same measured hammer-force channel. Time zero identifies the maximum of about 5.43 N, with the preceding samples visible.

Quality check, envelope and periodogram

  1. Import the Dataset. In Signal Analysis, choose plate_impact, set accel_m_s2 as the primary signal and time_s as the time coordinate. Leave Paired blank and the declared rate at 0.
  2. Click Analyze with quality.basic. Read n, peak, rms and crest_factor in the report. This method produces a scalar summary.
  3. Switch to analytic.envelope, click Analyze and publish with an unused ID, for example real.plate.envelope. amplitude still uses m/s². Compare sections of the decay on the same scale.
  4. Choose spectral.periodogram, click Analyze and publish with real.plate.periodogram. The full-record FFT has 32,768 points, 0.2604167 Hz bin spacing and 16,385 one-sided bins up to about 4,266.67 Hz. The Hamming window weights the start and tail differently. Use the peaks as indications of frequencies present in this transient record.
Signal Analysis with plate acceleration and the quality check method selected
The quality check returns scalar metrics for the full record before envelope and periodogram analysis.

Read the decay and frequency peaks

The summary reports an absolute peak acceleration of about 397.23 m/s², RMS of about 21.39 m/s² and a crest factor of about 18.57. The crest factor expresses the large difference between the peak and the RMS level of the full recording. The synchronized force trace locates the excitation. It has several lobes in the first few milliseconds, including a local peak of about 5.03 N at +4.10 ms. The first window used to compare the decay starts at +20 ms, after the strongest force lobes. The acceleration is still oscillating in this window.

Window after peak forceEnvelope RMS (m/s²)dB relative to the first window
[0.02, 0.07) s149.100.00
[0.50, 0.55) s16.94−18.89
[2.50, 2.55) s0.43−50.77

The four highest local peaks in the periodogram are at about 1,214.32, 972.14, 740.10 and 360.16 Hz, in descending PSD order. The reference results include the full frequency grid.

Acceleration envelope after impact on a logarithmic scale
The published envelope shows the decay after the force maximum. The last few samples are sensitive to boundary effects from the Hilbert transform.
Periodogram of the measured plate acceleration
The viewer published this periodogram. Peak heights depend on the response measured at point 5 and on record processing.

The observed frequencies characterize this particular impact and sensor pair. Modal characterization and frequency response function (FRF) estimation would require dedicated measurements and analysis, including several comparable impacts. This exercise uses a single recording with a force of finite duration. When interpreting the peaks, consider the measured shape of the excitation and the location of the sensor. CC BY 4.0, attributed to Saullo G. P. Castro

File attribution and license