Art Illustration
If you are a test engineer and you plan to perform vibration testing on your product, you already collected data from actual vibration environment. Then, what you need to perform next will be Signal Post Processing.
Since we are dealing with vibration, then we will be more looking at dynamic signal post processing. Unlike other measurement such as temperature, humidity, or air velocity, where frequency is less important, most of the time (if not always), we will be dealing with frequency parameter.
The type of post processing that we need will be quite depending on:
The primary objective of data analysis.
Failure mode expected to happen.
Type of sensors used
Accelerometer, Velocity Sensor, Displacement Sensor
When we are using these types of sensors, most probably we expect to see failure which is related to overall vibration level, such as broken solder joint, signal connection loss, or display malfunction due to excessive vibration. When we are dealing with high frequency, then normally we will use accelerometer since acceleration is very dominant at high frequency. In the other hand, we may want to try to use velocity sensor or displacement sensor for low frequency measurement.
However, there are many cases when we need to measure the displacement level while no displacement sensor is available or vice versa. These might be due to difficulty to install, limited space, or even due to limited budget. In these cases, we may want to perform Integration/Differentiation post processing to the signal. Though, it may not be as easy as we find it in the theory. When we want to perform Integration, we may want to pay attention to the high pass filter setup. In the other hand, when we are doing Differentiation, we may want to be sure that the floor noise is low enough for us to achieve acceptable result.
Accelerometer Velocity Sensor Displacement Sensor
Strain Gage
When we are using strain gage for measurement, normally we will be interested in the fatigue life calculation of our structure, expecting to see when the crack will start to initiate. Since there are three configuration of strain gage measurement: quarter bridge, half bridge, and full bridge, then we may want to make sure that our DAq is able to cover these three configurations. In complex structure, sometimes it is not easy to deploy half or full bridge strain gages.
When we are using one strain gage (quarter bridge), we will also face a challenge on determining the orientation of the strain gage. It will be quite helpful if we can get a hint from Finite Element Analysis result. When the strain gradient is low, we may want to try to use strain rosette in order to eliminate difficulties due to strain gage orientation. But then we may need to perform mohr circle post calculation.
Strain Gage (Gauge)
Dynamic Signal Post Processing
During vibration analysis, most common post processing feature we use will be Fast Fourier Transform (FFT), to check the frequency content (spectrum) of our signal. When we are dealing with random signal, normally we will use Power Spectrum Density (PSD). For these types of post processing, we may want to properly set frequency resolution, windowing type, averaging, etc to be able to obtain meaningful result. For even deeper analysis, we can go for Frequency Response Function (FRF/Transmissibility) or even Waterfall (Order Tracking) analysis. By doing this, we may be able to see the dynamic characteristic of our structure (resonance, damping, Q factor, etc). For short duration transient shock pulse, sometimes we want to use Shock Response Spectrum (SRS) to check the frequency content. For fatigue analysis, especially for random signal, we can perform rainflow counting analysis. Combining this with S-N fatigue curve, we may be able to calculate the fatigue damage.
Power Spectrum Density (PSD)
