Once the license terms have been confirmed, the application is ready for use (a). In the app, default settings have been selected that provide the most precise and high-resolution results possible for the most common devices in terms of sampling rate, without impairing smooth operation (b). Depending on the instrument and the measuring situation, adjusting the settings can improve the results considerably. A sample rate of 50 Hz is preset for the acceleration recording. For the acquisition of higher-frequency signals, however, an increase in the sampling rate is strongly recommended. The performance limits are with sensors with lower resolution, mostly these can be found with older or inexpensive devices, but quickly reached. By reducing the FFT resolution (resolution), which is preset to 256, the computing load can be reduced and the capacity freed up can be used in favor of a higher sampling rate.
By clicking the Start button, the measurement starts, the sensor data is acquired and displayed in the acceleration-time diagram. Furthermore, the corresponding frequency spectrum is displayed parallel to the time signal (c). Another click stops the measurement. The data can now be evaluated and analyzed. By pressing the table button, the results can be viewed in tabular form (d). The time steps of the signal sampling and the recorded accelerations in the direction of the three coordinate axes (e) are listed.
During the measurement, the maximum recorded acceleration amplitudes and natural frequencies are displayed directly in the diagram for the respective directions. After stopping the measurement, there are different evaluation options. By touching two points in the time signal, a range can be selected in which the attenuation is to be determined. A third tap determines the damping using the decaying amplitude method (c). This requires clearly separated eigenmodes. By clicking the Analysis button, the maximum acceleration and velocity amplitude as well as the maximum weighted vibration intensity KBFmax according to DIN 4150-2 can be displayed. In addition, an evaluation of the vibrations according to ISO 10137, VDI 2038 and DIN 4150-3 is integrated (f).
4. Saving and Loading the measurement
The measurements can be saved in csv format using the Save button (e.g. for an import into a professional vibration processing system). The file can be named as desired (g). By default, the system suggests date and time of the measurement. It is also possible to recall stored measurements using the Load function in the "iDynamics" app. [Abbildung Namensauswahl Speichern]
In order to achive clear and precise measurement results with (less sensitive) mobile devices, post-processing of the raw data can be advantageous.
Several filter algorithms are integrated in the app for the post-processing of the recorded measurements. Special algorithms for high-pass, bandpass and low-pass filters with Butterworth characteristics are implemented in accordance with the German standard DIN 45669-1.
The following figure shows the application of the filters to the vibration measurements of a frame with two degrees of freedom. With the filter option the vibration modes can be seperated and evaluated quickly and easily.
With small vibration amplitudes, the acceleration recordings of smartphone sensors show a relatively small signal-to-noise ratio. In order to reduce the disturbing influence of noise, a 5-point smoothing algorithm was implemented, which calculates the mean value of the four nearest values. Up to three smoothing iterations can be performed. However, caution is advised, as low sampling rates can cause high-frequency signals to be distorted and "smoothed out".
The following figure shows the difference between a measurement before and after smoothing. It can be seen that using the smoothing algorithm with a 1.8 Hz oscillation leads to excellent results, while with a 12.5 Hz oscillation it has a negative effect. The measurement shown here was made with a sampling rate of 50 Hz.
6. Current User Limits
Since Android version 6.0 (2015), Google requires the installation of a 12-bit accelerometer for smartphones with acceleration sensors and recommends the use of a 16-bit sensor. By adhering to these compatibility requirements, most recent smartphone sensors have a sufficient resolution. To compare the suitability of mobile devices for measurement the effective value of noise can be used. To evaluate this value, the smartphone should be placed in a very quiet environment without any disturbances from environmental vibrations (e.g. basement of a bunker), recording the acceleration for a long time (several minutes). In the app, under the tab “Analysis” the effective acceleration value aRMS(t = ∞) over the complete measuring time can be computed. The following table lists the effective noise for several typical smartphones.
A detailed description of the app with example applications and evaluation possibilities can be found in the German Bauingenieur Journal 05/2017*. A publication in English is in progress.
*Sadegh-Azar, H.; Feldbusch, A.; Agne, P.; Kögel, C.: Schwingungsuntersuchungen mit dem Smartphone und Tablet, Bauingenieur, Mai 2017