Research on Test Conditions of Sealed-Relay Remainders Based on High-Speed Video Technology - Intelligent Computing for Sustainable Energy and Environment - page 511

Table 3. Performance verification of h (mm)

measured values
actual values

1 2 3 4 5 6 7 8 9 10

9.7 (mm) 9.72 9.75 9.66 9.67 9.69 9.69 9.73 9.70 9.68 9.71

4.3 (mm) 4.35 4.32 4.33 4.27 4.31 4.30 4.30 4.33 4.28 4.27

1.5 (mm) 1.66 1.56 1.62 1.42 1.58 1.46 1.70 1.52 1.41 1.63

From the experimental results, the calculation of R is almost accurate; the error is
less than 0.06mm. The calculation of h is comparatively accurate when h>R , the error
is less than 0.05mm, but when h≤ R, the calculation error may more than 0.2mm as
the remainder will interfere the location extraction of the cavity in this case.

3 Study of Remainder Test Conditions

The movement of remainders in a sealed cavity is affected by the height of the cavity,
vibration acceleration, vibration frequency, the features of the remainders and energy
coefficient of restitution between remainders and the cavity, etc. Due to the gravity,
collision between remainders and the bottom wall is always stronger than that of the
upper wall, so collision between remainders and the bottom wall is mainly concerned
in experimental studies. In the following, the average relative speed before collision
between remainders and the bottom wall is denoted as y average ; the height of the cavity
as H, the diameter of remainder as R , the test vibration frequency as f, the test
acceleration peak value of vibration as a m, and further we assume the upward direction
is the negative orientation.

3.1 Study of Relationships between Test Conditions and Output Signals Based

on Transparent Cavities

1) Experiments with Varying Frequency
Experiments with varying frequency were carried out. Bearing Steel balls were
selected to be remainders, for which R is1.5mm and 1mm respectively; the dimension
of transparent cavities are 30mm × 30mm × 40mm and 20mm × 20mm × 20mm
respectively. Remainders are put into transparent cavities and tested with varying
frequency. The bandwidth is between 10Hz and 50Hz; f was step changed by 10Hz
each time; and a m is 5g;We have captured 200 collision images for each combination
of test conditions, remove images with h ≤ R, and then select 50 images with a
collision between remainders and the bottom wall for movement analysis.
It was found that when f is set below 20Hz, the vibration generator may become
unstable; if above 110Hz, the bounce height of remainders is less than its own
diameter due to small amplitude of the vibration table, which is disadvantageous for
detection. Therefore, the actual range of f is chosen between 20Hz and 110Hz. The
y average results of experiments with varying frequency are shown in Table 4, where the
experimental results of approximate periodic collision vibration are recorded in italic.