SOURCES OF PEAK SIGNAL VARIATION FROM A THIN PLATE IMPACT SENSOR MEASURING PARTICLE RADIUS
This thesis identifies and examines the sources responsible for the variations in the peak signal from a thin plate impact sensor used to measure particle radius. Steel spherical particles, dropped onto a thin aluminium plate that is simply supported on all sides, cause plate vibrations that are detected by a piezo-electric transducer. The peak positive signal from the sensor is used as a measure of the radius of the impacting particle. The thin plate impact sensor is a simplified version of a sensor from a commercial device known as the grain loss monitor. Earlier experiments using the grain loss monitor sensor to measure the radius of particles, reported an unexplainable peak sensor signal variation. An Experimental and a theoretical analysis of this problem using the thin plate impact sensor, shows that small changes in impact position result in large changes in the peak sensor signal. Experiments using six steel spheres, ranging in radius from 3/32 inches to 8/32 inches, impacting onto a square 1/32 inch thick aluminium plate, show that the peak sensor signal and the degree of sensitivity of the peak signal to changes in impact position depends on. the location of the impact site. Specifically, for the impact locations tested, this sensitivity exhibits a minimum at approximately 2 cm from the transducer and a maximum sensitivity at the transducer site. Therefore, using these results, a hypothetical change in the impact position of approximately 0.1% of the lateral dimension of the plate, or about 0.5 mm, results in the peak sensor signal exhibiting changes, for all the particles tested, that average 4.7%, 3.9%, 1.9%, and 4.5% of the peak at impact locations at approximately 0 cm, 1 cm, 2 cm, and 3 cm from the transducer, respectively. The peak sensor signal also shows an almost linear response to changes in particle radius at a.ll impact positions tested, indicating that the peak signal of this particular sensor configuration provides a measure of particle radius. A theoretical model and analysis of the thin plate impact sensor identifies the factors that influence the generation of the peak sensor signal and indicates that changes in the impact position are the most likely cause for changes in the peak sensor signal. Numeric simulations, using the theoretical model to approximate the experimental thin plate impact sensor, confirm that the sensor signal and its spectra change with impact location and particle radius to a degree that is similar to that seen in the experimental results. However, the simulated peak sensor signals are not accurately modelled and is due to the simple transducer model having difficulty reproducing the large signal response of the transducer. An extension of the theoretical and experimental results to the grain loss monitor sensor in sizing steel spheres suggests that unperceived changes in impact position of approximately 0.4% of the sensor's lateral dimension, or changes in distance of about 0.25 mm and 0.5 mm in the shortest and longest lateral plate dimensions, respectively, could result in a standard deviation of 16% of the mean peak sensor signal.
Master of Science (M.Sc.)
Electrical and Computer Engineering