This research investigates possible monitoring methods to non-invasively detect blockages in the cyclone of milk powder spray driers to avoid costly production shutdowns. Two possible solutions have been identified. These are guided wave and vibration analysis. This research focuses on vibration analysis method, based on variation in the natural frequencies and/or damping ratios caused by blockages. Experimental simulation studies the performance to assess the viability of the vibration analysis method in identifying blockage in the cyclone. To test this method, a 1.87m long vertical cylindrical steel tube with both ends simply supported was setup as a prototype and flour was used as the powder conveyed by the tube. Analytical, numerical and experimental methods were implemented on the prototype using vibration analysis techniques. Experimental resonant frequencies of the empty shell from the impact hammer excitation were compared with analytical and numerical solutions to analyse the modal shapes. The first bending mode was determined as 79 Hz, which decreased with the amount of added mass and location closer to the middle point. This was validated by the simulated mass experiment, which also matched with the beam-mass theory. Flour induced experiment further validated the variation of the first bending mode. Experimental 270Hz, 380Hz and 398Hz resonant frequencies increased with the amount of added mass, which was validated by the simulated mass and flour induced experiment. Besides the variation of the natural frequencies, the damping ratio was also studied and quantified using the Hilbert transform envelope curve method. Onsite monitoring of the vibration of the cyclone in a spray dryer was done at Fonterra Te Rapa and clear resonant frequencies were obtained. The research results from the prototype demonstrate that vibration analysis as a non-destructive method to detect the milk powder deposition or blockage in the cyclone is possible and promising. However, more work is required before industry application.
| Identifer | oai:union.ndltd.org:ADTP/289428 |
| Date | January 2008 |
| Creators | Li, Li |
| Publisher | AUT University |
| Source Sets | Australiasian Digital Theses Program |
| Language | en_NZ |
| Detected Language | English |
Page generated in 0.0013 seconds