(121g) Some Applications of Acoustic Emission in Fluidized Bed

As the acoustic measurement has many advantages such as high sensitivity, safety, non-intrusiveness, being real-time and on-line, it is employed in many fields, including applied engineering, electronics, signal analysis and chemo-metrics, besides, it is also helpful to realize quantitative analysis and real-time supervision of parameter. The results of many researches show that much information, such as particle characteristic, bubble characteristic, condition parameter, and so on, is included in the acoustic signals of fluidized bed. In our experiments, we achieved the real-time and on-line detection of flow pattern, bed height, the initial fluidization velocity, particle size distribution, agglomeration as well as particle velocity, by collecting the information from the acoustic signals of fluidized bed.

The experiments were performed in a gas-solid fluidized bed with f150 in diameter and 1.4 m in height. The distributor was a perforated stainless steel plate with 0.002 m orifice diameter and 2.6% orifice ratio. The fluidizing gas was transmitted by air in normal condition. Polyethylene powders were used as fluidized particles with the density of 944 kg/m3. On the other hand, an online measurement and analysis system (UNILAB AE2003) were used, including the AE sensor, narrow-band receptors (SR-15), pre-amplifier, signal conditioning system, together with data acquisition and computer. The AE sensor was easily installed on the wall of bed with the help of coupling agent. The frequency of AE wave was measured up to 500 kHz. Each measurement lasts one minute and the values listed in this study are obtained as the arithmetic average over one minute. Blank experiment shows that the sounds of background have no effect on the measurement of particles collision or friction in fluidized bed.

Solid fluidization generally releases various AE signals, which are consisted of particle-particle or particle?chamber collisions impact sound, particle-particle or particle-chamber friction sound, along with air turbulence in fluidized bed. Therefore, the measurement and analysis of AE energy can reflect the real-time particle activity, or even reveal the flow pattern, mixing of particles in fluidized bed. Meanwhile, different scales decomposed by wavelet reflect different frequent signals. To be more exact, the small particles correspond to the high frequent signals in small scales, while the big particles correspond to the low frequent ones in big scales. So it is reliable to measure the particle size distribution and agglomeration, utilizing the multi-scale analysis of acoustic signals.

As it is shown in Fig. 1, there are solid circulations with smaller size, as well as with larger size (main circulation zone). The movement of particles with low AE energy will be less active in the interconnection area due to counterwork of two circulation flows with opposite flow direction on the wall area. The region is named as stagnant zone at the position of 0.24 m above the distributor. It can be anticipated that worse transfer of heat and then agglomeration will be appeared at the stagnant zone near the wall. Much experience from commercial operators confirmed this prediction.

The changes of acoustic energy distribution in every scale reflect the changes of acoustic main frequency. In detail, since the acoustic energy produced by particles in different sizes will accumulate, the acoustic energy distribution is different in each scale. Moreover, it is possible to predict the PSD by multi-scale analysis of acoustic energy distribution. The AE-PSD Model was on the basis of the principles mentioned above. When AE-PSD Model was used to measure the PSD in the fluidized bed in different heights, it was found that the AARD between AE-PSD method and sieves method was only 5%.

The initial fluidization velocity was presented by the change of the energy piece-rate of wavelet packet analysis of the acoustic signals from the granulation effect of the fluidization bed based on the wavelet packet analysis. Being compared with the initial fluidization velocity measured by the pressure drop method, the AARD was 5.18%. Being based on wavelet analysis, the acoustic signals of particles with agglomeration in different mass fraction in fluidized bed were decomposed into 7 scales, consequently, the quantity and size of agglomeration could be got, according to the variation of energy and peak energy of a7 scale wavelet analysis. Therefore, the analysis of that wavelet of acoustic signals can be used to forewarn the agglomeration.