Experiments of Bubbling Fluidized Bed using ECVT Measurements

Abstract

Gas-solid bubbling fluidized bed (BFB) is widely applied in industrial processes such as combustion, polymerization, cracking, etc. In BFB, gas-solid flow depends on formation of bubbles and its characteristics such as shape, size and frequency. This study estimates bubble size in a lab-scale BFB (Figure-1a) at three different fluidization velocity using non-invasive measurements from electrical capacitance volume tomography (ECVT). The experimented bed is a cylindrical column of 20 cm diameter and 120 cm height. The measurement section covered by ECVT sensors is 35 cm height and placed at 5 cm above the distributor. Typically, ECVT or any other tomographic technique provides volume fraction distribution in the test section at a resolution that depends primarily on sensor configurations and reconstruction algorithm. In this experiments, the ECVT sensor has 24 electrodes (Figure-1a) which gives 3D image having 20 × 20 × 20 voxels. At such low resolution of the ECVT, estimation of bubble size strongly depends on a threshold solids volume fraction that is used to distinguish the bubble from the dense phase in post-processing of data. Previous studies used a single threshold value to postprocess the captured data. This study finds that use of a single threshold value to post-process all captured data leads to inaccurate estimation of bubble diameter and sometimes can’t even capture smaller bubbles. This has been a major drawback of previous studies on estimation of bubble size using ECVT. In this work, we conducted a series of calibration experiments of static bed with a known size beaker at the centre of the bed. The calibration experiments show that lower values of the threshold underestimate the beaker size, whereas higher value of threshold overestimate the size. It is found that higher value of threshold should be used to capture the small size beakers. The measurements from these calibration experiments are further used to develop an iterative algorithm to determine optimum threshold value for a given ECVT image. The developed algorithm is used to estimate bubble size under dynamic fluidization. It is found that variable thresholding from the iterative algorithm is able to capture bubbles as small as 2.3 cm in 20 cm diameter bed and also gave estimated bubble size consistent with those calculated from empirical correlations with ±20% error.