Estimation of bubble column hydrodynamics: Image-based measurement method

Abstract

Bubble column reactors (BCRs) have proven their worth in the chemicals industry by achieving high heat and mass transfer rates at relatively low cost. Limiting the design and scale-up, however, are the complex hydrodynamics that they exhibit, whose influence on transport characteristics are not fully understood to date. In addition, most studies in the literature focus on single phase system, whereas typical industry systems are multiphase.In this paper, a better understanding of the BCR hydrodynamics is obtained through investigating the coupling of Electrical Resistance Tomography (ERT) and Dynamic Gas Disengagement (DGD) technique. The increase in a Population Balance Model (PBM) simulation accuracy is achieved by allowing for the inference of bubble population property distributions such as bubble size and axial position. However, using a PBM to simulate bubble swarm phenomena under the influence of bubble coalescence and breakage requires accurately known boundary conditions. These are gas void fractions, the bubble size profile and the bubble number density distribution (BNDD).Electrical Resistance Tomography (ERT) is a non-invasive imaging technique suitable for creating images of changing gas void fractions in BCRs. In this work, a steady-state two-phase air-water system was set-up in a BCR. The ERT data for analysis of the local disengaging gas volume were captured over four rings of electrodes on a column of height 1.545m and diameter 0.29m. To facilitate measurements of bubble population distributions, a Dynamic Gas Disengagement (DGD) approach was used to induce transient gas holdup fractions which were captured by the ERT apparatus.The time profiles for the disengagement of gas void fractions locally were calculated for superficial gas velocities (ugs) in the range 5.0×10−3 to 1.0×10−2ms−1 (bubbly flow regime). From the known time profiles and height of the column sections, the bubble population properties were calculated. At ug of 1.0×10−2ms−1; the local average bubble rise velocities were found to range from 2.0×10−1 to 6.5×10−1ms−1; the local Sauter Mean Bubble Diameters (SMBD) were found to range from 2.3×10−3to 2.5×10−2m; and the axially averaged BNDD of bubble sizes were found to range from 3.6×10−3 to 7.6×10−6m−1.The smallest bubble size classes were found to be 2.3×10−3m at ug in the range 7.0×10−3 to 1.0×10−2ms−1 with relative error of 4% compared to the theoretical prediction of 2.2×10−3m . At ug in the range of 5.0×10−3 to 6.0×10−3ms−1, the smallest bubble size classes were determined to be 1.65×10−3m yielding the relative error of 37% compared to the theoretical prediction. The determined BNDDs were log-normal distributions and are consistent with both theoretical predictions and experimental findings. The low error values of the obtained results indicates the method is suitable for the development of PBM boundary conditions as well as the BCRs design and scale-up.