Experimental and modeling study on the hydrodynamics in multiphase monolith modules with different distributors

Highlights

• Three commonly used distributors were introduced into multiphase monolith reactor.

• The nozzle distributor exhibits the lowest pressure drop and largest liquid holdup.

• A discrete phase and continuous phase coupling 3D model was established for nozzle distributor.

• The gas and liquid velocities on the top of the monolith bed over a nozzle distributor was explored.

Abstract

Three commonly used distributors (i.e., nozzle, packed bed, and foam) were introduced into a multiphase monolith reactor in this work to investigate the pressure drop, liquid holdup and gas–liquid distribution performance by combining experimentation and CFD simulation. The experimental results showed that the nozzle distributor exhibited the better gas-liquid phase distribution and the lowest pressure drop. However, so far, the gas and liquid velocities at the top of the monolith bed, which are crucial for the transition from a single channel to bed scale, have not been studied. Thus, a discrete phase and continuous phase coupling 3D model using CFD simulation was established for the nozzle distributor, and the magnitude and distribution of the gas and liquid velocities at the entrance of the monolith module were investigated. The simulation results showed that the droplets’ velocity reaching the monolith module was very high and had a strong influence on the gas velocity distribution. The increase in superficial gas velocity could enhance the gas distribution to some extent.

Introduction

Because of their unique advantages (e.g., low pressure drop, good mass transfer performance, good operational flexibility, minimal axial dispersion, and easy separation of catalysts and products) [[1], [2], [3]], monolith reactors are widely used in the conversion of vehicle exhaust gas, flue gas denitrification, methane conversion, benzene alkylation to cumene [4], hydrogenation of 2–ethylanthraquinone to hydrogen peroxide [5,6], and other gas–solid or gas–liquid–solid catalytic reaction processes [7,8]. Monolith catalysts and reactors have regular and parallel channels, the most common of which is square with hydraulic diameters ranging from 1 to 5 mm. The active components are coated onto the wall of channels, and when gas or gas–liquid two–phase reactants pass through the channels, a chemical reaction occurs under the action of the catalyst.

In recent years, several authors [[9], [10], [11], [12], [13], [14], [15]] have studied monolith reactors from a single channel to a bed scale by means of both experimentation and simulation; however, most works have focused on the pressure drop, mass transfer, holdup, liquid slug length, axial and radial dispersion, flow distribution, and reactor performance. Liu et al. [16] observed five flow regimes, namely, bubbly flow, churn flow, Taylor flow, slug–bubbly flow, and annular flow in vertical capillaries through a high–speed video camera by varying both gas velocity and liquid velocity from 0.008 to 1 m·s^–1. It is worth mentioning that the gas bubbles had a hemispherically shaped top and liquid slug that alternated with each other, and a layer of thin liquid film appeared between the bubbles and the wall when the flow regime was classified as Taylor flow. In our previous work [17], we also investigated Taylor flow and three other flow regimes in a single channel using computational fluid dynamics (CFD) calculation. Liquid circulation occurred in the liquid slug, and a vortex appeared near the bubble cap. Moreover, the flow direction of the liquid film was opposite to that of the bubble velocity during Taylor flow, which increased the two–phase relative velocity and thus enhanced the mass transfer performance. Furthermore, the volumetric mass transfer coefficient of the monolith packing was improved by at least one order of magnitude compared with a conventional structured catalytic packing and pellet catalyst in the decreasing order monolith packing > structured catalytic packing > pellet catalyst [17,18]. To date, most studies have been based on the uniform profile for the single channel. However, it is not entirely applicable to mechanically apply the experimental and simulation results obtained from a single channel to the upscaled monolith reactors. The gas–liquid distribution at the entrance of the monolith channels is vital for the performance of monolith reactors, because the monolith reactors comprise straight channels. That is, once the gas–liquid phases enter the reactor, there is no possibility of redistribution. Thus, the selection of an appropriate distributor is essential for the actual operation of monolith reactors. This selection is in essence a bridge connecting single channel research and bed research.

At present, the experimental methods for directly measuring the flow distribution mainly include ECT (electrical capacitance tomography), CT (computed tomography), MRI (magnetic resonance imaging), the liquid collection method, and optical fiber probes. Although the liquid phase distribution in a reactor was studied via ECT by Mewes et al. [19], the influence of the gas velocity on the distribution was not reported. Roy et al. [20] quantitatively investigated the liquid distribution performance using three distributors (nozzle, shower head, and foam) for a gas–liquid concurrent flow system via CT and found that the best distribution was by nozzle, and the liquid flow had a strong influence on the nonuniform liquid distribution, while the gas flow had no obvious influence on the liquid saturation distribution. Sederman et al. [21] measured the flow patterns and bubble lengths in the channels of the reactor under the condition of Taylor flow via MRI and found that the bubble length did not change with the operation time. Reverse flow was also observed in several channels. However, this method was very expensive. Behl et al. [22] also evaluated the flow distribution using a customized collector with a gravimetric liquid collection method for a packed bed distributor of 2 mm alumina spheres 3.5 cm in height. It was observed that the increase in the liquid velocity made the distribution obviously uneven. Xu et al. [23] measured the gas holdup in nine channels at the radial location of the catalytic bed via nine optical fiber probes in order to estimate the flow distribution among the entire bed. This led to a large relative deviation using nine channels instead of the entire bed. More specially, Satterfield et al. [24] and Crynes et al. [25] took indirect measurements to judge the flow maldistribution, such as the pressure drop and reactor productivity, but there were no quantitative parameters to estimate the degree of maldistribution, and thus, a direct comparison among different distributors could not be made. To date, several studies on different kinds of distributors and the influence of the superficial liquid velocity u_L,s on the distribution have been reported, but reports on the effect of superficial gas velocity u_G,s are rare. Furthermore, some experimental methods are expensive and evaluations for the distribution of the entire bed on the basis of a single or a few channels have large deviation.

The objectives of this work address the following issues: (i) measuring the bed pressure drop and bed liquid holdup under three common distributors (nozzle, packed section with 1 mm glass beads, and foam) to investigate the relationship between these two characteristics; (ii) comparing the distribution effects of three distributors by observing the flow of the gas–liquid phase at the bottom of the monolith module; (iii) developing a suitable and valid CFD model for the optimal distributor; (iv) exploring the magnitude and distribution of gas and liquid velocities on the top of the monolith bed. One significance of this work is helping guide the selection of a suitable distributor for monolith reactor. Another significance is extending the knowledge of hydrodynamic properties from single capillaries to bed scale, especially the gas-liquid distribution.