Study on computer tomography and particle tracing of liquid-solid riser

Authors: Shantanu Roy, Chen, Sailesh B. Kumar, m. h. Al-dah Han * and M. P. Dudukovic [* indicating the meaning of the correspondent].

SETTING: Chemical Reaction Engineering Laboratory, Department of Chemical Engineering, Washington State University, St Louis, Missouri (63 130).

Abstract: Liquid-solid circulating fluidized bed is a potential reaction device, which is widely used in various industrial processes, such as oil refining, fine chemicals, petrochemical products and food synthesis. In these processes, the rapidly deactivated solid catalyst needs to be regenerated after the basic reaction is completed and recycled in the solids of the riser. This study shows that the computer-aided radioactive particle tracer technique (CARPT) can be used to establish the solid velocity model in the riser and the solid reflux model at the test fluid velocity. ? X-ray computed tomography (CT) showed that the solid concentration was slightly higher in the middle of the fractionator. This is contrary to the case of gas-solid riser reactor, which has higher solid concentration on the tower wall.

order

As an alternative reaction device, liquid-solid circulating fluidized bed has been rapidly popularized in various industrial processes such as fine chemical industry, petrochemical product synthesis, petroleum refining, etc. (Liang et al., 1995). This process is completed in a reactor, which contains liquid reactants (typical hydrocarbons at high pressure and low temperature) (Thomas, 1970) and solid catalysts (Corma and Martinez, 1993), which can be deactivated quickly. The basic reaction is completed in a vertical riser with high liquid-solid velocity ratio (in the riser, the solid becomes liquefied and can be carried by the liquid). The deactivated catalyst is regenerated in an independent treatment process, which is coupled by circulating solid and alkaline reaction in a continuous internal circulation. The design and assembly of this continuous flow liquid-solid system need to know the flow model and phase holdup distribution of each phase. The purpose of this work is to study the velocity and holdup distribution of solid phase in the riser of the flow model of laboratory-level circulating liquid-solid system through experiments.

experiment

The equipment diagram of laboratory-grade liquid-solid circulating fluidized bed is shown in figure 1. The riser is a plexiglass column with a diameter of 6 inches and a height of 7 feet. Tap water in the riser drives glass beads with a diameter of 2.5 mm to flow, and then flows back into the system through the plunger and ejector. The ejector (which has calibrated the solid flow rate as a function of the water flow rate) is used to control the liquid flow method to keep the solid material flowing in the riser. The total solid/liquid velocity ratio can be adjusted by the distributor at the bottom of the tower. The pump in the internal circulation and the circulating water in the water storage tank are used to maintain a constant high-speed water flow in the gas distillation tower and the injection port. The experiment was carried out on the CARPT and CT devices developed by the Chemical Reaction Engineering Laboratory of Washington State University in St. Louis, Missouri (Devanathan,1991+0; Kumar, 1994). It may be pointed out that the system used in this study has a compact structure and low viscosity, and only non-immersion fluid detection methods such as CARPT and ct can accurately measure the solid flow rate and concentration. The current equipment can install the riser on the CARPT-CT operating platform for this study. Long before the study of solid fluid mechanics, the residence time distribution analyzer of liquid phase was applied to liquid phase. After pulse rapid injection of potassium chloride solution, the conductance of liquid phase at a given position was measured. The results of this study have also been reported in other places (Roy et al., 1996), and we found that the liquid phase actually showed a concentrated flow potential with little dispersion effect. The two-dimensional variance of the E curve of liquid tracer particles is always less than 0. 1.

CARPT research of American chemical society (Devanathan,1991+0; Yang et al., 1992) introduced radioactive Sc-46 particles (emission wavelength is 350 Ci, half-life is 83 days) into a hollow aluminum ball, and the particle size and density matched with the glass beads to be mixed to prepare tracer particles. Using the exquisite CARPT calibration step (Yang et al., 1992), the particles were put into about 200-300 known positions in the reaction section to be detected, and the calibration diagram of the distance-density relationship of each detector was obtained. After the calibration is completed, the required liquid supercritical velocity is set and maintained, allowing solid particles to freely enter the flow field to simulate the movement of typical glass particles. After a long time (8 hours), the position of the tracer particle (represented by the number of photons obtained by the detector) is recorded as a function of time. Then, by discarding and processing the rough raw data, the average fluid composition and fluctuating fluid composition, viscosity coefficient and kinetic energy of solid particles can be calculated (Devanathan,1991+0; Larach et al., 1997). This is the first successful demonstration of CARPT technology in a system in which tracer particles periodically leave and re-enter the reaction part of the fractionator detected by the detector.

The CT scanner in the Chemical Reaction Engineering Laboratory of the Department of Chemical Engineering, Washington State University in St. Louis, Missouri, uses fan line geometry to measure? -Radiation attenuation after the ray passes through a given object in the riser. Then, the time average holdup distribution of each phase on the cross section is reconstructed with a coarse attenuation measuring instrument. The radioactive source is placed in the Cs- 137 isotope of 100 mCi, and the attenuation test is carried out with an angular array composed of 1 1 sodium iodide detector (maximum). The expectation maximization algorithm based on maximum likelihood principle (Lange and Carson, 1984) is used to reconstruct the image obtained in the projector. Kumar et al. (1995), Kumar and Dudukovic'( 1997) have discussed the details of the software and hardware of the CREL scanner. In this study, the liquid-solid riser was scanned at four axial positions along the tower.

Results and discussion

The experiment was carried out in the range of liquid supercritical velocity (12-23 cm/s). This study reports the typical results of the system operating at the supercritical velocity of 20 cm/s liquid. Glass beads with a diameter of 2.5 mm were used in all experiments, and the water flow rate of the ejector was 25 gallons/minute. The water velocity at the bottom of the riser is kept at 33 gallons/minute, so that the average liquid supercritical velocity in the tower can reach 20 cm/s.

Fig. 2 is a logarithmic average and time average radial solid holdup (solid concentration) distribution diagram measured at four axial positions at the supercritical liquid flow rate of 20 cm/s. It is observed that the order of solid holdup does not change much with the increase of radial position (the maximum change is 4%), but it decreases slightly with the change of axial position (the maximum change is 4%). Compared with the tower wall, the solid holdup in any given axial position is slightly higher than that in the middle of the tower. This is an interesting result because the opposite trend has been widely reported in gas-solid risers (Rhodes and Geldart,1989; Rhodes, 1990). The radial gradient of the solid content distribution reported here is also small.

Fig. 3 shows the solid velocity field estimated in the CARPT experiment. Fig. 3a is a velocity vector diagram. From the perspective of time averaging, it can be clearly seen that the solid phase has an internal circulation loop: the solid rises in the center of the column and falls on the wall of the column. Fig. 3b shows that the time-averaged axial components of solid velocity at four locations in the middle of the tower also have the same quantitative results. It should be pointed out that the downstream velocity of solid phase on the column wall is less than that of upstream fluid, and the total mass of downstream solid phase is still satisfactory (9.6% in this experiment). The solid holdup diagram at the height of 33 cm in the tower is generally orderly. This height is just above the distributor and injector in the tower (figure 1), which is part of the mixing zone and obviously has lower solid content than the height of 78 cm. The experimental results of CARPT also confirm this point: Figure 3a clearly shows that the direction of solid velocity vector is randomly oriented at this height, and a clear circulation circle appears at a higher position in the tower. Therefore, it is still necessary to consider the fluid at the height of 33cm in the tower, and it shows obvious deviation behavior compared with other parts of the tower. By a new method, the time distribution (RTD) of solid residues in the riser can be calculated indirectly from the CARPT data. Because tracer particles are considered as a typical dispersed component that can be recycled to the riser, the distribution of residence time in the riser is its RTD value. The "remaining time" obtained through these uninterrupted data collection is made into the histogram in Figure 4. An arbitrary assumption is proposed, which gives the RTD value of solid phase. Finally, in Figure 5, the average axial velocity of solids in the axial direction is expressed as a function of the supercritical velocity of liquids. Experiments under different conditions show that the velocity of the tower center line and column wall (downstream) is generally increased. Of course, this may also be due to the higher liquid modulus in the same cross section, which leads to the increase of solid modulus and the increase of solid average velocity. Based on these experiments, the results seem to show that with the increase of liquid supercritical velocity, the solid velocity tends to "saturation value". However, these results still need to be verified by further experiments in the future.

conclusion

Until today, the design of fluidized bed and riser still stays at the level of empirical rules. The actual phenomenon in this kind of system is far more complicated than the results obtained by heuristic approximation algorithm as the basis of design program. Therefore, users and designers of liquid-solid riser can get great inspiration from the basic understanding of fluid mechanics in this kind of system. The current research is only a small step towards the quantitative aspect of similar experiments. In CREL (the author's laboratory), the research work on riser configuration with different operating conditions and different particle sizes is under way. There are plans to study the static phenomena in this system in the future. The data will be further processed to calculate the kinetic energy, viscous shear stress and viscous dispersion coefficient of the solid phase. The overall goal of this research work is to understand some key variables that affect the efficiency of liquid-solid riser, and then study the more basic law of proportion increase. We hope that our experimental data can be used as a benchmark for computer dynamic simulation of liquid-solid riser fluid.

The title of the chart is translated as follows:

Figure 1. Liquid-solid riser equipment diagram

Figure 2. Distribution of solid holdup (concentration) at different axial positions at supercritical velocity of 20 cm/s liquid.

Figure 3. Solid velocity field at supercritical velocity of 20 cm/s liquid: (a) velocity vector diagram; (b) Axial average velocity diagram.

Confirm (omit)

References (omitted)