(38f) Experimental Investigation of the Liquid Discharge Behavior of Structured and Random Packings

Structured and random packings are state-of-the-art for ensuring excellent heat and mass transfer between liquid and vapor / gas phase in distillation or absorption columns. For their standard application, both academia and industry have continuously improved our understanding of how these packings influence heat and mass transfer, which has led to countless correlations (Wang et al., 2005). Nowadays, the performance of packed columns can be predicted via process simulation, although safety margins are admittedly still used for the actual plant construction (Pöschmann et al., 2023).

Unfortunately, most studies have so far only focused on how the column and especially the packing behave in the normal fluid dynamic operating window and only very few contributions investigated the dynamics outside this classical operating window. However, these dynamics are of high relevance in domains, such as operating modes with strongly flexible loads or process safety, where seconds can make a difference regarding serious damage to the plant or the environment, or even injury or death to people.

Research gap

Thus far, there are only some experimental studies and models on the dynamic discharge behavior of liquids from dense random particle packings, which occur in fixed-bed reactors or ore packings. Urrutia et al. (1996) investigated the liquid discharge of reactors with glass bead packings (2 mm diameter) and catalyst pellets (1.5 x 4.5 mm) under trickle flow conditions. They used gauging to determine the dynamic holdup in unstructured fixed beds. In their experiments, they observed a linear relationship between liquid discharge and time at the beginning. Afterwards, the discharge decreased due to the declining driving force. Ilankoon et al. (2014) studied glass bead packings and copper ore beds (porosity in the range of 35-40 %) without counter-current gas flow for packing heights of 300, 500, and 800 mm. Based on their findings, they proposed a model for the holdup dynamics inside the packing. Finally, Assima et al. (2015) studied the influence of the inclination angle of the column on the dynamic liquid discharge for glass bead packings (3 mm diameter, porosity = 0.4, packing height = 1.5 m, column diameter = 57 mm). They found that larger inclination angles result in larger liquid discharge. Despite the aforementioned studies, the transferability of their results to random or even structured packings with porosities of 96-98 % is questionable and must be verified experimentally. In addition, the structures in the cited literature were only investigated without the initial counter-current gas flow, which is inherent to distillation and absorption columns.

As a result, there are virtually no studies on the liquid discharge behavior of structured and random packings in the context of safety analysis in distillation and absorption. Examples of such safety-relevant situations are failure of a feed pump for the detergent, failure of the evaporator, failure of electric power supply, or blocking of the column. Such examples might result in the liquid running out of the packed section into the sump of the column, where light ends can boil up and increase the pressure in the column.

It is currently not clear how long it takes for the liquid to leave the system or how long the gas in an absorption column can be purified without violating concentration bounds. As a first fundamental step to answer these questions, this contribution investigates the fluid dynamics of the liquid discharge in a packed column for three different packing types relevant for distillation and absorption columns: a structured sheet metal packing (Montz B1-350MN), a wire gauze packing (Montz A3-500) and a modern metal random packing (Koch-Glitsch IMTP®).

Experimental setup

The experiments were conducted under ambient conditions in a DN300 glass column for the system water-air. The glass column is 4 m high from bottom to top, and the packing height amounts to about 1.5 m. Depending on the packing type, different support grids are installed to keep the packings in place. In addition, the column contains demister and liquid distributor. The latter cannot run dry due to its overflow type construction.

Overall, 7 temperature sensors and 7 pressure sensors are installed to measure liquid and gas phase. The liquid flow is measured with a Coriolis flow meter, whereas the gas flow is determined with a thermal flow meter. The installed pump and blower can provide liquid throughput of 1.5-50 m³/(m²h) and F factors in the range of 0.5-3.7 Pa^0.5, respectively. The plant is fully automated via a process control system.

Experimental procedure

From the possible operating window as given by the installed pump and blower, we selected 16 load combinations for which the column was operated. For each experiment, we ensured that the packing was fully wetted by setting the desired liquid flow and using the largest inlet gas flow possible. Afterwards, this gas flow was reduced to the targeted setpoint until a steady state was achieved. At this point, both inlet gas and liquid flow were shut down via the process control system and the outlet liquid volume flow was measured by gauging with a time resolution of 1 second. After the experiments for the first packing were completed, packing 2 and 3 were studied using the same procedure.

Results

The contribution compares the dynamics of the liquid discharge of the three investigated packings after both gas inlet and liquid flow are turned off. Therefore, we first analyze the impact of the initial holdup in a packing section as a result of the varied steady-state conditions for liquid and gas load before the shutdown. Here, we find an increasing speed of the liquid discharge with increasing initial holdups (larger initial steady-state liquid and gas loads). Afterwards, we compare the packings for the same steady-state initial conditions regarding liquid and gas load and show that the random metal packing shows the fastest and wire gauze packing shows the slowest liquid discharge. Finally, we propose a correlation for describing the liquid discharge dynamically.

References

Assima, G. P., Hamitouche, A., Schubert, M., & Larachi, F. (2015). Liquid drainage in inclined packed beds—accelerating liquid draining time via column tilt. Chemical Engineering and Processing: Process Intensification, 95, 249-255.

Ilankoon, I. M. S. K., & Neethling, S. J. (2014). Transient liquid holdup and drainage variations in gravity dominated non-porous and porous packed beds. Chemical Engineering Science, 116, 398-405.

Pöschmann, R., Paschold, J., Mueller, S., Harding, L.-S., Lachmann, N., Hiller, C., Ausner, I., Gäbler, A., Illner, M., Brösigke, G., Repke, J.-U. (2023). Will laboratory and pilot plant columns soon become superfluous? -- A concept for the determination of structured packing characteristics in a measuring cell under distillation conditions. Separation and Purification Technology, 325, 1-9.

Urrutia, G., Bonelli, P., Cassanello, M. C., & Cukierman, A. L. (1996). On dynamic liquid holdup determination by the drainage method. Chemical engineering science, 51(15), 3721-3726.

Wang, G.Q., Yuan, X.G., Yu, K.T., (2005). Review of Mass-Transfer Correlations for Packed Columns. Industrial & Engineering Chemistry Research, 44(23), 8715-8729.