Figure 18. Configuration of column scale-up tests.
It is also preferred where suspended solids create a high pressure drop, or dissolved gases create bubbles in the carbon bed. For a downflow or percolation system, an influent line should be installed at the top of the column, with an effluent at the bottom. To prevent the column from draining during operation, the effluent line from the last column should extend from the bottom of the column to above the top of the column. This will keep the column filled with liquid at all times during operation and prevent siphoning from occurring.
It is recommended that suspended solids be removed from the feed stream to a GAC column. If this is not possible in the scale-up design, then the effect of suspended solids should be included in the pilot run. In the upflow operation, most of the suspended solids work their way up through the GAC bed without a significant increase in pressure drop.
The carbon bed should be at least 60 cm deep with a 4 cm internal diameter. A smaller column is not recommended as the wall effect becomes significant. The carbon bed can be supported by glass wool, wire cloth, etc. Columns and fixtures can be constructed from glass, plastic, reinforced fiberglass, or metal. Borosilicate glass is commonly used. It is essential that all columns used in the pilot system have at least the same internal diameter.
Contact flow rates, Hourly Space Velocity (HSV), or the quantity of feed liquor, are expressed in the number of carbon bed volumes passing through the column per hour. A bed volume is the volume occupied by the carbon bed, including carbon volume and void volume. The recommended HSV range is between 0. 1 and 3.0 depending upon the degree of purification required, the type and concentration of impurity, the nature of the process liquor and the pressure drop. Generally, high levels of purification, high impurity concentrations, and/or high viscosities will require a lower HSV As an example - a typical HSV. for decolourisation of starch based sweeteners is 0.25. The carbon will perform more efficiently, but with a tradeoff in the maount of liquid that can be processed through a column in a given period of time. On the other hand, for the removal of traces of organics in drinking water and wastewater, a HSV range of 2 to 3 produces good results.
As the flow rate and quantity of liquor are the most important controllable variables in developing design data, a feed pump suitable for accurate and continuous flow is required. Depending on the size of the pilot column system, the use of peristaltic, diaphragm, piston-type or centrifugal pumps are recommended. The feed pump should be used in combination with a volumetric or gravimetric flow control
NOTE: To develop reasonably good data for scale-up to full plant design, it is important to have the operation of the pilot column system as near as possible to the anticipated plant conditions. The most critical factors, flow rate and feed impurity concentration, must be constant for the entire test run.
device. Before process liquor is delivered to the carbon, suspended matter should be removed, preferably by the same method planned for the plant system.
Adsorption to activated carbon is a function of diffusion rate. This means that the columns should be operated at a temperature at which the liquor approaches the viscosity of water but will not decompose or form too much color. If the process requires operation at elevated temperature, a jacketed tube should be used. The column temperature is controlled by circulating water of the required temperature through the jacket. This requires a thermostatic bath enabling accurate temperature control. When loading the column, care should be taken to avoid entrapping air in the carbon bed. Entrapped air can cause channeling during column operation, preventing complete contact of the process liquor with the carbon particles. In small columns, entrapped air can be avoided by pouring out the carbon in boiling water just before loading. Most of the excess water can be poured off, along with most of the fine carbon particles. About one quarter of the column should be filled with water before loading the carbon. As the carbon is added to the column it should be submerged in the water at all times. Sometimes it is useful to backwash the carbon prior to operation in order to remove remaining dust and entrapped air. It is recommended that a T-junction is used between dosing pump and the columns to allow any entrapped air to be released. Before starting the test, the complete system should be checked by running on water for several hours. After setting the appropriate flowrate, the liquor to be treated can be fed to the columns and this will displace the water. Several samples of the feed liquor should be taken over the duration of the test to highlight any drift in impurity concentration. Samples of effluent liquor after each column should be taken at regular time intervals or after a fixed number of processed bed volumes. If the accuracy of the analysis is affected by undissolved solids, all samples should be filtered prior to examination. For examination of purity levels of taken samples, your standard purity test can be used. When the effluent from the parallel columns or last column in series exceeds the purity requirement, the test should be stopped. It is not recommended to stop the liquor feed during the test. A procedure with intermittent daytime runs will give deviating test results in most cases. When you collect your data, tabulate the information on a spreadsheet. Effluent impurity concentrations for all columns should be plotted against elapsed time or processed bed volumes, generating "Breakthrough curves". The points at which the purity requirement is exceeded are defined as "Breakthrough points".
Was this article helpful?