Jun-2018

Packed bed performance analytics

Advanced analytics introduce a numerical and easily understood approach to tower gamma scanning.

LOWELL T PLESS
Tracerco

Article Summary

Refinery and chemical plant operations depend heavily on distillation and separation towers. Tower gamma scanning is well established in the process industries as a qualitative tool to help troubleshoot towers. Advances in data analysis have led to a quantitative approach in expressing gamma scan data in numerical terms easily understood by process and operations engineers.

For packed towers, a grid scan of three or four equidistant scans crossing through the beds of packing would typically be used to investigate the quality of liquid distribution. The conventional approach to ‘analysing’ a gamma scan has been to visualise how well the scan data from the individual scans matched each other or how well they ‘overlaid’ with each other. This is a totally subjective analysis lacking consistency, open to varying interpretation and does not translate well from tower to tower. Therefore, the resulting conclusions from this approach can be very ambiguous regarding the magnitude of any detected liquid maldistribution.

An alternative analytical approach, termed PackView, has been developed whereby a relative density scale is calculated from data that the grid scan provides. The density scale begins at the density of the dry or non-operating packing. The density scale displays the calculated density of liquid retained in the bed of packing based on the scan data results.

Another calculation by which to put the liquid distribution into perspective and to get a measure on the useful capacity of the packing is to calculate the liquid hold-up fraction or liquid volume fraction. If the measured liquid retention density is divided by the process liquid density at bed conditions (the liquid density at the actual operating temperature and pressure), liquid hold-up or liquid volume fraction can be established. A comparison of the liquid hold-up fraction to the packing operating capacity curves provides an objective appraisal of current operating capacity.

It is always easier to understand and discuss technical issues when quantitative information can be used to compare operational parameters with engineering design. This advanced analysis provides a new method for extracting quantitative information from gamma scan data to diagnose and characterise the operation of distillation and separation towers. It is our goal that the use of the advanced analytics presented improves the value of gamma scan data and facilitates improvements in the operation of mass transfer equipment.

Figure 1 shows the most typical orientations for conducting a gamma scan on a packed tower. These scans are conventionally called grid scans. Some of the process information that can be obtained for a gamma scan of a packed tower includes:
•    Condition of packed beds – 
elevation of packing, depth of packing elements, uniformity of packing elements, and so on
•    Collector and distributor liquid levels: Damaged? Overflowing?
•    Base or bottoms liquid level
•    Flooding present?
•    Foaming present?
•    Excess liquid retained in packing?
•    Liquid maldistribution?

The purpose of doing multiple scans of a packed tower, as shown in Figure 1, is to determine the state of the liquid distribution. The foundation for this analysis is the Beer-Lambert law:

I=Io e-ρµχ   
where
Io is the initial gamma ray intensity measured at a given distance with no material interfering with the radiation transmission
I is the radiation intensity (counts) measurement after passing through the tower
µ is the absorption coefficient of the material the radiation is passing through (material physical property)
ρ is the density of the material the radiation is passing through
χ is the thickness of the material the radiation is passing through.

Beyond questions concerning damage to internals and flooding within a packed tower, the next biggest concern is the state of liquid distribution through the beds. Historically, gamma scan analysis has relied upon performing two sets of parallel scanlines (given the tower diameter is sufficiently large) through the packing (see Figure 1a), commonly referred to as a grid scan. When gamma scanning a tower, Io remains fixed and µ is essentially a constant. For a grid scan, the multiple paths of radiation through the tower must be kept equal so χ remains a constant. With I being measured, the equation is solved for ρ, the material density.

On the basis that χ was indeed kept equal, then the multiple sets of scan data from a packed column grid scan could be simply visually compared to each other. Since all scan parameters were constant, particularly the length or path of radiation through the column (χ), uniform liquid distribution would be confirmed by all four scans detecting identical radiation. Figure 2a is a typical example of a grid scan showing all four scanlines matching, implying good liquid distribution.

However, when there was some variance in the radiation measurement and the sets of scan data did not seem to match very well, then liquid maldistribution would be diagnosed as the cause. Comparing the sets of scan data, lower radiation counts (higher density) indicated more liquid and higher counts (less density) indicated less liquid, therefore liquid maldistribution. An example of this type of conclusion is demonstrated in Figure 2b.

This is a totally subjective analysis lacking consistency, open to varying interpretation, and does not translate from tower to tower. Additionally, this quantitative analysis does not give any insight into the severity or quantity of liquid maldistribution. Therefore, the resulting conclusions from this purely qualitative approach can be very ambiguous regarding the presence and magnitude of any liquid maldistribution.

Advanced analytics: PackView
An advanced analytical analysis for gamma scan data from packed columns was developed to consistently analyse gamma scan data and reach a 
conclusive result. Densities are calculated based on tower dimensions, scan path length (variable χ) and gamma scan response through the packing. The results are used to superimpose a density scale onto the scan data. The baseline of the density scale is the dry bulk or non-operating packing density. Calculated densities greater than the dry packing density 
represent liquid and/or solids retained in the bed of packing. Figures 3b, 4b and 5b show examples of this liquid retention density scale.

As with the qualitative gamma scan analysis, if the multiple scanlines have matching liquid retention densities then the implication is the liquid distribution is good. However, if there is a difference between the scanlines, the retention density gives a numerical comparison from which to gauge the 
extent or severity of any liquid maldistribution in terms of liquid density.