Temperature and gamma scans identify problems in crude tower

Mayela Perez, Ricardo Duarte
Pdvsa Manufacturing & Marketing SA
Caracas, Venezuela

Lowell Pless
Tru-Tec Services Inc.
La Porte, Tex.

Using temperature profiles and gamma scans with traditional data-collection methods (that is, operating data and physical inspections) helps to reliably detect and identify the cause of maldistribution in fractionation towers.

In 1994, the use of temperature profiles and gamma scans shortened the downtime for the vacuum-distillation tower of Pdvsa's El Palito, Venezuela, refinery by more than 50%. In 1991, similar damage took 12 days to repair while the unit was offline; however, in 1994, the damage took only 5 days to remedy.

After studying temperature measurements, refinery personnel suspected liquid/vapor maldistribution on a packed crude tower. The refinery then used gamma scans to confirm maldistribution, identifying packing damage as the root of the problem.

The application of these two techniques gave the El Palito distillation team (technical, operations, and maintenance personnel) the advantage of advanced knowledge. The team was able to confidently determine the cause of the operational problem, plan the purchase of replacement internals, and schedule an orderly shutdown and repair plan before shutting down and inspecting the unit.

Additionally, replacement material was purchased via normal delivery rather than via emergency delivery at premium emergency rates.

Temperature profiles and gamma scans complement other means of data collection used to troubleshoot fractionation columns. Refiners should make field measurements of the operating column (especially temperature and pressure profiles), analyze unit heat and material balances, and, where appropriate, run computer simulations to pinpoint the problem.

This comprehensive approach was successfully applied to Pdvsa's El Palito vacuum tower.

Liquid/vapor maldistribution

Liquid/vapor distribution problems cause variation in the fluid composition at different points at the same height of the tower. The problem is reflected in a difference in temperature around the circumference of the tower.

El Palito's major problem in 1994 involved an under-performing packed, crude-vacuum tower. Usually, engineers first suspect liquid/vapor mal distribution as the cause of poor performance of a packed-distillation tower.

The question engineers must answer is, "Why is there liquid/vapor maldistribution?" Physical damage, poor distributor design, or extreme operating conditions affect a distributor's ability to uniformly regulate liquid and vapor flows.

Maldistribution can be detected in the field using the following techniques:

1. Water testing prior to start-up (Water testing the liquid distributor is done to ensure proper liquid distribution and level installation.)
2. Measuring the circumferential temperature at different elevations (bed depths)
3. Gamma-scanning techniques.

At Pdvsa's El Palito refinery, these techniques were applied effectively to troubleshoot the crude-vacuum tower.

Vacuum tower history

The vacuum unit at the El Palito refinery was designed to process 66,500 b/sd (10,000 metric tons/day) of reduced crude from the atmospheric crude tower. The unit consists of a fired heater, the vacuum tower, vacuum system, product, and pumparound circuits. The vacuum tower operates "dry," which means it requires neither velocity steam nor stripping steam. The tower has three sidecuts, two pumparounds, and overhead and bottom products ( Fig. 1 [142,818 bytes]).

The overhead product goes through an ejector system, which produces a net vacuum of 10 mm Hg. The first sidecut is light vacuum gas oil (LVGO), a portion of which is cooled and returned to the tower as the top pumparound. The remainder of the LVGO is blended with heavy atmospheric gas oil (HAGO) to produce diesel for the local market.

The second sidecut is heavy vacuum gas oil (HVGO), a portion of which is cooled and returned to the tower as pumparound. Another fraction of the HVGO is used as wash oil in the washing section; the remainder is used as feed to the catalytic cracking unit.

The third sidecut is slop wax, the majority of which is mixed with the feed to the vacuum unit. A small amount of slop wax is mixed with the vacuum residue, which is the bottom product.

The vacuum residue is pumped through a heat-exchanger train in the atmospheric crude unit (thermal integration between the crude and vacuum units). The vacuum residue is then split into two parts: One part returns to the vacuum tower as a quench oil stream and the other part goes through more heat exchange (simultaneous steam generation) to finally mix with light cycle oil (LCO) and slurry oil. The second mixture is sold as fuel oil to local and foreign markets.

In 1990, the wash section of the vacuum tower was modified to increase HVGO yield. A pall ring bed was replaced by a combined packed bed with structured packing on top of grid packing. The LVGO and HVGO sections remained the same with ballast rings as shown in Fig. 1.

In September 1994, during the startup of the crude and vacuum units, a high liquid level was observed in the bottom of the vacuum tower. Once operating conditions were stabilized, the operation of the tower was limited because HVGO color and metals content specifications were not being met. The overall yield of the vacuum gas oils was 9% lower than expected.

Pdvsa evaluated the vacuum tower in four ways:

1. Running an optimization test
2. Obtaining a temperature profile
3. Using gamma scans
4. Physically inspecting the tower.

Optimization test

An optimization test was done to study the effects of the main operating variables on the performance of the vacuum tower.

The tower pressure, HVGO hot reflux (wash oil) rate, and transfer-line temperature were manipulated. Process data and samples were taken from the HVGO and vacuum residue circuits. Using laboratory results, the process variables were adjusted to maintain HVGO product quality.

Table 1 [43,992 bytes] shows the production and laboratory results.

Results of the optimization test suggested that the hot HVGO reflux (wash oil) rate had to be reduced to improve the quality of the HVGO product. This was strong evidence of liquid entrainment from the slop wax (wash) section of the vacuum tower up to the HVGO-packed bed and draw tray.

Although reducing the HVGO reflux rate still did not bring the HVGO product on specification, it increased the yield of vacuum gas oil by 6%. This yield lowered the production loss from $48,000 to $20,000/day.

Temperature profile

Wall temperatures around the circumference of a column at different vertical heights can be measured with a surface pyrometer. A pyrometer with an extension probe can be used if penetration through the tower insulation is needed.

Temperature differences greater than 15° F. between different points at the same elevation indicate liquid/ vapor maldistribution.

This technique is particularly useful for situations in which there is a dramatic temperature gradient over the tower height. One limitation, however, is that the maldistribution pattern must exist along the wall of the tower. If the maldistribution is concentric, wall temperature measurements will be unable to detect it.

In any case, the wall temperature measurements do not provide information about the cause of the maldistribution. Thus, gamma scans of the tower are often used in combination with the temperature measurements.

In the El Palito refinery, temperature measurements were taken around the circumference of the vacuum tower at three different elevations across the slop-wax section of the tower: at the hot HVGO (wash oil) spray distributor, within the layer of structured packing, and within the layer of grid packing. The temperature profile is shown in Fig. 2 [105,836 bytes].

The profile indicated that there was liquid/vapor maldistribution, with a strong liquid bias on the east side of the tower and a strong vapor bias toward the west side. There was more than a 100° F. temperature difference between these diametrically opposite points.

Gamma scan

Gamma scans are used to diagnose operating abnormalities in towers, like flooding, foaming, missing and collapsed trays, collapsed packed beds, and liquid maldistribution through packed beds.

Uniformity of liquid/vapor distribution through packed beds is critical. Gamma scanning accurately and reliably detects nonuniformities in a packed tower.

For packed towers, the scan-line orientation follows a three-way or four-way grid pattern of equal-distant chords, like that shown in Fig. 3 [190,028 bytes]. The basic assumption is that uniform liquid/vapor distribution also has uniform or equal bulk density through the packing. Therefore, plots of all four lines should coincide, or overlay on top of each other, indicating equal density and uniform liquid distribution through the packing.

Problems are diagnosed when nonuniformity is observed from the scan data plots. Any divergence among the scan data can be an indication of maldistribution. Lower "radiation intensity" indicates excess liquid, and higher "radiation intensity" is associated with liquid deficiency.

The vacuum tower was scanned in the grid scan-line orientation shown in Fig. 3a.

Fig. 3b is a plot of the scan data taken from the slop-wax section of the tower. The colors and line patterns of the data curves correspond to the scan lines shown in Fig. 3a.

As can be seen from Fig. 3b, the black scan line (corresponding to the east side of the tower) is much denser than the other scan lines. The black line indicates a strong liquid bias on the east side.

In addition, the scan shows unequal heights of packing, indicating possible packing displacement. A higher height of packing was seen on the east and north sides of the vacuum tower.

The dense vapor area above the slop-wax bed on the east and north sides may also be a result of liquid being entrained upwards. The extra dense readings on the north scan line, just below the slop-wax bed, was caused by interference from a platform which is noted in Fig. 3b.

Chimney Tray P-3, the slop-wax draw tray, was found to be lightly loaded and operating dry. Operations indicated that the tray had a 7% level in the sump during the scans and was being operated dry to avoid a high liquid residence time, which promotes coke formation. Otherwise, the light, liquid loading would have been an indication of partial tray damage.

Fig. 3c shows the plot of the scan data through the HVGO bed. The HVGO bed showed signs of displaced packing. Uneven heights of packing were seen on all four of the scan lines. There was as much as 1.5 ft of packing above the design height on the east side of the tower.

The north side was lacking approximately 1.3 ft of packing from a total height of 5.5 ft. A very dense region or more liquid traffic was detected on the east side of the tower as well. The dense region in the vapor area above the packing on the east side was an indication of severe liquid entrainment or displaced packing.

Inspection and recommendations

When the vacuum tower was shut down for turnaround and inspection, massive damage to the slop-wax bed, hot HVGO spray distributor, and HVGO bed were seen. Fig. 4 [71,460 bytes] shows some of the damage to the slop wax section.

The El Palito distillation team offers the following recommendations based on its experience with this study:

1. Gamma-scan crews must not only have competent knowledge in the scanning techniques but also in distillation and tower internals.
2. The following troubleshooting strategy will expedite the solution of packed column problems:
   • Clear problem definition
   • Operational and laboratory data analysis
   • Verification of control room instrumentation with field measurements
   • Mass and energy balance of the tower
   • Use of temperature profiles
   • Use of gamma scanning techniques.
3. Temperature profiles and gamma scanning of towers operating at optimum conditions are recommended in order to establish baseline-comparison parameters for future problems in distillation towers.

The key to troubleshooting refinery distillation columns is reliable plant data. Field measurements of temperatures and pressures, a heat and material balance, and computer simulations can help pinpoint the problem.

Additional troubleshooting methods, such as temperature profiles and gamma scans, are useful for detecting and identifying the cause of the problem. In 1994, use of these tools resulted in significant savings by limiting repair costs and production losses at Pdvsa's Palito refinery.

Acknowledgment

The authors wish to acknowledge the work of Lance Freeman and Matt Moffett, Tru-Tec Services, who performed the gamma scans under adverse conditions.

Bibliography

1. Bowman, J.B., "Use Column Scanning for Predictive Maintenance," Chemical Engineering Progress, February 1991, p. 25.
2. Kister, Henry Z., Distillation Operation, New York, McGraw-Hill Book Co., 1990, p. 729.
3. Kister, Henry Z. "Corpoven's Crude Tower Problem," Fax dated Sept. 20, 1995.
4. Kister, Henry Z., Rhoad, Rusty, and Hoyt, Kimberly A., "Improve Vacuum Tower Performance," Chemical Engineering Progress, September 1996, p. 36.
5. Lieberman, Norman P., Process Design for Reliable Operations, Houston, Gulf Publishing Co., 1988, p. 253.
6. Strigle Jr., Ralph F., Random Packings and Packed Towers Design and Applications, Houston, Gulf Publishing Co., 1987, pp. 123-27.

The Authors