Level Measurement Inconsistencies

Nuclear, DP, and sight glass in a FCC main fractionator

To: "'asloley@distillationgroup.com'" <asloley@distillationgroup.com>
Subject: Foaming impact on level indication
Date: Wed, 14 Feb 2001 16:12:29 -0600

Andrew,

I have some questions related to the article you wrote that appeared on Hydrocarbon Online discussing problems associated with foaming. Your article described some of the symptoms that we have observed on our main fractionator.

We have four level transmitters measuring level in the bottoms section of the fractionator of differing technologies that have a consistent offset with respect to one another. Two of the transmitters are Honeywell sealed lead remote diaphragm D/P cells, while the other two transmitters are Texas Nuclear radioactive level transmitters. Each transmitter in the two pairs agrees with its counterpart, but the two pairs of transmitters have a 40% offset in their readings.

The radioactive pair are reading higher than the D/P pair, and we assume this is due to the affects of foaming. The D/P transmitters provide a level indication that is about 10% below the level that is indicated by visual inspection of the gage glass.

The condition that you described on page 5 of your article would tend to explain the difference in the indication of the D/P transmitters compared to that of the gage glass. Visual inspection of the gage glass allows you to witness the process fluid that has condensed running down the walls of the gage glass. You will also notice constant bubbling up through the process fluid in the gage glass.

We realize that we don't have the ability to adjust attenuation on the radioactive transmitters to filter out the affects of the foam layer in the column, but would like to determine which pair of transmitters should be utilized for level control for the column. Do you have any suggestions on possible methods to control or eliminate the foaming we are experiencing?

The operations personnel are not comfortable with any level indication that would not agree with the visual level indication offered by the gage glass unless strong reasoning is offered to explain the deviation. In the past there have been attempts made to determine the level in the column through the use of thermal scanning instrumentation, but this proved to be inconclusive.

The process engineer for this plant area does not believe that we are experiencing foaming, but can offer no alternative reasoning for the discrepancy. I handle the instrumentation engineering for this area and will need to bring the operations personnel and the process engineer into agreement with whatever explanation and solution that we propose for this situation. Any suggestions you could offer with regard to this situation would be greatly appreciated. Let me know if additional data is necessary for you to form an opinion on our situation.

Thank you,

J., US Refiner

Subject: Foaming impact on level indication
Date: Wed, 14 Feb 2001 4:42 PM

J.,

Found your question impossible to leave alone, so I went through it and have some questions for you.

1. Is the color of the liquid in the sight glass the same as that of a freshly caught liquid sample or are they both so black that you cannot tell if there was a difference?
2. What service is the tower in (FCC, coker...)?
3. When you say that the liquid level is 40% higher for the nuclear devices is this liquid level from the bottom of the vessel or is it by comparing the span of the instrument readings?
4. Are the level spans being measured identical for the DP cells, radioactive methods, and the sight glass?
5. Do you use purge gas for the top connection of the sight glass and purge liquid for the bottom connection of the sight glass?
6. 6. What purge liquid do you use for question 5 above? Is it stripped or does it have light ends in it?
7. Can you sketch or give dimensions for liquid level compared to the stripping tray level, stripping steam inlet, or flash zone inlet (what's relevant depends upon service)? A vessel sketch or elevation drawing (faxed or AutoCad or TIF or other image file) would be fine.
8. Does the sight glass plug often or ever?
9. Do all the levels respond about at the same rate when the liquid level is deliberately changed (allowing for instrument response speed)?

Andrew Sloley
DGI

To: "'Andrew Sloley'" <asloley@distillationgroup.com>
Subject: RE: Foaming impact on level indication
Date: Wed, 14 Feb 2001 16:38:42 -0600

This particular problem is located in the FCC plant.

J., US Refiner

To: "'Andrew Sloley'" <asloley@distillationgroup.com>
Subject: RE: Foaming impact on level indication
Date: Wed, 14 Feb 2001 15:16:19 -0800

Andrew,

I have added responses to those questions that I have responses for, but will investigate to obtain more detailed responses. The process engineer for this plant is out of the office this week, but I should be able to get responses to most of the questions by tomorrow morning.

Thanks.

J.

1. Is the color of the liquid in the sight glass the same as that of a freshly caught liquid sample or are they both so black that you cannot tell if there was a difference? The fluid in the sight glass is relatively clear, and has the appearance of a base oil stock. I will have to speak with the operations personnel to determine what appearance a sample typically has.
2. What service is the tower in (FCC, coker...)? This tower is in the FCC plant, and is listed as the Fractionator Bottom Section.
3. When you say that the liquid level is 40% higher for the nuclear devices is this liquid level from the bottom of the vessel or is it by comparing the span of the instrument readings? This deviation is a comparison of the span of the instrument readings. The gage glass shares taps with the D/P transmitters, and the nuclear transmitters scan the same range of the tower.
4. Are the level spans being measured identical for the DP cells, radioactive methods, and the sight glass? Yes, the measurement is looking at a four foot span of the column.
5. Do you use purge gas for the top connection of the sight glass and purge liquid for the bottom connection of the sight glass? There is a purge/flush connection for the transmitters and gage glass that utilizes LCO
   for this flushing. I will have to check the arrangement to tell you what is connected to the top connection.
6. What purge liquid do you use for question 5 above? Is it stripped or does it have light ends in it? The purge liquid is LCO, and the LCO is usually steam stripped. This can vary depending on the source of the LCO. I will check to verify the source at the morning meeting tomorrow.
7. Can you sketch or give dimensions for liquid level compared to the stripping tray level, stripping steam inlet, or flash zone inlet (what's relevant depends upon service)? A vessel sketch or elevation drawing (faxed or AutoCad or TIF or other image file) would be fine. I will pursue this in the morning.
8. Does the sight glass plug often or ever? The sight glass and the D/P transmitters used to have plugging problems, and this is the reason the flush system was installed.
9. Do all the levels respond about at the same rate when the liquid level is deliberately changed (allowing for instrument response speed)? The different level transmitters, as well as the gage glass do track each other when there is process movement only with an offset.

To: "'Andrew Sloley'" <asloley@distillationgroup.com>
Subject: RE: Foaming impact on level indication
Date: Wed, 21 Feb 2001 11:35:00 -0600

Andrew,

In speaking with the operations personnel regarding this issue, they agreed with the responses that I had given previously. They did indicate that they normally use MCO as the flush fluid, and only use LCO for start up. The operators did ask if the fluid seen in the gage glass could be diluted by the flush, causing the light appearance. I have left a message with the process engineer inquiring about the appearance of any samples drawn recently, and how it compares to what is seen in the gage glass. I will attempt to get a copy of the fabrication drawing for the vessel to provide you details on the internal construction and how it may affect the process.

Let me know if you need any additional data or more detail on any of the questions you posed previously.

Thanks,

J.

To: "'Andrew Sloley'" <asloley@distillationgroup.com>
Subject: RE: Foaming impact on level indication
Date: Wed, 21 Feb 2001 12:31:36 -0600

Andrew,

I will fax you a couple of drawings that will hopefully provide the detail you require on the column internals. I had to cut the inspection drawing in half in order to fax it, but this should not affect the section of the drawing we are interested in. I have also sent a portion of the P&ID that depicts the level transmitters and the gage glass. Let me know if you need additional detail with respect to the vessel internals.

The taps for the gage glass and transmitters are identified as L2 and L4 on the inspection drawing, and can be found just above the lower tangent line on the right side of the column. I spoke to our FCC process master regarding the fluid seen in the gage glass, and how it compares with what is present in the vessel. He indicated that the samples drawn would be the same as what is visually witnessed in the gage glass.

Let me know that you have received the fax, and whether you any additional data.

Thanks,

J., US Refiner

To: "'Andrew Sloley'" <asloley@distillationgroup.com>
Subject: RE: Foaming impact on level indication
Date: Thu, 22 Feb 2001 06:55:20 -0800

Andrew,

I finally heard back from the process engineer for this plant area and she indicates that the samples she has seen are black, and that she has never been successful determining the level in the gage glass. On the occasions when I have looked at the gage glass the fluid has an appearance like the base oil samples I've seen in the past. I will speak to her this morning at the morning meeting to get further details. I want to provide you with good data.

One of the operators did pose the question last week is we were actually seeing the flush fluid on the gage glass, but we didn't think much about it since we had never seen a sample drawn from the column.

J. US Refiner

Sent: March 8, 2001 11:07 PM
Subject: Re: Foaming impact on level indication

J.,

I've been out of the office troubleshooting a delayed coker. I will get a reply (I had an idea on what your problem might be) within a couple of days. I want to sketch out a couple of things first.

Andrew Sloley
DGI

Sent: March 17, 2001 2:09 PM
Subject: Re: Foaming impact on level indication

J.

The faxed figures helped immensely. The figures don't show a tower diameter for the FCC, so I've worked out an estimate based on a scale basis from the drawings. Figure 1 shows an approximately scale drawing of the main fractionator along with the liquid levels. The liquid levels on your drawings have been referenced to the L2 and L4 level tap elevations.

You have identified a basic level measurement problem. First, you sight glass reading does not agree with the DP cell. Both are measuring level across the L2-L4 connection range. Second, the nuclear level measurement does not agree with either the sight glass reading or the DP cell reading.

Based on your information, there seems to be a difference of opinion over if the sight glass can even be successfully used. However, assuming that it can be, it appears that the sight glass liquid is not the same liquid as in the tower. This would amply explain differences between the sight glass and the DP cell. The sight glass has a different density liquid than the density assumed for the DP cell. This problem may be exacerbated by condensation inside the sight glass.

The difference between the nuclear level gauge and the DP cell has two most likely causes. First, the density in the DP cell reading-to-level calculation is wrong. This should be checked. Second, the high velocity inlet vapors to the column are frothing, waving, of otherwise disturbing the surface of the liquid. From the sketch, the distance from the feed inlet to the normal liquid level is only approximately 0.95 inlet diameters. High velocity vapor impingement on the liquid surface is likely. This is a common problem in services with a high-velocity vapor inlet close to the liquid surface.

FCC bottoms service is not an area where foaming often occurs. While foam formation can never be ruled out, it is unlikely here.

Since you have them, using the nuclear gauges for level control makes the most sense. To meet your control needs, you need to look at three major objectives. First, have enough liquid head to meet bottoms pump NPSH requirements. Second, keep liquid from backing into the vapor inlet line. Third, have sufficient control range between the two limits to maintain stable control.

In reverse order, the nuclear gauge is probably the best measure of keeping the liquid out of the inlet line. Even slugs of liquid splashed into the inlet line and then entrained up the column by high-velocity vapor can damage equipment further up the tower. Mechanical damage to FCC MF is a major source of operating losses [1].

If a test shows that at the nuclear gauge LLL that the bottoms pumps do not have NPSH problems, then there is really no issue involved in using these gauges. They will protect the unit against high levels and give a suitable height at the LLL to prevent pump damage.

If the LLL, as measured by the nuclear gauges, is too low to protect the pumps, then the LLL level should be adjusted to a higher level. It appears from the sketches that you have a margin to adjust the NLL and the HLL up without causing a problem with liquid entrainment from impingement. This would probably require relocation of the nuclear gauges.

Of course, this does not resolve the differences between the nuclear gauge readings, the DP cell, and the level glass readings. Some differences between these are inherent in how they operate. Your difference is a little on the high end (1.6 ft), but not impossible given the high inlet velocities in the vapor feed on many FCCs.

References

[1] Sloley, A. W. FCC product recovery section reliability and operability issues. NPRA FCC Maintenance Conference, Houston, 4-5 August 1998.

Andrew Sloley
DGI