33 replies to this topic

[Fatema1987]

Hi,

 

I need ur support to point out our current issue with one of orifice plates at Refinery site. It is measuring LPG produced from Overhead of Platformer Stabilizer Column. The issue is unclear for us until this moment;

The orifice was calculated and its transmitter was calibrated for a full range of 222 TPD of LPG for delta pressure of 0.5 bar. The calculation was checked through software and confirmed that the data is suitable for orifice plate with bore size 27 mm.  The current issue is; the transmitter of this orifice plate is reading over range "IOP" at normal throughput and DP was checked at site to be found 0.7 bar which makes sense for having "IOP" reading at DCS. However, the actual flow based on other downstream transmitters are showing around 70 TPD coming from this line ! and it goes to the sphere as a final destination. So we are sure this flow should be lesser than what the transmitter is showing (>222 TPD) ! The high DP 0.7 bar around this orifice at site is making us without a clue! 

We took plenty of actions trying to understand why we have a high DP where; the impulse lines were changed, the transmitter was changed with a new one, a live pressure gauge was installed at site to measure DP around impulse line. With all of these actions, the DP is still high "~0.7" bar. 

What is missed here? Please advise,

I attached the calc sheet for this orifice.

Attached Files

•  Orifice Plate Calculation.pdf   110.96KB   46 downloads

Edited by Fatema1987, 08 March 2022 - 02:01 PM.

[Art Montemayor]

Fatema:

 

I never relied on orifice plates on LPG flow lines.  We always had troubles with the application and we had options that we used with success.  I can visualize that you're having similar problems.

 

I basically have two comments to your query:

 

1. The attached .pdf you cite as an orifice calculation is not that at all.  All it is is a brief specification sheet.  There are no calculations in it.  What's more important, is that it doesn't identify the fluid (LPG in this case) nor its composition and conditions.  I would have to assume that it exists as a saturated liquid - as most LPGs do.
2. My choice for a reliable and accurate LPG meter is found described in the attached .pdf - an Endress-Hauser Coriolis Flow meter.  This is what I selected and installed on LPG process and service lines with good results in the past - and no problems.  Endress-Hauser furnished expert technicians during our installations and applications and we never had any problems.  This is what I would use instead of an orifice.  My frank opinion is that with the state of the art existing today in process instrumentation, orifice plates are a thing of the past that can't compete against coriolis meters when maintenance, reliability, accuracy, labor, downtime, etc. are all considered.

I hope this experience helps you.

 

 Endress-Hauser LPG Coriolis Flow Meter_8FE_EN.pdf   256.79KB   21 downloads

[Bobby Strain]

Sounds like junk blocking the flowline.

 

Bobby

[Fatema1987]

Fatema:

 

I never relied on orifice plates on LPG flow lines.  We always had troubles with the application and we had options that we used with success.  I can visualize that you're having similar problems.

 

I basically have two comments to your query:

 

1. The attached .pdf you cite as an orifice calculation is not that at all.  All it is is a brief specification sheet.  There are no calculations in it.  What's more important, is that it doesn't identify the fluid (LPG in this case) nor its composition and conditions.  I would have to assume that it exists as a saturated liquid - as most LPGs do.
2. My choice for a reliable and accurate LPG meter is found described in the attached .pdf - an Endress-Hauser Coriolis Flow meter.  This is what I selected and installed on LPG process and service lines with good results in the past - and no problems.  Endress-Hauser furnished expert technicians during our installations and applications and we never had any problems.  This is what I would use instead of an orifice.  My frank opinion is that with the state of the art existing today in process instrumentation, orifice plates are a thing of the past that can't compete against coriolis meters when maintenance, reliability, accuracy, labor, downtime, etc. are all considered.

I hope this experience helps you.

 

Endress-Hauser LPG Coriolis Flow Meter_8FE_EN.pdf

 

Thanks Art Montemayor. I totally agree to use a reliable meter to measure liquified LPG and we do have it before the sphere. We have a turbine meter to measure LPG at the RunDown. 

However, we need this orifice flow to work properly because it is a flow controller slave for the level controller master of Overhead vessel. The liquid LPG is a distillate flow while Stabilizer reflux flow is controlled through a ratio set. I attached a quick scheme.

 

In every shutdown we inspect this orifice and all time it is found in excellent condition. The strange thing that we have another orifice plate measuring the flow of Reflux. We dont see this problem with Reflux orifice plate! The issue is only with distillate orifice plate measuring high DP.

Attached Files

•  LPG Flow Controller.JPG   77.65KB   1 downloads
•  Flow Calc Sheet.pdf   448.22KB   21 downloads
•  LPG LAB Data.png   6.7KB   0 downloads
•  Compositions of LPG.png   16.64KB   1 downloads

Edited by Fatema1987, 09 March 2022 - 05:14 AM.

[Fatema1987]

Sounds like junk blocking the flowline.

 

Bobby

would you further elaborate ?

Edited by Fatema1987, 09 March 2022 - 12:04 AM.

[breizh]

Hi Fatema ,

I notice that your calculation sheet is from 2008 , does it mean the problem is new ?

Breizh 

[Fatema1987]

Hi Fatema ,

I notice that your calculation sheet is from 2008 , does it mean the problem is new ?

Breizh 

Now we will complete two years since this issue started. Before two years, it was working fine and we can keep it on cascade with the level controller but now we can't. We did water wash for this loop couple of times to ensure it is clean from salt participation. However, the transmitter remains reading over range due to high DP around orifice. 

[breizh]

Hi,

Using your data and keying them in a software available on Internet , I found that your stream is a mixture with # 11% mol of gas  ( P =21.5 bar absolute and 38 C) . The model using is Peng Robinson .

Are the composition and Operating conditions the right one ?

 

Note : Software name : Canary by Quest consultants

 

Good luck

Breizh 

[Fatema1987]

Hi,

Using your data and keying them in a software available on Internet , I found that your stream is a mixture with # 11% mol of gas  ( P =21.5 bar absolute and 38 C) . The model using is Peng Robinson .

Are the composition and Operating conditions the right one ?

 

Note : Software name : Canary by Quest consultants

 

Good luck

Breizh 

 

Yes the given data for compositions and operating temp & pressure are the actual parameters. If the given data in your analysis sheet is true, would it make a difference in terms of orifice calc ? Do we have a special calc for mixed phases? 

 

Please note we have a control valve after the orifice plate followed by a pump to pump it to the next equipment.

[Art Montemayor]

Fatema:

 

I think Breizh's inherent curiosity caught the hint that I mentioned in my prior post: the LPG fluid seems to be saturated liquid.  If that is the case, what is the resulting effect if that fluid goes through a pressure drop?  Although this may not be the case, I would suspect a flash operation taking place.

 

That's one of the reasons I stayed away from orifice plates on saturated liquid service.  Again, this may not be your case, but I would follow Breizh's initiative and look into the composition of the fluid and its expected performance during a pressure drop through the orifice.  You may have a control valve downstream of the orifice, but it still has to generate a prescribed pressure drop in order to do its expected job.

[latexman]

If the flow rate is incorrect, is it consistent?  And, if it's consistent, why can't it be run in automatic/cascade mode?  Seems like it should be able to, if it's consistent.

[Bobby Strain]

Unfortunately the orifice is not shown on the drawing. But it is most likely at the pump discharge. If it's a Platformer, the basic engineering was done by UOP. And they know what they are doing. Until the orifice can be inspected, connect the level controller to the flow control valve without the cascade.

 

Bobby

[Fatema1987]

If the flow rate is incorrect, is it consistent?  And, if it's consistent, why can't it be run in automatic/cascade mode?  Seems like it should be able to, if it's consistent.

The flow is showing over range >222 TPD since max DP of transmitter is sat at 0.5 bar , so we cannot switch it into Auto/Cascade. Although, u gave me an idea to try increasing the transmitter DP into 0.9 bar and flow accordingly then we'll be able to see if the flow is consistent and can be cascaded but in terms of a measured value, the reading will be unreal.

[Fatema1987]

Unfortunately the orifice is not shown on the drawing. But it is most likely at the pump discharge. If it's a Platformer, the basic engineering was done by UOP. And they know what they are doing. Until the orifice can be inspected, connect the level controller to the flow control valve without the cascade.

 

Bobby

I attached the PID, the controller is "82FC024". U r right it was designed by UOP but unfortunately in the last recent years they send trainees more than experts. I got the same feedback when I worked at RFCC previously. Any way this is my opinion and can be wrong.

 

 

U r right abt cascading the level directly to the flow controller bypassing the transmitter and this is how we operated in last two years. This setup helped to maintain a stable level for a limit, but it caused Reflux controller to be operated on auto mode because both flow controllers should be cascaded to Level controller as a normal setup. If one controller is out of Cascade control, the other cannot be on cascade too. Since the reflux controller was forced to operate on Auto, we had incidents for Copper strip failure at LPG. This is why we want retain the normal Cascade setup for both controllers.

Attached Files

•  PID.JPG   118.53KB   2 downloads

Edited by Fatema1987, 10 March 2022 - 05:40 AM.

[Fatema1987]

Fatema:

 

I think Breizh's inherent curiosity caught the hint that I mentioned in my prior post: the LPG fluid seems to be saturated liquid.  If that is the case, what is the resulting effect if that fluid goes through a pressure drop?  Although this may not be the case, I would suspect a flash operation taking place.

 

That's one of the reasons I stayed away from orifice plates on saturated liquid service.  Again, this may not be your case, but I would follow Breizh's initiative and look into the composition of the fluid and its expected performance during a pressure drop through the orifice.  You may have a control valve downstream of the orifice, but it still has to generate a prescribed pressure drop in order to do its expected job.

Thanks for your input. I was searching for orifice calc for two phase case, Please recommend a website or share an excel. 

[Bobby Strain]

There seems to be something wrong with what you highlighted in red. You start at the vessel water draw, not the hydrocarbon draw. There appears to be other errors in the drawing, too.

To control the reflux, you should be able to change the control to a temperature somewhere in the top of the fractionator. UOP is not very good at this, so you would need to rely on your own resources. Or a qualified consultant.

 

Having both reflux and product on level control seems to be a poor way to control the overhead flow. I would query UOP about this setup. And talk to some experienced engineers in Des Plains.

 

Good luck

Bobby

Edited by Bobby Strain, 09 March 2022 - 11:16 PM.

[Fatema1987]

There seems to be something wrong with what you highlighted in red. You start at the vessel water draw, not the hydrocarbon draw. There appears to be other errors in the drawing, too.

To control the reflux, you should be able to change the control to a temperature somewhere in the top of the fractionator. UOP is not very good at this, so you would need to rely on your own resources. Or a qualified consultant.

 

Having both reflux and product on level control seems to be a poor way to control the overhead flow. I would query UOP about this setup. And talk to some experienced engineers in Des Plains.

 

Good luck

Bobby

Sorry I was in harry and highlighted the wrong line. I corrected it now in attach. As far as I know, having two temp controllers "Reboiler Temp Controller" and "Top tray Temp Controller" is normally not recommended, please correct me if mistaken. Two temp controllers for same column will keep the loop in oscillation and takes a long time to stabilize. 

 

Thanks,

Attached Files

•  PID.JPG   118.53KB   1 downloads

Edited by Fatema1987, 09 March 2022 - 11:36 PM.

[breizh]

Hi Fatema,

you may consider this paper to support your work .

 

note 1 : It should be a good option to change the technology , from orifice to mass flow meter . Worth for you to get contact with E&H or other vendors . Like Art  I 've been always very satisfied with E&H  on technology/ service and price .

note 2 : In case you will find a paper about distillation control ( Basic) 

 

Good luck

Breizh 

[Fatema1987]

Hi Fatema,

you may consider this paper to support your work .

 

note 1 : It should be a good option to change the technology , from orifice to mass flow meter . Worth for you to get contact with E&H or other vendors . Like Art  I 've been always very satisfied with E&H  on technology/ service and price .

note 2 : In case you will find a paper about distillation control ( Basic) 

 

Good luck

Breizh 

interesting material to read, really thanks. We'll try to check other alternatives to change the meter type in nearest opportunity.

 

I went through many pages of both documents and found that if we decided to cascade Reflux flow controller with the tray temp controller then we should meet following condition “Figure 8.19s. In this configuration the sensing point for column pressure control should be located near the temperature control point. Keeping the sensor locations close to each other helps to fix the relation between temperature and composition at this particular point.”

 

In our condition, the pressure controller is located at tray 4 while top tray temp is located at tray 34. The pressure of column is stable generally with Reflux flow controlled by a manual ratio sets by panel-operator. So I am not sure if recommended setup to control top temp by reflux flow wouldn't disturb the loop. 

[Bobby Strain]

All this discussion about 2-phase flow is irrelevant, a red herring. The LPG is sub-cooled, significantly, since it is pumped from the reflux drum where the LPG is saturated. I am curious about what events occurred when this problem was first encountered? Was it following a shutdown?

 

As to control, there are lots of alternatives to evaluate. But you need an expert, not us armchair engineers. Your situation without product flow control limits the possibilities. And you need a good process simulator to evaluate options. For upper column temperature control, rarely is the top tray very effective for control. Maybe you will just have to accept that you may occasionally have off-spec product. I would simulate the column to see if there is a tray temperature that indicates the need for increased reflux flow to maintain on-spec product. Then the operater can make informed adjustments.

 

Bobby

Edited by Bobby Strain, 10 March 2022 - 11:11 AM.

[Fatema1987]

All this discussion about 2-phase flow is irrelevant, a red herring. The LPG is sub-cooled, significantly, since it is pumped from the reflux drum where the LPG is saturated. I am curious about what events occurred when this problem was first encountered? Was it following a shutdown?

 

As to control, there are lots of alternatives to evaluate. But you need an expert, not us armchair engineers. Your situation without product flow control limits the possibilities. And you need a good process simulator to evaluate options. For upper column temperature control, rarely is the top tray very effective for control. Maybe you will just have to accept that you may occasionally have off-spec product. I would simulate the column to see if there is a tray temperature that indicates the need for increased reflux flow to maintain on-spec product. Then the operater can make informed adjustments.

 

Bobby

Really we couldnt find a clear reason when this issue started in terms of operating parameters but the compositions might changed. I mentioned top tray temp because it is a strong indication for salt participation in the column and normally operator keeps an eye on this.

Attached Files

•  Data Trend.JPG   79.83KB   0 downloads

[PingPong]

Hi Fatema,

 

I saw your message and assume that you were referring to this topic.

 

It is very difficult to troubleshoot operation via the internet by persons who do not have access to all design data and all operating data. At the moment we don't even have a P&ID of the whole column.

 

I get the impression from the part of P&ID that we see that there is some form of Reflux Ratio control intended by UOP and the ratio can be set by the operator. Is that correct?

In that case you would indeed need a reliable product flow measurement for that RR control to work.

 

You can however simply control the drum level directly with the LPG product control valve i/o via the FC slave and operate the reflux simply on flow control. It's not ideal but that's how it has been done for decades before all kind of fancy control schemes were invented and implemented. It requires however that the operators understand what they are doing. They should keep an eye on the flow indicated by the turbine meter at the RunDown and adjust reflux flow accordingly. Or maybe it is possible to incorporate that meter into the RR control?

 

Having two temperature controllers on a column is not a good idea. Some process control specialists will claim that it is possible if a lot of Advanced Process Control (APC) is added but be careful before you spend time and money on implementing that. In any case top tray temperature is not suitable for that, you would need a TC a few trays lower.

 

Problem with top tray temperature is that the top tray gets hot vapor from below and relatively cold liquid from the side or center (in case of DXF trays). These do not mix instantaneously, like in a process simulator, however at different points on the tray the temperature will be different, and it can even differ at different locations in the downcomer, where I assume the TI is located. So that measurement is of little use, even for only monitoring salt formation risk, because a TI in the downcomer is not at the coldest location(s) on the tray.

Edited by PingPong, 20 March 2022 - 11:13 AM.

[Fatema1987]

Thanks PingPong for your input.

1- Reflux Ratio control intended by UOP and the ratio can be set by the operator. "YES"

2- control the drum level directly with the LPG product control valve i/o via the FC slave and operate the reflux simply on flow control. "YES, this is how we operate since this issue started.

3- Or maybe it is possible to incorporate that meter into the RR control? Turbine meter is located at offsite for final product, so there is a big time-lag between Stabilizer column and turbine meter. Besides, GTU "Gas Tail unit" is located in-between to receive this LPG distillate  and separate Fuel gas from LPG before send it to turbine-meter R/D.

4-  Advanced Process Control (APC) is added but be careful before you spend time and money on implementing that. In any case top tray temperature is not suitable for that, you would need a TC a few trays lower. I do agree regarding APC where we implemented it at NHT stabilizer column and we endup by reducing the priority of R/R CV and sometimes we switched off this controller in order to maintain btm temp stable. 

 

5- For Salt participation we monitor tray 34 and reflux nozzle is located at tray 40.

Attached Files

•  Stabilizer.JPG   119.37KB   2 downloads

Edited by Fatema1987, 21 March 2022 - 04:18 AM.

[PingPong]

5- Due to your statement "top tray temp is located at tray 34" I was under the impression that 34 was the top tray.
If the TI is 6 trays below the top tray then I wonder why the operators keep an eye on this as this is not the coldest tray. But that's really a different subject.

 

On the Stabilizer P&ID a GLC (Gas Liquid Chromatography) analyser is visible on the column overhead.

If that also measures C5 content then maybe you could use that to control the reflux.

 

Edited by PingPong, 21 March 2022 - 08:44 AM.

[Fatema1987]

5- Due to your statement "top tray temp is located at tray 34" I was under the impression that 34 was the top tray.
If the TI is 6 trays below the top tray then I wonder why the operators keep an eye on this as this is not the coldest tray. But that's really a different subject.

 

On the Stabilizer P&ID a GLC (Gas Liquid Chromatography) analyser is visible on the column overhead.

If that also measures C5 content then maybe you could use that to control the reflux.
 

 

Unfortunately, we dont have GLC analyzer at the site. It failed since a long time and demolished accordingly. We have Siemens analyzer for C5 control at NHT unit at Splitter Overhead but frankly speaking it is failed most of time. However, we will consider this as a part of next required improvement. I really like this recommendation.

 

To update you, today we increased the Range of transmitter from 0.5 bar into 0.9 bar in order to see flow pattern. As you see in trend we were able to cascade FC024 controller since its reading is consistent. However, tomorrow we'll conduct step test at set point of the level to see the respond of FC024. 

 

My next question, will increasing bore size of orifice plate help to minimize the effect of vapor portion in the DP around the orifice plate?

Attached Files

•  Trend.jpg   83.68KB   0 downloads

Edited by Fatema1987, 30 March 2022 - 10:55 AM.