19 replies to this topic

[rohollah]

Dear members

According to attached sketch,there is a control valve that control the level of a tower.as you can see outlet line is drained to an oily water pit.since P3 < vapor pressure cavitation occure.the problem is that,the outlet pipe is being damaged due to cavitation.
How can we solve this problem?

Regards
Rohollah

Attached Files

•  Book1.xlsx   13.55KB   82 downloads

[ankur2061]

Rohollah,

When you say damage to piping due to cavitation, does it mean that the cavitation is causing a hammering sound and / or vibration in the pipe just after the 2nd isolation valve in the control valve circuit?

If hammering and vibration is the issue than the following can be done:

1. Check the piping support for the line after the 2nd isolation valve for damage and misalignment.

2. Instead of a fixed support use a spring support to accomodate lateral or axial pipe movement due to hammer and vibration.

If your damage is in the form of erosion of the pipe material due to cavitation then replace the carbon steel material with stainless steel after the 2nd isolation valve which will be much more erosion resistant.

One thing that puzzles me, is how are you draining hot water of 139°C in an oily water pit which I presume is at atmospheric pressure. If your drain pipe terminates above the liquid surface of the oily water pit you will flash almost all the hot water to steam. Is your drain pipe terminating as a dip pipe into the oily water pit below the liquid level? Please clarify?

Regards,
Ankur.

[kkala]

To my understanding, globe valve between manometers P2 and P3 is responsible for the downstream flashing. Upstream pressure P2 indicates only liquid, since water vapor pressure is 3.51 bara.
It may be worth while replacing the mentioned globe valve with a gate line size valve*, just for the control valve isolation (when needed). And locate the globe valve on the same pipe, as close to the oily water pit as possible. In this way two phase flow will occur is a shorter pipe length, probably (I do not have a picture of how shorter this length could be).
Of course new piping downstream of the globe valve location should be designed for two phase flow. By noting it on the PID, piping usually knows the precautions required. Post No 2 also refers to them.
Also try to check flow regimes (no slug flow) and velocity, which should not be erosional (otherwise change of downstream line diameter will be needed).

* Probably similar to the gate valve upstream of the Control Valve.

Edited by kkala, 12 February 2012 - 09:59 AM.

[ankur2061]

Rohollah,

I doubt that isolation valves to the control valve are globe valves as presumed in post # 3. At least to my knowledge, you never provide globe valves for control valve isolation. The pictorial reperesentation you have provided does not indicate that a globe valve is provided downstream of the control valve.

Probably local practices in Greece ask for providing globe valves as isolation valves.

Regards,
Ankur.

[rohollah]

Ankur

You are absolutely right. A globe valve should not be used as an isolating valve. Before that, the downstream isolating valve was a gate valve, Now it has been replaced with a globe valve to keep P2 above the vapor pressure. When it was a gate valve, cavitation and erosion occurred right after the control valve. Now by throttling the globe valve, P2 can be controlled.

Regards
Rohollah

[rohollah]

Ankur:

First of all, thanks for your quick response.
Secondly, I have to apologize because the outlet line does not drain into the oily water. It is drained into a separate pit.

Actually, we have upgraded the pipe material to SS316, but erosion is still continuing.
Other recommendations are appreciated.

Regards
Rohollah

[kkala]

Rohollah, I doubt that isolation valves to the control valve are globe valves as presumed in post # 3. At least to my knowledge, you never provide globe valves for control valve isolation. The pictorial reperesentation you have provided does not indicate that a globe valve is provided downstream of the control valve. Probably local practices in Greece ask for providing globe valves as isolation valves. Regards, Ankur.

Although the matter has been clarified (the valve is globe), following is pointed out.
1. A globe valve is indicated between P2 and P3 on "book1.xlsx", same as the bypass valve of the level control valve (LCV). Valve upstream P1 is a gate valve and Post No 3 says so.
2. Isolation valves of LCV are expected to be quite open in normal operation. ΔP of valve downstream LCV (between P2, P3) is almost 4.9 bar. It is not an isolation valve.
3. Hint about local Greek practices is irrelevant. Style does not promote right answers.

[S.AHMAD]

Salam Rohollah

1. The main problem is "flashing" not so much of "cavitation". If the piping experience EROSION then it must be caused mainly by flashing.
2. The first solution that you should think of is to cool the liquid to below 100C. This is a permanent solution.
3. If cooling is not viable, then you have to live with the situation by changing the drain pipe as the need arises and keep the pipe as short as possible to reduce the replacement cost. Relocate the "globe valve" as close as possible (practicable) to the pit.
4. You must not change the globe with gate valve since you require to maintain pressure at P2 above the vapor pressure to ensure that flashing does not occur at P2. This is done by regulating the globe valve. Gate valve is not effective for regulating
5. If P2 is lower that the VP, the problem will be transferred to the control valve. More costly to replace the control valve.
6 If flashing (in this case it is cavitation) also occur at the control valve, you should increase P2 further up such that the control valve is about 80% opening and P2 is kept as high as possible.

Edited by S.AHMAD, 12 February 2012 - 07:34 PM.

[S.AHMAD]

Rohollah

1. If the whole piping (upstream and downstream of CV" is insulated,temperature may be reduiced to some degree by replacing insulation with wire gauze for human protection. This may provide some degree of cooling eventhough not good enough to aleviate the problem.
2. Alternatively, before discharging to pit, installed a flash drum being kept at 3.5 barg (or slightly lower) where steam is discharged to atmosphere and liquid to the pit.
3. You may also considered increasing the diameter of the drain pipe. This will lower the velocity of flashing steam to a more acceptable figure.

Edited by S.AHMAD, 12 February 2012 - 08:30 PM.

[DB Shah]

Flashing for 46.6 Barg & 139°C to 0.6 barg will give just ~ 5.% of flash steam. Utlity of this heat source at such a low pressure will be not viable.

The problem of material damage at downstream is due to improper sizing of the header downstream of control valve. (46 barg condensate line (single phase) and 2phase line downstream of control valve are of same size ie 6"). As per design intentions, even globe valve should not be required to maintain single phase (liquid) at control valve down stream.

My recommendations will be-
1. Resize and increase the control valve downstream line. This will increase downstream isolation valve size also.
2. Relocate CV & bypass vavle nearest to the pit. If possible flash the content right on the pit, ie no extra piping downstream of existing globe valve.
3. If you properly size the downsteam line, isolation valve can be converted back to gate valve (no need of controlling single phase at CV down stream,CV is intended for flashing service)
4. Keep it simple!, I donot want any operator to go to the globe valve, adjust it in such a way that it creates a backpressure P2 which is > vapour pressure.
Compare it with distillation column reboiler outlet condensate line going to dearator (operating in ~0.5 barg), we size the condensate line properly and in a running plant operator never bothers to look at condensate pressure or the reboiler chest pressure.

[breizh]

I 've added a paper about cavitation and flashing ,
Hope this helps

Breizh

[ankur2061]

Rohollah,

These days control valves of the type: "anti-cavitation multi-trim" are available for services where large pressure drops need to be addressed across the control valve. This could be your permanent solution to the cavitation problem. Refer the links below:

http://www.dresserma...tion-Solutions/

http://www.flowserve...alve-Trim,en_US

http://www.documenta...d102310x012.pdf

Hope this helps.

Regards,
Ankur.

[Bharathiraja]

Rohollah,

This problem has nothing to do with Cavitation. It happens only because of Flashing.... the fluid flow changes to two phase at pressure P3. For two phase flow, you must size your line in such a way that, the mean velocity of the mixture shall stay within the erosional velocity calculated. You can very well follow the guidelines given by Mr. Kkala and DB Shah.

[GS81Process]

I expect that cavitation is occurring across the level control valve because the pressure drop of 41 bar will result in a pressure at the control valve's vena contracta less than the vapour pressure (3.51 barg). Cavitation will occur rather than flashing because the pressure downstream of the control valve recovers to above the fluid vapour pressure. Flashing is also definitely occuring across the globe valve.

The maximum allowable control valve pressure drop can be estimated using the method shown in Ch.5, p.115 of the Fisher Control Valve Handbook: http://www.documenta.../book/cvh99.pdf

I estimate a maximum allowable pressure drop of ~3.6 bar across a single pressure reducing device to avoid cavitation.

Rohollah: I would have to know more about the specifics of your tower operation before recommending a solution, however I would recommend a two-stage pressure reduction in this level control line to pit in order to avoid cavitation (could be a control valve and regulator in series, control valve and orifice in series, etc). The other option is a cavitation trim design as recommended by Ankur. However, in my experience cavitation trims are never quite as good as advertised because they are subject to such harsh erosional impacts from the collapsing of bubbles happening so rapidly.

Flashing is inevitable if the liquid is to be discharged to atmosphere at temperatures above 100C. What is the normal flowrate in this line? Is it enough to justify recovery of flash steam and are there any sources nearby requiring heat? If not, may I ask why there is not a cooler to allow for safe temperature discharging to the pit? Is it possible to design and install a heat exchanger for this purpose?

Edited by GS81Process, 14 February 2012 - 08:39 PM.

[S.AHMAD]

1. The problem posted by Rohollah is not with the "control valve" but it is the piping downstream P3. Cavitation may occur at the control valve BUT that is not the problem
2. As long as the temperature is above 113C (sat temperature at 0.6 barG), flashing will occur. That is why cooling the stream below 100C will be the permanent solution. That is by eliminating the root cause (temperature higher than sat temp)
3. It is VERY obvious that the problem is associated with two phase flow.
4. At the pressure of 0.6 barG, 1 m3 of water turns into steam of 1091m3. This means that the velocity increases by 1000 times. Let say as per DBShah that the flashed steam is only 5%, that mean the velocity increases by 50 times for every 1 m3 of water flashes.
5. If the current velocity is 1.5 m/s then the average velocity will be 75 m/s. It is even worst if the current velocity is 3 m/s (150 m/s). The normal practice is to keep below 45 m/s for steam. The recommended velocity should be even lower for two phase flow and if the waste water is acidic.
6. The actual velocity is expected higher than 50 times since the cross sectional area available for the steam is much smaller than the pipe cross sectional area that is mostly occupied by water.
7. If cooling the stream is not viable, then the next best solution is to size the drain pipe correctly.
8. The solution should be simple but cost effective.

Edited by S.AHMAD, 14 February 2012 - 07:55 PM.

[rohollah]

GS81process

As Ahmad mentioned, in this case installing a heat exchanger is not cost effective.

Rohollah

[GS81Process]

Ok. I agree with Ahmad's comment that the problem downstream of the globe valve is caused by erosional damage due to flashing. However, I would still be tempted to address possible cavitation as well (up to you).

Have you trended the level controller output to estimate flow rates (and velocities)?

Edited by GS81Process, 15 February 2012 - 06:39 PM.

[kkala]

Let us look into what part of piping has been damaged. The damaged part may be:
α. The pipe downstream of the level control valve (including P2, globe valve, P3, on book1.xlsx). Then LCV cavitation is also significant.
β. The pipe downstream of the globe valve (pipe between LCV and globe valve does not have indications of damage). Then LCV cavitation is not significant.
I have assumed that option (β) is valid, as understood from the posts. Can rohollah confirm it, or advise accordingly?

[kkala]

....
4. You must not change the globe with gate valve since you require to maintain pressure at P2 above the vapor pressure to ensure that flashing does not occur at P2. This is done by regulating the globe valve. Gate valve is not effective for regulating
5. If P2 is lower that the VP, the problem will be transferred to the control valve. More costly to replace the control valve...

As same post indicates, meaning of above is that the globe valve located closer to the pit should not be replaced by a gate valve.
A gate valve, located where the globe valve is now, is needed to isolate control valve (e.g. for maintenance, when the bypass globe valve will function).
Of course proposal by DB Shah in post No 10 concerns a more radical way out. This is what would be applied in a new project, I believe.

Edited by kkala, 18 February 2012 - 12:02 PM.

[rohollah]

Let us look into what part of piping has been damaged. The damaged part may be:
α. The pipe downstream of the level control valve (including P2, globe valve, P3, on book1.xlsx). Then LCV cavitation is also significant.
β. The pipe downstream of the globe valve (pipe between LCV and globe valve does not have indications of damage). Then LCV cavitation is not significant.
I have assumed that option (β) is valid, as understood from the posts. Can rohollah confirm it, or advise accordingly?

Kostas

You are absolutely right.

Rohollah