11 replies to this topic

[ankur2061]

Dear All,

I have a wet steam separator which separates wet steam (80% steam quality) to dry steam and condensate.

The system is such that the condensate line has to be taken to a destination 2.5 km away from the separator.

I need to provide a LCV for the separator liquid level control. In my opinion the LCV needs to be located close to the termination point of my condensate line which is 2.5 km away in order to prevent 2-phase flow for a 2.5 km long condensate line with all the issues associated with 2-phase flow.

However, I am not sure of the contollability issues related to locating the level control valve 2.5 km away from the separator. What kind of controller lag or sluggishness of the control can be encountered in such a scenario?

For easier understanding, I am attaching an excel sheet showing two schemes where Scheme 1 is my preferred option with the level control valve located 2.5 km away and Scheme 2 where the level control valve is just located at the liquid outlet of the separator.

I need some expert opinion on the controllability issues related to the Scheme 1 as per my attachment. Also if someone has a preference for Scheme 2 than I need a technical justification for the scheme including issues related to 2-phase flow and controllability of LCV.

Art/Doug/Zauberberg please have a look.

Regards,
Ankur.

[Art Montemayor]

Ankur:

I would not install the LCV 2.5 km distant. I would need a lot of further information and basic data before accepting that option. One important detail that would need clarification would be the criticality of preventing any condensate buildup in separator due to instrument or other failure and the time increment available to remedy the situation. I suspect that any liquid condensate carryover into the overhead steam is considered a risk and a hazard to downstream equipment (steam turbines?).

I would install an appropriately designed condensate drain valve at the separator because:

1) I cannot place confidence on having a required self-venting condensate drain line for a length of 2.5 km. In order to ensure that the LCV can receive saturated liquid condensate, I must ensure that that the line is self-venting and contains NO HIGH/LOW spots throughout the entire length of 2.5 km. This is a requirement for instantaneous, steady flow throughout the length of the line. Although we don't know the additional details of the line's layout and installation, I suspect that this will be almost impossible to ensure due to the required expansion joints and the length involved. Correct me if I am assuming wrong basic data.

2) The distance involved places a special caveat or condition on the type, class, and reliability on the LC controller and LCV signal transmitter. Any "accident", breech, or failure of the level signal along the entire distance will put the separator in danger of an overfill. I do not consider this as acceptable (although I don't know all the scope of work).

3) This will be an automatic, un-manned operation at remote sites (I assume) and, as such will need to be fully reliable 100% of the time and instrumented with redundant fail-safe instruments. Any increase in separator level due to unknown or controllable reasons will cause an immediate shutdown of the steam main. With this type of scope condition, I cannot justify putting 100% reliability on a remote LCV. I can place confidence on a static, 2.5 km long condensate line, but a remote signal and its interconnections requires additional engineering and back-up.

4) Expanding high pressure, saturated condensate is a problem. However, this is done everyday in high pressure power plants using special LCV design and construction. Fisher Controls publishes a special manual that covers this very subject.

5) Handling a 2-phase flow for 2.5 km is also a problem - one that I would prefer to avoid, if possible. However, it CAN be done - albeit with additional capital and expense. Here, I am putting the criticality of the separator ahead of the capital expense as important items to bear in mind for the engineering decision. I would guess that the majority of the flash steam would condense as it travels underground for 2.5 km. You still have to provide some separator at the condensate's terminal end - in either case. However, when you resort to handling 2-phase flow in the U/G line, you are freeing many conditions and restrictions from the actual field construction engineering and crew. You no longer are constrained with ensuring positive liquid drainage and a continuous pipe incline. This is a big item for the construction crew and could add up to be a very big capital expense savings when compared with locating the LCV at the condensate terminal.

The above are my immediate thoughts and opinions based on what you have given as basic data and scope of work. I hope these are of some help to you.

[ankur2061]

Art,

First of all my thanks for your prompt response.

I forgot to add that the dry steam from the separator is for joining to another bigger steam header at a distance of 2.5 km which in turn is used to feed wells for Enhanced Oil Recovery (EOR). Only the temperature of the steam and the pressure of the steam are important. No steam turbines are involved. A certain carryover of water in steam line is not a detriment to the process, however there are issues related to water hammer and the mechanical integrity of the steam pipeline due to water hammer if water is carried over in the steam pipeline.

As suggested by you I will provide an adequate drain for liquid draining and emergency isolation valves to cut-off the separator in case of level build-up in my engineering P&ID's.

Another thing that I need to clarify is that the condensate line is running overground on sleepers at a elevation of 500 mm from grade leve. Buried piping is not envisaged.

I would like to point out a small error in my spreadsheet. In scheme 2 the delta P shown in the line downstream of the LCV will be lower than what is shown on the spreadsheet since the flow will be 2-phase flow for the entire length of 2.5 km. It means, that I will require a higher pressure than 7.13 barg downstream of the LCV to have a termination pressure of 5 barg.

The basic question still remains. The controller response and whether any instrumentation vendor will be able to provide the right level control valve with proper controller response.

Regards,
Ankur.
Ankur.

[benoyjohn]

Ankur,

Please upload the P&ID or indicate the other controls associated with the scheme. Also can you indicate the separator dimensions and level settings. This will enable a better understanding of the controllabilty issue .

regards,
Benoy

[Zauberberg]

Hi Ankur,


It's hard to say anything on top what Art has already said, although I would give myself a freedom to put the LCV in remote location - that is something which is done quite frequently in desert gas plants, having control room few kilometers away from the plant (especially in the case of high-pressure sour facilities, where QRA dictates minimum distances). The speed of electronic signal from DCS to the LCV is really not an issue, in my opinion. It's all the same - what if control room is equally away from the separator and the final condensate destination?

If you put an LCV at remote location, the condensate line will be 100% full of liquid - down to the LCV. I believe ambient heat transfer effect will overweight the impact of pressure drop on condensate flashing, and the line will probably run liquid full, which is something you can calculate based on pipe geometry, liquid velocity and ambient conditions. What is the purpose of insulating condensate line?

If the line is 100% liquid full, there are no serious controllability issues. In addition, since you have mentioned that condensate carryover into the separator vapor line will not cause any upsets downstream, there is no reason to have additional concerns regarding the reliability of separator level control. Having two-phase flow downstream of LCV (if it is located close to the separator) will probably cause frequent hammering, due to subsequent steam condensation. That is something which I remember from my Operations days, when saturated condensate passes through the control valve and goes to distant location - hammering will be happening for sure.

This is my common-sense thinking, based on previous experiences and also considering few existing plants in remote locations. I completely agree with Art's comments, and the only difference is my additional observation that I have seen LCVs located quite far from vessels at which the level was controlled, and they were working quite fine.

[ankur2061]

Dear Zauberberg,

Thanks for your reply. Indeed, for the scheme 1 I have checked in HYSYS that there is absolutely no flashing occurring and the stream composition shows 100% aqueous phase at the LCV inlet. Essentially it means that for the Scheme 1, the inlet to the LCV is 100% liquid.

Based on what you have mentioned about remote location of LCV from your personal experience, it seems that there may not be such an issue of controllability and the LCV can be tuned to get the desired response for level control.

As far as insulation is concerned, the condensate (two-phase) after the LCV is sent to a LP separator to generate LP steam and we don't want to lose heat for the purpose of generating LP steam.

Thanks once again and I intend to stick to my "Scheme 1" unless somebody can prove it is an engineering blunder.

Regards,
Ankur.

[mishra.anand72@gmail.com]

If two phase flow is not required, scheme 1 is preferred scheme. It does not affect controllability.

[sathyan srinivas]

HI
WHAT I SEE HERE IS OPPORTUNITY TO SAVE ENERGY. YOU TAKE THE CONDENSATE AFTER NEARLY 60 BAR DROP. ANOTHER THING IS NORMALLY SCADA IS USED FOR LARGE DISTANCES AND PARTICULARLY UPSTREAM INDUSTRIES WITH LOCAL PLCS. NOWADAYS WITH SMART CONTROL VALVES U CAN BRING DOWN THE RESPONSE TIME. ALSO DCS WOULD NOT HELP HERE AS THERE WILL BE SIGNFICANT SIGNAL LOSS, EITHER U NEED BOOSTERS IN BETWEEN AND ADDITIONAL COST. SINCE IT IS USED FOR ENHANCED OIL RECOVERY , CONSIDER USING EXCHANGER FOR HEATING UP THE CRUDE WHICH ENERGY EFFICIENT.

[joesteam]

I am not sure of the answer to your question about the lag time if you have the control valve far away from the separator. However, did have two thoughts.


[*]If both your steam and condensate need to travel 2.5KM, why not locate the separator at that point and save on all of the condensate piping??
[*]I did a quick check on pipe sizing, the 8" for two phase flow seems good, but I think you can reduce the pipe size if you put the control valve at the end of the run.

Just my 2 cents.

[ankur2061]

I am not sure of the answer to your question about the lag time if you have the control valve far away from the separator. However, did have two thoughts.


[*]If both your steam and condensate need to travel 2.5KM, why not locate the separator at that point and save on all of the condensate piping??
[*]I did a quick check on pipe sizing, the 8" for two phase flow seems good, but I think you can reduce the pipe size if you put the control valve at the end of the run.

Just my 2 cents.

Hi Joe,

First the separator is being fed wet steam from a steam generator which is located 2.5 km from the final destination of the steam and condensate. This is how the location has to be and I don't have layout choices.

Second, the flow upstream of the LCV is single phase (liquid only). Two-phase flow occurs only after pressure reduction from the LCV. The upstream line of 8" Sch. 80 gives only a pressure drop of 0.5 bar (7.3 psi) for an equivalent pipe length of approx. 3.0 km. The line downstream of LCV which sees 2-phase flow will be sized as 12" to account for the two-phase flow and the resultant higher pressure drop and velocity.

Hope this explains the situation.

BTW, when you say "Just my 2 cents', is it a New York phrase or typically American.

Regards,
Ankur.

[joesteam]

Ankur,

I understand now, and agree that you will need two pipe sizes for just liquid or two-phase flow. Many people do not uderstand this, that the gas phase takes much more room than the liquid. Just the cost of the piping of 2.5km would make me want to put the control valve far away from the separator, and get a good controller that and instrument guy can tune to take into account any lag in the flow/time.

JMTC or JM2C - just my 2 cents, is American, just like IMHO (in my humble opinion). Comes mostly from text messaging and these boards where you want to say something that may not be 100% correct and up for discussion.

Look at: http://en.wikipedia....o_cents_(idiom)

Joe

[Zauberberg]

The cost of two-phase piping downstream of LCV (if it is located close to the separator) would be justified in case you have very expensive equipment being fed by the steam - this is equivalent to supplying the steam turbine, for example, as pointed by Art. In such case, you wouldn't want to take the risk of feeding the turbine with condensate, and you would accept associated extra cost of two-phase piping. If I have turbine in such location, I would definitely ask for such arrangement. However, in Ankur's case, there is no reason to go for that option; level control system can be configured to function quite decently, and any condensate carryover can be tolerated.