Control philsophy for 10 pump sets in series

[stanier]

I have a application where there are 10 sets in series of 2 in parallel pumps. The units may be up to 1.6 miles apart and 300ft head. The connecting pipework is PE with a pressure class of 320ft.

The client is reluctatn to rely on telemetry and would like a standalone control system. Each unit has a variable speed drive. This is to limit discharge pressure to protect the PE and also to ensure that the NPSHr is not < NPSHa.

One concept is to run the first set up to speed and use pressure control on the suction and discharge to limit speed of each set. The suction control would ensure that the speed did not exceed that which compromises the NPSHr and the discharge to limit the speed so as not to over pressurize the PE. MOdelling of the system shows that hydraulically this would work with the pumps selected.

The other scenario is to use individual flowmeters set at the desired speed and use the above pressure control as set point cascade control to protect pumps and PE pipe.

The pump sets run at a fixed speed in campaign mode, only changing bi annually, what problems do you see?

The VFDs are provided as the pump sets could be moved to different locations, heads, distance apart and must have infinitely variable characteristics and set points. Fixed speed with pulley changes were looked at but were considered to be unable to cope with the number of scenarios required.

Fluid contains solids so control valves were discounted.

Positive displacement pumps were considered but discounted because of cost.

Geoffrey D Stone FIMechE C.Eng;FIEust CP Eng

http://www.waterhammer.bigblog.com.au

 

[BigInch]

Stainer,

That's an interesting system!

Is this a high static head system, or is flow/diameter ratio high? (why the 10 stations?) WAG its hi flow in small pipe.

My opinion is that it would run better using pressure control. I don't find flow control effective in pipeline work, except as overrides at supply and delivery points. I can't see the need from here to have flow control at each station. Probably a waste of money and certainly of commissioning time.

I agree with the client on the individual distributed controls. Otherwise its a major SCADA system he'll have to get running and then keep it running. Not something to take lightly.

Problems may be start timing of the 10 "pump stations" without getting a "hammer", which is why I'll suggest the next option.

Do you have bypasses with check valves at each station? Usually helps startup, but may have static pressure implications with a failure and backflow to low elev.

Is this a high static head application? Something tells me it is not, but its hard to tell with 10 stations.

Have you done any transient analysis? Both for Start-up and for a 1 pump trip and a station trip?

The movable part worries me a little. Any chance of getting slack flow at high points now or in future?

VFDs might work for you, given you probably have flow turndowns of 2 to 1, so that's a good ratio to start with, as long as all pump stations don't have high static heads, but they also fit in with the future mobility plans.

I'd like to know more about this project. I would appreciate if you could find my e-mail (profile tab) on my virtual pipeline page and tell me anything you would like to (or have time to) about it.

Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[BigInch]

Forgot this, damn near the most important part.

Do you have recirculation back to suction at the pumps? Sometimes just an on-off valve helps (thinking of your solids flow). If you can start the pumps and keep them running at low rpm without heating up and without having them running out of NPSH and tripping offine, it will help a lot getting the lot of them up.

Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[stanier]

Thanks for the responses.

The static head is 45 to 100m depending upon the pump set deployment and mine plan. Design is for worst case of 100m. That is why the VFDs are on the system.

Transient analysis is yet to come when the order is received. Pipelines are PE so celerity is reduced hence pressure increase. Given the high static head, column separation is not likely to be a concern unless the pipe is not continuously rising. Velocity in the line is reasonable < 1 m/s.

At present there is not recirculation back to the pump suction. Experience shows that this doesnt prevent fluid heating up. There are check valves at each station. Designed for the solids in the fluid. Soft flap types will also reduce hammer as they are short travel qucik closing.

A high degree of labour involvement is required to start the system up if telemetry is not used. Radio contact and operators at each station. Telemetry is preferred, if not for control, but for one person to monitor what is going on. It has been used before successfully but It is agreed it does cost money and involves technical back up.

Damage from cavitation is not an instant thing and the pump sets are operated on a strategy basis so some snoring can be tolerated. Pumps have hard allo impellers to resist wear from solids. PE can tolerate some over pressure . So both suction and discharge control loops can be slow and lazy.

Theiodea is to fill the line at night. PE can get very hot and lose strength when empty. So slow filling and cooling is the order of the day. All this under manual control. When flow is about 75% the pump sets can be switched to auto and control wiht a clamp on suction conditons limiting speed and discharge pressure allowing the pumps to operate up to design rating of PE. As flow increases form first station it increases the suction clamp on successive stations.




Geoffrey D Stone FIMechE C.Eng;FIEust CP Eng

http://www.waterhammer.bigblog.com.au

 

[BigInch]

Thanks Stainer,

I certainly understand why transient analysis is "on hold". Better to wait.

My question about recirculation was primarily related to having some ability to run just a short time at lo rpms as a station starts while the others are still line packing, but with the bypass checks there, looks like you can probably bring on a new station without too much trouble. If recycle lines are short or small diameter, there's not much heat protection. If heat does turn out to be a problem, you could recirc back to a fiberglass tank or something.

I recently designed a very hi static pressure system where bypasses could not be used at 2 stations. Too dificult to open a check with the pressure above, which made for long hot ramp up times.

Hmmmm, manpower vs SCADA. SCADA comes with a much more expensive form of manpower. Just keep it in ... mine.

The rest sounds ... interesting. Good thing the pump stations arn't a hundred miles apart.



Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[Ashereng]

Allow me to put my 2 cents worth in.

Most of the pipelines that I work on are drivent by pressure controls as well. For each pump station, protection is based on pressure. This, I am in agreement with BigInch.

On the telemetry, or SCADA side, my experience is that SCADA is preferred over manual communication using telephone or radio. I guess since I am in I&C, there is a bias. However, all my clients use SCADA for their pipeline. The systems require some expertise, but this is par for all of my client since SCADA and PLC/DCS are very similar technology, and the same group usually handles both. The benefit of SCADA is the quickness of data transmission, and data collection. Of course, SCADA costs money, not just for the hardware, but the radio antenna, tower, pile(s) and all the configuration/programming/maintenance.

If your client is wary of SCADA's cost, then that is their decision of course.


"Do not worry about your problems with mathematics, I assure you mine are far greater."
Albert Einstein
Have you read FAQ731-376 to make the best use of Eng-Tips Forums?

 

[BigInch]

Sounds good, but don't get me wrong about SCADA systems. They do have their uses. I'm just afraid it might be overkill here, as well as I've never seen them work right out of the box. Some seem to take a year or more to get them up and running. It will not be a minor SCADA system either. With 10 (probably relatively small) pump stations, I think it could wind up to be an investment way out of proportion to both total cost and benefit. When in doubt, I tend towards "KIS". It could be added later. Besides, getting the job will be tough with it included.

The deciding factor may be how long this system can keep going once its started. I tend to think that the manpower issue will work out to be a lot more people in the field keeping the system running than is currently thought, with SCADA or without and there will probably be someone around to push the right button when necessary anyway. It would make a good op-reliability simulation.

Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[stanier]

Thanks guys.

Re the SCADA debate. Labour is a dollar a day. Men with a red flag on top of every hill is a lot cheaper than the technician to maintain the SCADA. Atmospheric interference with radio telemetry has lead to unreliable performance and a basic mistrust of such technology.

As this is a dewatering exercise, if the pumps go out for a while it doesnt interfere with much. Just need to pump harder to make up the difference.

Geoffrey D Stone FIMechE C.Eng;FIEust CP Eng

http://www.waterhammer.bigblog.com.au