Pipeline design with pump station and gravity flow

[farhadsh]

Hi, I am looking for some advise on the reverse flow for below scenario:

I am designing a pipeline and pump station to transfer water from low elevation to high elevation and for simplicity I use imaginary numbers:

-Pipeline size 16”
-Pipeline length 5500m
-Flow rate 0-610 m3/hr (variable pump speed by VFD)
-Elevation change 500m (elevation 0 to 500)
-Two pump station selected: first PS at the elevation 0 and second PS at elevation 250m
-Friction loss calculated to be 50m
-Each pump station will generate 275m head. (assume friction loss equally absorbed by each PS)
-Assume pipeline with constant slope from bottom to the top.
-Water supply is from pond to first pump. water is released to pond at destination.
-Pumps and pipeline are directly connected and controlled by process. No buffer tank included in system;

Scenario : During emergency power failure the pump stops and flow has to be drained from same pipeline back to the sump to avoid freezing inside the pipeline. There is bypass around each PS for this purpose. Flow is reversed in pipeline and drains through same pipeline by gravity.

Question1: Is during reverse flow the pipeline full with liquid?

Question 2: If Yes to questions 1 I have to assume hydraulic pressure at pipeline bottom equals the static head (ignore friction loss during reverse flow) and pipeline has to be designed for 500 m head. Is this statement correct?

Question 3: If No to question 1, I assume pipeline is partially empty and under atmospheric pressure so pipeline design pressure with reverse flow is atmospheric. Is this statement correct?

Question 4: How to calculate time the pipeline will be fully drained in either full flow or partial flow concept? I can use the (2*g*Elevation change )^0.5 formula but since head is time dependent I look for a more accurate way.

Thank you,

 

[1503-44]

1 No, if you do not siphon from the discharge pond. When you start, yes. As you drain, no, if you have a vent valve at the top.
2 Yes. If bypass can open when the 5500 m length is full.
If you could first drain the lower segment, then drain the upper segment into the lower after you open the bypass at PS2, then no. You would only have 250m head max. You should have a pressure relief system to insure no high pressure is created by accidental bypass opening.
3. If air can enter through a vent, atmospheric. If no vent can open, then it will create a vacuum as the water drains out the bottom.
4. Calculate the flow rate in the pipe backward flow. Use 0 psig at the top of the water column and the P = head x density as the pressure at the bottom. P will change as level decreases through time. Use a time step that changes with the calculated flow rate. Do the calculation for each time required to lower 1m elevation. The total time of all time steps will be reasonably correct.

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[LittleInch]

farhadsh,

I realise this is theoretical but a few issues leap out before you I answer your questions.

With a high static head to friction loss ratio (10:1), A VFD is a complete waste of money and power. To do 1m3/hr your PS will need to create 250.5m of head.
To do 610, you'll need 275. That's less than 10% difference so the change in speed for a centrifugal pump is lass than 5% so the flow will only vary by 5% as well.

Direct connection can be good, but means you need both pumps working at all flow rates - doesn't sound like a good plan to me. Think seriously about a break tank if you want to work this way - makes life a lot simpler and prevents one pump starving the other if it gets out of sync.

1) As mr 44 says - at the start yes, at the end no and in between the air will enter from the top hopefully
2) Maybe. Depends on how you design and set up the bypass around the booster station. Design for 500m is the lowest risk approach, but you could introduce a pressure regulating vale on the bypass to reduce d/s pressure to say 1 bar (10m). But you would need to make this quite reliable and maybe need to have some expensive controls. Or drain the top section into a break tank (I told you it made life simpler. Then each section could be designed for max 275m head ( design for 300m probably to take account of shut in head from the pumps)
3) NO Totally FALSE
4) You will need to control the flow via a control valve otherwise you'll end up with a huge flow and velocity as you have initially maybe 500 or maybe 250m head compared to your 50m for friction loss.

If you want a time stick this into a transient analysis program or break it down into steps of head of say 50m as the pipeline empties. Just use normal flow equations for head difference, pipe size, roughness etc.

But much easier to control it via a control valve.

It must be said that if you're designing this, your questions indicate a low level of knowledge and experience. Is there not someone there you can ask these questions to?



Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[1503-44]

We did exactly that in Abha, but with gasoline, diesel and jet fuel.
Two pump stations, both 2500m head, with a 3012m elevation head.
Pump station 2 had a bypass with block and check valves and also a pressure relief to a tank, in case the check valve broke. We didn't ever want to see 3000 to 5000+m of head in the lower pipeline.

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[farhadsh]

Thank you very much 1503-44 and LittleInch for your tips.

let me first discuss with 1503-44, I will add another post for LittleInch:

I have to add more assumptions and repeat my Q1. Assume there is a electrical isolation valve at discharge pond and vent and vacuum valve at high points in system. Once there is power failure, pumps stops and flow reverses and discharge valve closes. There is no siphon here.

Q1- Is Pipeline during reverse flow drained in sections full with liquid? of course the liquid level will get less and less until all flow is drained. but if pipeline is drained in full flow sections the pressure at bottom will be 500m in the beginning. if the pipeline is not drained in full flow sections there will be air entering to pipeline faster than liquid draining the system so pipeline will be atmospheric, but I think the likelihood of this scenario is minimum.

For your reply on Q2 I like the idea to sequence drainage so lower segment is first drained, I just have to make sure once upper segment is opened, by time water travels 250m down it will not cause huge impact on pipe. (new topic for myself to investigate )

Q3-based on my new assumptions on Q1 do you still think pipe bottom will be atmospheric?

Q4- understood.

 

[1503-44]

Electrically operated isolation valve. Got it. No siphon. (A check valve there could ensure it never happens.)

If you open the bypass when 5500m of pipe is full, you will immediately get 55bar pressure at PS1.

When draining the pipeline, both the reverse flow rate at every point in the pipeline and the air entry rate at the top would have to be carefully controlled. If water outflow is balanced with air inflow into the pipeline above, you will have atmospheric pressure at the top. But...

As the draining progresses, there is at least some possibility that steady flow is not maintained everywhere, all along the portion filled with water as time goes on. If you have some up and down slopes variations in the elevation profile, it can get complicated. Anytime atmospheric pressure at the top is not enough to push water over a hump in the pipeline below, vacuum conditions could develop at the hump and water will get trapped in the upslope immediately above the hump. Vacuum conditions can develop in any portion of the pipeline below such a hump. You would need to carefully examine the elevation profile to know if it is possible and where that might happen. A Transient flow analysis of some kind is needed. You could not automatically assume that steady state flow during draining will be happening at all places.

The bypass should probably never be open when there is water in the top section. Check valve in the pump discharge is a good idea to keep high head off the pump, reduce high motor starting torque and prevent spinning the pump backwards when starting. You may also need a recirculation around the pump to enable pump cooling while starting and reaching pressure sufficient to open the discharge check valve. Depends on your pump driver. If diesel driven, that may take some warmup time. Electric drive is a faster start. Don't forget to look into providing pressure relief to tank at PS1 and possibly other points that may be subject to overpressure should the bypass get opened at the wrong time, or if the check valve at pump discharge at PS2 get damaged and reverse flow in the pump.

You will probably never have atmospheric pressure at the bottom while there is any water in the pipeline.

I would ensure that the pipeline is designed for
1) highest pressures possible in each segment, either 25bar in both sections, or 55 bar at PS1, depending on the PS2 bypass operation.
2) full vacuum everywhere.
3) pressure relief at high pressure points, if the pipeline design pressure is lower than any higher potential pressure (including any possible accidental conditions.)

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[LittleInch]

Thinking about this a little more, why do you want or need to drain this pipeline?

Buried lines are usually below the frost line and don't freeze. The above ground stuff might, but wrap it in insulation and stick on some trace heating.

To fully drain, the line needs to have a constant slope.

Beware that re start will require good control over the flow rate as the pipe will initially have no resistance to flow which will then increase as the line fills up. If you don't your pumps will just trip on end of curve operation

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[farhadsh]

Discussion with LittleInch:


whether variable speed pump selection for this example makes sense or not is an excellent topic you raised.

1-I will investigate further once we move to the next phase but for now I assume VFD for capex.

2- Systems with high static head to friction (10/1 here) use of affinity law may generate errors as pump efficiency in not constant with lower speed. To reduce the head by 10% the speed needs to be 5% reduced but flow will be reduced much more than 5%.
3-for systems with high friction loss to head affinity law is valid. I do not yet know if there is any golden rule when VFD economically makes sense. I will switch to throttle control valve concept if VFD is waste of money.

4- It is on-grade steel pipeline and due to cold weather risk of freezing so need to drain. EHT might be costly but an option.

5- If I use hdpe instead of steel pipeline chance of freezing will be reduced.

 

[LittleInch]

Ok

1) - OK, but don't forget also the extra OPEX. VFDs eat about 8% of the pump power as heat in their cabinets
2) - within the speed range that is going to work the efficiency will be pretty close. The issue is that the total flow range is over about 5% of the pump speed range. That makes it very difficult to control flow with the pump alone.
3) Each system has its own economics, but unless your flow range is quite wide and you are operating at lower flows for long periods, then with extra CAPEX and OPE it often doesn't make sense. Remember lower flow with a control valve is lower power on fixed speed motor
4) Really? Why don't you bury it? Surface laid is not usual outside of deserts or within a fenced area. Too much risk of damage, snaking, thermal issues etc etc 5km plus sounds like a long surface line to me. You ight start to need expansion loops if its empty and the sun starts to shine
5) Beware of temperature derating issues with PE plus pressure ratings generally limited to about 16bar, 2o at a push.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[georgeverghese]

During flow reversal, control valve will see 50barg approx initially. And if this FCV (or HCV) fails open (or operator accidentally opens it 100% at control room), velocity/momentum will be so high pipeline may break. So suggest adding a safety restriction orifice downstream of FCV that will keep max drain velocity below 14m/sec, equivalent to rho-v2 = 200e3 (SI units) even when FCV is wide open. Pressure upstream of drain FCV should also account for max head in discharge pond at elevation 500m.
Suggest 2 check valves in series at PS2 pump discharge. In any case, without a break tank at PS2 suction, think you may have to ask for design pressure of 50barg for PS1 pump.

 

[1503-44]

Forget the VFD. It will not work in a high constant head application.
At RPM lower than rated operating speed of the pump, VFD will not develop the 250m head.
Head is proportional to RPM^2.
At 90% speed, you only get 81% of the rated head.
At 80% speed, you only get 64% of the head.
At 50% speed, you only get 25% of the head.
At 40% speed, you only get 16% of the head.

When you always need 100% to reach 250m, you can only run at 100%

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[farhadsh]

I will add another reply for georg.

Discussion with 1503-44 and littleinch:

Please let me know if you are in agreement with my notes below:

The PS1 Head range is 250-275m
Rated condition 610 @ 275m
100% rpm 275m 610m3/hr
‎94٪؜ rpm 250m 573m3/hr

The actual flow will a bit less than 573 but not a lot.
If I combine VFD + throttle valve, it might do the job, but appears to be waste of money and energy.

It drives me to forget about vfd and to control the flow by throttle only.

With throttle I do not consider to put any recirculation line after each PS to source pond. Do you see any design concern here? As long as min pump flow is checked it sounds to be okay.

 

[1503-44]

Looks workable, except for vfd. Totally forget about them. Yes, really.

Plan on recirculation at the pumps for now. You won't get any significant flow into the pipeline until check valves open at a pressure of 25-27 barg is reached.

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[farhadsh]

Hi georgeverghese
Can you explain further about formula you discussed and where 14m/s is coming from.

 

[georgeverghese]

For general piping, rho-v2 should be kept below 200e3 (kg/m3).(m2/sec2). Beyond that, additional pipe support reinforcement will be required to avoid mechanical damage. Given that water density is 1000kg/m3, in this case, v = 14m/sec.

 

[farhadsh]

Georg
thanks for clarification. Using RO at bypass can be help to reduce pipe design pressure at PS1, but I will have to calculate time the pipeline is drained before it freezes in winter.
I think for now I am good with assumptions. thank you all.

 

[Nutzman]

As Littleinch states, Why drain the line? Bury it. I did a 2.6km 300mm PE lined steel pipe, surface laid (mining tailings disposal operation), with a temperature range of 5-50 deg C. We ended up installing expansion loops. nightmare of a design.
Another downside for HDPE lines, has the maximum excursion limit been considered? Meaning, when you have a power failure and the column of water comes back at PS1, "water hammer" will occur. Now the 500m becomes a load more. If the line is designed to withstand the "over pressure", can it tolerate the "under pressure"? Dropping below 50% of the line (pressure)rating? Remember that water hammer travels at the speed of sound in water. (depending on the gas content) this velocity can vary considerably?
This design is more than you can do "on the back of a fag packet"?

 

[1503-44]

At a velocity of 14m/s (thanks for that George) there will be potentially severe issues with water hammer.

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[georgeverghese]

@1503-44,
Am assuming there will be sequence trip valves upstream of the pond inlet at el 500m, and downstream of PS2 and PS1 discharge line check valves, that would close before the PS1 bypass drain valve opens. So I think there shouldnt be any water hammer concerns.

@farhadsh, pls note the RO on the PS1 bypass/drain line cannot help with reducing design pressure of PS1 pump. The dual check valves on PS2 discharge also cannot help with this, since the drain bypass around PS2 would be teed off the main pumping line from downstream of these check valves.

 

[1503-44]

The sequence trip valves are exactly my concern, as well as loss of pump power. All valves will have to close very slowly to avoid hammering.

The high segment of pipe must be drained before opening the bypass. In which case, I'm not sure why a bypass is even needed. Any bypass should have a high pressure prevent lock, pressure relief and perhaps two block valves.

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."