Sump Pump Issue

[Engineering1012]

New engineer here, this bounces off a question I posted a bit ago on recirc valves and got some help on. Bare with me on the information below.

By design, there are three vertical pumps that pull waste from unit buildings and send it to are large atmospheric waste tank. They pump from a sloped sump that contains two settling chambers and a pumping chamber, separated by a weir. This helps the large particles that get through the initial strainer basket to settle, thus allowing the cleaner fluid to dump over into the pumping chamber. Out of the three pumps, one is a primary (constantly running) and another is on standby to switch on when the tank rises above 40%. The third serves as a backup along with an external diesel pump for emergency situations. The three vertical pumps feed into a common header where there is a throttling control valve. We've had numerous costly failures over the years before I got here, mostly dealing with failed bearings that cascade into impeller failures...etc.

I believe the root cause to be that the pumps are operating off their curve. Typical operation requires the primary pump to run at about 350 GPM which is around minimum flow for these pumps. However when they run in parallel due to rain...etc they need to produce 2000 GPM. Jist of this is, we have oversized pumps for normal operation.

I am working on a system curve and struggling in this area. Opening the throttle control valve 1000% the flow was 970 GPM with a discharge pressure reading of 22psig (~51ft). The entire system is tricky though and this is where I struggle.

1. So would that 22 psig on the discharge be my frictional losses in the system?
2. I can't really get an accurate measurement of my discharge tank level since it dumps into the bottom of a tank downhill over 1000 feet away. Not sure how to address my static head?
3. Would I be able to fit a parabola to the two points to fit the curve?

My preliminary options are.
-One small pump as primary, can be submersible or not however it is a diluted nitric acid harsh environment. (you may run into issues if needing to go parallel for more flow)
- VFD on the primary
- Larger recic line (Existing small one not sufficient, however large line may stir up debri in sumps and kill pumps)

All of these stem around first getting a reliable system curve. Any advice on this as well as my project in general would be much appreciated.

Thanks in advance.

 

[Artisi]

Do not understand you concern about a system curve, you know the sump depth, pipe sizes, pipe lengths, elevation for minimum sump level to the discharge point-- calculate a theoretical system curve to start with for 1,2,3 pumps in operation. Do not factor in the by-pass line initially until you understand the system and pump operating points for the various flow conditions.
Alternately employ a competent engineer to do it for you - as we (eng-tips) can't do it for you as we have insufficient information available and we are here to help not act as consulting unpaid engineers.
As for the submersible pump and the pumped product, contact likely pump manufacturers or reputable pump suppliers.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[LittleInch]

Ok, A sketch would be nice to make sure we understand both the system and the profile, but in answer to you questions

1) No. It is the discharge pressure of the pump into the system. thus it is single point on the curve which then represents both frictional losses and static head losses or gains ( difference between your pump centre line and the liquid level in the discharge line)
2) Accuracy is maybe not so important as understanding what this number is and could vary by. So your system curve may need to have both a min tank height curve and a max tank height curve. How far "downhill" is this tank liquid level - this is why a profile sketch would be very useful.
3) I would say you need at least four points to make a curve that is of any use.

If you have a control valve and flow meter, all you need is discharge pressure d/s the control valve into your pipeline. The you should be able to generate 10 points on the curve when you get enough water to pump or if you can up the flow rate for short periods - no idea how big your initial holding tank is compared to flow rate.

Smaller pump sounds like a good plan
VFD might work but you need to see what the system curve looks like
Recirc line using a control valve or ARV might be good, but make sure the exit is suitably dispersed to avoid the dirt stirring up.


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

 

[Artisi]

If you can supply one accurate point of performance - consisting flow, discharge pressure together with an accurate static head, it is possible to generate a reasonably accurate system curve - however this can only be as accurate as the figures you supply.




It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[Engineering1012]

First off, thank you for the input.

Here is the pump curve.

Out in the field I had the operators open the discharge control valve completely with one the primary pump running. This results in a reading of 22 psig which is roughly 51ft of head (considering its mostly water) at a flow rate of 970 GPM. Since the pressure gauge on the discharge is about 11' up from the center line of the vertical pumps impeller, I would add that to my total correct? So my pump is producing roughly 61ft of head at 970 GPM? Even then, it is still off the curve a bit, could this be due to impeller wear?

I've also attached a rough layout. Another question was when you have multiple lines that come into the main 12" waste header is it ok to disregard them since they are just 3" lines mostly?

-LittleInch what did you mean by d/s?

Thanks again guys. Just trying to size create an accurate system curve so I can move forward. Not asking anyone here to do it for me.

 

[Artisi]

The head will be 22 psi plus water level (not pump length) to the gauge point.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[pumpingtips]

Frank217. You can´t produce a system curve with the control valve variability. As the control valve closes, the system curve gets steeper. It´s not a fixed "system". Firstly though you need to determine you static head. At least that point won´t change. It´s on the left hand axis at about 40-45 feet I suppose looking at your sketch. The system curve sweeps up from there. The throttling on one pump will reduce the flow to your minimum, and the pump discharge pressure will increase along the pump curve to the left, as the valve closes.

Interesting to know whether you have a vent at the high point. The downhill run static head then won´t play a part. If the downhill run is a "launder" type flow with enough air getting in, then likewise, the static head after the high point does not count. You only need to worry about static to the high point. This is the limit of the pumping duty, "probably" (there is still missing information in your sketch).

Don´t worry about the 3 inch branches feeding in. They are nothing.

You may be right about oversize pumping for the minimum flow case. But it may be that the control valve is causing issues. I would have preferred a fixed restriction in the form of a separate bypass line around the main control valve for when ONE pump is operating (could be orifice plate, or small bore line). Then when you operate two pumps you can start using the main control valve again. Just a thought and a bit of a guess without having all the data.

 

[LittleInch]

pumping tips - I don't agree. the system curve is dependant on the pipe, it's elevation and resistance to flow. The curve doesn't get steeper. A control valve acts on the pump or discharge pipe to reduce the head to the system curve at the flowrate being experienced. You can produce a system curve by varying inlet pressure and flow. The pressure though needs to be downstream (d/s) of the control valve so you can probably use that one you show saying 21 psig.

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

 

[Engineering1012]

So my static head would be the high point of the system (adding on the suction light height so 35+10)? I've gotten numerous feedback on this, some say it would be just the vertical distance (difference) between the suction tank level and discharge tank level. I don't know this precisely yet but that value would be very small if not negative. Any clarification? Thank you for the feedback again everyone.

 

[LittleInch]

It's a bit odd, but to get any flow, you need to have at least the head to get over your high point. The problem then comes as flow increases.

You need to get a proper profile drawing, distance along the x axis and height on the y. Then plot the head loss line from your pump to the surface of the tank at the far end. If this line is above the high point, then the "static head" at that flow is your negative figure.

If your hydraulic line intersects the ground level then you are probably working in slack flow and the static head is still the height of the high point. I don't have a scanner wher ei am at the moment to sketch this out, but

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

 

[Engineering1012]

I'm quite unfamiliar with these hydraulic profile lines. Doing some homework on it now. When sketching the profile, are they different points just calculated at different elevations throughout the system?

 

[LittleInch]

For liquid flow you only really need the hydraulic head at the start point and the head at the end point and then draw a straight line. For the ground elevation profile to be drawn on the same graph you need as many points as you can get.

So e.g. if your pump elevation is at say 50m above sea level (asl), you start at 50 at kp O and then add the hydraulic head ( say 40m ) = 90m. That's one end. At the other end the elevation is 40m and you arrive at a head of 10m. The other end of your line is at 50m asl( kp 3000m). Just draw a line between the two.

It should always be above the ground profile.

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