well field pump sizing

[drain]

Does someone with experience in parallel pump operations, using well fields and two discharge heads have time to look over my procedure below and comment. Or direct me to resources where I can check myself. I haven't done a parallel operation with 2 discharges before and may have missed a few steps. Thanks for your precious time.

My data is below for use of two constant speed pumps discharging directly into two tanks:

Well #405 -- rated Q = 175 USgpm, static lift range 440 - 525 ft

Well #407 -- rated Q = 250 USgpm, static lift range 395 - 450 ft

Both connect to a common main some 500 ft downstream, then discharge to two tanks at different elevations. Here's what

Planned procedure is as follows:
1)calculate suction and discharge losses up to the manifold for each pump. Plot separately.
2) calc losses for combined flow from manifold onwards to each tank with high and low C-values
3)plot max and min system curve for static lift range from manifold to discharge for combined flow, then superimpose on plots in 1)
4)Add the pump curve to each plot, then subtract the losses in 1) from the pump curves to re-plot pump curves (explained in Pumping station design, 2nd Ed)
5)Add revised pump flows together to create a curve for combined pumps
6) determine operating points for each pump individually, then combined
7)check that pumps are working at 80% of shut off head

 

[d23]

Drain:

First I would like to define terms so we are both on the same page about wells.

Static Fluid Level: this is the point that fluid level or bottom hole pressure achieves when the well is not producing.

Producing Fluid Level: This is the fluid level or bottom hole pressure in the well at a fixed flow.

PI Productivity index: This is GPM divided by (static in feet minus producing fluid level in feet.)

Being “Civil/Environmental” engineer you may be pumping flood water etc. with an almost infinite “PI”, but I just wanted you to understand my perspective before answering.

I am not sure about the term “suction Losses”. You need to be very careful to make sure the suction or pump intake has plenty of fluid where “losses” need not be calculated. If you are starving the intake for fluid you will be cavitating your pump. There is the other end of the spectrum. If you have a high intake pressure assume that your pump is a dp device (I’ll catch hell for that ). Pressure is additive.

I think you have most of the rest right. Calculate the flow for each pump separately. Then calculate the flow for both pumps together with the adjusted friction loss from the header on. This will require some iteration until it all matches.

I’m not sure about the 80% flow commit. Pumps have a minimum and maximum operating range for the pump. You need to get that from the manufacture and stay within their recommendations.

If it’s a submergible motor / pump you need to maintain X amount of fluid past the motor for cooling. You also need to consider the amount of thrust on the thrust bearing with this type of configuration. With 500 more or less feet lift required the thrust bearing should not be a problem, but you should consider it just in case.

One last thought is that for a centrifugal pump the API specifications are to be plus or minus 5% for head and flow. While doing iterations you may want to spend a few extra minutes and calculate “what if.” If the pump manufacture can provide actual test data it would help the accuracy of your calculations.

Hope this helps!

 

[drain]

d23
What is a dp device?
If we choose pumps based on the rated well capacity, will a 250 gmp pump in well 407 throttle a 175 gpm pump in well 405? The two wells join in a common header about 500 ft away.
Very helpful. Thanks.

 

[d23]

Drain:

You are correct there has to be throttling. The 250 GPM pump will throttle the 175 GPM pump and the increased head or pressure from the 175 GPM pump will also throttle the 250 GPM pump. That is why you will need to make several iterations until you find the balanced system. With both pumps running the friction loss will increase so the flow will decrease for each.

dp device:
I think a better way to say dp device would have been to state that you are only lifting from the producing fluid level (for water) not the pump intake. In most cases when I design pumps in the petroleum industry I consider fluid level in terms of bottom hole pressure rather than feet from the surface. This proves to be more accurate with multi phase flows. Lighter crudes will achieve higher fluid levels than say saturated brine with a constant bottom hole pressure for each case. You then consider the dp of the pump plus the intake pressure not feet from surface. Sorry for the confusion.

 

[DLANDISSR]

DRAIN:
Some additional information.

The static requirements and friction loss should be calculated for each pump from the well liquid level through the pump to the system discharge. This will be the head each pump must over come. A plot of this data will be the system curve for that pump.

The pump curve for each pump should then be over laid onto the system curve for that pump. This will show the relationship between the system and the pump curve.

With both pumps running the flow through one pump will affect the flow through the other since with a centrifigual pump flow varies with head.

The static requirements and friciton loss should also be calculated for each pump with that pump only running. By over laying the pump curve on this curve you will see that with only one of the pumps running the flow from that pump will increase. In this case that well may be over pumped.


You will need to either reduce the pump capacity so that when only one pump is running the well for that pump will not be over pumped or you will have to put a flow restriction device on the pumps so that the wells cannot be over pumped.

A Griswold valve is a possible candidate for a flow restricting valve. Griswold valves maintain the flow through them by adding restriction to the system if the flow increase. They are accurate to +-5% or less. There is a website for Grisold valves.

The pumps should be certified confirming that they will produce the flow and head specified. After installation the shut off head can be checked and the flow at required head measured. The head at required flow may or may not be at 80% of shut off head. This depends on the pump and where the design conditions have been selected on the particular pump curve.