Overall Pump System Efficiency (level control)

[rkwolf]

Hi all,

I am currently working on a project that requires evaluation of overall pump system efficiency, specifically in cases of waste water treatment facilities. My goal is to determine what pump combination is more efficient for a given incoming flow rate in order to maintain the same level in the wet well.
Just for the sake of discussion, let say I have the following:

two 100 hp pump (pump_1 and pump_2)
one 250 hp pump (pump_3)
one 350 hp pump (pump_4)

in the same wet well and I want to maintain the same wet well level for any given incoming flow into the well (since fluctiation in incoming flow is reasonable for a wwt facility). In addition let say that the two 100hp pumps are VFD.

The question then becomes how to determine which combination is the best combination from pump efficiency stand point for a particular incoming flow rate? Would it be better to run two VFDs at the same time at the same speed (ex: 60%) or stage them with pump_1 running at 100% and then pump_2 running at something like 50%?
And if the flow continues to increase, what would be the best way to turn pump_3 on? Would that involve turning pump_2 off as pump_3 is turned on and then mathing head with pump_1 by adjusting VFD (pump_1)?
I realize that's quite alot of questions and since I'm more of a thermo guy, I am a bit confused. Also please consider the fact that these numbers are just for discussion sake. Any help or tips would be greatly appreciated.

 

[BigInch]

Forget the level in the wet well, you're simply trying to get pump flow to equal inlet flow, so you're just trying to find out what is the most efficent way to pump some given flowrate using a combination of x pumps.

First, throw out the VFDs, that will only initially tend to confuse the situation and everyone involved. I doubt you will be using them in the end anyway, especially if you have any static head to contend with. As a guide for possible consideration later, if you have some kind of "typical" system static head, you must need to be able to run at 70 to 85 percent of design flowrate for a LOT of the time to be able to use VFDs, or no static head and have to be able to meet WIDELY varying (20 to 90% of design flowrates all the time.

The only way to run pumps in parallel is with equal discharge head, adjusting the flowrates to each pump to get it and keep it that way.

To be able to halfway attempt to answer your additional questions, it would be more helpful to know the design flowrate of each of those pumps as well as their power. Generally I try to target the running of all pumps between 80 and 110% of their rated flowrates, choosing the number of each model of pump to facilitate matching the resulting operating bands as best one can to the required process flowrates.

It will be fairly obvious which configurations will burn the lesser amounts of power. If you think VFDs will tune that more favorably, then start adding them, but very carefully. You'll probably be wrong unless you meet some very special additional conditons.

Let your acquaintances be many, but your advisors one in a thousand’ ... Book of Ecclesiasticus

 

[rkwolf]

I see how introduction of VFD's complicates things, but what if throwing them out is not an option. Let's say that like in the example above I have two 100hp VFD pumps and then one 250hp constant speed pump.

The way I understand this, the best way to run would be the following:

1) Pump1VFD speed is varied to handle the incoming flow
2) If flow reaches the max flow that Pump1VFD can handle, then Pump2VFD comes on. Both pumps' speeds would be equal and set to handle the incoming flow (same head/flow per pump).

From what I've gatherd so far, that would provide best overall efficiency, rather than continuing running pump1VFD at 100% and then running pump2VFD at some other speed. I understand this so far.
But what if the flow continues to increase and pump3 has to be turned on? What is the best way to go about doing that? Do I need to implement some sort of a valve to match head with two VFD pumps? What would happen to the flow in that case?
I do not know flowrates of these pumps or their power. I am supposed to come up with a way to "predict" best combination of pumps for any given flowrate and any given combination of pumps (w/ some of them being VFD, but not all). I does seem like I need more information that is available to me. This "prediction" may have to be designed differently for each given waste water treatment plant (their individual wet well set-up). I do appreciate any help/advice that is given.

 

[BigInch]

I meant throw them out "figuratively" at first, until you understand how pumps run in your system without VFDs. Then you will be able to more easily see how VFDs might help, if they can help at all.

Running one pump at 50% with a vfd means that you will only generate 25% of head. If you can flow at 50% with 25% head, then you probably can operate as you plan to, if you can't flow 50% with 25% head ... well, then you just can't operate as you plan to. If you must have more head than that 25% that the vfd will give at 50% speed (50% flow), you must not use a vfd, but instead should consider a flow control flow valve.

Obviously when you are at max flow that one pump can handle, you are at 100% speed and 100% head. Bringing on a second pump will require you to quickly change that first pump's settings to (and bring on the second pump at) 50% speed, for which you will get 25% head. As you can see, you were first doing that flowrate with one pump at 100% speed and 100% head, but now you want to do it by quickly changing the pumps to give you the same flow, but at only 25% head. You must know how you will also control your system to make that possible. Can you adjust your system to make it possible to carry the same flowrate, but do it now with the pumps delivering only 25% of head?

Without the specific knowledge of the particular pumps and the precise system characteristics, this is a fools erand. What you propose is a difficult problem to solve when you do know these things. I would suggest trying to develop a genetic algorithm to do it. I think that might have the best chance to arrive at a solution, but as I'm sure you can understand, the complexity of this thing probably increases geometrically with the addition of each pump and probably exponetially with the addition of each VFD. Good luck.

Let your acquaintances be many, but your advisors one in a thousand’ ... Book of Ecclesiasticus