pump configurations for variable flow system 1

[marcoh]

Am looking at a variable flow pumping system, with a total flow rate of 200L/s controlled by VSD's and a pressure sensor in the pipework.

Currently the system comprises 1 pump at 200L/s plus a low flow pump of 60L/s (plus a standby 200L/s pump). Normally the system runs with the single 200L/s pump.

The alternative is to provide 3 pumps at 100L/s (ie 2 duty + standby) in lieu of the first combination.

Am looking for pros and cons for both and any control issues that may arise?

 

[BigInch]

First answer this question.
Why do you want to change the current configuration?

The reason should include a big necessity to pump different flowrates between some minimum flowrate and some maximum flowrate.

Please indicate the minimum and maximum flowrates.

You should also have some idea of how much time you will be pumping at various flowrates in between the minimum and maximum. If you know that, please try to describe.

If you will spend 50% of the time pumping between 50% and 85% of a system target flowrate, VSD may be a good option, provided that you do not have a high static head (high injection pressures at both low and high flowrates).

If you only need to deviate from a target flowrate by more than 25% for a limited amount of time, you should probably forget about VSD and work with fixed speed pumps.

When you have the right system and pumping requirements to select a VSD, they can save operating expenses, however you must be very sure about selecting a VSD system over fixed speed pump configurations, or it is very possible VSDs will actually decrease your net pumping efficiency and increase energy consumpiton well over that needed for fixed speed pumps. In some cases the advantages of one over the other are relatively obvious, in others only a detailed analysis can determine the best possible selection.








**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[marcoh]

Its a building chilled water system and flow rates will vary from about 30L/s at night up to 200L/s at maximum design conditions and everything in between. Realistically maybe 1% of time would be spent at 200L/s and most of the time at 60-140 L/s

The pumping requirement is to keep a constant pressure in the pipe so the control valves operate correctly and the change in flow rate is a result of this.

2 pumps at 50% duty is simple but you flatten out the pump curve with pumps running in parallel.

1 pump would just happily pump away and the low load pump would operate when the flow goes below the minimum flow for the large pump.

 

[BigInch]

Its pretty much a waste to have control valves and VFDs. Since head is proportional to speed^2, a VFD won't give you the same pressure at all speeds. If you can get the pressures and the flowrates you need at all speeds, VFD might be the way to go. If you have to run a VFD at a certain frequency to get enough pressure, just so you can operate the control valve to get the right flow, something's wrong there. If that's the case, it leans more towards a fixed speed solution, perhaps using constant speed pumps, but maybe with different flow capacities.
If a vfd is to be used to its maximum effectiveness, you should be able to eliminate the control valve.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[raronen]

"The pumping requirement is to keep a constant pressure in the pipe."

To follow up on what BigInch wrote, I think this pretty might answer the question.

If it is a water system for a building, then the pumping system needs to be capable of overcoming the static head, which I assume is the height of the building. Slow down the pump and you lose pressure very rapidly. Conversely, if you speed up the pump, it will increase the pressure rapidly.

Using the pump affinity laws, flow is directly related to speed. Double the speed, double the flow. As BigInch pointed out, head is ^2 speed. Double the speed, 4X the head.

So, if you have a pump capable of overcoming the static head and producing your 30 L/sec at night, it will have to increase in speed nearly 7X to pump 200 L/sec. At that point, you will have 49X the head (and pressure) in your pipes. I suspect that would be undesirable.

Sounds like a low-flow pump/high flow pump situation...

 

[mtngreen]

BigInch - system pressure will remain constant but flow will vary...the key to making a VSD work is to get a pump with a high rise to shutoff. Granted the drive will probably only operate from 80%-100% capacity. See attached.

http://files.engineering.com/getfil...9ee5-47cb-b4ac-f32be1d2aae4&file=VSD_Pump.doc

 

[BigInch]

That's exactly my point. WHY use a VFD if you only pump between 80 to 100% of BEP? You should get a pump with a BEP of 90% flowrate of what you were previously thinking and run between 90 and 110% of that flowrate at fixed speed with a control valve.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[mtngreen]

Why? Energy savings. Power is proportional to cube of speed. A 20% reduction in speed = +/- 50% reduction in power.

An arguement can be made that if you run the pump back on the curve you'll be running at a lower BHP...so the above statement is rough.

 

[BigInch]

Its not a 50% reduction in the power BILL unless you stay at 80%. If I wanted to do that, I'd buy a pump with a BEP_2 of 80% BEP_1.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[mtngreen]

I wasn't inferring a 50% reduction in a power bill. It will, however, result in a net savings over operating a pump at a fixed speed 24/7/365.

 

[BigInch]

What you have there is the theoretical energy savings, which is really not interesting to anybody. What operators really care about is how much you save on your power bill. Your statement is meaningless unless it evaluates true when considering a reasonable operating schedule. Just assume some reasonable operating schedule, say 1/2 of a normal distribution of flowrates starting from the 80% you mentioned and going to 100% BEP (the highest part of the curve falls over the 100% BEP). Keep pressure constant as in the above description of the OP's problem. We will say the affinity laws are valid. Now, the key part, assume an operating schedule of 1/2 of a normal distribution of 24/7 time pumping each flowrate. OK, now try to prove that a VFD saves money in that system over a pump w/o VFD using a control valve. Let's say the VFD efficiency remains at an average 95% over all speeds and we forget about any possibility of nasty harmonics.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[BigInch]

We need a pump curve. I picked this one, first one I found on the internet.

I give you maximum advantage. Assume the system curve exactly matches the head calculated by the affinity curve head - speed relationship.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[quark]

I don't think the OP gave a correct description of the system and its requirements. If this is a chilled water system for HVAC applications then the system resistance is predominantly dynamic (if the system is carefully designed then it is only dynamic resistance) and thus reduction in speed doesn't effect the head requirement as the head capacity reduces by the ratio of squares of flowrates so as the resistance.

In majority of cases, system exactly matches the head calculated by the affinity laws if proper control scheme is applied. The pressure control as suggested by OP is useless and DP across the headers is the way to go. If we have pressure control then there are more likely chances that control of flowrate will never happen as the reduction in resistance due to decrease in flowrate will be compensated by partially closing control valve.

The control valves are required to maintain constant temperature of the air leaving the cooling coil.

Bell&Gossett has excellent system that controls parallel pumps. One can feed in the pump characteristics into the control system and then it controls the entire scheme for optimality, including shutting off pumps under low load conditions. However, this is pretty expensive and I got similar thing done from Danfoss about 7 years back. The chilled water system control is pretty established, IMHO, and there are many discussions in HVAC forum along with some good links.

 

[BigInch]

OK, getting back to the point of the OP, I maintain that he must evaluate the frequency of time pumping at each flowrate. That's the only way to make the choice.

If flow load is something like a normal distribution, for example, about 120 L/s with a minimum at 60 and a maximum at 140 L/s, then he might consider a 125 L/s BEP pump, with extra motor power to cover the high extreme of 140 L/s for a couple of hours during the afternoon, but return to 120 for the bulk of the time, then eventually get as low as 60 L/s for a couple of hours at night. If the flow distribution is radically different, ie 50% of pump time at 60 L/s and 50% of time at 140 L/s, a wholely different pump configuration may be applicable, such as 1 pump w BEP at 60 and another with BEP at 140, still with no VFD. Without knowing the exact flow histogram, anybody's guess as to what the optimum system should comprise is as valid as anybody elses. My point is that you cannot make general statements of whether VFD saves money or not without a careful, through, and detailed study of possible configuration alternatives matched to your operating duty loads. The theoretical energy saving calculations you see everywhere just do not hold water when it comes to paying the power bill at the end of each month. Do the math.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[massimovirgilio]

dear people I'm the firts time in the forum, I've a question about hidraulic balancing in a barrel flowserve pump.
The question is, the pump have 13 impellers in tandem arrangement and in the end pump discharge side have a drum balancing, in the end flange in corrispondence of balancing drum there is a connection flange connected to the discharge line of the pump whit an check valve unfortunatly normally closed.
Evry two months we change the trust bearings and the free end mechanical seal.
I want to know if the balancing line is completely wrong (it needs to connect to the suction syde); if the line is correct but the check valve it remain normally open. Please help!!

 

[BigInch]

massimovirgilio, please start a new thread.
This one might get very busy.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[marcoh]

The system is a closed system chilled water pumping system and VSD saving are real and doesn't need in-depth analysis.

As the system is a closed system the building height does not affect the pump head (only the casing pressure)

At 50% building cooling load, water flow is approx 50% with all control valves on average 50% open (some more, some less, average 50% flow). At 50% flow the system resistance is 25%, therefore pumping power is 12.5%, ie a 87.5% saving on average operating conditions. In practice this saving is generally not totally realised but should achieve a 50% to 75% enery saving depending on how well the control valves and cooling are selected.

The question that has been raised is whether 2 pumps at 50% flow (ie operate in parallel at high load) is better than a large pump plus a small pump.

Personally I think either can work fine but interested in other people's views.

Note i posted the same question in the HVAC forums.

 

[BigInch]

If you have VSD, I feel that the 2 pump vs 1 pump question becomes largely a matter of deciding if you need standby capacity, in case one pump requires maintenance, or not. If you think about it, you can easily verify that logic.

1 VSD pump can do any flowrate 0-100% at presumed best efficiency.

2 VSD pumps can do any flowrate 0-100% at presumed best efficiency. That's equal.

The disadvantages for 2 pumps (for smaller pump sizes) is that approximately 2 X space is required, 2 X the wire is required, 2 X the local pump piping, fittings and valves are required, 2 X the probability of all those connections failing, etc. The only real advantage you have is the ability to run 1 pump, if the other is down for maintenance.

With all those disadvantages, my opinion is, if you don't NEED 2 pumps for some specific reason, you should only have one. I gave you a lot of reasons NOT to use two; how many reasons can you list showing that you MUST have two or more pumps? Unless you have two truck, ship, aircraft fueling or product loading points, where one pump must still fill one airplane, even if the other is broken, or a similar situation, a whole lot of reasons are usually hard to come up with.

OK, so now the size question. Should they be the same size or different size. The only question remaining to ask there is, "If one pump breaks", what capacity MUST I provide
to the system to keep it running? Will 50% flow be useful to the system? If in the worst case, the answer is yes you can run at 50%, then 2 X 50% could work. If you have sized the 2 pumps at 75% and 25% and the 75% pump breaks down, will 25% capacity in the worst case do you any good? If that's a yes, you have another option. So that works out to sizing one pump for the minimum flow you
must provide to run the system and the other can be any size from 100% - first pump capacity ... to infinity.

To run both pumps simultaneously, and the pumps have different capacities, you should be especially careful to match the heads produced by both pumps as closely as possible to their respective % rated flows, if both pumps will be run from the same VFD controller (get the same rpm signal) to share load equally. If each has a separate controller, you must use discharge pressure feedback so that each pump runs at the appropriate rpm to deliver equal discharge pressures into the same header.

Where there is no VSD being consider, multiple pumps can add efficiency to the pumping operations, as selecting the best efficiency combination from an infinite number of pumps would in effect approch being able to pump any flowrate at a very good net BEP of the pumps selected for duty. In effect the same thing you have with 1 pump on VSD.
Fixed speed pumps require more pumps to efficiently cover a required flowrate range.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[quark]

Marcoh,

Splitting the total flow rate into two pumps is beneficial if your cooling load is equal to or less than 50% for maximum hours in a day.

For example, if your flow rate requirement is 100 cu.mtr/hr and correspondingly the head requirement is 100 meters then if you select 2 x 50 cu.mtr/hr pumps then the head capacity of each pump should be equal to the head requirement at combined flowrate. That is each pump should be designed for 100 mtr head.

If your total heat load reduces to 50% then your flow rate requirement is 50 cu.mtr/hr. Flow rate wise, you can half the speed of each pump. But, the system resistance corresponding to 50% flow rate is 25 meters. Since you have to match head developed by each pump equal to that of combined flow rate(25 m), you have to run the pumps at (25/50)^1/2 = 70%.So, when you want only 50% of the flow rate, you are forced to run the pumps at 70% flow rate. This will, infact, be more than 70% as the system resistance corresponding to 70% flow rate will come into picture.

So, one bigger pump with variable speed can work under optimum conditions (including the inefficiency of motor at low loads). If your heat load is less than 50% for a predominant period, then you can switch of one pump thus saving more energy.