Running 2 small pumps in parallel vs one single pump 1

[planck121]

Hi,
Is it possible to show that running two pumps (smaller) is cheaper than running one single centrifugal pumps. I believe the energy savings are better in running two smaller pumps (esp given process variability in a running plant) since they are able to handle this better than running a larger single pump and then throttling flow at the discharge to meet requirements. However, I would like to know if there is some way to show based on general calculations from pump performance curves that running two smaller pumps is more economical than a single pumps. I do understand reliability is a different story (2 pumps provide better reliability in case part of the flow is useful) compared to a single pump. However if 100% flow is required for a particular process than a single pump would be a better option.

In this case however I am trying to look at it from a cost stand point and I would assume it makes better sense to look at it further from the lifecycle costing stand point of the pump rather than just purchase cost.

Thanks

 

[Artisi]

Look at in power costs, just run the power calculation for 2 pumps versus 1 pump. I pump properly selected should be streets ahead of 2 two pumps, that is unless some of the time 1 smaller pump will be run for reduced capacity.

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.)

 

[JJPellin]

Artisi states the energy issue well. Other lifecycle costs favor one pump. Purchasing and installing one larger pump is likely to cost much less. The cost of ongoing maintenance would tend to be higher with two smaller pumps. If this is a critical service where production cuts will result from a pump failure, then two small pumps could offer an advantage. The cost of spare parts might be lower with small pumps.

Get your pump vender to quote it both ways. Run the numbers for all of these areas. Allow for the time value of money over the likely lifespan of the project. Add it up and see which is better.

Johnny Pellin

 

[DubMac]

It may not apply to your situation, but another consideration to look at might be whether or not using one pump pushes the motor size above some critical limit (above NEMA??). In some facilities the cost of going above NEMA (more rigorous specs, etc.) may skew the total cost in favor of staying under 250HP.....

 

[TD2K]

If you have two versus one cost assuming the small pumps are half the flow of the big one for the same head then it's going to come down to efficiency of the two pumps and I don't think smaller pumps necessarily equal high efficiency.

If your plant has a lots of ups and downs then two small pumps might make sense but you'd have to likely turn down the plant a lot to be able to run on 1, 50% pump. If you have lots of rate changes then I would be looking at a VFD drive rather than multiple pumps.

Finally, I see lots of 2, 100% pump installations. I see very few 3, 50% pump installations or 4, 33% installations and where I have seen those, it's usually because the flow is such that 2, 100% pumps aren't practical, the client doesn't want to get into the higher voltage requirements that fewer pumps would require, etc.

 

[rmw]

I am kind of in an 'all of the above' situation. I just received bids for 3@50% vs 2@100$ and the price difference was negligable. That was probably because in one of the 2 pump sets, 5XXX frame motors were necessary instead of NEMA frames. 3 smaller pumps put the motors well under the NEMA frame size limit. That is why I had them bid the alternates. Plus our application will require pulling back to flow that get to the left of the minumum flow line with just one pump pumping.

rmw

 

[ScottyUK]

Don't forget that the equipmet costs for a small medium-voltage starter will be higher than a large LV starter. You might want to speak to your electrical engineers to see where you make the jump from low voltage to medium voltage. Normally it is in the 250 - 400 HP region.

 

[EdStainless]

There are other 'smaller pump' issues. Smaller pumps tend to like running off BEP less than larger pumps with flatter curves do.
If it were me, unless going one pump created serious risk of outage or raised installation cost (wiring, controls, motor) too much, I would stay with one properly sized pump.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[planck121]

Thank your for all the good responses. A couple of points from my end.... One of the main reasons for going for a 50%X2 rather than 100%X1 is the fact that our standard eliminates the need for force feeding of the lube oil for power below 750kw. So when we go for a 2 pump arrangement we reduce the need for lube oil skid. This itself boils down to cost savings...

- Maintence costs due to all this additional aux lube oil system. There is certainly issues with cooling water fouling and getting ride of the exchangers on the lube skid also avoids un-necessary maintenence.

- Smaller pumps in my opinions are better at accepting flow flucations than bigger pumps. In regards to EdStainless point about lower margin from BEP for smaller pumps. I do not feel this to be true, since the BEP margin around a pump is governed by Nss (suction specific speeds) the higher the Nss the lower margin of movement around the BEP.

- Smaller pumps can provide better efficiencies if rightly selected compared to larger pumps due to better/tighter running clearances. On a large pump the L3/D4 ratio of a shaft would naturally be higher compared to a smaller rotor. Therefore starting or stopping the pump can lead to more premature wear due to higher weight ratio of the rotor. Where as in smaller pumps this effect is lower and over time if the pumps are rightly operated we can expect a lower maintenence time compared to larger pumps.

- I am not sure about mechanical seal reliability for a larger versus smaller pump. I would assume again on my previous point that since smaller pumps will have better shaft stiffness ratios...vibration related concerns would be handled better in smaller pumps compared to larger pumps leading to better mechanical seal life on smaller pumps....seal being one of the largest contributors to pump failures and being the weakest link.

So maintenence costs along with energy costs would tend to be higher on larger pumps I assume. Certainly in our case with the lube oil skid system....But again as Johnny mentions it would be hard to define it excatly without the right calculations and life cycle numbers.

Regards

 

[brandonbw]

We design pumps for residential drainage. This one time we had a site that wasn't critical for a duplex system. But I was still curious on cost. Turns out the way a certain municipality wanted the setup for the duplex system we would need a custom control panel. We ended up going with a simplex system. Figured I would throw the control panel thing out there, since they said the panel would cost too much to make it worth doing.

B+W Engineering and Design
Los Angeles Civil Engineer and Structural Engineer

http://bwengr.com

|

http://bwstructuralengineer.com

|

http://bwcivilengineer.com

 

[Artisi]

planck121: Two points I would be interested in hearing your comments on from a design / hydraulic perspective would be:

1. Smaller pumps can provide better efficiencies.
2. Smaller pumps would have better shaft stiffness.

Maybe I have missed something over the years.


I would suggest it's time, if not already done, to make contact with a couple of major companies capable of engineering pumps of this size and ask for some budget pricing / pump selection / recommendations. From here you can start analysing and looking at your ideas re, efficiency, shaft stiffness etc. and running actual operating costs over the life of the plant.

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.)

 

[ccfowler]

I very much agree with Artisi. One large pump is usually a better choice if only initial cost and operating & maintenance costs for the pump are being considered. The use of multiple pumps is normally to address reliability concerns or to address widely varying flow rates.

The most important factor is the proper selection of the pump for the actual duty. Most problems and excessive costs arise from errors in matching the pump to its actual duty. Excessive "margins" or "safety factors" in the design and selection processes usually conspire to cause the selection of a pump that will almost never operate at its best with the flow and head conditions that will be imposed upon it.

Valuable advice from a professor many years ago: First, design for graceful failure. Everything we build will eventually fail, so we must strive to avoid injuries or secondary damage when that failure occurs. Only then can practicality and economics be properly considered.

 

[Artisi]

Oh yes, "safety margins" - how many times have I seen this. The designers calculates X flowrate and Y head, the plant guys just add a little in case of upset conditions, the owners decides maybe we need to upgrade soon, add a bit more, the sales engineer adds a bit more just to make sure he hits the required duty and the manufacturer adds "just a touch" to the impeller diameter to cover any shortfall in performance. So we end up with a pump selection with anything from 10 - 30% oversize. No problem until the pump doesn't meet duty when put into operation - and who's fault is it, the pump manufacturer / supplier of course.

Ring - Ring, hello, Hi, we just started the pump you supplied, it's way over on head and pumping too much, can you come and fix your problem. Yeah yeah, I know it was tested in your test facility but it doesn't work here yadda yadda -



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.)

 

[planck121]

Artisi,
I concur with some of your comments. My basis and in order to substantiate my statements these are my thoughts :
- Firstly we as engineers should strive to attain a balance between what the calculations show and what actually happens in the plant. I do agree with you on the benefits of one pump vs 2X50 but let us give a thought of refelection on a few points

As far as efficiency goes (the ulimate verdict on the efficiency would be the vendor preformace pump curve)in theory but lets look at it in pratice

- The friction rate doubles for each resistance increase. So for higher flow pumps we would expect to have more fritional losses in the incoming piping and also with in the
pump. Basically leading to higher temperature rises with-in the pump and including the stuffing box. So seal will have to be kept more cooler to sustain the higher temps
over time.
- We have all agreed atleast based on our trailing posts that 2x50% is better equiped to handle flow flucuations than 1X100%. We all know how it is in the plant even if process
says that a pump should pump a given flow rate. When it comes to production environment everything boils down to $$$ so how often do we find pumps running at or close to BEP.
In actuality depending on the industry (say oil and gas) pumps are actually running to meet an ever increase production demand and when things break down in the plant. Pumps
are turned-down without consideration to the BEP. Smaller pumps would be more efficient (energy wise) to handle such situations
- Let us also look at vibrations limits etc (larger pumps more vibration generally, because of larger bearing clearances), smaller pumps less vib. This is relative vib. I am
not speaking about absolute vib. Lower vib leads to lower seal issues or atleast reduces the probability.
- While not directly related to efficiency. Let us look at the maintenence and turn around time to repair a larger pump compared to a smaller pump.
- We can dicuss about stuffing box efficiencies as well on a larger pump (higher flow rates,head) versus lower flow rates/heads on smaller pumps in general. Try considering
the effect of this on the throat bushing of the pump. Do you feel for two pumps larger and smaller in the same application the wear rate on throat bushing (increasd clearence)
will be higher on large than small. This again effects seal relibility.
- Large pump, leads to larger seal. So more BHP for large seal dia, energy has to be expent to drive a large seal fluid flim compared to small seal drag.


Above are some practical considerations, again the best course would be the calculat the LCC.


On my comment about L3/D4. I was reffering to single stage overhung types. L3/D4 does not come into play much in BB type pumps. However lower L3/D4 helps seal relaibility along with maintaning a proper seal environment. So again maintenence since seals are a single most reason most pumps are brought down and repairs associated with seals should be also taken into account during LCC.

I shall try to hunt for the pump curves and just for a read.


Thanks

 

[Artisi]

Why would you have higher friction losses in the inlet pipe for a larger pump, surely the pipe will be sized to suit the flowrate and why would you get higher losses thru' the pump when the pump will be physically larger to suit the higher flow?

Yes, 2x50% pumps can possibly handle turndown of flow better than 1x100% unit, but then we don't know your flowrates or turndown flow nor the heads involved -so we have idea of what you are trying to achieve.

With pumps of the size you are now discussing - stuffing box losses would be well down the list of priorities in my evaluation, it could well be that the losses in 2 stuffing boxes will exceed the losses in 1.

I would agree that plant operation is never easy and pumps are often required to operate at other than their BEP - this can be left or right of BEP. However a single or multi pump installation that has been carefully evaluated stands a much better chance of operating more efficiently and with less maintenance problems than just choosing an installation on "gut-feeling" that one is better than the other.

Of course the whole exercise at the moment is guess work, pump duties / application are unknown, as is pump configuration etc.

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.)

 

[planck121]

Artisi,
Some data that I have been able to extract.

In the 1 pump config:
Flow (rated) = 643 mcub/hr
Efficiency: 71.1 % at flow of 708.5 mcub/hr
Flow ratio (rated/BEP) = 90.75%

In the 2 pump config
Flow(rated)= 322 cub/hr
Efficincy = 68.9 at flow of 327.1 cumb/hr
Flow Ratio (rated/BEP) = 98.37%

The slope of the efficiency graph around the BEP for the 2 pump case is pretty flat (less efficiency change for a fair amount of flow fluctuation), where as for the 1 pump case the efficiency slope is reltively steeper (for a smaller change in flow efficiency drops more rapidly)

Now given the pratical case we are all aware in the plant that pumps hardly operate at BEP. Would we expect more power savings in the long run using a smaller pump that is less sensitive to flow changes as far as efficiency goes, or a larger pump that is more sensitivy to flow as far as efficiency goes, I would argue that a system (any system) that can remain stable for changing process paramaters in rotating machinery has a better chance to produce economic gains.

In a turbomachinery presentation once I came to hear a very prudent saying, the lecturer said why is that a jet engine that has millions of parts does not break down often, or a car that has thousands of parts does not also break often, but then again we look at centrifugal pump in a refinery with only 100 or lesser parts and that keeps breaking down very often and regularly.....
The answer it seems is changing conditions, process conditions, operator intervention, etc....change especially dynamic is not good for machines. I would like to select machines that are less sensitivty to process changes esp (if that is one thing I can try elimating) at the minimum.

 

[DubMac]

Planck,
If you are running the 2 in parallel, you may want to look at those "flat" curves....can cause problems. You should look at the "2 pump" performance curve overlaying the system curve.
The efficiencies don't seem all that high; have you looked at several manufacturers? Maybe you have other constraints, NPSH etc...

I lived across the street long ago from a master mechanic for Chrysler, 30 or more years. He swore a lawnmower engine was much tougher to keep running than anything he touched.

 

[Artisi]

Still floundering around in the dark.

Initially you inferred that the 1x100% pump was in excess of 750kW "One of the main reasons for going for a 50%X2 rather than 100%X1 is the fact that our standard eliminates the need for force feeding of the lube oil for power below 750kw. So when we go for a 2 pump arrangement we reduce the need for lube oil skid. This itself boils down to cost savings..."

The efficiencies quoted for these pumps seem to be very low for pumps of this size, unless of course you are talking about slurry pumps.

Still unsure of what the actually application is and what you are trying to achieve - how about a fresh start if you are looking for meaningful assistance?
application
Flows / heads
pump configuration
pump curves
system curve

With any misapplied pump you can expect problems " .. but then again we look at centrifugal pump in a refinery with only 100 or lesser parts and that keeps breaking down very often and regularly....." I have seem pump applications that have been correctly selected and applied that have operated for years trouble free, on comes to mind that had run for so long pumping sea water that the motor bearings wear so worn that the rotor could be lifted up and down and the pump case/ back over (back pull out pump) were corroded together - the pump still operated without any problem and the owner was considering only in replacing the motor - we however talked them into a new all S.S. pump suggesting from the experience they already had with the existing CI/SS/SS a life of 20-30 years was more than possible.

Guess the lesson is, if you want short pump life - select the wrong equipment.

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.)

 

[planck121]

Artisi,
Yes lets try this again, I was a bit suprised with your initial proposal about 1 larger pump is better than 2 smaller pumps when correctly sized, perhaps I had not provided enough information in my initial post. In fact in my past experience, I have seen pump selection to be so open that even two pump vendors (designers) argue about the right pump for a same application. In my personal experience I have seen 2X50% work well depending on the situation compared to 1 large pump. I would argue the amount of time required in my case to fix the larger pump and the motor as well if the need arises would be much higher than fixing the smaller pump or the small motor (671 kw compared to 1.3 MW).

The cost savings are still there as far as the lube oil skid goes due to the reasons mentioned above, (motor HP and our standard) but I was trying to quantify energy costs in the long run. In my dicussions with peers and other people from the pump world, I have realised that to quantify the actual LCC is a very difficult excerices due to various paramaters present. Nonetheless, I have been trying to understand if we would see any costs savings from a energy/efficiency stand point from the 2 vs 1 pump senerio.

I will attempt to provide more information as below:

Type of pump is a two stage between bearing for both types, Application is Lean Amine service. Rated flow for all practical purposes can be taken as 643 m3/hr. Max Temp 83 C

1 pump senerio
In addition to the prior information
Motor HP would be 1.3 MW
This situation would entail going with sleeve bearings and force lube oil system
Flow conditions as given in my prior msg

2 pump senerio
Motor HP is 671KW (each)
Pump configuration would still be two stage between bearing but in this case no forced feed lube system.
Flow as given in my prior msg (322.0 m3/hr)



1. Process requirement I would say is that there would be turn-down on the pumps depending on when the need arises based on process upsets and other reasons. In which case I would expect more operation on the left side of the cruve than the right. I mean when the production is required ofcourse any pump is pushed to the right of the curve. But in realtity esp in our plant I have seen more process upsets and issues where the pump gets pusehd to the left.

2. Another thing is heat load on the mechanical seal. I do not have the calculations off hand, but it seems for the smaller pumps the heat load on the seal is lower than the bigger due to throat bushing clerances (less flow of process side into the stuff box) in the small pump compared to large.


Dubmac,
The flat curve are not the pump curves but efficiency curves. What I was wanting to say is that the efficiency curve is relativiely flat around the BEP for the 2 pumps at around 68% and its more steeper on the left of BEP in the 1 pump case senerio. Which means that if I were to move around the BEP for the smaller pumps it would be more efficienct than moving around the BEP for the single pump from an overall pump efficiency stand point.


Regards

 

[planck121]

Just providing another link from an experienced author on pumps and other rotating equipment. At the very bottom of the page the author suggests that the running two parallel pumps could be cheaper in certain senerios.

http://www.mcnallyinstitute.com/15-html/15-01.htm

For the curious minds.