Strategy for pump/flow turndown on a cooling water system? 3

[JoeySoap]

Hi all, I am looking at a fair size cooling water system for a new process plant and I would appreciate any views you have on a strategy for dealing with a significant turndown requirement in terms of flow throughout the year. The proposed system flow varies between around 150 and 600 metres cubed per hour throughout the year with the highest flow in the summer months and the lowest flow in the winter months. We are maintaining the water feed temperature set point constantly throughout the year with air cooled chillers (there are reasons for this that I won’t bore you with, but this is the best strategy in this case). We haven’t got a firm demand profile yet but the maxima and minima are as stated here winter/summer. The system is a little unusual for a process plant in that the source of the cooling water is actually on the roof of a building, so the water circulates to and from highest part of the system via the process plant equipment. This obviously means that the system resistance (piping and heat exchangers) represent the requirements for the pumps with no delivery static head requirement as such. We are presently characterising the system resistance at the various flows but basically the system consists of a feed and return loop of around 600m feed and 600m of return piping with take offs to various heat exchangers along the way. There is one main plant heat load (seasonal) that represents about 60% of the flow requirements with many smaller ones in addition that are much less seasonal. There appears to be a great chance to design a really energy efficient system here since we’re starting with a blank piece of paper and I think this could be a good solution in which to use variable frequency drives. What do the experts believe is the best strategy to achieve these aims? By the way, the number and capacity of pumps is as yet undecided but we need to provide this duty 365 days a year 24 hours with a likely bi-annual overhaul for two weeks so system reliability needs to be pretty high. Thanks a lot for your input. Best regards.

 

[Artisi]

If I understand this, you are saying that you have basically a closed loop system with the only variable being the flowrate which equates to a variable head across the pump due to the change of friction loss and you are currently calculating these system resistances.

Once you have the upper and lower resistances it is simply a matter of overlaying these onto a series of variable pump curves capable of the flowrates to find the best fit for the 60% duty and the best compromise for the other duties.
There will need to be some juggling to make the best selection as you will have to evaluate the various duties which may fall on other than ideal points on the curve ie, too far to the right or the left of BEP.

Operating in other than ideal areas of the pump curve impact of reliabilty - so sound engineering principles must be used to evaluate, capital costs, running costs, maintenance costs and long reliabilty.

If turn-down is too much for one pump to handle then 2x50% or even 3 x 50% pumps might be a good selection which can give you a great variety of flowrates at high efficieny as well as good stand by capacity in case of failure.

 

[JoeySoap]

Thank you Artisti, your interpetation is correct. I am weighing up whether it is best to use 2 x 50% or a different multiple of pumps to best match the system characteristic. Your opinion is welcomed.

 

[ccfowler]

JoeySoap,

The 4:1 turn-down ratio that your initial post suggests should be very easy for an adjustable speed pump system to handle. A pure circulation system with significant variations in flow rate is the ideal application for realizing energy savings during periods of reduced flow needs. The pumps can and should be sized to operate very close to their BEP for maximum benefits.

I would want to pay close attention to the variations in flow requirements throughout the piping network to be sure that there are no potentially unpleasant surprises lurking in this seemingly ideal arrangement due to flow balancing issues.

Since all equipment and systems eventually exhibit operating problems (usually at the most inconvenient times), I would not be inclined to use anything less than a 3 x 50% arrangement. When the costs associated with a pumping system failure are considered against the costs of redundant pumping equipment (lost production, equipment damage, damaged product, safety problems, etc.), you may find that a 4 x 50% arrangement may look quite reasonable.

 

[JoeySoap]

Thank you ccfowler. I have identified that there are areas and operating scenarios within this circulating system where we will have to introduce some way of balancing pressure drops across certain pieces of equipment in order to make the most of this system. Along with one of our process engineers I am going to build a flow model of the system to optimise our pump selection. I agree that a 3 x 50% pump selection looks good on the face of it. Thanks again.

 

[wimple]

I used to offer for HVAC application 2x50% twin pumps equipped with VFD, quite std approach...for sure double check again with your op. conditions

 

[JoeySoap]

May I ask please wimple, what advantage does running 2 x 50% pumps offer versus 1 x 100% pump please? If one of the 50% pumps failed when running at its BEP rated flow then that would result in double the required duty on the remaining 50% pump which would possibly trip on overcurrent when trying to maintain double its BEP rated flow?

Using a 3 x 50% arrangement, or a 2 x 100% then the failure of a single pump is covered by an auto-start standby pump. Thanks again.

 

[Artisi]

As you are not running at 100% call year it is far better to run 1 x 50% pump when demand is low rather than 1 x 100% well-off BEP. Yes when running at 100% load (2 x 50% pumps)and 1 x50% pump fails you could have a problem unless you can turn down the demand until such time as you can get the second pump up and running again.

This is why I consider 3 x 50% pumps as a better option, alternately 3 x 33% pumps could be used, it probably give more choice for varying flows but the same problem if 1 unit fails during a 100% operation period - you only have 66% flow.

 

[wimple]

JoeySoap

Operation in 2x50% means two pumps operating in parallel to cover peak loads. Twin pumps are also designed for stand by operation with changeover flap to switch 1x100%.
I can give you link to some Mfr, if you want to check capacity, etc, but depends from your system curves.

Regards

 

[786392]

Usual Observation reported a maximum of around 1.3X(unit pump's capacity) if the flows join each other at discharge header during simultaneous operations.

Not definitely if, the suction and discharge pipelines diameters are extra-ordinarily larger than optimum to feed the pumps and

allow for 'almost Nil back pressure' on each other's flow at pumps discharge; then actual situation may be derived through generating practical on-site data indeed I believe!

Best Regards
Qalander(Chem)

 

[JoeySoap]

Thank you all for the thoughts. I'm still working on characterising the system but I think that 3 x 50% pumps makes a lot of sense at this time. I appreciate the input.