Flow requirements 2

[vjr0512]

The design basis of client spec says that for all the pumps the Rated flow shall be 10% more than the Normal flow. I could understand that this margin of 10% is to take care of the errors in design of the pumping line design, routing, losses calculation and also wear and tear in due course. My view on this 10% can be applied during the FEED stages to take care of the above issues. But during Detail Engineering, where all the pipeline routings are finalised and Isometrics are drawn with the sizes, should this 10% margin be required?

Is there any optimisation required to reduce this margin for such design errors consideration?

For all the pumps, the selection of the pump model do comply with the API 610 requirements including the BEP , allowable and preferred operating range etc.

Thanks to provide some guidelines.

 

[MortenA]

Usually you will go out and purchase all your rotating equipment fairly early after FEED and thus the final ISO design would not be finished at this time?

 

[EdStainless]

And now you know why no system runs at BEP, everyone in the process ads a fudge factor and you end up with pumps that are 20-30% oversize and inefficient.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[Artisi]

EdStainless: spot on, I have been beating that drum for years.

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]

EdStainless and Artisi are absolutely correct that over-sizing pumps results in efficiency and maintenance burdens (stars to both!). I fully agree and have made similar statements many times.

In my view, the apparently simple 10% margin requirement is not bad on its own because that is a reasonable expression of normally tolerated production pump performance uncertainty. For relatively small pumps, getting closer performance guarantees from the manufacturer is unlikely to be practical. If it is a relatively expensive, high power pump, then the manufacturer may be able to assure a somewhat tighter performance uncertainty (for an additional price that may be justifiable for the overall economics of the installation). I've seen cases where the overall design margin proves to be far greater by the time the actual system comes into reality (30%, 40%, even 50% due to compounding margins provided on pumps, piping, and other system components are no great surprises).

In a realistic defense of including significant, but realistic, design margins by engineers, consider the career killing problem of a nice, dandy, big, EXPENSIVE, new process system that, when installed and running, can't be made to satisfy its required performance by anything short of "Divine Intervention" because one or more critical components have been hopelessly undersized as the result of an overly aggressive "right-sizing" design philosophy. I've seen some of these, and they can be a serious nightmare! The engineer(s), and often others, involved are never in a happy place in their career(s) when that reality materializes. When this happens, Engineer X (in charge of creating these problems by design) will become eternally famous within the facility. Whenever anything goes wrong, or is even just a bit of nuisance, the troublesome pump, pipe, valve, motor, or whatever else, will be very unkindly described or discussed as an "Engineer X" pump, or "Engineer X" pipe, ... regardless of whether or not the current problem is really in any way the fault of the now defenseless "Engineer X." You can be certain that nobody wants to be that "Engineer X."

Alternatively, "marginally" greater operating costs due to modestly excessive design margins are just a problem, and to some extent, the resultant costs may actually be somewhat compensated by somewhat greater than required system production capabilities (read greater net revenues).

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]

It works like this, by the time the owner puts a little margin on, plus the designer, operations, pipe work designer, and the consultant all add their "just in case fudge factor", the pump company could be sizing a pump with 40% over flow capacity and maybe 20% over stated head, they of course will add a little to ensure meeting a ridiculous tight spec. on flow and head can be met. You now end up with too larger pump and driver and everyone is now ducking for cover trying to cover their arse.

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.)

 

[ashtree]

Gents,
I agree with everything you have said. That's why VFDs were invented. It initially stood for Very Fast Dollar but that was just way to mercenary for the marketers so it was changed to Variable Frequency Drive.

How many bad or "just in case" pump selections have been made to work by adding a VFD?
I am not saying its right or even okay but it is how many in the industry do it, so its reality.

Regards
Ashtree
"Any water can be made potable if you filter it through enough money"

 

[georgeverghese]

Perhaps whether a 10% margin is indeed required may lie in the derivation of "normal flow" in this case.

 

[shvet2008]

What is the problem with 10% margin? Actually so many post and opinions for so little question.
You always can use lathe and remove this margin. Instead of this if you have no 10% margin you can do nothing - only to purchase a new one.

 

[lilliput1]

If there will be diversity in the load but the total gpm flow is based on simultaneous peak, no additional safety margin is required.

 

[pumpking]

We run quarterly centrifugal pump training courses to discuss these very issues, excess margin, pumps running beyond BEP, effects on power increase, effects on reduced reliability, effects on the poor mechanical seal manufacturers (its hardly ever their seals that are bad, it is the operation point) and you will be amazed how many pumps are in operation running at really bad points of their curves for this very simple reason (you remember the motor specification of being sized for EOC max Impeller that simply masks true pump problems).

Thankyou VFD's, your input is so very very valuable...

http://www.cdrpumps.co.uk/pump-training/

Ash Fenn

http://www.cdrpumps.co.uk

 

[LittleInch]

The issue here is that design of pumps, as with many other components and systems has a variety f inputs which can in reality or even in design, change over time.

Process engineers in general work on the design limits and seem to have (very sweeping statement here) a desire to be correct 100% of the time, instead of being right 90~% of time and at the extreme edges achieve 90% of duty requirements

To take a simple example - you pump out from a tank 20m high (operating from 19m to 1m) to achieve a flowrate of 1000 m3/hr. Viscosity can vary with temperature which varies from 10C to 30C. The end of the pump system ends at a facility where the arrival pressure varies from 5 barg to 10 barg. - how is it possible to define "normal flow" from this?

The process engineer will size the pump based on min head in the tank (1m say),highest viscosity at lowest temperature and highest arrival pressure. Then some dumbass spec says lets add another 10% to this just to make really really sure we can maintain 1000m3/hr in every possible circumstance. They don't often do the opposite calculation which is what happens when the tank is full of nice warm contents and the arrival pressure is low. Often quite illuminating.

Sure if 1000m3/hr in any set of circumstance is what you need then fair enough - you pay for extra capacity you don't need. In most case it would be acceptable if 1000m3/hr was possible in 90% of cases or averaged over say a complete tank emptying. At the start of a project I try to get out from the client what is it that is crucial and whether flowrates for example are instantaneous or averaged over a day or a month. Do they want "catch-up" capacity to cope with an maintenance or breakdown? Their money, their choice.

I have heard of a recent plant where they added a third or fourth identical train. The designers were able to go and monitor all flows, pressures, levels etc and then designed the next one to do its nameplate capacity and not a molecule more. Came as a bit of a surprise to the owner who was expecting nameplate + 10%....

Lets not forget here that it's not just the pump that gets oversized - it's the motor, the cables, the starters, the switchboard, the transformers or even worse the generator sets all of which either cost more than they should or run inefficiently.

So each pump and each system will be different. If you have very tight control or variance on the inlet conditions and outlet conditions then you can size your pump dead right. Real life in may places doesn't allow that and as ccowler puts it very well, the consequences of undersizing a single component in a plant system can be catastrophic. So the bit more, bit more plus 10% thought process becomes ingrained.


Remember - More details = better answers
Also: If you get a response it's polite to respond to it.