Optimal number of redundancies & spares?

[shockking]

Is anyone familiar with any research or rules thumb about optimal quantity of spares or redundancies, especially as a function of system size or cost? Are you better off installing a few large things, or lots smaller things? And as the system size or quantity increases, at what point does it make sense to start keeping additional spares?

This is a general question that could apply to anything -- electric grids, industrial processes, IT networks, etc. etc.. Today we were talking about a pumping station. Something like $20M will be spent upgrading this facility. We know it will have at least 1000 hp of pumps with N+1 redundancy, but should it be fewer large pumps or more smaller pumps? I.e. For 1500-hp of total pump horsepower and for minimum capacity equivalent to 1000 hp, is it better to do (3) 500-hp pumps or (5) 300-hp pumps? Both cases have N+1 redundancy and the same total capacity. Or what about (11) to (15) 100-hp pumps, perhaps leaving one or more as spares (i.e. not physically installed, just stored on site)? All cases have at least one redundant unit, such that any single pump can fail and we maintain the 1000-hp minimum capacity. If the pumps are small and we have a spare on site, we can replace it ourselves fast (of course with more pumps there's the more opportunity for failure). If the pumps are too large for our equipment and we don't have a spare on site, who knows how long it could take to procure and replace (of course with less pumps, there's less opportunity for failure). The (11) 100-hp case is interesting, since it minimizes the total capacity and total idle capacity. Anyway, just curious what people think.

TIA,
Tom

 

[TallTimber]

Just my $0.02, but as long as the site is accessible enough that replacing a large motor is not a logistics issue, I would vote for a few larger motors. The more motors you have the more likely you are to have a problem. VFD drives on water pump motors are very safe and have a good lifespan when properly installed/programmed. Fewer motors will also require less civil work on the manifolds and piping which will save a good deal of money. Also, if your redundancy is fixed at N+1 then providing the largest motors also provides you with the largest amount of "flow" on stand by.

 

[drktmplr12]

there is no rule of thumb. it depends on the process turndown ratio and the mechanical system.

to pick on a specific application for a moment, municipalities don't often have funds to throw around for "nice to have" stuff. just what is needed to get the job done. if one were to tour 100 wastewater plants in a region, they would see that nearly every RAS pump station has 2 or 3 pumps and most clarifiers are no larger than 50 ft. it boils down to the process requirements and economics. there isn't a reason to have a clarifier larger than 50 ft, at that point they just build another.

a process/mech engineer will perform a hydraulic analysis of the in and the out of the pump station over the course of an entire day, over an entire year of anticipated demand. this will take into consideration the suction head, discharge pressure and flow required at all times, and all fixed frictional losses due to pipes, etc. this will determine the maximum and minimum HP required. they will also consider areas of sustained demand in between the min. and max.

for example, if it needs to supply only 50 HP of power at 4 AM, but slowly ramp up to 1000 HP by 11 AM, maybe the best move is to provide one 50 HP jockey pump and six 200 HP pumps (or four 250's). But if the system services only large users, it may only need a minimum of 400 HP-then it makes more sense to do 3 x 500 HP.

this also depends heavily on the pump being coupled to the motor because they have specific areas that need to operate to (a) establish flow into a pressurized header and (b) operate efficiently. a pump operating "off the curve" will often draw more current and cause additional wear and tear on the moving parts.

get this wrong and you will have a system that has no mechanical "sweet spot" and will short cycle and be a maintenance nightmare.

the other thing is space. having 11 to 15 100 HP pumps would require far more space and have lots of piping, valves, etc. It would also mean that there is a sound engineering reason to operate 8 pumps versus 9 for example. also, you better be CERTAIN that all the pumps have the same curves. this becomes problematic as the pump station ages. I have a hard time imagining a pump station that needs that kind of precision of control.