Specific Speed Limit Issue 2

[virata]

Hello gents,

Our project specification has a specified maximum limit for suction specific speed which is 12000 (for less than 30kW) and 11000 (for greather than 30kW). Our vendor had offered Nss over the specified Nss for 4 pumps and told us that these are their best selection. Is this acceptable? 3 pumps (with 1 VSD) have fluid of tempered water, while 1 pump has caustic liquid. What are the parameters I should consider to meet our specification?

Need your thoughts on this.

Thanks,
V

 

[GPRnD]

It depends. If you want the history of how the 11,000 limit came into existence, see the link to the paper below.

http://turbolab.tamu.edu/proc/pumpproc/P29/LectureP05.pdf

I'm not in favour of blanket suction specific speed (Nss) limits that don't consider the specific speed of the pump. If you look at page 14 of the paper, you'll see Figure 23 which has a recommended Nss limit vs. pump specific speed.

The reasoning behind this is simple: If the pump impeller is a low specific speed, suction recirculation is not a problem and more aggressive suction specific speeds can be tolerated. Conversely for high specific speed impellers, suction recirculation is a big problem and Nss needs to be lower for reliable operation. See the picture below

 

[GPRnD]

Opps sorry I messed up the picture link, since it wasn't a picture

Just click here for the picture I was talking about

 

[1gibson]

I would accept it for the VSD pumps. The suction recirculation (related to reliability issues) is only a problem at low flows. If you are able to operate near the best efficiency point (which you should be targeting with the VFD) then it is a non-issue.

If you need to operate for extended periods of time near the pump's minimum flow, then you might consider a lower Nss pump.

 

[virata]

My apologies, I couldn't open the link..

Here's i can give for now:
- the impellers are open type
- the rated flow of one pump is 90m3/hr running at 3550rpm while the 3 other pumps are 400m3/hr running at 1780rpm.

 

[1gibson]

What is the rated head for the VSD pumps with open impellers? Varying the speed has a large (rpm^2) effect on head.

Head affects the pump thrust, pump thrust affects the shaft stretch, shaft stretch affects the actual impeller lift, impeller lift affects the pump efficiency. Basically, if you have high head and a large difference in running speeds, you may need to compromise on impeller lift. You need to avoid contact at full speed, but set the lift low enough that the pump is still efficient at reduced speeds.

This should not be an issue for small, low head pumps.

 

[Artisi]

Have you asked why these limits have been set - maybe someone just pulled the numbers out of another completely unrelated specification without any further thought - certainly wouldn't be the first time it's been done.

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.)

 

[virata]

Gents,

Only one pump has VSD. 3 pumps are running in parallel with 2 of them in fixed motor while the other one is in VSD. The 4th pump is stand alone with fixed motor.

The differential head of these 3 pumps are all 56 m.(400m3/hr, 1780rpm), while 4th pump has diff. head of 40 m.(90m3/h, 3550 rpm)

These limits are probably set because of the thumb rule for its reliability purposes and also that if these suction specific speed are too high, it could cavitate, right? especially that the 3 pumps running in parallel are handling water (tempered water to be exact).

What are the other recommended solution for this if any? The vendor told us that this is their best selection. Might be changing the model to double suction? or making them all VSDs?

Hope to hearing your thoughts.

Regards,
V

 

[Artisi]

Cannot understand why people are so hung-up on specific speed, it really has no part in correct pump selection / application. In my 30+ years in the pump industry and having selected 1000's of pumps I have never once used specific speed as a criteria for pump selection, as basic initial / correct pump selection is or should be based on:
1. correct pump style for the application
2. operating at or near best efficiency
3. NPSHa/r
4. speed
Materials, pump reliability / design and etc. are other engineering factors to be considered along with initial selection.

How about loading the pump details (end-suction horizontal, single suction HSC etc.) and curves of the pumps under discussion allowing people to actually view what is offered instead of crystal ball gazing.

Is the "vendor" the pump manufacturer or a go-between agent?



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.)

 

[1gibson]

"Best selection" is often code for "I sorted by efficiency and this one was at the top."

 

[Artisi]

How about a change of supplier, 400m3/h @56m head should be achievable by any respectable pump manufacturer, actually it's only a bit of a toy pump. For interest I looked at one major pump manufacturer for a rough selection - took all of 2 minutes to come up with what is probably a good selection for the duty 400/56/1780.

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]

I have seen suction specific speed play a very important role in reliability. This seems to come up for heavy vacuum gas oil regularly. We have had two sets of HVGO pumps that had very good NPSH margins, very good operating points relative to BEP but still had chronic high vibration and poor reliability. For both pairs, the solution was to change to a pump with a lower suction specific speed. For one pair of pumps, that involved going larger and slower. For the other pair, it involved going with a higher NPSHr to get a lower Nss.

I treat suction specific speed as one more variable that can impact reliability. This is truer for services where we have no choice but to occasionally run well below BEP flow. I do not enforce a strict upper limit on Nss. But, if the Nss is higher than 12,000 (US units) for hydrocarbon or 11,000 for water, I want to have a conversation about alternatives.


Johnny Pellin

 

[virata]

Thank you guys for your input.

Indeed Nss plays an important role in pump's hydraulic criteria.

I have another question, one vendor offered an alternative pump for those 3 pumps (running in parallel) with Nss 10450 US units (very acceptable) yet the head rise to shut off will be around 7% - this is not advisable for HRSO percentage running in parallel which requires at least 10% (as per PIP and project specs).Also, the motor power will also increase which will be more expensive.

Do you guys think this is more acceptable than the previously offered pump with above 12000 Nss? With regards to the issue of low % HRSO, is it advisable to use an orifice plate to increase the HRSO?

By the way the NSPH margin of these 3 pumps are all 35 meter so i think the issue for recirculation would not be a problem as cautioned by high Nss.

Let me know what is the best option.


Thanks,
V

 

[GPRnD]

3 of your pumps have 1500 Ns, the 4th is 1800 Ns. They are low power in the scheme of things. Personally I'd go with the higher Nss offering and get on with installing them.

For added assurance you could assess the design quality using Fraser's equation for the onset of suction recirculation. They should preferably be below 65%.

Just to be clear, recirculation and cavitation are two separate things (although significant recirculation can raise NPSHr and increase cavitation). 99% of all pumps operate with cavitation present. It isn't so much a question of "is there cavitation ?" (there is), but rather "is it damaging cavitation".

It sounds like your NPSH margin is more than sufficient.

 

[virata]

Thanks,

I will check that frazer's formula.

By the way, the 3 pumps (running in parallel) have rotative speed of 1800rpm. The Nss issue of the 4th pump (stand alone) has been resolved.

Additionaly, I have a question regarding 3 pumps which are operating in parallel. As mentioned above, one of the pumps is operating in VSD and its speed is to be controlled. Will this be a problem? Since most pumps operating in parallel shall be operated at the same speed, right?

Let me know your thoughts,
V

 

[Artisi]

GPRnD; I fully agree, 400/56 is really a toy pump, standard production for many pump companies probably available off the shelf.
However a pump curve would be nice to see - but ............

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.)

 

[Artisi]

The VSD pump can run at a different speed, however the flow will only be that which the pump can generate at the imposed head by the system. Even the 2 fixed speed pumps will have slightly different flow when compared to each other.

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.)

 

[virata]

Here is the Pump curve.

The alternate (with lower Nss) pump has lower effiency.

 

[Artisi]

Looking at the pump curve I would suggest there is a better pump available than the one selected - for the flow / head duty you have I would be looking for something with efficiency into the mid 80's not less than 80 as this unit is, also the duty point is not well matched to BEP. That alone would prompt me to look elsewhere for a better hydraulic fit and not even consider Nss.
But your call, unless you are stuck with this manufacturer / pump selection I would be searching the websites of the major pump manufacturers looking for a better pump.

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.)

 

[GPRnD]

Artisi,
I agree and you were right to ask for a curve.

BEP appears to be 520 m3/h vs 400 rated which is 77% of BEP. In addition because of the 12% trim, actual impeller shockless flow is probably >600 m3/h (it may be higher still given the NSPH curve shape), making the selection quite far from optimum.

The other thing that concerns me is that the MCSF line is located at approximately 50% of BEP. For a low power pump with that specific speed, I'd expect MCSF to be closer to 20-30% of BEP. So checking inlet recirculation onset with Fraser would be prudent.


virata,
How well these pumps work in parallel depends on your system curve. The pump curves are reasonably steep and stable which helps. You would want to look at how the 2 fixed speed pumps might vary in HQ (depending on their test tolerance, keeping in mind they will be different).

Having 2 fixed speed and 1 on a VSD can be problematic since the fixed speed pumps often end up being forced to operate somewhere they shouldn't. I've seen quite a few installations with this setup and they are seldom executed well (in my opinion).