10% head rise to shutoff

[SDRL]

Hi,

i would be grateful if you can help meon below case:

According to API 610, for parallel operation 10% head rise to shutoff shall be considered. We have ordered a pump last year which showed smoothed curve with exactly 10% head rise from rated point to shutoff. However, one week ago on performance test the head rise to shutoff was 4%. Vendor claim that since according to API tolerance table (Table 16 of ed. 11) shutoff can be decreased to -5% and also rated point can be increased +3% so it is acceptable. There is a note also in API table in which stated that if a rising head flow curve is specified (see 6.1.11), the negative tolerance specified here shall be allowed only if the test curve still shows a rising characteristic. My question is that whether it is necessary to pump follow 10% head rise on test or not.

thanks

 

[BigInch]

10% head rise is recommended enable ease of flow control and balancing pump load amongst parallel units. That's why it was specified.
If it was specified, than I would presume it is necessary.

 

[LittleInch]

Section 6.1.11 is your key issue with the vendor.

So, in your spec / data sheet did you actually specify (not implied or discussed but actually written down)

a) This pump was designed to be used in parallel operation
b) a continuous head rise to shut off is required.

The vendor is, IMHO, just trying it on as despite the fact that the table you refer to does allow this sort of thing, it should not overwrite a requirement of the main specification "...shall be at least 10%." It does though depend on the answers above....

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[SDRL]

OK, we have written in our spec that parallel operation is required, however vednor stated that 10% head rise is for offered curve on proposal and on test stand tolerances mentioned on API (-5%) may lead to 5% head rise. I want to know is this claim acceptable acc. to API 610 or not?

 

[LittleInch]

Your issue is that there are two conflicting statements in API 610, with no identification of which takes precedence.
Your claim is in section 6.1.11 with actual text below:
"If parallel operation is specified, the head rise from rated point to shut off shall be at least 10%."

The vendor is relying on this para in the testing section 8.3.1.2
"Performance tolerances shall be in accordance with table 16."

So there are two "shall be's" in those two clauses. You will claim the former is the key clause, he will claim the second...

Without any other overriding specification from you or notes to the contrary it now descends into a commercial fight between the pair of you. Is it technically acceptable to 610 - Yes I think so, is it a good idea - don't think so, especially for parallel operation.

My advise would be to "officially" reject the pump, but then ask what would be required to modify the pump to more closely match what you thought you were buying (probably a bit of impellor trimming) and then try and meet somewhere in the middle as to the impact on cost and schedule, recognising that you both have a case.

All depends on your commercial standing with the vendor (size matters). If yours is a one off then probably hard luck, if you're a customer they want to keep then your hand is strengthened.



My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[Artisi]

Please clarify one point, is the performance test you are discussing a factory test in a certified test facility for order acceptance?

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]

The 10% head rise to shutoff is not an absolute pump characteristic, it depends on where your rated flow is on the curve. If you have 10% head rise to shutoff based on rated flow of 500 GPM, then you might only have 5% head rise to shutoff if your rated flow happened to be 300 GPM. It's still the same curve, and will act the same when operated in parallel.

In my opinion, asking to satisfy this requirement when the rated flow is near minimum flow, is just onerous and is not necessary.

 

[BigInch]

I don't think it is onerous and unnecessary if you plan to control your flow and pump loads. A slight or flat slope will give widely varying and almost unpredictable pump surges with only a very small change in system curve resistance.

 

[1gibson]

Oh it absolutely needs to have continuous rise to shutoff, but the 10% is based on the rated conditions. If you're at 9% head rise, and you revise the nameplate for a higher flow / lower head that takes you to 15% head rise, the pump doesn't become any easier to control.

Often, if you don't meet the 10% head rise then it means the best fit was a pump operating near it's minimum flow, so you're already in a less than ideal situation.

You can make the pump curve steeper by not under-filing the impellers, at the cost of some efficiency. This would require a larger impeller compared to one with the standard underfile. So, probably too late there.

Separate issue from the fact that you may not really need the 10% head rise, but if the vendor's argument is that they only had to show you a proposal curve that met the spec, and don't really have to meet it on the test stand, then that is messed up. If the pump vendor has online selection software, select your pump through there and see how that curve compares to your proposal, and your test results. Curves can sometimes get a little tweaking to meet the specs, then you receive a customized curve in your submittal documents, but in reality there were no modifications to the pump, so it will perform as per the standard curve. This tactic is also known as "lying."

 

[BigInch]

Right, however I maintain that 10% is the minimum useful %rise to have any control. Anything less must basically be run at constant speed and whatever the pump wants to give you in terms of both pressure and flow. No controlling it.

In any case 10% is not very much to work with.
1000 ft BEP head pump (discharge pressure around 230 psi + suction) based on gasoline. A 10% rise = 23 psi over the pump's BEP flow capacity. If that is 1000 gpm, then it is 0.023 psi/gpm. If I have a pressure sensor accurate to within +/- 1%, then it's +/- 2.3 psi, so the pump can go drifting off by 100 gpm before I even get the first indication that it might be drifting. If my sensor is +/- 5%, not unusual in many situations, it's 20% change in Qbep. Now the pump is probably outside of best operation range before I even know about it.

10% is not very much to work with. If you need any kind of control, you need at the very least a 10% rise.

 

[1gibson]

10% from what? That's exactly the issue I am trying to bring up, head rise is not from BEP (that would be great if it were, very useful thing to compare curves.) Instead it is from the rated point. Because of this, a blanket statement that "10% minimum is required" is difficult to agree with.

If pump "A" BEP is 500 gpm @ 100 ft head, and shutoff is 150 ft head, that is 50% head rise from BEP. No problem to operate.

But my rated point for pump "A" is 200 gpm @ 140 ft head, so I have ~7.2% rise, less than the 10% rise to shutoff from rated, which is the API 610 definition. All of the sudden I have trouble operating pump "A"? No. I can't really operate much lower because I'm near minimum flow.

Now I have a different pump "B", BEP is 500 gpm @ 100 ft, and shutoff head is 111 ft. My rated point is at BEP, nice selection. 11% head rise to shutoff. Is pump "B" easier to operate in parallel than pump "A" with a rated flow of 200 gpm? Of course not.

I agree you need a certain amount of head rise to operate safely in parallel. I disagree with an arbitrary standard that is taken as an absolute requirement.

 

[BigInch]

Can't argue with you there about no absolute arbitrary standards. I also try to get rating points much nearer to BEP conditions. By reducing the flow "rating point" in your example, you have effectively, and in fact drastically, increased the slope; doubled it, no. Possibly reducing the need for a specified 10% rise to shutoff as well, because in a flow range that short, just about any pump of or over ANSI#300 would have a rise to shutoff well above 10%. Being that this is an API 610 pump, I will be willing to venture that the target flowrate is greater than 200 gpm, and the pump is possibly ANSI#600 pressure range, and he needs the 10%, especially if variable speed (I hope not), but I assume that mostly because 10% was what the spec said anyway. Get out to 2000 gpm and higher (a common range in my experience) and you really need to stick to 10% minimum, as those curves get to be relatively long and very flat from shutoff to BEP.

I am only trying to illustrate and emphasize the probable need to stick to the spec in this case, nothing else. I well realize that there are many reasons why you don't always need to do something for one reason or another, but here it is written into the spec. There is no ground to do otherwise without the customer's full knowledge, understanding and permission.

 

[1gibson]

Some more information is definitely needed. Is it really a flat curve, or is the operating point just far left of BEP? My gut says it's the latter, not a great selection to begin with, so meeting that 10% still won't help the situation much. Neither will rejecting the pump because not much can be done about it that late in the game, so realistically it is what it is.

One thing to note, the test setup will likely use a butterfly valve, there is probably some leakage at shutoff so it isn't really zero flow. A few GPM leakage might not be picked up by the flowmeter, and would skew the shutoff head reading to be low. This can be significant if it is a low flow, high head pump. The shape of the curve should tell you if this is happening or not. Depends on the smoothing, but if shutoff is not the highest head displayed (there is a peak shortly before zero flow) then this is probably happening, and the situation might not be as bad as it seems.

 

[BigInch]

I think we are all pretty far out on the curve right now, so to speak.

 

[Artisi]

Agree about being "out on the curve". I would be a good idea for the OP to post the so-called test curve - then we might just have something to discuss.

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.)

 

[SDRL]

Thank for you support, it helped a lot. I have attached pump curve. Take note that rated flow is 41.25 cubic meters per hour and total head is 1055 m. it is a low capacity high head pump as 1gibson said. As you can see from curve pump is operating at 70% BEP. I was thinking about rejected pump performance tested and force vendor to have some repairs on pump ( maybe have a slightly reduced discharge connection with a slightly larger impeller simultaneously to have a steeper curve), but since yoy are saying it may be not the case I am in doubt to do so. What would you recommend?

 

[BigInch]

You want a pump with 52% Efficiency?
Then you're going to operate it at 70% of BEP at 42% efficiency.

First, before you reject the pump, you should reject your pump selection procedure.

 

[LittleInch]

it would have been good to also mark on what the test results were, but I think we get the drift.

Seems to me like my learned posters above have alluded to which is that the vendor has been working on the edges of his design envelope and offered you a pump which is in reality a bit too "big". This will make your operation more expensive in operation (lower efficiency), but I don't know what your other options were.

I think the vendor is trying to pull a fast one myself, but the things to consider are this:
What leverage do you have with the vendor?
You say this is for a parallel pumping arrangement. Is the other pump(s) already there? or are you buying two from him?
If the latter then is your other pump the same?

Apart from the current response quoting the spec, have you had any other discussion with the vendor? you may need to go above your current contact to explain your situation and get some options. The contract engineer just wants to get this job complete and move onto the next critically urgent delivery, but the sales manager doesn't want to unnecessarily p*ss off a customer, so don't take a first refusal to compromise as the final statement - go round the obstruction if you can.

In any event, how are you / were you planning to control the flow through the pumps to equal it out. This forum is full of parallel pumping issues and in your case the flat curve will make it difficult to have equal flow within 10% without some form of flow monitoring or flow control to equalise between them.

So called identical pumps are not in reality that identical but you're not in great shape being significantly left of the BEP with a pretty flat curve.

We don't know where you stand on these issues therefore I cannot really provide any recommendations, just give some general advice.

Hope it helps.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[BigInch]

That's exactly it.
He's worried about 10% rise yet wants parallel operation of pumps working at 70% BEP, 42% efficiency. Crazy.
Start over using a new selection procedure.

 

[LittleInch]

I think it's a little late for that if he's at the point of testing them....

OP says it is parallel operation planned, but they only seem to have ordered one pump??

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way