Pump Field Performance Test: BB3 Multistage Pump

[PumpGuy101]

Would appreciate your thoughts. I am reviewing return to service results for a BB3 Multistage Pump. Flow meter located on discharge, suction and pressure gauge located more or less at respective flanges. No elevation correction required. This is a high head pump (>1300m, ~10-13 L/s), no need to consider velocity head, and the friction correction is also unnecessary (no minor losses between gauge and flange and short piping length). I reviewed all flow paths and because there is a TEE upstream of the main flow element, I added the flow from two separate flow elements to get the total flow.

After taking flow and pressure readings at operating point, I found the TDH for OEM curve comparison. Here is where I would appreciate some input.

Case A. If I assume the field flow reading is correct, and I find the TDH on the OEM pump curve at that flow rate, I find the curve TDH at measured flowrate to be within 4% deviation of the field TDH. This is within 7% therefore acceptable for a safety related pump (no evidence of degradation).

Case B. However, out of curiosity, I also did the following: if I assume the TDH field measured is correct, and I find the flow on the OEM pump curve at that TDH, I find the curve flowrate at measured TDH to be ~35% deviation from the field flowrate.

I believe Case a is the standard practice for field performance review. However I am confused as to why case b curve deviation is so much larger. To that point:

Case C: I measured BHP and estimated TDH and Flow from the BHP and Pump Curves. TDH predicted by this method was acceptable, and though estimated flow was closer to that estimated using TDH from field, it was still More then 20% deviation from measured.

Questions:
a. Any explanation for the mismatch? Why is assuming Flowrate or BHP as measured to find expected TDH on OEM curve for deviation review much closer to predicted then assuming TDH/BHP to find expected Flowrate on OEM curve for deviation review where the result indicates a large deviation from field to curve?
b. Do you concur that correct practice is performed per Case A?

 

[Snickster]

Attached is pump curves for BB3 multistage water injection pumps on a project I worked on. Notice the curves are fairly flat especially towards the lower flows. I think perhaps the flowrate measured with a meter is more accurate than the pressure measured with a pressure gage or BHP measurements. So if you are very accurate with your field flow measurements then if you go up from flow to the curve and to the left to differential head then you will be more accurate than if you take your measured differential pressure in the field and go from the head on the right to the curve and then go down to flow. This is because since the curve is flat, if you are just a little off on your field measurements of differential pressure you will be off a lot on your flow. I would go with the flowrate measurements in the field rather than the differential pressure measurements. Just my opinion as I never did such field testing.

 

[pierreick]

Hi,
Make sure your Gauges & meter are calibrated prior to test. Flat pump curves are not going to help you in your analysis.
A variation on RPM will affect the analysis.
Good luck.
Pierre

 

[Artisi]

Following Pierre's comment, discharge pressure changes at the square of the speed change and flow at the ratio of change, this can make a difference between OEM curve and operating curve, plus OEM test measurements are probably a lot more accurate than your site measurements.

 

[LittleInch]

I'm confused by your statement here "because there is a TEE upstream of the main flow element, I added the flow from two separate flow elements to get the total flow." Without a diagram I can't see where upstream refers to, but is this not double counting?

Anyway I agree with the others, the key parameter is flow and the differential head being generated at that flow.

Without seeing the pump curve it not possible to say exactly, but the diff head change can be very small for a large change in flow. So as noted above a small change in diff head could result in a large change in flow.

BHP is usually an output from the test, not an input as there are a number of things which could affect it including other losses in the pump, a different efficiency of your pump compared to the ideal pump. So for me it would be more, what is the BHP of the pump at flow of X and compare that to the OEM curve. Again, BHP curves can be quite flat so a small error or change can result in an inaccurate assessment of flow.

So yes, Case A is the correct way to do things IMHO.

 

[EdStainless]

Do a little sensitivity analysis.
Assume that your readings are off by 3% in either direction.
Now recalculate your results.
You will find that this creates a large range of results.
Normally people watch the pressures for signs of changing performance, but not to verify actual output.

 

[PumpGuy101]

Appriciate the responses. The slope is not entirely flat at this are of the curve but it is not large either. This result is similair to several other identical installations. It would appear that curve shape in this instance must indeed favor Case A in this instance (all measuring devices within calibration). In another pump model I found either case or b was accurate, curve was more steep in area of testing in that instance.

To answer above commenter: The tee upstream of the main flow element measuring device is such that I need to account for Total Flow by adding both the branches which are outlets from the pump header.

I will continue to follow case A and do agree with consesus.

 

[georgeverghese]

With this pump being high head, a flattish Q-h profile can be expected. NPSHr would also be high. For a flat Q-h curve, min flow recycle should be a dedicated flow control loop. Right now the FE is in the wrong place :- it should be upstream of this existing tee so it reads only flow from this pump. And the tee off to the min flow recycle FCV should also be upstream of this tee. Add a check valve immediately downstream of the take off to this new min flow recycle tee, which should also be upstream of the current tee.