Why the pump curve on the datasheet is not the one you get

Every curve on every datasheet was measured on clean, cold water, at a fixed speed, with a new impeller, on a test rig with straight pipe either side of the pump. Your installation is none of those things.

What the test rig does not have

Suction conditions are the first divergence. The rig is fed from a header with generous NPSH available; a real wet well is fed by whatever the level happens to be that morning, through a bend somebody added during construction. Below a certain margin the pump is not producing the curve at all — it is cavitating quietly and losing head you will attribute to something else.

The second is wear. A pump that has moved abrasive-laden water for two years has wear-ring clearances that no longer match the ones the curve was measured with. Recirculation through that gap costs flow at every point on the curve, and it costs most at the duty you actually run at.

Reading a curve properly

Take the duty point you need and find it on the curve. Then ask what the efficiency is at that point rather than at the best efficiency point, because those are rarely the same place. A pump selected to the right of BEP runs hot and wears fast; one selected well to the left recirculates internally and does the same thing for a different reason.

The useful question is not "does this pump make the duty" but "how far from BEP will it sit for the ninety per cent of the year that is not the design condition".

What to measure instead

A pressure gauge either side and a flow meter will tell you where you are on the curve today. Trend that over a year and the wear shows up as a slow migration down and to the left, long before anything fails. That trend is worth more than the curve it is measured against.