Part Load Efficiency Values (PLEV) Part 4: End of Curve Selections

By Chad Edmondson (JMP) and Norman Hall (RLD)

When looking at Bell & Gossett pump selections we often find higher part load efficiency (PLEV) pump choices are to the right of BEP or best efficiency point. How close to end of curve should a prudent engineer choose?

The following shows a Bell & Gossett selection table sorted on PLEV or part load efficiency value.


The most efficient part load selection is the e-1510-3AD with a PLEV (part load efficiency value) of 81.79.  This selection is 3500 RPM and not the traditional 1750 normally scheduled. That 3AD selection also shows it is 90% of end of curve (EOC). How important is EOC or end of curve?

Is it safe to select pumps to the right of BEP?

ASHRAE suggests pumps be selected between 66% and 115% of BEP or best efficiency point. In this example, the ASHRAE satisfactory range is between 473 GPM and 824 GPM. This selection meets that criteria.


ASHRAE’s selection criteria is concerned with the operating efficiency of the pump selection. Of course efficiency is important but concerns about running off the end of curve is even more important.

This example has an end of curve (EOC) point of 891 GPM at the design speed of 3550 RPM and a 6.625” impeller. Since the pump is selected at 800 GPM which is 90% of end of curve, could it be a problem? The answer, like so many when we get close to the edge, is, it depends!

Issues Surrounding the EOC Point

When ASHRAE suggests a maximum of 115% of BEP there’s the assumption that you selected the head of the pump exactly with no guesses, rounding up, or “rules of thumb.” There is also an assumption that there is no safety factor. What affect might these things have on a pump too close to EOC (92.5% in this example)?

Here is the same curve showing a control curve of 30% of design head and variable speed.


Where would the operating point be if there was 15% added for safety factor and 20% overestimate of pump head?

Think that is ridiculous? How many of us use 4.0 feet per hundred as a default in pipe friction when the B&G System Syzer® would show 2.7? How many of us automatically use 6” pipe when the system syzer suggests 5”? How many of us use 25% or 50% for fittings as an estimate? How many times did we round up an equipment pressure drop and then add safety to it? We play it safe in pump head selections and, in general, we are all guilty, including us!

This selection (see below) is off the curve at design flow rate!


Another issue is the balance of the system. Engineers often specify 10% balance tolerance so many of the balancing procedures start with the pump at 110% of design flow and the discharge triple duty valve or speed is raised to maintain that point during balance. It is not unusual that when the balance is complete, the speed is never adjusted or the impeller trimmed as ASHRAE suggests in the 90.1 standards. This makes the selection go even farther to the right.

Diversity is an expectation that all control valves on terminal units will not be open at the same time. When using diversity in the design, the pump GPM is lower than the total of the terminal units. But what happens when there is a system startup or night/weekend setback and the there is a call for the system to change room temperatures all at once? Diversity disappears as all the valves go open at once. This is yet another scenario which could cause the pump to move farther to the right of design.

Finally there is the assumption that the controlled system curve remains stable during operation. Many articles have appeared showing control curve shift under various combinations of 2-way operation. Some of those shifts cause the control curve to drop and operate with more flow and less head than design, again moving the operation to the right on the pump curve.


Best Practice for Selections

All of these possible outcomes should be weighed when deciding whether to move to the right of BEP with your selections.

When tempering pump part load efficiency with the possible unintended operational points, we suggest standardizing on selections at 85% of EOC or less, unless more analysis of a possible system curve drift is reviewed. This seems to be a comfortable compromise between the old sage statement of “always select far to the left of BEP” and a newer cry of “efficiency trumps all.”

Next week we will conclude this series on Bell and Gossett PLEV with a few comments about part load and part flow.