In simple terms the curve runs on the same graph of flow versus head.

At no flow the head might be positive, zero or even negative depending on where the static pressure is.

So if e.g. the end of the pipeline or pipe is an atmospheric tank 100m higher than the pump outlet, then you will need generate 100m of head before you get any flow. A similar applies if the you're pumping into a tank or vessel under pressure.

Then on top of the static head, you start to add frictional losses from your pipe system. These normally increase in proportion to a square of the flow rate so gradually rise as the flow increase, more so as you get more flow.

what you need to remember is that the pump curve is a differential head, so be sure to add on the suction head so you're all working from the same zero m point ( usually pump centerline)

So if your inlet pressure is equivalent to a fixed 50m add this to the pump output head to get you pump curve.

It is relatively simple, but always has complication in particular circumstances.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

We don't normally run experiments to determine the system curve. Instead, we make calculations.

For a simple system consisting of reservoir->pipe->pump->pipe->reservoir, the static head will be the difference in elevation between the two fixed head reservoirs. Then, calculate head losses (pipe friction losses and minor losses) for a series of flow rates (e.g. 100 gpm, 200 gpm, 300 gpm, etc.) and plot the static head (fixed) + head losses (variable) for each of the flow rates you calculated head losses for. Connect the dots and you have the system curve. I usually use the Hazen-Williams Equation for calculating pipe friction losses, but I sometimes use Darcy-Weisbach.

Sometimes it's that simple, but often it is not. I have designed sewer lift station and storm drainage lift stations that are about this simple, but I have also designed water wells and municipal booster pumping stations where the situations were more complex. For example, when you have water tanks in your system, the static head will vary because the water level(s) in the tank(s) will vary. For municipal water systems, you have various demand levels and fire flows to consider and possibly control valves, check valves, etc. For water wells, you may need to plan for a declining water table.

That being said, the fundamentals are still the same. A simple system is not too hard to calculate by hand, but a complex system is best done using water modeling software. I have recently answered a couple questions about EPANET (free software) and one of the answers involved building a model to do what you are taking about. Here is the thread: https://www.eng-tips.com/viewthread.cfm?qid=458044. Please note that I attached a file for developing system curves for a simple pumping station. If you have EPANET, or want to download it and give it a try, this example should help you understand the fundamentals.

See also my comments and other's comments here https://www.eng-tips.com/viewthread.cfm?qid=325862 and here https://www.eng-tips.com/viewthread.cfm?qid=320430 and here https://www.eng-tips.com/viewthread.cfm?qid=325862.

Fred

==========
"Is it the only lesson of history that mankind is unteachable?"
--Winston S. Churchill

System curve is graphical representation of resistance offered by the system.It
It shows that the pressure developed in the pump is high enough to overcome the resistance offered by the system.
It will start from 0 head and 0 flow if there is no static lift otherwise it will have some head atleast.

System curve is the curve on which power line, efficiency line, pump curve intersects and generate operating point of pump

It is very theoratical concept

A well phrased question has answers in itself!!!

No, it shows the calculated pump operating performance when pumping on a theoretical system.
Used initially to estimate the pump / motor and hydraulic requirements for a pump installation.
It is theoretical approach, but in the hands of an experienced pump application engineer can result in an accurate assessment.

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.)

Thank you everyone for replying.

When a pump is tested by manufacturer and pump curve is drawn based on that test, what is the system resistance? Doesn't it mean that pump manufacturers should provide both pump and system curves so consumers have better idea about how a pump will perform?

I think the pump curve should be plotted assuming zero system resistance. A system curve should be plotted assuming zero pumping force (zero external pressure). And then the two should be plotted on a same graph and a complete system can be analyzed. Please clarify. Thanks

"When a pump is tested by manufacturer and pump curve is drawn based on that test, what is the system resistance?"

Whichever resistances he applied to measure the differential head x flow to plot multiple points and draw a line.

"Doesn't it mean that pump manufacturers should provide both pump and system curves so consumers have better idea about how a pump will perform?"

What use would you have of the manufacturers testing rig's system curves? Are you evaluating his test rig or a system where you want to install a pump?

"I think the pump curve should be plotted assuming zero system resistance. A system curve should be plotted assuming zero pumping force (zero external pressure)."

What?

"And then the two should be plotted on a same graph and a complete system can be analyzed. Please clarify"

You have the pump's capacity curve that the manufacturer supplied to you. You calculate your system curve and plot it on top of this pump's capacity curve. You then see where both intersect and your system is thoroughly analyzed. This is of course assuming you calculated you system curve with 100% accuracy (which you didn't) and that the pump that will be supplied to you has a capacity curve exactly like the one on the catalogue (which probably doesn't). But this is another issue that we can discuss later.

If you don't know how to calculate a system curve, the following link may assist you.

https://www.engineeringtoolbox.com/pump-system-cur...

Just to add a small point to danschwind great post, where a manufacturer tests their pumps, the test results reported are corrected give the pump performance from inlet flange to discharge flange.

Why is this? Because
anything beyond the pump inlet /outlet flange is beyond the control of the manufacturer.

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.)

danschwind:

Thank you for the excellent answer. Your last paragraph made things crystal clear to me.