How do I size a centrifugal pump for a closed-loop cooling system?

[Addison40]

I'm trying to size a centrifugal pump for a closed-loop cooling system, and I keep getting stuck on the system curve calculation. My theoretical head requirement seems way higher than what feels practical for the pipe length and fittings I've laid out, which makes me question if I'm properly accounting for all the minor losses.

[Madison52]

Yeah I got stuck there too. I built the system curve by static head plus the friction losses, and the numbers looked crazy until I counted every little loss from fittings. Those few 90s and tees added more head than the pipe length suggested. I used equivalent lengths for each fitting and then ran Darcy-Weisbach with the pipe roughness for the material. It made the curve creep up and matched the pump curve better.

[Thomas_R]

I kept chasing the numbers until I realized the target flow might be driving the head up. If you pick a relatively high flow for cooling the minor losses become a bigger share and your system curve climbs fast. Try sanity checking the design flow vs actual cooling needs and see if you can operate at a lower flow without hurting temps.

[AddisonP]

Are you sure you are counting the static head and losses correctly across the whole path including the heat exchanger and any filters? A missing valve or a filter can flip the picture?

[BrandonLJ]

One concrete thing I did was measure pressure drop with a simple manometer at a couple of flow rates and compare to the calc. It helped me see if I was overestimating losses. If you can, run a small test loop and log flow and head, then plot the curve.