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

[Zoey90]

I'm trying to size a centrifugal pump for a closed-loop cooling system, but I keep getting stuck on the system curve calculation. My static head is negligible, so it's all about the friction losses, and I'm second-guessing my equivalent length of pipe fittings.

[PenelopeR]

I once chased a near zero static head loop and learned the friction losses steal the show. We weathered it by picking a target flow, then lining up the pump curve against the pipe losses with real fittings included. The biggest surprises were the elbows and tees—they ate into head more than I expected, and a lot of it came from long radius fittings rather than a few sharp ones. We ended up adjusting lengths to hit the right point on the curve rather than trusting the spec sheet.

[Jerry70]

It might be overthinking it. If the static head is negligible, maybe the real limiter is the cooling duty and how much flow you actually need to keep the coolant temp within range. Friction loss is a factor, but a loop can run fine with less flow than you think as long as the temperature target is met. I’ve seen people chase a perfect system curve and burn a lot of time.

[OliviaHM]

I did a quick check using the Darcy Weisbach style sum: compute hf = f (L/D) (v^2 / 2g) with f from Moody for the pipe material, add K values for each elbow and valve, and then convert to equivalent length for fittings to keep the numbers in one lane. The result swings a lot if you tweak the Reynolds number a bit or the roughness you pick, so I stopped trusting a single pass and kept rerunning with small changes until the pump curve met the losses at the needed flow.

[Charles19]

One time a leak turned the whole thing on its head mid-build and I had to recalc on the fly. The curve looked different after the fix, and I started doubting whether temp changes or density shifts are insignificant. Honestly I’m not sure we solved the root cause, just got a number that fit the spec for a moment.