How do I calculate total dynamic head in a closed-loop system with minor losses?

[Gary.R]

I'm trying to size a centrifugal pump for a closed-loop cooling system, and I'm getting stuck on the system curve calculation. My issue is that I can't seem to accurately account for the total dynamic head once I factor in all the minor losses from the valves and fittings in my piping layout.

[LarryFL]

I spent days chasing the total dynamic head and learned that minor losses from valves and fittings change with flow. I built a small test rig and pulled K values from the catalog and added them in as equivalent lengths. It helped but there was still quite a bit of uncertainty.

[Emily_G]

In the plant we split the losses into friction along the pipe and then treated each valve or elbow as a small hurdle. The curve moved a lot when a valve was opened a little and again when it was fully open.

[Savannah27]

We kept a rough system curve by using the pipe diameter and roughness to estimate friction and then added a fixed fudge for minor losses until the pump curve crossed at a reasonable flow.

[Nora.R]

A second approach was to group fittings by type and assign a single K value from the catalog and treat that section as a lump head loss.

[Frank63]

Could it be that the real issue is not the losses but the assumption of a single flow rate for the loop?

[Emily.L]

Sometimes when I hit this I end up setting a conservative spread on the pump curve and call it a win if the heat exchanger stays within spec.