Engineering context

Wherever a pipe divides or two streams combine, a tee or junction sets how flow and pressure are shared between the legs. The split is not arbitrary: flow distributes so that mass is conserved at the junction and the pressure is balanced across the connected network. The fitting loss depends on flow direction, the branch-to-total flow ratio, and — critically — which leg is designated as the branch versus the straight run.

The loss coefficients for the run-through path and the branch path are different, so designating the wrong leg as the branch changes the predicted pressure drop and, in a connected network, the flow split itself. This is also the most common cause of the Unable to Prepare Network error in FluidFlow.

Engineering workflow

1. Place the junction element and connect every leg — for a tee, all three pipes must be joined to the junction node with no loose ends.
2. Define which pipe is the branch and which legs form the straight run. This assignment drives the correct loss coefficients.
3. Confirm the flow regime: dividing (one inlet, two outlets) or combining (two inlets, one outlet). Loss behaviour differs between the two.
4. Choose the fitting-resistance method for your case. FluidFlow calculates K using Idelchik, Miller, Crane, and SAE methods; SAE is available for gas-junction work.
5. Assign pipe data to each leg — diameters, lengths, roughness — so the relative resistances of the paths are correctly represented.
6. Set the network boundary conditions and the fluid.
7. Solve the steady-state network. The solver enforces mass balance at the junction and pressure balance across the legs, returning the flow split and pressure drop through each path.
8. Review the split, branch and run velocities, and any design alerts before accepting the result.

Why branch assignment and the whole network both matter

Branch designation is not cosmetic — it feeds directly into the K-factor calculation and therefore into the predicted pressure drop and split. The correlation choice (Idelchik, Miller, Crane, SAE) affects the fitting loss magnitude, which matters most when junction loss is a significant fraction of total route resistance. Solving the connected network resolves the split directly and stays correct when anything downstream changes — a valve position, a pipe addition, a load change.

How FluidFlow helps

FluidFlow models tees and junctions as part of the connected steady-state network. You connect the legs, designate the branch, and select the fitting-resistance method; the solver balances mass and pressure to return the flow split and per-leg pressure drop. FluidFlow provides tee, wye, and cross junction elements, and the Idelchik, Miller, Crane, and SAE fitting-resistance methods for liquid and gas cases respectively.

Go deeper