Modeling a Tee
by Engineered Software, Inc.
Users often inquire about the correct way to account for the losses through a Tee fitting. PIPE-FLO Professional utilizes the fitting methods described in the Crane Technical Paper 410: Flow of Fluids Through Valves Fittings and Pipes. The TP-410 provides just two L/D values for Tees that are use to calculate the K resistance values; 20 for Flow Thru Run and 60 for Flow Thru Branch. The TP-410 offer no further information about how to apply these values as that is all of the information needed to solve for a single flow path. The confusion comes about when trying to determine how to incorporate these values into a PIPE-FLO model where we are solving multiple paths simultaneously. Ultimately, the choice of method is based in the engineering judgement of the user. For some applications, the losses through Tees are insignificant and for other applications the TP-410 method may be too general, and more geometry specific methods are desirable. I will cover some options to get these Tees modeled in PIPE-FLO.
1) The typical recommendation is to account for the resistance of the Tee in both of the pipes that represent non-combined branches of the Tee. This ensures that each of the two flow paths sees the resistance for the path that it is involved in, without any path seeing the resistance for two fittings. he branch with the combined flow assumes that the resistance is accounted for already in the other two paths. If the flow distribution changes, the location of the fittings should be changed to match. Using a generalized value for K will introduce some degree of inaccuracy to the model, but in many cases the variance in loss at the Tees is considered negligible compared to the overall losses in the system.
2) Calculate the losses through the Tee any of the other existing methods (Idelchik, Miller, ASHRAE, etc...) and enter the losses as custom loss items in the PIPE-FLO model; Fixed K fitting, Fixed Cv fitting, Fixed dP Device, or Curve dP Device. This is a more time consuming process, but provides losses specific to the geometry and flow profile of the actual fitting rather than a generalized K value for all fittings. System type and sensitivity of the application may dictate this more rigorous approach.
3) For some applications, the losses due to Tees and some other valves and fittings are insignificant compared to the losses from pipe runs and other devices. Some choose to ignore the fittings in these cases, though it is always recommended that you keep your model as close to the reality of a system as possible.
As you can see, there are several different approaches to modeling the resistance effects of tees. The approach you decide to use is up to your preference, application, and ultimately, engineering judgment.