based on the flow velocity in the vertical leg. This loss is shown on Fig. 2 by starting the p-u curve for the flow entering the vertical leg from zero velocity instead of from the actual velocity that exists and constructing the curve to show a loss of one-half the dynamic head. The losses specified in this way are consistent with the two limiting cases of no flow into the vertical leg, where static pressure should exist, and flow from zero velocity in the straight tube, where the loss based on the inflow dynamic head is taken because of the poor inlet configuration. The solutions can now be carried out by assuming different shock strengths and then using the continuity relation at the T juncture to determine the relative area ratios appropriate to the conditions assumed. If a particular area ratio is desired, different conditions can be assumed until the required area ratio is obtained. The conditions shown correspond to equal areas in all legs of the T. Calculations for a T configuration with equal area legs have been made for various shock strengths and the losses as assumed previously (Fig. 3). Compared with experimental measurements made in the reference, the agreement is reasonable for both subsonic and supersonic cases. Other values of the loss coefficients, particularly higher values of the loss coefficient based on the inflow dynamic head for the supersonic cases, would give better results. Measurement of these losses by steady-state experiments would be useful. For initial shock strengths that give Mach numbers near 1 (P3~ 56 psi), no solution is possible using the loss coefficients selected, indicating that they are not correct for these Mach numbers. The extension of this model to L junctures and other configurations is obvious and will not be considered here. One use of this model is to correlate the experimental data, such as in Ref. 1, in terms of a few loss coefficients and to provide a tool for extrapolating to other values of initial shock strength. Its more important use is to provide a boundary condition at the juncture to use in unsteady flow calculations by the method of characteristics. Empirical data will probably never be available to provide the detailed information of what happens as waves of different types (such as multiple reflections due to the downstream configuration of the two legs) impinge upon the T and some analytic model such as this one is required to provide a self consistent boundary condition.
No takes yet. Share an insight, caveat, or question.
KENNETH E. FRENCH (1964) studied this question.