Discussion: “Performance of Radial Flow Turbines Under Pulsating Flow Conditions” (Kosuge, H., Yamanaka, N., Ariga, I., and Watanabe, I., 1976, ASME J. Eng. Power, 98, pp. 53–59)

Abstract

1 The limited time allotted for the preparation of the written discussion prevented me from obtaining reference material dealing with the subject problem. Hence, I cannot adequately compare the authors' approach with others that have been developed.2 The use of the dynamometer for power measurement and the location of the flow measuring device upstream of the rotary valve demonstrated a good research approach in obtaining valid measurements of the two turbine parameters.3 The paper essentially presents a check on the old quasisteady flow assumption using an "equivalent amplitude" approach instead of a "pulse frequency" approach. The authors do not present a real comparison of their approach with that obtained by other investigators. The conclusions should make a stronger statement with regard to the approach developed by the authors. The major conclusion should be stated that the quasi-steady flow analysis is still not satisfactory in view of the large errors that approach 40 percent.4 The relevance of the exhaust pipe studies is not clearly stated with regard to the quasi-steady flow assumption with respect to the equivalent amplitude approach.The authors have presented very interesting results on the comparison between the predicted quasi-steady flow and the measured nonsteady performance of a radial turbine. The nonsteady flow performance of a turbine depends on the mean pressure and temperature levels at entry, and the shape, amplitude, and frequency of the pressure pulses. The concept of an equivalent amplitude (A7r*Ar) is very useful as it can take into account the amplitude of the pressure ratio variations and the pulse shape. However, the comparison between the performance of the turbine evaluated on the basis of the quasi-steady flow assumption and that measured directly under the nonsteady flow conditions is not as good as might be expected. A major cause of the discrepancies is probably the rotary valve which the authors used to generate the desired pressure pulses.The pulsating flow can be of two types: (i) nonsteady continuous flow; and (ii) nonsteady discontinuous flow. Usually the ports of the rotary valve housing can be designed to achieve the desired shape of the pressure pulses, but the flow would be discontinuous because of its geometry. From the sketch, it can be seen that the valve would remain closed for 180 -(8 + <j>)lu> s where w is the angular velocity of the valve in radians/s. Hence, ignoring leakage the flow would be discontinuous but the decay rate of the pressure downstream of the valve would depend on the volume of the manifold intervening between the valve and the turbine and on the flow characteristics of the turbine. It is clear that in quasi-steady flow