Considerations for the design of swirl chambers for the cyclone cooling of turbine blades and for other applications with high swirl intensity

摘要

The focus of this study lies on turbulent incompressible swirling flows with high swirl intensity. A systematic parameter study is conducted to examine the sensitivity of the mean velocity field in a swirl chamber to changes in the Reynolds number, swirl intensity and channel outlet geometry. The investigated parameter range reflects the typical kinematic flow conditions found in heat transfer applications, such as the cooling of the turbine blade known as cyclone cooling. These applications require a swirl intensity, which is typically much higher than necessary for vortex breakdown. The resulting flows are known as flow regime II and III. In comparison to flow regime I, which denotes a swirling flow without vortex breakdown, these flow regimes are characterized by a subcritical behavior. In this context, subcritical means that the flow is affected by the downstream channel section. Based on mean velocity field measurements in various swirl chamber configurations, it is shown that flow regime III is particularly sensitive to these effects. The channel outlet geometry becomes a determining parameter and, therefore, small changes at the outlet can produce entirely different flow patterns in the swirl chamber. In contrast, flow regime II, as well as flow regime I and axial channel flows, are much less sensitive to changes at the channel outlet. The knowledge about the sensitivity of the flow in different flow regimes is highly relevant for the design of a cyclone cooling system. Cooling systems employing flow regime III can result in a weakly robust flow system that may change completely over the operating range. As a remedy, the swirl intensity needs to be decreased so that flow regime III cannot be reached, which, however, reduces the maximum achievable heat transfer in the cooling system. Alternatively, the flow has to transition back from flow regime III to flow regime II or I before the flow leaves the swirl chamber. Two practical methods are presented. These findings can be directly applied in the design processes of future cyclone cooling systems, and other applications of swirling flow.