Requirements for aeronautical gas turbine engines for helicopters include small size, low weight, high output, and low fuel consumption. In order to achieve these requirements, development work has been carried out on high pressure ratio compressors with high efficiency. As a result, we have developed a single stage centrifugal compressor with a pressure ratio of 11 for a 1000 shp class gas turbine. This report presents a study on the internal flow of a high pressure ratio centrifugal compressor impeller. The centrifugal compressor is a high transonic compressor with an inlet Mach number of about 1.6. In high inlet Mach number compressors, the flow in the inducer is a complex transonic flow characterized by interaction between the shockwave and boundary layer, while the flow in the middle of the impeller is a distorted flow with a low energy region. In order to ensure the reliability of aerodynamic design technology for such transonic centrifugal compressors, the complex transonic flow and formation of the low energy region predicted by CFD must be actually measured, comparison must be undertaken between the CFD results and the actual flow measurement, and the accuracy and other issues pertaining to CFD must be clarified. In a previous report [12], we elucidated the flow in the inducer of a high transonic impeller by means of LDV and unsteady pressure measurement. That report showed that, in the flow of an inducer with a Mach number of approx. 1.6, the oblique shockwave in the middle of the impeller throat interacts with the blade tip leakage flow, and that reverse flow occurs in the vicinity of the casing. Furthermore, although CFD predicted a low energy region in the splitter portion, this could not be detected in actual measurement. In the context of the current report, comparative verification of the CFD and LDV measurement results was undertaken with respect to the formation of the casing wall surface boundary layer in the transonic flow within the inducer. In this conjunction, inducer bleed was introduced to control this boundary layer, and the effect of the inducer bleed on the flow was ascertained through actual measurement. It was also sought to additionally confirm the “low energy region” in the splitter. Accordingly, the flow velocity distribution was measured at two sections, thereby clarifying the characteristics of the actual flow in the region. The impeller for which measurement was performed has the same specifications as that in the previous report (see Table 1). In the present report, so as to measure the flow under conditions encouraging the formation of a boundary layer accompanying substantial inducer deceleration, measurement was conducted at 95% of design speed and a relative Mach number at the blade tips of about 1.5.

No abstract

No abstract