In this paper, for the first lime, a set of guide-lines are presented for the systematic design of mixed flow and centrifugal compressors and pumps with suppressed secondary flows and a uniform exit flow field. The paper describes the shape of the optimum pressure distribution for the suppression of secondary flows in the impeller with reference to classical secondary flow theory. The feasibility of achieving this pressure distribution is then demonstrated by deriving guide-lines for the design specifications of a 3D inverse design method, in which the blades are designed subject to a specified circulation distribution or 2πrV¯θ. The guide-lines will define the optimum choice of the blade loading or ∂rV¯θ/∂m and the stacking condition for the blades. These guide-lines are then used in the design of three different low specific speed centrifugal pump impellers and a high specific speed industrial centrifugal compressor impeller. The flow through all the designed impellers are computed numerically by a 3D viscous code and the resulting flow field is compared to that obtained in the corresponding conventional impeller. The results show consistent suppression of secondary flows in all cases. The design guide-lines are validated experimentally by comparing the performance of the inverse designed centrifugal compressor impeller with the corresponding conventional impeller. The overall performance of the stage with the inverse designed impeller with suppressed secondary flows was found to be 5% higher than the conventional impeller at the peak efficiency point. Exit flow traverse results at the impeller exit indicate a more uniform exit flow than that measured at the exit from the conventional impeller.

In this paper, for the first time, a set of guidelines is presented for the systematic design of mixed flow and centrifugal compressors and pumps with suppressed secondary flows and a uniform exit flow field. The paper describes the shape of the optimum pressure distribution for the suppression of secondary flows in the impeller with reference to classical secondary flow theory. The feasibility of achieving this pressure distribution is then demonstrated by deriving guidelines for the design specifications of a three-dimensional inverse design method, in which the blades are designed subject to a specified circulation distribution or 2πrVθ. The guidelines will define the optimum choice of the blade loading or ∂rVθ/∂m and the stacking condition for the blades. These guidelines are then used in the design of three different low specific speed centrifugal pump impellers and a high specific speed industrial centrifugal compressor impellers. The flows through all the designed impellers are computed numerically by a three-dimensional viscous code and the resulting flow field is compared to that obtained in the corresponding conventional impeller. The results show consistent suppression of secondary flows in all cases. The design guidelines are validated experimentally by comparing the performance of the inverse designed centrifugal compressor impeller with the corresponding conventional impeller. The overall performance of the stage with the inverse designed impeller with suppressed secondary flows was found to be 5 percent higher than the conventional impeller at the peak efficiency point. Exit flow traverse results at the impeller exit indicate a more uniform exit flow than that measured at the exit from the conventional impeller.

The application of a three-dimensional (3D) inverse design method in which the blade geometry is computed for a specified distribution of circulation to the design of turbomachinery blades is explored by using two examples. In the first instance the method is applied to the design of radial and mixed flow impellers to suppress secondary flows. Based on our understanding of the fluid dynamics of the flow in the impeller, simple guidelines are developed for input specification of the inverse method in order to systematically design impellers with suppressed secondary flows and a more uniform exit flow field. In the second example the method is applied to the design of a vaned diffuser. Again based on the understanding of the detailed flow field in the diffuser obtained by using 3D viscous calculations and oil flow visualizations, simple design guidelines are developed for input specification to the inverse method in order to suppress corner separation. In both cases the guidelines are verified numerically and in the case of the diffuser further experimental validation is presented.

The overall performance of shrouded and unshrouded identical impellers of a centrifugal compressor were tested and compared. A closed loop test stand with Freon gas as the working fluid was employed for the experiments. The inlet and outlet velocity distributions of both impellers were measured using a three-hole cobra probe and a hot-film probe to determine the velocity distribution and unsteady flows due to wakes and inlet stall. IntroductionBoth shrouded and unshrouded impellers are widely used in centrifugal compressors. The high-pressure, single-stage impellers used for turbochargers and gas turbine compressors are unshrouded because of the high stress requirements and the problems of machining and precision casting. Senoo et al.[1] report that the performance with unshrouded impeller is affected by the clearance between the casing and the impeller, especially in the region where there is a large reduction in relative velocity in the impeller channel.Multistage compressors requiring high reliability for industrial use and gas processing are widely used with shrouded impellers. However the performance difference between shrouded and unshrouded impellers in centrifugal compressors has, up to now, come under little investigation. Howard et al. [2][3] have measured the effect on internal flow caused by a shroud in a 9-in. diameter clear plastic impeller both with and without a shroud, using the hydrogen bubble method and a hot film probe. They reported that with unshrouded impellers the secondary flow in the impeller channel varies from the pressure side of the blade to the suction side due to leakage across the clearance, and that the presence of a shroud affects stall conditions at partial capacity. However, since the rotational speed was as low as 140 rpm, their study did not discuss the overall performance difference between shrouded and unshrouded impellers.In the study covered by this paper, the overall performance of shrouded and unshrouded backward impellers was compared. The circumferential Mach number of the tested impeller was varied up to unity, and a three-hole cobra probe and hot-film probe were used to measure the velocity distribution and unsteady flow at the impeller inlet and outlet.

In order to improve the operating range of a centrifugal compressor, computer-controlled variable inlet and diffuser vanes were attached to a compressor with a pressure ratio of 2.5. Low-solidity cascade vanes capable of controlling the vane angle up to 0 degrees from the tangential direction were used for the vaned diffuser. The compressor’s overall performance was then tested using a closed-loop test stand. By automatically adjusting the diffuser vanes to the most suitable flow angle, pressure fluctuations caused by the unstable flow in the diffuser during low-flow operation of the centrifugal compressor could be suppressed, and the compressor could be operated nearly up to the shut-off flow rate without any surge. The author experimentally confirmed the critical operating range of both the impeller and diffuser at two different tip speeds and five inlet guide vane angles. Furthermore, a three-dimensional viscous flow-analysis method was applied to the impeller, and a three-dimensional momentum integral analysis method was applied to the diffuser. Then the critical operating ranges obtained in the experiments were qualitatively validated. The operating range of a centrifugal compressor under low-flow conditions, which has until now been limited because of surge, dramatically improved in this study, thereby demonstrating that it may be possible to develop a surge-free centrifugal compressor.

For the axial turbine stage, the design of circulation rVθ¯ distribution between the nozzle and blade has an important effect on the stage performance, because it determines the work distribution in the blade, the stage reaction and the twisting shape of the blade. This paper describes the new method of full 3D design for axial turbine nozzles and blades by applying the 3D inverse design method in which the blade geometry can be determined by specified distributions of circulation rVθ¯ and blade thickness. In this 3D inverse design method, spanwise work distribution of the turbine stage is controlled by specifying the rVθ¯ distribution of the nozzle exit. In this design procedure, rVθ¯ distribution at the nozzle exit and 3D stacking condition are both controlled by 3D inverse method so as to suppress the nozzle secondary flows effectively. The desirable rVθ¯ distribution and 3D stacking shape which were obtained by the 3D inverse method were confirmed by Dawes’ 3D Navier-Stokes analysis. The results shows that the secondary loss is reduced when the design rVθ¯ at the mid-span is set larger compared to that near the endwall. In addition to the control of the rVθ¯ distribution, 3D stacking shape added only in the front part of the nozzle is very effective to suppress the secondary flows, although this 3D stacking shape is very simple compared to a conventional bowed type stacking. Moreover, when this stacking shape is used, spanwise distribution of work does not change from the design condition unlike the case of conventional bowed type stacking shape. The results of single stage performance test conducted using an air turbine facility show an improvement in efficiency compared to the 2D designed stage and prove viability of the 30 inverse design of axial turbine blades.

No abstract

No abstract