Theory and tunnel tests of rotor blades for supersonic turbines

Abstract

In special circumstances where a large work output is required from a turbine in a single stage it is necessary to us e high pressure ratios across the nozzle blades, thus producing supersonic velocities at inlet to the rotor. As part of an investigation into such turbines, several designs for the inter-blade passages of the rotor have been tested in a two-dimensional tunnel, a design theory being developed concurrently. The first design, featuring constant passage width and curvature as in steam-turbine practice, but having thin leading and trailing edges, was found to suffer from focusing of the compression waves from the concave surface, with consequent flow separation from the opposite convex surface. It gave a velocity coefficient of 0.929--based upon the area-mean total pressure as measured at an inlet Mach number of 1.90 and turning angle of 140 deg. The measured value compares favourably with values from previous steam tests; where the results have been in the range from 0.65 to 0.92. From theoretical reasoning, and from additional test observations, a subsequent passage was designed having an inlet transition length of small curvature, leading to a free-vortex passage of double the transition curvature; a small amount Of contraction was incorporated. Schlieren photographs showed the flow in this passage to be almost shock free. A thin region of low-energy air existed close to the convex surface, but liquid-injection tests located only one small bubble of reversed flow. Pressure traverses at exit indicated a velocity coefficient of 0. 952, based on the area-mean total pressure. When allowance is made for turning angle and Reynolds number this result appears to compare quite favourably with previous work. It would seem that the optimum blade pitching in a turbine would be about 20 to 30 per cent closer than in a two,dimensional cascade. However, the resultant pitching tends to become very close, except at very large turning angles, with the result that in some applications difficulties could arise in the practical design and manufacture. Several uncertainties remain and the present design must be regarded as still experimental.