An investigation of the flow in an axial compressor inlet casing and its effect on the performance of the first stage of the axial compressor

Abstract

Summary The present report is a continuation of an interim report dated 8th August, 1951. The latter- report dealt with the flow in a model casing whereas the present reports deals with a similar investigation on the ‘Reavell’ compressor inlet casing which is to be part of the 9in x 9in supersonic wind tunnel installation, In the basic condition the flow had a circumferential instability* caused by the splitting of the inlet flow around the central shaft fairing. In addition the main flow through the casing was under-turned by approximately 200, Consequently the flow at the exit annulas was unsteady and changes of 300 in the exit flow direction were found. It was also found that at least four guide vanes were stalled in any one flow configuration. The instability was prevented by a flat dividing plate placed behind the central shaft fairing, but smooth flow at exit, in the vicinity of the plate, was only possible when curved fairings were added to both sides of the plate. A gauze screen, of 30 wires per inch, placed at entry, partly reduced the unsteadiness at exit but it had little effect on the flow direction and did not unstall the guide vanes. A perforated plate, having 15. inch diameter holes and 50 per cent blockage, similarly placed, had little or no effect. The guide vanes were unstalled by adding curved extensions to their leading edges. The large changes in the flow direction across the annulus at exit were, however, only reduced by modifying the internal contour of the casing. With these modifications flow angle variations of between 5° and 10° were obtained at exit. The velocity at exit was not uniform over the exit passages between the guide vanes, but the variations, which were not unduly large, were caused by secondary flows and the wakes from the vanes. This investigation has clearly shown that in the basic condition the unsteady, non-uniform flow at the exit from the inlet casing, which is immediately upstream of the first stage of the compressor, will tend to cause periodic stalling of the rotating blades in that stage. Since the relative velocities are high the danger of running into stn3ling flutter cannot be overlooked, especially as the cascade effect may reduce the critical flutter speed of the isolated blade. Stalling flutter, which usually takes place in the fundamental torsional mode, may be aggravated by a high order resonance occuring at the given rotational speed of the compressor. The suggested' modifications to the design of the inlet casing should prevent stalling of the first stage blades 'of the compressor caused by upstream, unsteady, non-uniform flow.