Tests concerning novel designs of blades for axial compressors Part 1 blades designed for increased work at root and tip Part 2 blades designed to operate in a parabolic axial velocity distribution

Abstract

Investigation of the flow between the blade rows in a research compressor with conventional Half-Vortex blades have shown a rapid stage-to,stage build-up of the annulus boundary layer. The axial velocity distribution under these conditions bears no resemblance to the straight-line distribution which is usually assumed in design so the incidence angles at both root and tip probably give rise to considerable secondary losses. Two sets of blades were designed therefore, with the object of minimising these effects. The first set was designed with an increased work input at the root and tip of the blades in order to re-energise the boundary layer and prevent axial-velocity profile deterioration. The second set was designed using as a design assumption a mathematical approximation to the axial-velocity profiles found experimentally at various stages in a conventional compressor. The results obtained from low-speed tests of the first set of blades show that although the boundary-layer thickness was considerably reduced, this was achieved only at the expense of 2 per cent in maximum efficiency as compared with conventional blades. The drop in efficiency is attributed to the stator blade rows which stall at incidences well below the design value, but the effect is probably not uniquely associated with variable work done. The blades designed for a variable axial-velocity profile also failed to give an improved performance. The maximum efficiency was 3 per cent below that of the equivalent conventional blades and the axial-velocity profile instead of remaining approximately:as assumed, deteriorated still further. It is considered that designing for a poor axial-velocity profile is probably wrong in principle. The Variable Work Done blades, however, would almost certainly have succeeded in producing an axial velocity substantially constant over the greater part of the annulus if the work had been symmetrically distributed about the mean radius with a ratio of mean to minimum work of about 1.1. Under these circumstances it is likely that there would have been a reduction in secondary losses and an increase in efficiency.