A wind-tunnel investigation of entry loss on propeller turbine installations. Parts 1 and 2

Abstract

The report is in two parts, following a general introduction. Part I describes wind-tunnel tests on (a) a series of models of annular entries, with and without propeller, in the 5-ft tunnel ; (b) a set of large circular blade roots on a full-size nacelle in the 24-ft tunnel. The models were based on two representative propeller turbine engines of different sizes. Various shapes and sizes of spinner and duct were tested, including 'vertical' and 'sloped' entries and 'elliptical' and 'conical' spinners. The work follows on from past tests, model and full-scale, on entries for radial air-cooled reciprocating engines. The smaller engine tends to have the higher entry loss, owing to the blade roots being relatively thicker. In a typical case, under cruising conditions, the total entry loss on the model is 25 per cent of free-stream dynamic head, of which 18 per cent is caused by the blade roots. Scale effect is likely to be small. In these circumstances a large diameter spinner gives the best result. Sloped entries are not recommended. From a generalised analysis of the results empirical rules are suggested for the estimation of spinner loss, duct loss, and blade-root loss, making up the total entry loss in flight. The additional duct loss which is usually present in ground running is also considered in general terms. Part II describes wind-tunnel tests on models of a number of alternative ducted spinners for a typical engine, and, for comparison, one annular entry similar to those tested in Part I. It is shown that the ducted spinners give 90 to 95 per cent total head in cruising flight compared with about 75 per cent for the annular entry. Most of the gain is in a reduction of blade root loss from 17 per cent total head to about 2 per cent. The results are not sensitive to the shape of the blade root fairing. Low velocity must be maintained as far as possible, both in the spinner itself and in the rear duct. Expansion of the duct in the neighbourhood of the leak should be avoided, however. The leak gap should be kept small, to minimise the extra flow taken through the spinner. A short cowl version, in which the outer cowl of the spinner terminates just ahead of the propeller, is satisfactory for practical purposes, and has the advantage of being lighter in weight than a long cowl spinner with nose entry. A detailed analysis of the loss is given, using methods evolved in Part I.