FlightGlobal.com
Home
Premium
Archive
Video
Images
Forum
Atlas
Blogs
Jobs
Shop
RSS
Email Newsletters
You are in:
Home
Aviation History
1948
1948 - 1904.PDF
580 FLIGHT NOVEMBER IITH, 1948 Internal Combustion Turbines Six Lectures by Staff of the National Gas Turbine Establishment A NOTABLE service has been rendered by theInstitution of Mechanical Engineers in arranging,on November 3rd and 4th, six lectures on various aspects of the gas turbine. They may be regarded as asequel to an earlier series of lectures given before the Institution in February, 1946, and which have since beena valuable source of reference. The present series, de- livered by Stafi members of the National Gas TurbineEstablishment, Farnborough, discloses the findings of the research and development in the intervening period. As is implicit in the titles they do not deal specificallywith the aircraft gas turbine, but with efforts to determint1 the fundamental data essential for design in all spheres.As the lectures were, in the main, closely reasoned expositions of interrelated investigations into variousphenomena affecting the performance of gas turbine power units or their components, they are scarcely amenable tointelligible condensation and should be studied in extenso. They will be published by the Institution of MechanicalEngineers at an early date. Performance of Axial Flow Turbines By D. G. Ainley, B.Sc. IT was generally considered at the time of the early develop-ment of gas turbines for aircraft propulsion that the turbine design would probably present less difficulty than the combus-tion process or the compressor. Nevertheless, the early tur- bines benefited from advances made in the study ofaerodynamics and did not follow the established methods of steam engineers. Dr. A. A. Griffith conceived that the nozzlesand moving blades of a turbine should be designed to permit the gas stream to conform to a flow of constant angularmomentum in the spaces between blade rows instead of neg- lecting completely the centrifugal forces of the swirling gas.A. Cdre. Sir Frank Whittle arrived independently at a similar conclusion and designed bis turbines accordingly. With theseflow assumptions, the distribution between reaction and im- pulse turbines, as used by steam engineers, lost a great deal ofits meaning, as the reaction of a turbine stage now varies radially along the blade height, with the point of least reactionoccurring at the root. It is commonly accepted by aerodynamicists that energylosses of a fluid flow in a curved passage are considerably less when an acceleration is imposed on the flow than when theflow is accompanied by diffusion. Accordingly it was expected that an efficient turbine design of small weight and size wouldbe somewhat easier to achieve than an efficient design of axial flow compressor. So far, test results suggest that this viewwas somewhat optimistic and present-day axial compressor efficiencies are often as high as those of the related turbine.As a consequence, aerodynamic research into turbine perform- ance has been accelerated to elucidate questions such as: (a) The effect upon turbine efficiency of blading of high orlow deflection, with varying degrees of reaction, and the part-load characteristics. (b) The effect of blade profile form and blade pitch.(c) The effect of fluid compressibility and Reynolds number on the aerodynamic characteristics. (d) The effects of secondary induced flows, tip clearance,shrouding, axial spacing of blade rows, length /chord ratio of blades and other factors. A number of these factors may appear trivial, but in gasturbines a 2 per cent increase in expansion efficiency may lead to a 5 per cent saving in fuel consumption and the effect ofeach variable must be appreciated in order to attain maximum efficiency. Most of these problems can only be approachedby the expensive process of trial and error. The few full scale tests that have been made provide only the nucleus of ageneral fund of data that is badly needed. Apart from full- scale tests the cascade tunnel can furnish much useful datacomparatively quickly and cheaply. Cascade tests show that the range of incidence through whichboth an impulse blade section and a high-reaction blade section will operate without excessive variation in loss is quite largebut the reaction blade has a greater working range of incidence and the lowest loss. When a large family of turbine bladesections is tested it is found that the minimum loss co-efficient of a cascade invariably increases as the reaction of the blade is decreased. In this sense " reaction " is a qualitative expres-sion referring to the acceleration imparted to the gas as it flows through the blades and the accompanying drop in staticpressure. Such tests show that for a particular cascade the spacing of the blades must be chosen with much care to obtaina minimum profile loss. In practice turbine blade thicknesses are often more than 10 per cent of the chord at the mean bladeheight and maximum thicknesses are sometimes 30 per cent of the chord at the blade root. As the Mach number is increased, the minimum profile-losscoefficient varies only slightly until an outlet Mach number of 0.7-0.8 is approached. At this point a small local shock waveappears on the convex surface of the blade some little way inside the blade passage, causing a thickening of the boundarylayer and a consequent increase in loss. As the Mach number increases further the shock wave moves towards the trailingedge with a corresponding decrease of the length of blade ex- posed to the thickened boundary layer and results in a decreaseof the loss. The performance of a turbine is influenced by the Reynoldsnumber for gas flow and, at a fixed incidence, the losses in the turbine increase rapidly as the Reynolds number is reducedbelow 1.2x10'. Cascade tests show that when the Reynolds number is less than about 1.5 x io5 the profile loss for a par-ticular blade is approximately proportional to Re-'-'. In a tur- bine, where the mean turbulence is likely to be greater than inany of, the tunnels, the critical value of the Reynolds number may be rather less than 1.5x10*. The turbine'experimentalpoints suggest a critical Reynolds number of approximately 1.0 x 10s. This leads to the conclusion that the performance ofsmall-scale turbines or turbines intended to work with very low gas densities (as on aircraft turbines operating at very highaltitudes) may be seriously affected if the Reynolds number falls below about 1.0 x 103. • It is interesting to compare the performance of an impulseand a reaction turbine both in terms of the variations of horse power and efficiency with speed, when a constant pressure-ratio is maintained, and in terms of the variation of efficiency with horse power as a constant speed. These two conditionscould refer to an independently mounted power turbine in a gas turbine cycle which is used as a power unit for directtraction drive or as a power source to operate a constant-speed unit such as an electric generator or a constant-speed propellor. The relative variation of efficiency of an unchoked reaction turbine, a choked impulse turbine and an unchoked impulse turbine are similar at constant pressure-ratio, although the absolute values of efficiency are higher for the reaction turbine. At constant pressure-ratio, in a reaction turbine or a chokedimpulse turbines, the point of a maximum power coincides with the point of maximum efficiency, but the unchoked impulseturbine obtains its maximum power at a speed some 20 per cent less than the speed corresponding to maximum efficiency. Thiseffect is associated with the increase of mass flow with reduction in speed at a constant pressure-ratio through an unchokedimpulse turbine The change from "unchoked" to the DESIGN POINT (M0K. EFFICIENCY) REACTION TURBINE IMPULSE TURBINE (CHOKED) | IMPULSE TURBINE UNCHOKED) O2 O4 O6 O6 I-O 1-2 IR- RELATIVE SPEED Fig. I. Axial-fljw Turbine Performance at Constant Pressure-ratio. B 24
Sign up to
Flight Digital Magazine
Flight Print Magazine
Airline Business Magazine
E-newsletters
RSS
Events
