The Cooled Gas Turbine

Abstract

The cooling of turbines offers the alternative advantages of the use of higher cycle-maximum-temperatures than are possible with uncooled turbines, with consequent increase of thermal efficiency and power output for a given size of compressor, or the use of lower quality materials for a given turbine inlet-temperature. Two main methods of cooling the blades, which are the most difficult elements of the turbine to cool, have been investigated in the past. They are internal water-cooling, offering a simpler technical problem in cooling due to the great heat-removing capacity of water as compared with air; and air cooling, offering an inherently simpler “plumbing” system as compared with water cooling. Experimental and theoretical information on both these classes of cooling methods si given, and, in addition, experimental information on a new method of cooling is presented, in which a water spray is injected directly on to rotor-blade exterior surfaces. Work at the National Gas Turbine Establishment (N.G.T.E.) has been concentrated mainly on the application of fundamental heat-transfer information to the design of internally air-cooled blading, and measurements on an internal air-cooled cascade blade showed that the blade could be maintained at temperatures several hundred degrees Fahrenheit below gas temperatures, for small cooling-air quantities. Cooling was non-uniform, and difficulties due to consequent thermal stressing may well appear with blades of the type tested. A great deal of improvement over the performance obtained with this first cascade blade is possible, however, and some information on a type of blade which has theoretically a much better performance is given. All cooling methods so far tested show the feature of non-uniformity of cooling, and the present lack of knowledge on the allowable extent of this non-uniformity is an obstacle to the assessment of the gas temperatures at which the various cooling methods will allow a turbine to be run. Probably the air-injection methods, that is, “effusion” cooling, by effusing air through a permeable blade wall, or “film” cooling, by injecting air through slits pointing backwards along the blade surface, will permit the highest possible gas temperatures, since in both these methods there is an insulating action by the injected air. Results of theoretical work at the N.G.T.E. are presented, showing the need for a higher effusion velocity near the blade edges than elsewhere, for effusion-cooled blades. Information is given from tests on a four-stage turbine whose rotor blades were internally water-cooled by the free thermosiphon method, with steam production. Whilst it is likely that effective cooling was obtained in the sense that metal temperatures were kept low, difficulty was experienced with corrosion of the mild steel rotor, a blade eventually corroding through. Certain difficulties with vibration were experienced, but they are thought not to be inherent in the method of cooling. Another method of liquid cooling, that of spraying a small quantity of water on to the rotor blades (of a Whittle jet-engine) from tubes in the nozzle blades, showed unexpectedly great cooling action. The investigation is as yet, however, in a preliminary stage. A discussion is made of the effect of blade aerodynamic design upon the heat which must be extracted from a stage in order to cool it. It appears from the extrapolation of present blade heat-transfer data, that a high axial-velocity (or, more accurately, a high ratio of axial velocity/peripheral velocity) is advantageous and should lead to low heat-extraction quantities. This high velocity ratio will lead to increased volute losses, and there are other factors tending to cause loss of work and efficiency. Some of these are discussed, but mainly not quantitatively owing to lack of systematic knowledge.