504 



NA TURE 



[February 25, 1909 



required a high-speed prime mover to drive them, and 

 these provided encouragement to the worliers on steam 

 t'urbines ; thus between 1884 and 18S8 we find the practical 

 and successful realisation in altered and correct propor- 

 tions of ideas and suggestions of previous workers, the 

 compound steam turbine in 18S4 applied to the direct 

 driving of dynamos, and the single-stage impulse wheel 

 in 1888, of very high velocity, played upon by the expand- 

 ing steam jet, both types possessing great ratios of ex- 

 pansion. 



All steam turbines now in practical use expand the 

 steam usefully over nearly the whole range from the boiler 

 pressure to the pressure in the condenser, and their designs 

 are based on the principles involved in the construction of 

 their prototypes of 1884 and 1888. 



There is, first, the compound turbine, the character- 

 istic feature of which is the gradual expansion of the 

 steam by small drops of pressure at each turbine of a long 

 series of turbines of gradually increasing volumetric 

 capacity, as in the Parsons, or a somewhat less gradual 

 expansion with greater drops of pressure at each stage, 

 as in the Rateau, Zoelly, and others. 



Then there is the exfansion by the divergent jet in one 

 stage, as in the de Laval, or an expansion in a relatively 

 small number of stages by expanding jets playing upon 

 rows of buckets with intermediate rows of reaction guides, 

 as in the Curtis and Reidler-Stumpf. 



Then there are combinations of the first and second, 

 where the first stage of the expansion is affected by, say, 

 a Curtis element, and the rest by a Parsons, and many 

 other combinations have also been proposed, too numerous 

 to mention here. 



Let us consider these principal examples of the turbine 

 principle more closely. In the compound turbine the veloci- 

 ties of the steam are low ; at each passage through the 

 blades it expands a little, yet it obeys, as regards the 

 velocity of efflux, approximately the laws of flow of fluids ; 

 but the aggregate of the small expansions soon becomes 

 apparent, and has to be taken into account when reckoned 

 over a considerable number of the series of elemental 

 turbines. For instance, if the expansion ratio for a single 

 turbine of the series be as i to 1-03 in volume, a 3 per 

 cent, expansion, then after passing through twenty-three 

 turbines its volume will be doubled, and the vclocitv of 

 flow through the guide blades and moving blades (pre- 

 suming they are of equal area of passage way) will be 

 about 230 feet per second. The velocity of the blades is, 

 generally speakinc, about half the velocity of the steam 

 at issue, and will therefore, in this case (which I have 

 taken as common in marine practice), be about 115 feet 

 per second. 



The difference in velocity of the steam and the blades is 

 smoothed over largely by the curvature of the blade, which 

 somewhat resembles a shallow hook around which the 

 stream lines in the steam arrange themselves with very 

 little shock or eddying in the steam, so that the coefficient 

 of efficiency is high. 



In turbines for driving dynamos and other purposes 

 where higher speeds of revolution are permissible, steam 

 velocities up to 600 feet and blade velocities up to 300 

 feet per second at the exhaust ends are general. 



In turbines of the Rateau, Zoelly, and other types with 

 multiple discs, each disc carries one row of blades onlv, 

 and works in a cell, through the w^alls of which the shaft 

 passes in a steam-packed gland ; nearly the whole drop 

 in pressure takes place at the guide vanes, and very little 

 at the moving vanes, which are of cup form ; the velocities 

 of the steam generally range from 900 feet to iioo feet 

 per second, and the velocities of the blades from 350 feet 

 to 450 feet per second. In turbines, however, "of the 

 de Laval single-wheel and of the Curtis and other types 

 with a relatively small number of pressure stages, higher 

 steam velocities are used, ranging from 4200 feet per 

 second in the single-wheel down to 1500 feet in a seven- 

 stage Curtis turbine. The jets used in the single-stage 

 turbine are of very divergent form, but when the expansion 

 is divided over seven stages very little divergence is 

 necessary. 



In the single-stage turbine, blade velocities so high as 

 1200 feet per second are adopted, the discs being of taper 

 form and of the strongest nickel-steel ; but even this high 

 NO. 2052, VOL. 79] 



velocity is insufficient to obtain a very good coefficient of 

 efficiency from the steam, and when the disc is made large, 

 so as to reduce the immense angular velocity incidental to 

 the high peripheral speed, the skin friction of the disc 

 and the prime cost and weight increase rapidly. 



In the Curtis five-stage the blade velocities are about 

 460 feet per second, and the steam velocity about 2000 feet 

 per second, and by the passage of the steam through two 

 rows of moving and one row of guide blades between them 

 at each wheel the steam is brought nearly to rest before 

 passing on to the next succeeding chamber, and by this 

 sinuous treatment of the steam efficiencies are obtained 

 comparable to those of the compound turbine. 



From the commencement of turbine design in 1884 I 

 have avoided the adoption of high steam velocities on 

 account of their cutting action on metals when any water 

 is present. The cutting has been found to be due, not to 

 the impact of gaseous steam, but to that of minute drops 

 of water entrained by the steam, and hurled by it against 

 the surfaces. The drops, formed like fog, consequent on 

 the e.xpansio/i of saturated steam, are sufficiently large to- 

 cause the erosion. To test the effect in an extreme case, 

 a hard file w'as placed ,opposite to a jet of steam issuing 

 at lOo-lb. pressure into a vacuum of i lb. absolute 

 pressure; in 145 hours it was found to be eroded to the 

 extent of about 1/32 inch, as if it had been sand-blasted. 

 The calculated velocity of the issuing steam in this case 

 is about 3800 feet per second, and the striking fluid pressure 

 of a drop of pure w'ater at this velocity about ninety tons 

 per square inch. Owing, however, to the receding velocity 

 of the blades from the blast in all turbines, tlie erosive 

 effect is much reduced. In multicellular turbines of few 

 stages, though the erosion is slow, yet provision is neces- 

 sary for renewal of blades at intervals. In turbines of 

 many stages it is still slower, and in the compound turbine 

 erosion is, practically speaking, absent, and renewal of 

 blades unnecessary. This absence of the tendency to- 

 erosion in compound turbines permits the use of brass or 

 copper blades, which are found to preserve their polish and 

 are not liable to corrosion or rusting, and preserve their 

 smoothness of surface and the initial economy of the engine 

 unimpaired for many years. 



It is now just fifteen years ago, and exactly ten years 

 from the commencement of work on the compound steam 

 turbine, that the results obtained on land w^ere thought to 

 justify an attempt to apply the turbine principle to the 

 propulsion of vessels. These results lay in the fact that a 

 condensing turbine engine of 200 horse-power, with an ex- 

 pansion ratio of 90 volumes, had been found to have equal' 

 economy to a good compound piston engine, and that, 

 besides, there w-ere within sight reasons to hope for still 

 better results. .-X commencement was made, and by the 

 end of 1897, after three years of hard work and experi- 

 ment, the Turbinia was completed. Her trials were 

 usually made on the measured mile in the North Sea, but 

 occasionally, when the sea was too rough, runs at speeds 

 up to 31 knots were made on the Tyne, where the legal' 

 limit of speed of steamships was 7 knots, and by the 

 magnanimity of the Tyne Improvement Commissioners the 

 completion of the Turbinia was greatly facilitated, though 

 it is fair to say great care was exercised and no harm 

 done to the public. In her the problem of adapting 

 the turbine to the screw propeller was worked out. The 

 result was a compromise between the two. The turbine 

 had to be made short and broad, so as to revolve as slowly 

 as possible, and the screw had to be made with finer 

 pitch and wider blades. The result in propulsive efficiency 

 was found to be good, and the problem satisfactorily solved' 

 for fast vessels of 16 knots and upwards, and it was also 

 seen that the faster the vessel the more favourable would' 

 be the economy of the turbine as compared with the re- 

 ciprocating engine. 



The destrt^vers Fi^cc and Cobra followed. The next 

 step was the application of the turbine to vessels of com- 

 merce. 



Dumbarton was the scene of many conferences. Mr. 

 .\rchibald Denny was deeply interested in the problem, and' 

 so was Captain John Williamson, with the result that the 

 first passenger vessel, the King Edward, was built in tool 

 at Dumbarton to the joint ownership of Captain Johrt 

 Williamson, Messrs. Denny, and the Parsons Marine- 



