April i, 1909] 



NATURE 



145 



foundation. They all agree in attributing to the greater 

 or smaller temperature-differences between gas and wall 

 which occur in practice the higher or lower rates of heat 

 transference which are met with, and in ignoring any 

 effect upon that rate which may be produced by a varia- 

 tion of the speed of gas flow. 



In 1S74 Prof. Osborne Reynolds brought before the 

 Literary and Philosophical Society of Manchester a paper 

 entitled' " The E.xtent and Action of the Heating Surface 

 of Steam Boilers." In this paper, starting with the laws 

 then recently discovered of the internal diffusion of fluids, 

 he endeavoured to deduce from theoretical considerations 

 the laws for the transmission of heat. His formula ex- 

 pressing this law is 



H = (A+Bpu)(T-e), 

 where .-V and B are constants, p, «, and T are the density, 

 speed and temperature of the fluid, and B is the tempera- 

 ture of the wall. For small values of h this becomes 

 Newton's law of cooling; for large values the A-term is 

 less important, and the formula becomes one which is 

 applicable to steam boilers. 



No further investigation of the subject was made until 

 1S97, when Dr. T. E. Stanton made a series of experi- 

 cnents to test the truth of the views advanced. He found 

 (hat the amount of heat transferred when water forms 

 both the heat-conveying and heat-receiving medium is 

 nearly proportional to the speed of flow, and that Osborne 

 Reynolds's views were abundantly confirmed. 



In 1899 Prof- Perry, in his book on the " Steam 

 Engine," wrote a chapter on " How Fluids give up Heat 

 and Momentum "; and, in discussing the efficiency of steam 

 boilers, he finally remarked : — " It seems to me that 

 when a good scrubbing action is established on both sides 

 of the metal there ought to be at least ten times, and may 

 be more than too times, as rapid an evaporation per 

 square foot of heating surface as has yet been obtained 

 in any boiler." 



There is no record that up to the present time any 

 British or Continental engineer has paid serious attention 

 to these pregnant words, or has realised the immense 

 possibilities which lie behind Prof. Osborne Reynolds's 

 statements, should their truth be experimentally demon- 

 strated. 



The author attached so much importance to the matter 

 that, in i8g8, after reading Dr. Stanton's paper, he con- 

 structed an apparatus in his laboratory at the McGill 

 University for the purpose of further study. His removal 

 to Manchester in 1899, and his subsequent occupation by 

 other work, had, however, prevented his further taking 

 up the question until 1905. 



Since that time three series of experiments have been 

 made by himself, and one series by his pupil, Mr. H. P. 

 Jordan, on the subject. 



The apparatus used was usually of a fairly simple type, 

 consisting of two long concentric tubes through which 

 (a) warm compressed air and cold water, (b) superheated 

 steam and cold compressed air, (c) superheated steam and 

 cold water, and (d) products of combustion of coal and 

 boiling water, were passed in opposite directions through 

 the two pipes at various rates of speed. 



An anaIysis_of all these results, together with those of 

 Pptiet and GeoPfroy, and Henrv and Mari^ upon locomotive 

 boilers, have led to the formula : — ■ 



■==[4*"|(-,^)'.".](-) 



for the heat flow Q in B.Th.U. per sq. ft. per hour when : 



T = temperature of ga? (°F.), 



e = „ wall (°F.), 



Pi = density of gas (lbs. /cu ft.), 



/!■] ■= speed nf gas flow (feet/sec), 



ip - 4(T + fl) = mean film-temperature, 



w;, = hydraulic mean depth of gas-flue (inches). 

 According to this formula, the rate of heat transfer for 

 a given temperalurc-differeiiec between gas and wall 

 depends upon (a) the mass-flow of the gas, lb. per sec. per 

 sq. ft. of flue-section ; (b) the average temperature of gas 

 and wall : and (r) the smallness of bore of the tube or 

 width of channel conveying the gas. 



The usual rate of heat transfer in steam boilers is (from 

 XO. 2057, "^'OL. 80] 



3 to 10^ about 5 B.Th.U. per sq. ft. per hour per degree 

 difference of gas and water (i.e. wall) temperature. In 

 the author's experimental apparatus he succeeded in trans- 

 mitting more than 300 B.Th.U. per sq. ft. per hour per 

 degree difference, although the air was only about 30° F. 

 hotter than the water. 



In an experimental boiler he was able also to produce 

 evaporations at the rate of from 30 lb. to 50 lb. of steam 

 (as from and at 212° F.) per sq. ft. of indirect heating 

 surface per hour when the gas temperature was at about 

 1500° F. and the water at 300° F. In both cases this 

 high rate of heat transference was principally due to the 

 very high gas velocities employed. These varied from 300 

 to 550 feet per second, whereas in ordinary boiler practice 

 it never exceeds 150, and is usually from 10 to 30 feet per 

 second. 



The way in which the new experimental facts should 

 influence practice in steam generation may be stated as 

 follows . — 



In the first place, since the amount of heat that can be 

 transmitted for a given temperature difference is almost 

 directly proportional to the speed of the gases, a reduc- 

 tion of area of the heating surface in steam boilers-^to 

 one-half, one-quarter, or even one-tenth of what is now. 

 usual — can be made without the chimney temperatures 

 being raised or the efficiency lowered to any material 

 extent. Or, otherwise, if the surface be kept the same, 

 but the cross-sectional area through the flues be reduced, 

 in order to obtain the necessary high gas speed, a very 

 much lower chimney temperature and correspondingfy 

 higher efficiency can be secured than is now available. 



.Accordingly, drafts of 10 or 20 inches of water gauge, 

 induced by fans, should always be employed for really 

 economical working. 



In the case of boilers of usual construction, in which 

 the gases pass through flues or tubes and leave the boiler 

 at a point where the temperature on the other side of the 

 heating surface is that corresponding to the steam pressure, 

 a limit will soon be reached as to the amount of draft- 

 suction which can be employed, and this from one or 

 other of two causes : — ■ 



(a) Whilst the fire need not be forced, even when these 

 high drafts are used, to a greater extent than that now 

 usual, the high speed of entry of the glowing gas at the 

 furnace end of the tubes will cause leakage, and some other 

 construction for such purposes than that now generally 

 emploved must be sought for. 



(b) The fall in temperature of the chimney gases due 

 to the small flue-section and accompanying high speed 

 will, as intimated above, certainly provide an additional 

 amount of evaporation which can be drawn upon to cover 

 the extra power required for the fans, but the margin 

 between present chimney temperatures and the lowest 

 which are possible in ordinary designs of boiler under the 

 above conditions is not very great when the steam tempera- 

 tures are from 350° F. to 400° F. A limit will therefore 

 soon be reached beyond which it will not pay to pass. 

 The author, after a careful investigation of costs and 

 running expenses, has good reason to believe that this 

 limit of draft-pressure is, even for the ordinary type of 

 boiler, much higher than that now in common use. 



In the second place, since the author's experiments have 

 shown that with a counter-current flow of gas and water 

 it is possible to lower the gas temperature (at such high 

 speeds as he used) to within 20° F. of that of the enter- 

 ing feed, we have here the evidence that chimney tempera- 

 tures of 100° F. to 150° F. can be reached and maintained 

 provided only high gas and water speeds are resorted to 

 and the boiler is designed on the economisor principle, with 

 strict attention to counter-current methods of flow. 



Now such a low chimney temperature as this corresponds 

 to a transmission efficiency of more than qf, per cent. ! The 

 margin of additional evaporation available for the supply 

 of the fan-power is, accordingly, so much greater than is 

 required that not only can higher economies be obtained 

 on this system than have hitherto been thought possible, 

 but they can be effected with boilers having smaller areas 

 of heating surface, smaller total volumes, smaller floor 

 areas, lighter weights, and lower first costs than those we 

 ordinarily employ. 



Finally, since the enhanced evaporative efficiency of the 



