12 QwKViV, Measurement of Air Velocities. 



is concerned), an angular error of even lo" is not of very 

 great importance. Of course for velocity readings the 

 end of the facing gauge must not be displaced by more 

 than about one degree. 



III. Measurement of Air Volumes. 



When the average air velocity over any given section 

 of a pipe is known the volume of air passing that 

 section per second can easily be determined by measuring 

 the pipe. But the average velocity cannot easily be 

 measured ; for even where accuracy in velocity measure- 

 ment can be obtained it is a long and trying process to 

 take measurements all over the section in order to obtain 

 the average. To avoid the necessity for this, it is very 

 usual to measure the velocity at the centre of the pipe 

 and to allow a coefficient of contraction thus : — 



Area of pipe section X velocity at centre X ^ = volume 

 of air per second where t= coefficient of contraction 

 allowed. 



The value of c seems to vary considerably with 

 different authors, although a small change in its value 

 is of enormous importance in measuring, for instance, the 

 efficiency of a fan. The values adopted by different 

 authorities seem to vary between 065 and 09. The 

 latter, I believe, is adopted by Davidson, while the former 

 is used by Innes.* 



In order to test the accurac}' of such figures in pipe 

 circuits, I took a series of tests in a pipe 9 in. in diameter 

 by means of a Brabbee tube and micromanometer. For 

 each set of tests about ten positions of the tube were tried, 

 and velocities ranging from 6 to 9 metres per second were 

 adopted. 



The result is shown in Fig. 4, whilst Table I. gives 

 the ratio average flux per unit area -f central flux per 



* " 1 he Pan," first edition, p. 36. 



