25.01.2015 Views

The American Society of Mechanical Engineers

The American Society of Mechanical Engineers

The American Society of Mechanical Engineers

SHOW MORE
SHOW LESS
  • No tags were found...

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

T urbulence an d E n erg y D issipation<br />

T his paper incorporates a study o f th e origin and d issipation<br />

o f turbulence energy w hich m akes possible a better<br />

understanding o f th e m echanics o f energy losses th a t are<br />

introduced by various flow -disturbing devices such as<br />

expansions, bends, valves, etc. T he im p ortan t param eters<br />

w hich characterize turbulence are th e root-m ean-square<br />

values o f the fluctuatin g velocity com p on en ts, th e len gth<br />

factor proportional to th e size o f th e sm all eddies responsible<br />

for th e dissipation o f energy, and th e len gth<br />

factor proportional to the average size o f th e eddies. <strong>The</strong><br />

effect o f variation o f th ese param eters on th e energy losses<br />

occurring in turbulent flow are discussed, and also th e<br />

change in these param eters in th e decaying turbulence<br />

beyond turbulence-producing devices is in dicated. D ata<br />

are presented show ing th e variation in th e k in etic energy<br />

o f m ean flow and th e turbulence energy in a 15-deg conical<br />

divergence. Visual studies o f th e start o f turbulence at<br />

rounded entrances to sm ooth conduits seem to in dicate<br />

that there is a regular vortex form ation a t th e boundary,<br />

and th e dispersion o f th ese vortexes in to th e m ain fluid<br />

stream gradually establishes norm al tu rb u len t flow.<br />

T<br />

H E flow <strong>of</strong> all real fluids involves energy loss due to the<br />

frictional resistance <strong>of</strong> the fluid. <strong>The</strong> terms “energy loss”<br />

or “energy dissipation” are used in this paper to describe<br />

the eventual changing into heat <strong>of</strong> the energy producing flow.<br />

Energy dissipation in a continuous fluid for either lam inar or<br />

turbulent flow is due to the action <strong>of</strong> the viscosity <strong>of</strong> the fluid,<br />

and such terms as “shock loss” or “im pact loss” do not describe<br />

correctly how energy losses occur in continuous fluids.<br />

Hydrodynamics indicates quite definitely how energy is<br />

dissipated in a fluid in viscous flow. <strong>The</strong> rate a t which energy<br />

is dissipated in viscous flow can be predicted from a knowledge<br />

<strong>of</strong> the flow pattern and the characteristics <strong>of</strong> the fluid. However,<br />

this is not the case for turbulent flow. I t is ordinarily<br />

possible by the use <strong>of</strong> semiempirical formulas to calculate the<br />

total over-all pressure drop and thus the total rate <strong>of</strong> energy loss,<br />

however, we know very little about the mechanics <strong>of</strong> how this<br />

energy loss takes place in the turbulent fluid itself. <strong>The</strong> purpose<br />

<strong>of</strong> this paper is to analyze more thoroughly the mechanics <strong>of</strong><br />

energy dissipation in a turbulent fluid, and to relate such analyses<br />

to various practical hydraulics problems. Liberal use will be<br />

made <strong>of</strong> the fundamental ideas regarding the mechanism <strong>of</strong><br />

turbulence which have been put forth by L. Prandtl, G. I. Taylor,<br />

Th. von K&rm&n, and their co-workers. <strong>The</strong> principles <strong>of</strong> the<br />

statistical theory <strong>of</strong> turbulence will be extended to the practical<br />

problems <strong>of</strong> energy dissipation in liquid flow.<br />

F u n d a m e n t a l C o n c e p t i o n s<br />

In order th at the meaning <strong>of</strong> various terms and expressions<br />

used in this paper may be clearer they will be defined at this<br />

stage.<br />

By A. A. K A LIN SK E,1 IOWA CITY, IOWA<br />

1 Assistant Pr<strong>of</strong>essor <strong>of</strong> Hydraulics, College <strong>of</strong> Engineering, State<br />

University <strong>of</strong> Iowa, and Research Engineer, Iowa Institute <strong>of</strong> Hydraulic<br />

Research.<br />

Contributed by the Hydraulic Division and presented at the<br />

Semi-Annual Meeting, Milwaukee, Wis., June 17-20, 1940, <strong>of</strong> T h e<br />

A m e r i c a n S o c i e t y o f M e c h a n i c a l E n g i n e e r s .<br />

N o t e : Statements and opinions advanced in papers are to be<br />

understood as individual expressions <strong>of</strong> their authors, and not those<br />

<strong>of</strong> the <strong>Society</strong>.<br />

41<br />

When a fluid flows in a straight conduit a t distances far enough<br />

beyond bends or transition sections, so-called normal conditions<br />

become established. In such a case, the am ount and nature<br />

<strong>of</strong> turbulence present is unvarying and the shape <strong>of</strong> the meanvelocity<br />

distribution remains constant. For this condition <strong>of</strong><br />

flow, the rate <strong>of</strong> creation <strong>of</strong> turbulence is in equilibrium with its<br />

rate <strong>of</strong> dissipation. <strong>The</strong> kinetic energy <strong>of</strong> mean flow per pound<br />

<strong>of</strong> fluid flowing is ordinarily designated as Um2/2g, where Um is<br />

the mean velocity in the cross section. <strong>The</strong> total kinetic energy<br />

<strong>of</strong> the mean flow is then equal to QyUm2/2g where Q is the discharge<br />

and y is the specific weight. Because <strong>of</strong> the variation<br />

in mean velocity, all particles <strong>of</strong> fluid passing any section do not<br />

have the same kinetic energy, therefore the foregoing expression<br />

can only be exactly correct for uniform velocity distribution.<br />

For a circular cross section, the correct expression for the total<br />

kinetic energy <strong>of</strong> mean flow is<br />

where p =<br />

V =<br />

U =<br />

unit density <strong>of</strong> fluid<br />

distance from center<br />

radius <strong>of</strong> cross section<br />

mean velocity with respect to time a t point y<br />

In true turbulent flow, the velocity a t a point varies irregularly<br />

with time in direction and magnitude. However, in vortex motion,<br />

as for example in the von Kdrm&n vortex street produced beyond<br />

a body immersed in a flowing fluid, the variation in the velocity<br />

vector with time is quite regular. Whenever the term “turbulence”<br />

is used it will refer to a condition <strong>of</strong> flow where there is no<br />

regularity in the variation <strong>of</strong> the direction and magnitude <strong>of</strong> the<br />

velocity vector, except in the probability sense. I t is convenient<br />

to designate the velocity vector a t any instant by the components<br />

U, V, and W along the x, y, and z axes, respectively.<br />

<strong>The</strong> x axis is in the direction <strong>of</strong> mean flow, the y axis is normal<br />

to the mean flow, and the z axis is the other normal axis.<br />

<strong>The</strong> kinetic energy per pound <strong>of</strong> fluid a t any instant is actually<br />

equal to U2/2g. Separate the fluctuating component U into two<br />

parts, such th a t U = U + u, where u is the fluctuating part.<br />

Obviously, u = 0; (the bar indicating an arithm etic mean).<br />

Thus the mean kinetic energy is in reality equal to (U + u)2/2g<br />

or to ( t /2 + u2)/2g and the p art u2/2g is called the kinetic energy<br />

<strong>of</strong> turbulence. <strong>The</strong>refore, the total kinetic energy due to the<br />

turbulence per pound <strong>of</strong> fluid will be<br />

<strong>The</strong> total kinetic energy <strong>of</strong> turbulence for the fluid flowing in a<br />

circular conduit will then be<br />

E, = ttp U(u2 + v2 + w2)ydy [3]<br />

Another im portant fundam ental idea is the concept <strong>of</strong> shear<br />

in a moving fluid. For viscous flow the shear per unit area at<br />

any point is equal to n dU/dy, or the product <strong>of</strong> the coefficient <strong>of</strong><br />

viscosity and the velocity gradient. <strong>The</strong> mean shear in a tu r­<br />

bulent fluid is equal to<br />

<strong>The</strong> term e is the so-called diffusion coefficient having the dimen-

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!