Prediction of batch heat transfer coefficients for pseudoplastic fluids ...

Prediction of batch heat transfer coefficients for pseudoplastic fluids ... Prediction of batch heat transfer coefficients for pseudoplastic fluids ...

archives.njit.edu
from archives.njit.edu More from this publisher
04.11.2014 Views

42 tween 0.25 and 0.58. L~~l had reported ~n 18 percent difference in Nusselt nu..l11bers for an impeller in hw different locations a..D.d a fev1 authors had mentioned that hi'>.:her heat transf~r rates Here achieved for certain impeller locations but the effect had not been studied qua.nti ta ti vely. The effect of height variation may have a considerable signific8~ce since changes in the impeller height cause minimal if not neglirible changes in pOHer requirements. The model proposed in Chapter 3 results in a dLmensionless equation which acclJ.rately characterizes the batch heat transfer system as sh01~m by the good fi t achieved. Thus the model itself may be inferred to be a fairly accurate portrayal of the :mechanismof heat transfer in an agitated vessel. To be more specific~ the center core of tIle fluid is in turbulent flm! Hi tll. thorough mixing of the fluj e3 in this region. In addi tion to tlJrbulence in the impeller region, the impel10r produces bulk fluid .flow, largely axial. This results in fluid flov~ alon!:. the cylindrical heat tra..Ylsfer surface in the vertical direction. At the wall surface the flu:td is motionless. There is a velocity gradient in the radial direction. Heat is transferred across the stagnant fluid layer at the Hall by conduction 8Jld is then transferred by diffusion and bulk floH into the turbulent core. TI'le controllinG factor in the rate of heat transfer is thus a stagnant layer at the Hall. An increase in the bulk rImA)" rate through the eye of the impeller (by increasine; im-

LfJ peller diameter, l,~idth, or speed) causes an increase in the velocity of the fluid near the lfJall.. This results in greater momentlL111 transfer in the radial direction Itli th a subsequent decrease in the thickness of t.c"I-J.e stagnant layer .. Comparison of Correlations * 1 .. w « ' p ... . _$ ",' -'1' 'M - -"- The data was correlated by equations representing tvw different approaches, theoretical and semi-empirical. The best equation of each type 1~ill be compared v,Ii th each other later in this chapter, but first their common characteristics v-rill be mentioned.. One characteristic is that they both revert to the cornmonly accepted correlations :for NeHtonian :fh:tids for the case of n equal to unity.. This is not the case lvi th ma..ny correlations, an eXE'uuple being the correlation of Blanchard and Chu (22) for the prediction of batch heat trans:fer coefficients. The accuracy of the cOl'"'relation in reproducing the experimental data is very good, the average error for all :fluids is in the ra..n8e of 9 to 14 percent Hi th the greatest average deviation being :for the most pseudoplastic fluid in the range of 13 to 20 percent. In no insta..nce is the average error o:f the correlation in representing the data greater than the error in the determination of the heat transfer coe:f:ficient. Both the theoretical and semi-empirical correlations are based on a more fundamental fO~Uldation than are the previous correlations :for the prediction of batch heat

LfJ<br />

peller diameter, l,~idth,<br />

or speed) causes an increase in<br />

the velocity <strong>of</strong> the fluid near the lfJall..<br />

This results in<br />

greater momentlL111 <strong>transfer</strong> in the radial direction Itli th a<br />

subsequent decrease in the thickness <strong>of</strong> t.c"I-J.e stagnant layer ..<br />

Comparison <strong>of</strong> Correlations<br />

* 1 .. w « ' p ... . _$ ",' -'1' 'M - -"-<br />

The data was correlated by equations representing tvw<br />

different approaches, theoretical and semi-empirical. The<br />

best equation <strong>of</strong> each type 1~ill<br />

be compared v,Ii th each other<br />

later in this chapter, but first their common characteristics<br />

v-rill be mentioned..<br />

One characteristic is that they both revert<br />

to the cornmonly accepted correlations :<strong>for</strong> NeHtonian<br />

:fh:tids <strong>for</strong> the case <strong>of</strong> n equal to unity..<br />

This is not the<br />

case lvi th ma..ny correlations, an eXE'uuple being the correlation<br />

<strong>of</strong> Blanchard and Chu (22) <strong>for</strong> the prediction <strong>of</strong> <strong>batch</strong> <strong>heat</strong><br />

trans:fer <strong>coefficients</strong>.<br />

The accuracy <strong>of</strong> the cOl'"'relation in reproducing the experimental<br />

data is very good, the average error <strong>for</strong> all :<strong>fluids</strong><br />

is in the ra..n8e <strong>of</strong> 9 to 14 percent Hi th the greatest average<br />

deviation being :<strong>for</strong> the most <strong>pseudoplastic</strong> fluid in the range<br />

<strong>of</strong> 13 to 20 percent. In no insta..nce is the average error<br />

o:f the correlation in representing the data greater than the<br />

error in the determination <strong>of</strong> the <strong>heat</strong> <strong>transfer</strong> coe:f:ficient.<br />

Both the theoretical and semi-empirical correlations<br />

are based on a more fundamental fO~Uldation<br />

than are the<br />

previous correlations :<strong>for</strong> the prediction <strong>of</strong> <strong>batch</strong> <strong>heat</strong>

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

Saved successfully!

Ooh no, something went wrong!