Why Are My Heat Exchanger Test Results Different?

[almarzooq]

hi ever1,

im final year student doing my final year project at uni. It is basically finding justifications to many abnormal outcomes of the test. I was given 2 standard correlations for Nu and F and these are:

Nu=0.023 Re^-8 *Pr^(1/3)
Fr=0.064/Re^0.2

But when carring out the test I am getting different correlation. Beside, there are two graphs (Fr vs Re) and (Nu vs Re) .. the first graph is straight line with upwards trend and the other graph is straight line acting downwards. The two graphs I am getting out of the test are compeletly different. anyone suggest why is this difference ?

Many thanx

 

[Cyrus]

Thats because heat transfer is more of an art than a science. When you calculate the Nu number if you are even with 30% error it is considered "very good" results.

We did a lab where we found the Nu number to be off by more than 50%.

 

[almarzooq]

that the whole point, i have to do some work to solve this error. I suggested that this is bcoz the mass flow rate which is involved in the calculations is obtained from another experiemental work and it is found by this formula:

m= 4.25/10^3 * SQRT ( pressure drop from inclind mamometer )

so i have to find another way of getting correct value of mass flow rate ( i tried using pitot-tube + orifice) another thing is the instruments used to measure the temperatures ..

Any other suggestions ?

 

[FredGarvin]

What are the flowing media?

Did you do an energy balance between the hot and cold streams to see if they are relatively equal?

If you don't have a properly sized orifice, it could be off on flow rate by over 25%. I would start with the inclined manometer and use Bernoulli to solve for the velocity and thus volumteric flow rate.

 

[almarzooq]

The media is air and it is cooled by water..

I used the energy balance to get the mass flow rate, but I am not convience with the results ..

I am now trying to use squar-edge orifice to calculate the pressure drop and hence the mass flow rate of air, but don't know which one to use ?

The Re range is from 10000 to 30000 , and the pipe diameter is 27 mm. I am looking for an orifice (using BSI ) to fit this case.

 

[eaboujaoudeh]

i would go with the manometer, it is more accurate.
There is an error in the formula u r given, no wonder ur off target. the Nusselts number is DIRECTLY proportional to Reynolds number, so u have to lose the -8, actually the equation is: Nu= 0.023Re^0.8 x Pr^n. where n is 0.4 for heating of the fluid, and 0.3 for cooling. This formula is usually correct for fully turbulent flows with Pr between 0.6 and 100. A correction to this formula would be 1. Nu=0.0214(Re^0.8 - 100)Pr^0.4 for (Pr between 0.5-1.5, and Re between 10^4---5x10^6) or 2. Nu=0.012(Re^0.87 - 280)Pr^0.4 (for Pr bet.1.5-500) and Re bet. (3000---10^6). Remember all those equations go for SMOOTH tubes, so if u want to account for mismatches, try to find a correction for unsmooth tubes. i have in front of me now the correction formula used to account for friction losses, if u want it just let me know, and anyway u can find it in any heat transfer book.

 

[almarzooq]

yes please i would b greatful if u just write it down here. Thanx a lot

 

[eaboujaoudeh]

ok follow this up, i hope ur familiar with the Staunton number: here goes
Eq.1: St*Pr=f/, f is the friction coefficient defined by:
Eq.2: Delta P (pressure difference)=f (L/d)(Rho)(Um^2/2g)
Um is the mean velocity
f=1.325/[ln(epsilon/3.7d)+5.74/Re^0.9]^2 for tubes..
Another approximation u can use is the following:
f*Re/4 = 16000...in this approximation if the tube is under constant heat flux then Nu =4.364, if constant wall temperature Nu=3.657
Check those up, i hope they help