Verification Manual Input Data 121 (VM121) contains a fully developed laminar flow problem with constant heat flux. On the basis of theory in that problem, Nusselt number will be 4.36 and corresponding heat transfer co-efficient will be 7344.4 w/m2-k. This theoretical result is discussed in the "Temperature Solution" section of VM121.
But heat transfer coefficient determined by ANSYS (PLNSOL,HFLM) ranges from ~100 - ~600 and averages about 250 W/m^2C, which of course is far lower.
Why is this so? Is the FLOTRAN-calulated result incorrect?
From the equations in VM121 write-up:
H = Nu*k/D = (4.36)(8.54 W/mC)/(0.005m) = 7450 W/m^2C
The FLOTRAN-calculated film coefficient ranged from ~100 - ~600 and averages about 250 W/m^2C, which of course is far lower.
The reason for the discrepancy is discussed in the attached presentation. To summarize, the theoretical film coefficient is defined differently than that from FLOTRAN.
q` = HtheorydA (Twall - Tmix)
q` = HANSYSdA(Twall ` Tbulk)
q` = heat flux = 5000 W/m2
dA = surface area m2
Twall = fluid temperature at wall (C)
Tmix = average fluid temperature over flow cross section (C)
Tbulk = arbitrary temperature prescribed by user using FLDA,TEMP,BULK,value
Both definitions give same heat flow but the corresponding values of Hfilm are VERY different. Note that defining different values of Tbulk in FLOTRAN with [FLDA,TEMP,BULK,value] will result in different calculated H values.
Attached is a presentation and modified version of the input file for VM121 which explains the reason for the large difference between film coefficients from theory and FLOTRAN. They are both valid and legitimate, since they are defined differently.