What problem you are doing and what kind of geometry you are using for heat transfer problem? It is possible to answer after knowing these situations. However, in normal single-phase convective flows when conjugate effects are pronounced, that is, when conduction heat transfer in the geometry is significant as compared to convection, then this kind of wall temperature profiles are observed.

I understand the formation of bubbles/froth from your question.  If that is true, the formation is associated with higher energy.  The energy is taken from the liquid which can also be referred as homogeneous nucleation.  Since the length of the tube or the flow velocities are not presented, the absorption of energy for vapor formation might be reflected in the decrease in wall temperature.

It is not clear how the wall is heated. Does not look like it had a uniform temperature at the beginning since it increased initially and then decreased. However latent heat of vaporization of the fluid may possibly be a good reason.

Balkrishna Mehta@ we have plate heat exchanger with constant heat flux along the channel

K. V Sharma@ I dont know if it is the bubbles formation, the channel dimension is (0,5*2,5*5) cm, and we have same profil for several velocities.

Subhasish Mitra@ the heat flux is constante along the channel on both sides.

check the heat flux.  Is the value of the order of 10^6?   Check the pressure drop.  Is the value higher than estimated with single phase fluid?  These calculations can give you insights into the problem.

Thank you very much 

A question. What is the transversal dimension on this channel?

the channel dimension is 5*2,5*0,5 cm

As I have thought!

.

You have a problem of development of the thermal boundary layer.

If you calculate how the ratio of the smaller (0.5cm) with measurement points you will have to end a relationship channel (L / D = 1), in this case the effects of the input can reach the exit.

.

In your case you should take into account more variables than simply the condition of diffusion (molecular - if laminar boundary layer - molecular + turbulent - if you reach the turbulent boundary layer) and the input condition.

.

If you are working with a high Reynolds number, that allows a long reach a turbulent flow conduit, you can put at the entrance of the channel something precipitate the formation of turbulent boundary layer. If you are working with a low Reynolds number, you can calculate analytically (or numerical method) the development of the layer.

.

Important in both cases must take into account the thermodynamic properties of the fluid, to be able to calculate the Prandtl number and Nusselt number, which are required as appropriate.

.

If you are working near the transition will be is unfortunate!

In your work "Experimental study of the upward forced convection between two heated plates." When you compare with Li Xiao (2010) there is no reference to the conditions of entry (eg relationship between the diameter of the conduit that feeds the canal and canal diameter ). In addition there is another problem, the Reynolds uses is 1200 and Xiao Li (2010) uses 1500 is a slight variation, but you are within the transition zone!

.

You write in your text "Qualitatively, the curves are similar. The quantitative difference is due to the fluids Essentially eldery properties are the same." But forget that the conditions of ingress may be different.

.

But when you compared with Yasar (2004) curves seem more similar. As the experience of Li Xiao (2010) the conduit is long enough to achieve a developed and calculating Yasar boundary layer numerical model is relatively short, perhaps there is no error between the three works, the problem is simply the conditions of entry .

Thank you very much Mr Rogerio Maestri

about entry effects you're right, !!

For the Reynolds number it is 1427 for me and 1500 for to Xiao Li (2010)

About the flow regime, from the pressure measurement, we have turbulent flow,

my question was about wall temperature versus position, and how can we interprete the descreasing of wall temperature at the outlet of channel (knowing that it is the same profile for severale namber of reynolds and heat flux values)?

Ahmed

.

The channel is SHORT, just the influence of input and OUTPUT plays in the velocity profiles. If your realizing it the instructions for measuring flow, for example, equipment is always required minimum length of 10 diameters at the entrance and at the exit 5.

.

The viewing channel has a length of 50mm and a width of 5 mm, or roughly> 10 diameters (25mm whereas the other dimension), in summary does not develop the boundary layer.

.

If you do not develop the boundary layer is not possible to compare without adjustment data with other studies, the results are restricted to similar geometries.

Without knowing how the internal development of the flow, it is extremely difficult to try to imagine what happens in terms of heat transfer, then you can restrict your analysis to the central area, where at least the influence of input and output is lower.

.

Another thing to compare with similar reduction assays and the expansion of the area at the beginning and end of the channel is important, the figures 1 and 2 of the work can not know what conditions are like approach.

.

Question 1: Is it (a) a foil heater (b) a solid wall heated by ohmic heat generation, or

(c) a wall heated by external electrical coils?

As has been pointed out in other posts, when constant heat flux is supplied but the heat extraction (latent heat vis-à-vis sensible heat) is larger than the supplied heat, the temperature will drop.

Therefore,

Question 2: What happens when you increase the wall heat flux?

Question 3: What happens when you vary the flow rate?

I see that the drop in wall temperature coincides with the fluid reaching saturation temperature.

The wall temperature drop could also be synonymous with an increase in local heat transfer coefficient, the reverse of what happens at dryout.

@Vijay Raghavan, the  wall is heated by external electrical coils, and the curves are simular for several values of heat flux and flow rate.

Then it must be due to an increase in the local boiling heat transfer coefficient.

Thank you very much

In the case of pool boiling, Rohsenow (1954) has reported a drop and a raise in the wall temperature.  

In flow boiling also similar nature of the wall temperature profile is obtained.

It is due to the raise in local heat transfer coeff (due to bubble departure).  Later due to higher heat flux, the wall temperature shall have an increasing trend.  

Dr.S.Ravindran

thank you very much Dr Ravindran.