[Allankey]

Hi, I am trying to model a heated aluminium plate that is cooled by a water pipe on the back.

The fluid I am using is water and the mass flow rate is 0.009kg/s. My pipe diameter is 8mm. If I manually calculate the Reynolds number I get a value of 1915. The flow in the pipe is fully developed so with a Nusselt number of 4.36 I am expecting a heat transfer coefficient of around 350W/m2K however I am getting much higher values of around 1000W/m2 and more depending on how much contact the pipe has with the plate. Only the top portion of the pipe, around 32% of the circumference, is in contact with the plate.

For the top surface of the plate I have specified a Heat Flux in, as well as a temperature driven loss coefficient out under the energy sources tab. This is to ensure that there is some temperature variation across the plate.

To calculate the heat transfer coefficient I am taking the difference in average temperature of the plate and the fluid. Q is the energy transferred to the water (outlet-inlet*cp*massflow) and I then divide by the contact area to convert to W/m2K.

I am wondering what is causing the high values of heat transfer coefficient? The solution converges fine for all residuals including heat transfer and energy balance. I have added some contact resistance at the interface to lower the values of heat transfer coefficient but I don't think this is the best way to go.

Is it possible that the flow is somehow been overridden and more heat is being driven into the fluid than would naturally occur? Or are the results I am getting normal?

Any help or advice would be much appreciated, I hope I have been clear but if there is anything else let me know.


Thanks

[ghorrocks]

Before you model the fluid solid system as you have done I would do a benchmark simulation on flow in a short section of pipe only. Make the outside of the pipe a constant temperature and the fluid enters the pipe at a lower temperature; and the flow rate is the same as your fluid/solid model. On this simple model I would do a careful accuracy study and work out what convergence, mesh size and time step is required for accurate heat transfer model.

Once you have established what you require for accurate heat transfer modelling in the simple model you can have some confidence that using the same settings int he fluid/solid model will give an accurate answer.

At the moment you do not know if your issue is due to inaccurate simulation or if it is real.

[Allankey]

Thanks, I will try this.