Hi,
In the attached 3D model model I am investigating fundamental thermodynamic physics (usually I do that to understand/check how COMSOL handles the physics before I gonna apply it to my complicated model).
The model briefly:
- I have a cylinder filled with some liquid (all thermophysical properties are set correctly w.r.t the material data sheet).
- In the middle of this cylinder a heat source dissipating total heat of 160 j/s (assume in reality, cold electronics). Here that is represented by an evacuated sphere (i.e. the sphere is cut from the cylinder).
- The outer surfaces (boundaries) of the cylinder is thermally insulated, hence all dissipated heat is maintained, only, in the liquid.
- The simulation run in Time dependent mode for 3600 s in 720 s time step.
- Just for simplicity (you can extract these parameters also by yourself), these are the thermodynamic properties of the liquid at initial Temp. (169 K):
> liquid density = 1772 kg/m^3 (is Temp. dependent, since it is not huge time change so you can keep it const.)
> liquid Cp = 0.90246 kJ/kg.k
> liquid vol. = 6.197 m^3.
Now, after one hour, I am integrating the thermal energy over the volume of the liquid (the domain) (this can be exported by COMSOL Derived values node --> total internal energy). This gives the tot. int. E in J/kg unit. Then I calculate the change in volume energy after one ĥour (tot. int. E at time 0 - tot. int. E at 3600 s) multiply this by the mass of the liquid (density*vol) in kg then dividing that by the time (i.e. 3600 s) I expected to get back the 160 W (J/s) of the heat source, but this not happening. I get 20-30% more (over all time steps)??
I then checked my calculations a bit simpler, I took the Average volume Temp. of the liquid (Tf after 1 hr and at each time step). Hence DeltaT = Tf - Ti (Ti = 169 K). Then using the relation:
DeltaE = Cp.M.DeltaT
Then DeltaE (J)/total time (s)
This should give me the input power, i.e. 160 W, but that gave me the same numbers as the previous calculations.
Excuse me for the details (if you need more just let me know), but since this is basic/fundamental physics problem so I thought I have to explain it in detailed. I also could miss fundamental issue here.
I will be very glad if you can take part of your time and check what is the problem or what mistake I do here.
Thanks a lot,
Tamer
*P.S.: This model takes a bit less than an hour on 16-core machine and it does not need too much memory (~ 6-7 GB RAM).
In the attached 3D model model I am investigating fundamental thermodynamic physics (usually I do that to understand/check how COMSOL handles the physics before I gonna apply it to my complicated model).
The model briefly:
- I have a cylinder filled with some liquid (all thermophysical properties are set correctly w.r.t the material data sheet).
- In the middle of this cylinder a heat source dissipating total heat of 160 j/s (assume in reality, cold electronics). Here that is represented by an evacuated sphere (i.e. the sphere is cut from the cylinder).
- The outer surfaces (boundaries) of the cylinder is thermally insulated, hence all dissipated heat is maintained, only, in the liquid.
- The simulation run in Time dependent mode for 3600 s in 720 s time step.
- Just for simplicity (you can extract these parameters also by yourself), these are the thermodynamic properties of the liquid at initial Temp. (169 K):
> liquid density = 1772 kg/m^3 (is Temp. dependent, since it is not huge time change so you can keep it const.)
> liquid Cp = 0.90246 kJ/kg.k
> liquid vol. = 6.197 m^3.
Now, after one hour, I am integrating the thermal energy over the volume of the liquid (the domain) (this can be exported by COMSOL Derived values node --> total internal energy). This gives the tot. int. E in J/kg unit. Then I calculate the change in volume energy after one ĥour (tot. int. E at time 0 - tot. int. E at 3600 s) multiply this by the mass of the liquid (density*vol) in kg then dividing that by the time (i.e. 3600 s) I expected to get back the 160 W (J/s) of the heat source, but this not happening. I get 20-30% more (over all time steps)??
I then checked my calculations a bit simpler, I took the Average volume Temp. of the liquid (Tf after 1 hr and at each time step). Hence DeltaT = Tf - Ti (Ti = 169 K). Then using the relation:
DeltaE = Cp.M.DeltaT
Then DeltaE (J)/total time (s)
This should give me the input power, i.e. 160 W, but that gave me the same numbers as the previous calculations.
Excuse me for the details (if you need more just let me know), but since this is basic/fundamental physics problem so I thought I have to explain it in detailed. I also could miss fundamental issue here.
I will be very glad if you can take part of your time and check what is the problem or what mistake I do here.
Thanks a lot,
Tamer
*P.S.: This model takes a bit less than an hour on 16-core machine and it does not need too much memory (~ 6-7 GB RAM).