Transient heat transfer 2-D cylindrical

[kseng]

Hi All

I'm embarking on a model of a large physical system.
I have a reasonable concept of the model but need advice on which simulators or programming languages to use.
I've developed a (simple) explicit finite difference algorithm in excel for 'a few metres' and 'a few timesteps'.
I have experience programming in Fortran and am confident I could expand the algorithm into Fortran.

I am unsure if this is going to be an efficient tool or method. Please advise on method (explicit FD) and tool (Fortran) given the model scale.
Some details of the problem below.

The system is a wellbore of nearly 1km depth, subsea.
This determines the need for cylindrical coordinate system.
The boundary conditions are mostly straightforward
The sea temperature fixes the upper plane temperature.
Geothermal gradient determines the initial dT/dz and also fixes T at r=infinity.
Geothermal heat flux will be approximated by fixing temperature (say) 20-50m below the wellbore for all t.
The temperature of the cylinder centre will cool and heat as fluid is injected/withdrawn.
The rock thermal properties are consistent.
The wellbore consists of several material layers (cement, steel pipe, brine, nitrogen).

The aim is to track temperature variation over a period of several months.

I may have to dovetail this with a reservoir thermal simulation.

 

[gtoguy97]

Any advice on tools/methods appreciated.Thanks in advance.For this type of problem, you may want to consider using a finite element or finite volume discretization approach. These methods are more efficient than explicit finite difference algorithms and can handle complicated geometries and material properties. Additionally, these methods provide higher accuracy and produce more stable results. You can use a variety of open-source programs such as FEniCS or OpenFOAM for this purpose. If you are comfortable with Fortran, you can also use the library called FEMM (Finite Element Method Magnetics).