combined heat and mass transfer

[masoodmoghaddam]

I think we should this consider this item in order to implement coupled heat and mass transfer: 1-we want two network: first network is that one we create for paper reproduction (i.e. 2D circle with hierarchical porosity ...) second network what exist in Cu conduction example

2- each network has its own geometry, phase, algorithm and etc. 3- solid reads the reaction rate from gas and calculates the heat source term 4- gas reads the temperature from solid

[samanjalilian]

how the networks are placed relative to each other?

[masoodmoghaddam]

how the networks are placed relative to each other?

Once this models is defined on physics, you then specify to the algorithm object to add this source term using the 'add_source_term' method. With this done, you then just run the di»usion simulation as normal, but the algorithm will automatically incorporate the kinetics, including iterating non-linear terms. For your purposes, you'll eventually want to include heat generation and transport but this is where things will be tough, since multiphysics is still largely by hand in openpnm. Given the reaction rates from the previous step, you'll calclaute the heat generation, then solve a heat transport problem to get the temperature distribution, then recalculate all kinetic values, and repeat until converged.

[masoodmoghaddam]

how the networks are placed relative to each other?

I write Jeff answer above. I think placing of two network is not the matter. You can consider a disk over the another or you can consider a disk in you pc and another in my pc. You simulate the diffusion-reaction in your pc and you reaction produce heats should be removed by solid. So you consider heat sinks in pores to removing spatial value of heat in each pore. I receive your sink heat as source term in my Fourier conduction simulation.

So if you agree with me, we should determine the value of heat sink in your pores. In addition we should declare that were I have not pore and throat and what is your pore counterpart in my network.

[samanjalilian]

i found this page which explains how to generate Dual networks and can be useful in our work : https://github.com/PMEAL/OpenPNM-Examples/blob/master/Topology/generate_dual_cubic_lattice.ipynb

[masoodmoghaddam]

In order to implement heat transfer simulation we can: thermal conduction in gas phase is neglect able. So we should consider heat conduction just in solid phase. Solid phase is not porous! How we can simulate heat conduction in non porous media with openpnm?

[samanjalilian]

In thermal conduction simulation in the plate, we have a non-porous media, But openPNM is a porous media simulator. If we consider pores and throats, it means that we have porosity(not acceptable!) For considering non-porous media we have two way: 1-Throat length considered equal to zero (for this propose a geometry modification should be done with generic geometry) 2- Throat length is calculated from stickandball geometry but thermal conductance of throats considered infinite. Note: Normal Stickandball geometry with normal physic will cause error because it is for a porous plate.

generic.txt stick.txt stick(error).txt

[samanjalilian]

I wrote the heat conduction and mass diffusion code, the procedure of the code seems ok but it has lots of errors which prevent the code to give the correct answer, so we have to work a lot on the code,

1. it seems that the throat endpoints create an error while generating geometrys
2. we have to give the correct heat source term to the code
3. we have to give the algorithms in correct way

mass-Heat.txt

[masoodmoghaddam]

Good start! Go ahead ...

[samanjalilian]

two codes are attached below, one only for Heat transfer and the other one for coupled mass and heat transfer, the results for codes are different! we have to find out why? Heat.txt mass-Heat.txt

[samanjalilian]

here is a code which includes heat and mass transfer and reaches the solution by doing some iterations because we know that mass transfer and heat transfer are coupled together. first we consider a initial temperature ( 298 k ) then the Knudsen diffusion coefficient is calculated based on the initial temperature then the Fickian diffusion algorithm will calculate the total rate and the rate in each pore as well. then by having the rate of mass generation in each pore and the enthalpy of reaction we can add heat source term ( rate of heat gen=rate of mass gen * enthalpy) then the Conduction algorithm will calculate the new values of temperature which will be used in first step till convergence. there is a problems in this work because we supposed that the void and the solid are in a same location . so we should not forget dual networks!!

Iterative solving heat and mass.txt