# About the gravity and initial stress in OGS6

Dear all,
I’m working on a THM process (initialization is realized by TH process) to investigate the fault stability in 2D influenced by EGS stimulation. I have a question about the gravity and initial stress in OGS6, mainly in THM process:

part of prj code (y is from 0 to -7000m):
<initial_stress>sigma0</initial_stress>
<specific_body_force>0 -9.81</specific_body_force>
...
<parameter>
<mesh>THM_test_initialization</mesh>
<name>sigma0</name>
<type>Function</type>
<expression>26280*y*0.65</expression>
<expression>26280*y</expression>
<expression>0</expression>
<expression>0</expression>
</parameter>

Question in THM process:
What is the function of gravity in this sentence? <specific_body_force>0 -9.81</specific_body_force>
Would the gravity conflict with the initial stress I have set? Because the initial stress I set here is produced by gravity. Does the above prj code mean that I set gravity effect twice?

Dear Rui Feng,

as long as you don’t tell the programme otherwise, you have to specify an initial state that is in equilibrium. This means, yhat whatever you give for the initial stress is interpreted as the initial effective (!) stress and, in conjunction with your definition of bodyforce (gravity in your case) and your boundary conditions has to fullfill \partial_{x_j}(\sigma'_{ij} - p \delta_{ij}) + \varrho g_i = 0. So it’s not redundant but you have to make sure your initial effective stresses, initial pore pressures, initial body forces and initial boundary conditions are compatible. If you don’t want that, there are also options to specify non-equilibrium initial states and calculate from an arbitrary state (usually fine for elastic settings).

1 Like

So, how can I get a compatible and balance state? Does carrying out an initialization work?

Of course, you can perform an equilibration run. However, you have to ensure that the final equilibrated state is the desired state. If you perform an initialization by a TH run first, I assume that you want to start with the final pressures and temperatures of that run, so the equilibration should have a significant effect on the displacement (and thereby stresses) only. It is probably easier to set the initial stresses according to the pore pressure and g_i from the intialization based on the equation above.

Got it, thanks! I will try it these days. By the way, is there any guidace or manual to help me run this initialization? I’m just a newbie.

Thanks for your valued reply. I have known the problem of my model, that is I need to obtain an equilibrium state before THM process. However, I didn’t find any benchmark to teach me how to get a TH equilibrium state before running THM process. So, I want to inquire whether there is any example to teach me how to run this pre-process or called initialization?

I don’t get your problem, sorry. You said you want to start with a TH initialization. But I have no idea what the purpose of this intialization should be, so I can’t tell you how to do that. To plug it into the THM simulation, you can take any state, it doesm’t has to be necessarily in a steady state (it just depends on what you want to do with and this I don’t know) However, if you want a steady state, you just need to simulate as long as nothing happens anymore on the desired time scale. The equilibrium Thomas and I were talking about can’t be reached in a TH simulation as you don’t calculate displacements/stresses there. If you want perform a THM-run to reach equilibrium than you have to play around with dirichlet conditions and look that T and p stay nearly constant, but this might be a non-trivial task depending on you problem, and there is no general guide. Which is why it is much easier to calculate the initial stresses by hand.