In this tutorial page the main concept and approach for scaling salt viscosity in a MEM workflow will be demonstrated. To this end a single axisymmetric element test will be used. A top surface stress is applied and the response of the salt will be monitored via a history point (see figure below).

 

 

Schematic of model geometry and boundary conditions

 

 

Salt rocks have a visco-plastic rheological behaviour that is associated with creep driven by deviatoric stress. The creep deformation facilitates deviatoric stress relaxation (decrease in stresses) and consequently the stress within salt rocks is often close to isotropic / hydrostatic. The strain rate at which salt creeps is controlled by salt viscosity, which defines the resistance of a given material to flow (creep) and is expressed in units of [ stress · time ]. Thus for a given viscosity and a given load generating deviatoric stress the salt will be able to creep at a given rate. Consequently, the length of the time frame during which salt creeps will determine the amount of strain and hence the amount of deviatoric stress relaxation. It should be noted that to model salt properly within MEM workflows the targets are:

 

1.To achieve a quasi isotropic stress state within salt during initialisation

2.To capture an appropriate/realistic salt response during production which is elastic with creep at production time scales

 

To achieve the above target #2, appropriate salt properties must be defined in the material model (e.g. properties that may be measured in the lab). However, a quasi isotropic stress state during initialisation may not be achieved with these properties due to the relative small time frame. Initialisation over a larger time frame would facilitate salt creep and deviatoric stress relaxation. However this would require either; too many time steps (resulting in excessive CPU time) or a sudden change in time step size between initialisation and production (which is not recommended to avoid stability issues). Hence the recommended approach is to initially define salt viscosity in time units that will facilitate creep ("dka" which corresponds to 1 simulation time unit = 10000 years), perform the initial initialisation stages and then during subsequent initialisation stages gradually scale the time units to achieve production time units (days).

 

To demonstrate the effect of such approach in salt response 2 single element tests are used. In the fist one the time units will be gradually updated as recommended for MEM initialisation whereas in the second one the same time units will be maintained during the simulation. The files are provided in: MEM_002\Creep_Test\Data

 

The five stages for the test have a duration of 1.0 simulation time units and are defined as follows:

 

1.Load application (1·106 Pa)

2.Creep with initial time units (dKa)

3.Creep with time units of cyears

4.Creep with time units of years

5.Creep with time units of days

 

Note that a case in which all stages occur at dka time units is provided to allow comparison of the results.

 

 

Initial Salt Properties

 

 

 

Change of time units

 

 

Results