[CIG-SHORT] nondimensionalization scales in PyLith v1.4.2

[Brad Aagaard]

Faqi Diao wrote:
>  I adjusted the value of the length_scale to 1.0m and 10.0m ,
> but it seems that this change can not increase the velocity of 
> the convergence clearly. How can I get the appreciate value of
> the length scale and the shear modulus? Is there any criterion?
> For example, my mesh scale is 2000*2000*300 km, so what is the 
> appreciate length scale?

The pressure scale used in the nondimensionalization is the shear
modulus. An appropriate value is the average shear modulus in your
domain. The proper choice for the length scale in the
nondimensionalization is not as clear. There are two important length
scales, the size of the features being resolved and the amount of
deformation. Because the amount of deformation (for example, slip) is
closely tied to the Lagrange multipliers, in most cases it should be
used as the length scale. For simple problems there is usually little
difference in the rate of convergence between choosing a length scale of
1 m and 1 km.

Regards,
Brad