5.1. Inversion parameters¶
For each of the experiments conducted, most of the inversion parameters were held constant for consistency. Some of these parameters are include:
- octree_mesh
159,772 cells on an underlying base mesh which is 128 x 128 x 64 cells and has a smallest cell size of 30 x 30 x 15 m.
- reference model
A uniforom halfspace with a conductivity of 0.001 S/m and a chargeability of 0.0001.
- initial model
Same as the reference model
- active cells
No topography is used and all model cells are active in this example. So the topography and model active cells are both set to ALL_ACTIVE.
- cell weights
No cells weights are used (NO_WEIGHTS)
- interface weights
No interface weights are used (NO_FACE_WEIGHTS)
- beta
Beta values are set to DEFAULT
- alpha’s
Alpha coefficients are: \(\alpha_s = 1E-5, \alpha_x = \alpha_y = \alpha_z = 1\). This corresponds to a length scale of approximately 316 m in all directions.
- chifact
1
- tol_nl
1E-2
- mindm
1E-3
- iter_per_beta
2
- tol_ipcg
1E-2
- max_iter_ipcg
15
- reference model change
The reference model is not changed at each beta step (NOT_CHANGE_MREF)
- smoothing
The SMOOTH_MOD_DIF option was used in all inversions.
- bounds
For the DC data, NO_BOUNDS were applied. For the IP data, a positivity constraint was enforced using constant bounds (BOUNDS_CONST 0 1).