Modeling workflowsyou can rerun.
A modeling instance is a saved, reusable set of parameters for one estimation or simulation method. It reads your samples, writes results into your block model, and runs on the Annapurna compute engine.
Cu_SGS_2026 · Resource modeling
Completed (00:01:07)| Parameter | run_UE_1 | run_UE_2 | run_UE_2_LVA |
|---|---|---|---|
| method | sgs | sgs | sgs |
| cd_filename | Comps_2026 | Comps_2026 | Comps_2026 |
| cd_attribute | NScore_Cut_UE_1 | NScore_Cut_UE_2 | NScore_Cut_UE_2 |
| cd_selection | ue_cut == "1" | ue_cut == "2" | ue_cut == "2" |
| bm_filename | Base_PPMT | Base_PPMT | Base_PPMT |
| bm_selection | ue_cut == "1" | ue_cut == "2" | ue_cut == "2" |
| bm_realizations | Nscore_Cut_SGS_Iter8 | Nscore_Cut_SGS_Iter8 | Nscore_Cut_SGS_Iter8 |
| bm_etype | · | · | · |
| bm_variance | · | · | · |
| bm_cv | · | · | · |
| transformation_settings | custom | custom | custom |
| transformation_boolean | no | no | no |
| decl_weights | · | · | · |
| variogram_models_settings | automatic | automatic | automatic |
| nugget_effect | 0.0241 | 0.0853 | 0.0853 |
| number_structures | 2 | 2 | 2 |
| st_1_it | Exponential | Exponential | Exponential |
| st_1_cc | 0.2169 | 0.3933 | 0.3933 |
| st_1_ang1 | 69.0 | 75.0 | 75.0 |
| st_1_ang2 | 68.0 | -59.0 | -59.0 |
| st_1_ang3 | 168.0 | 116.0 | 116.0 |
| st_1_a_hmax | 30 | 70 | 70 |
| st_1_a_hmin | 30 | 50 | 50 |
| st_1_a_vert | 100 | 40 | 40 |
| st_2_it | Exponential | Exponential | Exponential |
| st_2_cc | 0.7563 | 0.5118 | 0.5118 |
| st_2_ang1 | 69.0 | 75.0 | 75.0 |
| st_2_ang2 | 68.0 | -59.0 | -59.0 |
| st_2_ang3 | 168.0 | 116.0 | 116.0 |
| st_2_a_hmax | 50 | 400 | 400 |
| st_2_a_hmin | 100.0 | 100.0 | 100.0 |
| st_2_a_vert | 200 | 190 | 190 |
| st_3_it | Exponential | Exponential | Exponential |
| st_3_cc | 0.0 | 0.0 | 0.0 |
| st_3_ang1 | 69.0 | 75.0 | 75.0 |
| st_3_ang2 | 68.0 | -59.0 | -59.0 |
| st_3_ang3 | 168.0 | 116.0 | 116.0 |
| st_3_a_hmax | 0.0 | 0.0 | 0.0 |
| st_3_a_hmin | 0.0 | 0.0 | 0.0 |
| st_3_a_vert | 0.0 | 0.0 | 0.0 |
| search_plans_settings | custom | custom | custom |
| local_anisotropy | no | no | yes |
| a_max | 100.0 | 100.0 | 100.0 |
| a_semi | 100.0 | 100.0 | 100.0 |
| a_minor | 100.0 | 100.0 | 100.0 |
| cd_min | 5 | 5 | 5 |
| cd_max_simulated | 16 | 16 | 16 |
| cd_max_original | 16 | 16 | 16 |
| global_azimuth | 0.0 | 0.0 | 0.0 |
| global_dip | 0.0 | 0.0 | 0.0 |
| global_tilt | 0.0 | 0.0 | 0.0 |
| local_azimuth | · | · | lva_azimuth |
| local_dip | · | · | lva_dip |
| local_tilt | · | · | lva_tilt |
| additional_settings | custom | custom | custom |
| discretization_x | 1 | 1 | 1 |
| discretization_y | 1 | 1 | 1 |
| discretization_z | 1 | 1 | 1 |
| num_reas | 50 | 50 | 50 |
| data_to_node | no | no | no |
| random_seed | 998877 | 123423 | 123425 |
Illustration: one SGS instance with three runs: units 1 and 2, and unit 2 again with locally varying anisotropy. Scroll the table to see all 61 parameters.
From drillholes to resource model,in three steps.
Build the geology, define the domains, then estimate or simulate grades. Each step builds on the one before, in the same project.
Geological modeling
Turn drillhole logs into closed 3D solids for every rock unit.
- Choose your drillholes, the model area and the level of detail
- Order units from youngest to oldest: deposits, erosions, intrusions and veins
- Guide the model with lines, surfaces or points where drilling is sparse
- Save the solids, review them in 3D and export them
Methods: Implicit modeling, chronology rules, interpretation guides
Domain modeling
Give every block in your model the domain it most likely belongs to.
- Build domains straight from your drillhole samples
- Keep sharp or gradual contacts, as the geology requires
- Create several equally likely versions to see where boundaries are uncertain
Methods: Nearest neighbour, indicator kriging, plurigaussian simulation
Resource modeling
Estimate or simulate grades inside each domain.
- A best-estimate model for reporting and mine planning
- Simulations that show the range of possible grades and the risk behind them
- Save every setup and rerun it when new data arrives
Methods: Nearest neighbour, inverse distance, kriging, sequential Gaussian simulation
Prefer a hand with any step? Our geostatistics experts can build or review it with you. Work with our specialists
From a one-off model to a repeatable process
Monthly updates, scenario testing and audits all depend on being able to run the same workflow again, exactly.
Parameters as a table
One row per parameter, one column per run. Compare variants side by side.
Run in the cloud
Runs execute on the compute engine. Close the browser while they work.
Execution log
Follow every step, see run times and the reason for any failure.
Clone and reuse
Clone an instance, download it as CSV, reuse it in another project.
Repeat every update
Next month, load the new data and run the same workflow again.
Drive it from Python
Load, edit and save instances from Python, inside the platform or with the SDK.
Need a workflow built for your deposit?
Our consulting team sets up and calibrates workflows on your data, from monthly models to stockpile updates, and hands them over to your team.
Your models,anywhere, anytime.
Tell us about your project, or register to get started.