Version History
Every criteria change and ranked result set is snapshotted. Restore a prior version as a new run without losing the current design.
Automated Design builds pumping unit, pump, and rod combinations for a well and simulates each one. It returns the top 100 ranked on performance and installed cost.
Manual iterations
~25 designs
Automated Design simulates
the full design space
How It Works
From your equipment options and well data, Automated Design builds pumping unit, pump, and rod combinations, simulates each, and keeps the top 100.
Installed Cost
Every ranked design carries installed cost on your cost basis. On one well, the top five can differ significantly. Pick on performance and price before you commit.
The Stack
RodSim designs one string. Dynamics sweeps how it runs. Automated Design searches the full space and returns the top 100 - same physics, larger search.
In the Workspace
Ranked results are the start. Version history, side-by-side comparison, and Dynamics charts keep every iteration auditable before you accept equipment.
Every criteria change and ranked result set is snapshotted. Restore a prior version as a new run without losing the current design.
Compare pumping unit, pump, rod string, loads, and installed cost across candidates in one grid before you accept a design.
Open any candidate for min, avg, and max loads, rod taper breakdown, and dynacard charts from the same RodSim Dynamics engine.
Platform
Pull wells from Snowflake, Databricks, WellView, or file import. Run Automated Design. Open the ranked list in the workspace, or the same job over REST, MCP, and agents.
Data in
Data in
PetroBench
Automated Design
Search top to bottom, same engine as RodSim
Data out
What production and engineering managers ask before an Automated Design pilot.
Yes. Every candidate runs through the same RodSim engine your engineers already use. The search expands how many designs get simulated. The physics stays the same.
Feasible designs are scored on performance against your constraints, then ranked with installed cost on your cost basis. The Top 100 is the shortlist your team reviews in the workspace.
A well record with geometry, production context, and the equipment catalogs you design against. Pull from Snowflake, Databricks, WellView, or file import, same paths you already use in PetroBench.
Yes. Runs use your org equipment libraries and templates. The search stays inside gear your team is willing to order, not an open catalog of everything.
Your team. Cost basis is configured for the org so every ranked design carries installed cost on numbers you already trust for procurement.
No. Automated Design returns a ranked shortlist. Engineers pick the design to issue. Ordering and field work stay on your existing approval path.
Runs complete in the workspace after the search finishes against your constraints. Exact timing depends on well complexity and how wide you open the equipment set. We walk a real well on the demo.
Yes. Most teams begin with a well they already know, review the Top 100 with installed cost, then expand to a 5-well pilot with success criteria.
Same PetroBench platform. Open ranked results in the workspace, or call the same job over REST and MCP. Vector can work on the same wells under your scopes, writes stay behind engineer signoff.
Yes. Ranked candidates sit next to the installed design on the same well so engineers can see where cost and constraints moved, not a separate spreadsheet export.
Next
Start with a well you already design. We run Automated Design and go through the top 100 with installed cost.