Complete AI Training
Sign inGet my AI kit

Your job's AI kit

Get your AI kit

Tell us who you are and what you do. We show you your kit right away and email you the link: skills, prompts, AI agents, MCP servers and courses for your job.

500+ jobs ready, and we make a kit for any other job. No payment needed to look.

Share

AI agent for physicists

Simulation Convergence Check Agent

Converged, verified simulation results

Simulation Convergence Check Agent: what goes in, what the agent does and what you get

What it does

A simulation result means little unless it has converged and the numerical error is known. This agent runs or reads your simulations at several resolutions or time steps, checks that key outputs converge, and estimates numerical error. It also checks conservation of mass, energy or momentum, and compares results with known analytic cases, so a code bug does not hide behind a smooth-looking answer. If results have not converged, it suggests finer settings and launches new runs within your compute budget, then checks again. It writes a short convergence note with plots for each case. You decide which results to use and approve any run that would exceed the budget. Edge case: a run that converges but violates energy conservation is flagged rather than accepted.

How it works

Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.

Start and resultWhat it doesA check on its own workWaits for your OKGoes back and retries
Yes, continueApprovedNo 1 STARTS WHEN Simulation campaign completes 2 USES A TOOL Read outputs at several resolutions 3 DOES Check convergence and conservation 4 CHECKS THE RESULT Are results converged and conserved? If not: launch finer runs within budget. Back to step 2. 5 DOES Compare with analytic cases 6 YOU APPROVE Physicist approves results 7 RESULT Verified results
Read the steps as a list
  1. Simulation campaign completes
  2. Read outputs at several resolutions
  3. Check convergence and conservation
  4. Are results converged and conserved?If not: launch finer runs within budget. Back to step 2.
  5. Compare with analytic cases
  6. Physicist approves resultsThe agent waits here for your OK.
  7. Verified results

How it decides

It accepts results when outputs change less than the tolerance between resolutions and conservation checks pass.

  • Flag conservation violations
  • Stay within compute budget
  • Compare with analytic cases

Make it yours

Every agent is a starting point. You choose these settings for your own situation.

  • Convergence tolerance
  • Compute budget
  • Test cases
  • Output quantities

What keeps you in control

It always asks you first

  • Using results
  • Exceeding compute budget

Hard limits

  • Never exceeds compute budget

It stops when

  • Done: results verified
  • Stop: budget exhausted

Set it up

We guide you through the set-up, step by step

Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.

10 minto set it up in your AI
5 AIsChatGPT, Claude, Copilot, Gemini, Grok
  • One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
  • The agent then walks you through connecting your own data, one source at a time
  • A downloadable copy with the flow chart, the rules and the full guide
Get access to this agent

An example run

What happensA fluid simulation campaign finished with runs at 128, 256 and 512 grid points. Drag converged within 0.4%, but the agent's conservation check found total energy drifting 3% over the run at 512. It went back, reread the outputs and traced the drift to a time step above the stability limit. After a rerun at half the step, energy held within 0.1%. The analytic test case matched. The physicist approved the results.

More agents for physicists