Slow or Stuck Simulation
Fix slow or stuck RodSim simulations in PetroBench. Understand the progress stages, tell a slow run from a stuck one, and speed things up without losing accuracy.
Your simulation has been running for a while and you're wondering if something's wrong. Most of the time it's still working. This page explains what's normal, how to tell a slow run from a genuinely stuck one, and what to do about it.
What's Normal
A Static run usually completes in 15 to 30 seconds. The progress modal steps through these stages, in order:
| Stage | What's happening |
|---|---|
| 1. Loading Simulation | PetroBench opens your saved design and its inputs |
| 2. Importing Variables | Your five input tabs (Preferences, Tubing, Well, Rods, IPR) are pulled in |
| 3. Processing Input Variables | Inputs are validated and prepared for the solver |
| 4. Setting Rod String | The taper string is placed in the wellbore geometry under static conditions |
| 5. Simulating | The solver establishes steady-state system dynamics. This is the longest stage |
| 6. Calculating Output Variables | Loads, stresses, gearbox torque, and production are computed |
| 7. Creating Report | Results and charts are assembled for the Results tab |
If the modal is moving through stages, it's working. Stage 5 (Simulating) takes the most time, so a pause there is expected.
Dynamic runs are different. RodSim Dynamics sweeps your design across ranges of pump fillage, pump intake pressure, and stroke rate, running many simulations in parallel on cloud compute. It finishes in minutes, not seconds, and that is normal. See Dynamics.
When to Worry
The key question is whether the run is still progressing or has actually frozen.
Still progressing (keep waiting): The stage label is changing, or you're on Simulating (stage 5) on a complex design. This is normal, especially for the reasons in the next section.
Stuck on the same stage for over 2 minutes: Something may have gone wrong. Refresh the page and run again.
Progress bar frozen or not moving at all: Your browser may have lost connection to the simulation engine. Check your internet, then refresh and re-run.
Lost connection message: If you see "Unable to connect to the simulation engine," disconnect any VPN and retry. See Troubleshooting for engine connection errors.
Ends in an error: If the run finishes with a failure instead of results, see Simulation Errors.
Diagnosis Steps
Work through these in order:
- Refresh the page and click Save All, then run again. Many stuck runs clear on a second attempt.
- Check your internet connection. An unstable or dropped connection is the most common cause of a frozen progress bar. Try a stable, wired, or reliable Wi-Fi network.
- Check status.petrobench.com for a reported outage or engine maintenance.
- Try a different browser. Safari in particular can be slow with RodSim; Chrome is recommended. See Browser Issues.
If it keeps failing, duplicate a working simulation and change one parameter at a time. That isolates whether a specific input is causing the slowdown.
Why Some Simulations Take Longer
Some designs legitimately take more time. These are the usual reasons:
Complex or multi-taper rod strings: More taper sections mean more calculations. A six-taper string takes longer than a three-taper one.
Small step length: Step Length (on the Simulation Preferences tab) sets the calculation step size. Smaller steps are more accurate but slower. The default 50 ft is a good balance and works for most wells.
Deviated wells with detailed surveys: More directional survey points mean more geometry calculations. A highly deviated well with a dense survey runs longer than a vertical one.
Dynamic instead of Static: A Static run is a single steady-state solve. A Dynamic run sweeps three sensitivities across ranges of conditions, so it does far more work and takes minutes.
Optional detail turned on: Extras such as IPR deliverability and per-rod guide calculations add work to the run.
Speeding Things Up Without Losing Accuracy
- Keep Step Length at the default 50 ft. Only drop it lower when you specifically need finer resolution. Going smaller than you need is the easiest way to make a run slow for no benefit.
- Use Static for iteration. Static returns dyno cards, stress, loads, and gearbox torque in seconds. Reserve Dynamic for when you actually need the sensitivity sweep.
- Simplify where appropriate. For a first pass, skip optional IPR data and add it later once the design runs cleanly.
- Right-size the survey. Use a representative directional survey rather than the densest one available when the extra points don't change the geometry meaningfully.
When to Contact Support
Email support@petrobench.com if:
- A simulation that ran fine before is now consistently slow or stuck.
- Runs hang even after refreshing, switching browsers, and checking status.
- The status page shows no outage but the engine still won't respond.
Include the well name and simulation name, a screenshot of where it's stuck, and what you've already tried.