Closed tonyinme closed 7 months ago
Following up on our conversation with @psakievich @lawrenceccheung @michaelasprague , here are the input files for a case with less number of actuator points (60) and coarser epsilon (epsilon/R=.1). The same issue is still present even when using less actuator points.
This issue is stale because it has been open 30 days with no activity.
@tonyinme @lawrenceccheung and @mbkuhn is this still an issue or did the recentish updates to the ALM fix this?
Hi @marchdf , yes, this is still an issue. It only happens for the first time-step, so it is not stopping our simulations from running or affecting the results after some time.
Hi @marchdf , yes, this is still an issue. It only happens for the first time-step, so it is not stopping our simulations from running or affecting the results after some time.
Do you know if this is an issue upon restart?
That's a good question. I haven't checked restarting. I see it in the beginning of a simulation.
On Fri, Feb 9, 2024, 12:03 PM Michael B Kuhn @.***> wrote:
Hi @marchdf https://github.com/marchdf , yes, this is still an issue. It only happens for the first time-step, so it is not stopping our simulations from running or affecting the results after some time.
Do you know if this is an issue upon restart?
— Reply to this email directly, view it on GitHub https://github.com/Exawind/amr-wind/issues/750#issuecomment-1936454590, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACULCAPTS5XR5VQJNNOYPZTYSZXHRAVCNFSM6AAAAAASZ2SYOSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZWGQ2TINJZGA . You are receiving this because you were mentioned.Message ID: @.***>
@mbkuhn , I tried running a restart simulation with the test above and I can confirm that this only shows up in the very first time-step of a new simulation. When restarting the simulation, we do not see the same problem.
When running a simulation using OpenFAST and ALM, there is a problem with the velocities being passed to openfast. This is a one-blade turbine simulation (fixed wing) that we are using to debug. These velocities have a problem shown in the file below. On the very first time-step, these velocities should all match the initial condition (1 everywhere).
This issue not seem to depend on number of cores and can be reproduced with only 1 core.