[rmotwan]

Hey all,

After going thru the two approaches for configuring multiple injections in CONVERGE, I'm facing a challenge in successfully implementing either of the approaches:

1. Single injector, multiple injections (SIMI) - by providing an injection rate shape and other injection parameters of inject_start_time, inject_duration, and inject_mass, CONVERGE solver throws the following error:

(Rank 25) FATAL_ERROR: [ 589] Temporal inject_start_time is not allowed.

It's not just inject_duration but all the three injection parameters ("inject_{}")

How are the injection parameters "temporal" when the input files (attached) have them in a "tabular" form?

2. Multiple Injectors, Single Injection (MISI) - by configuring the desired number of injectors, CONVERGE solver throws the following error:

terminate called after throwing an instance of 'std::runtime_error'
what(): Unexpected HDF5 datatype encountered: Dataset name INJECTOR_VELOCITY, Expected Dataset size = 2, Actual size = 1
[node2097:2992244] *** Process received signal ***
[node2097:2992244] Signal: Aborted (6)
[node2097:2992244] Signal code: (-6)

I'm also including the "spray_{}.in" files for respective modeling approaches. What would be the appropriate next steps to solve these issues? Or what are the things that I'm failing to understand on how to set up multiple injections?

Thanks for your help!

[ksrivast]

Hello Rahul,

1. Single injector, multiple injections :
Using profiles for inject_start_time, inject_duration and inject_mass is only applicable for variable RPM cases. The tabular entries are provided for each injection cycle/engine cycle. If your case is indeed a case with variable RPM, then the tabular entries are not set up correctly. Please refer to our v3.0 CONVERGE manual, Chapter 11 and 23 discussion on variable RPM cases.

From my understanding, this is not what you're after. So for your case, you would have a single injector and you do not require the profile option. The start time will be the start time of the first injection. The duration will be the total duration of all injections. And your rate shape profile will include the profile of (and delay between) all injections. Your attached rate-shape profile is of the correct format. The injected mass will be distributed among the multiple injection events based on your profile. It is recommended that you check your spray_rate_inj0.out file, which is created at the start of the simulation, to ensure the multiple injections are set up correctly for the injector. This can also be verified using the Spray Rate Calculator tool in Studio.

With this approach, if you change the injection timings/mass, you often need to update the rateshape profile as well. As such, this is not our recommended approach. We recommend having an individual injector for each injection. This gives the user flexibility to vary each injection event independently of others.

2. Multiple Injectors, Single Injection (MISI) :
For the error you are seeing, are you perhaps using UDFs for your case? The INJECTOR_VELOCITY dataset name is encountered within custom UDFs. Could you check the UDFs you have with you? If you continue to face issues, please reach out to us at support@convergecfd.com. Please use your official email for all correspondence with Convergent Science. Please mention your issue, attach your case setup and add the cfd-online thread, as reference. I don't see any issues with your spray.in file presently. We can get a better idea if we look at the entire case.

Hope this helps.

Sincerely,

[rmotwan]

Hey Kislaya,

Thanks for the helpful answer! Over the weekend, I did end up figuring out the issue.

The problem wasn't with defining the injection process; it was because I was restarting my simulation with a restart file belonging to a different simulation with one injector. The 'INJECTOR_VELOCITY' error makes a lot more sense now.

Thanks again!