Generating Time Series Data from ML training

[Soul1087]

Hey fellas, I have created model of a engine but I want to generate time series data from it so that I can create and train a Machine learning model on the generated data can anyone suggest any idea or solution for the same.

[Kenneth-T-Moore]

Hey, Soul1087. I am not sure what you mean by time series data here. Pycycle does thermodynamic cycle analysis for an engine at design and off-design conditions, so it doesn't really do any kind of transient analysis. The result is assumed to be steady-state. If you wanted a time history of thrust and fuel consumption over an aircraft trajectory, you could combine your pycycle engine model with something like Aviary (https://github.com/OpenMDAO/Aviary). Probably the easiest way to do that would be to create an engine deck, which is a table of thrust, fuelburn rate, and emissions over a range of mach numbers, altitudes, and throttle settings. Still, I'm not sure there is anything you would gain from training the ML on time histories if the underlying engine is all steady-state.

[Soul1087]

thanks for the suggestion I was thinking of calculating at various steady state in a time step and then integrate for the gap. Will look into your suggestion of using Aviary

[Jhawk414]

Yeah not sure what/how OP plans to do ML with transient data, but I think it is possible to do some rudimentary transient cycle work using Dymos. The time history of must of course be compared to real transient test data or time-stepped NPSS output or else it's just guesswork.