dymoval 0.9.0

Dymoval is a Python package for analyzing measurements data and validate models.

Build status

Tools

---

What is it?

Dymoval is a Python package for analyzing measurement data and validating models.

Dymoval validates models based only on measurements and simulated data, and it is completely independent of the used modeling tool. That means that it does not matter if a model has been developed with Simulink, Modelica, etc., dymoval will only look at the produced data from the model.

If you are tracking your models changes in a CI/CD environment, then dymoval API can be easily used to run tests in Jenkins or GitHub Actions pipelines as it enables unit-testing on models.

Finally, dymoval provides a number of functions for for handling measurements data, addressing common issues such as noise, missing data, and varying sampling intervals.

Installation

Dymoval exists on both pip and conda, so you can choose between the following:

pip install dymoval
conda install -c conda-forge dymoval

Getting started

Suppose you want to validate a model and you have the simulated out y_sim the measured input u_meas, and the measured out y_meas time-series sampled with period sampled_period. Just call the following function:

from dymoval.validation import validate_models

validate_models(
    measured_in=u_meas,
    measured_out=y_meas,
    simulated_out=y_sim,
    sampling_period = sampling_period
)

to get something like the following:

  Input whiteness (abs_mean-max)      0.3532
  R-Squared (%)                      65.9009
  Residuals whiteness (abs_mean-max)  0.1087
  Input-Res whiteness (abs_mean-max)  0.2053

           My_Model
  Outcome: PASS

Congrats! Your model passed the test!

But what if the test didn't pass? Don't worry, it might not be the model's fault.

For example, you could be dealing with noisy measurements, over-sampled signals, missing data, and other factors that might affect the results. Take a look at the tutorial to learn how to address such issues:

  import dymoval as dmv
  dmv.open_tutorial()

The above commands create a dymoval_tutorial folder containing all the files needed to run the tutorial in your home folder. All you have to do is to run Jupyter notebook named dymoval_tutorial.ipynb. You need an app for opening .ipynb files.

And if you want to discover more on model validation and Dymoval check out the docs.

Main Features

Model validation

• Validation metrics:
  • R-square fit
  • Residuals auto-correlation statistics
  • Input-Residuals cross-correlation statistics
• Coverage region
• MIMO models
• Independence of the modeling tool used.
• API suitable for model unit-tests

Measurement data analysis and manipulation

• Time and frequency analysis
• Easy plotting
• Missing data handling
• Linear filtering
• Means and offsets removal
• Re-sampling
• Physical units

License

Dymoval is licensed under BSD 3 license.