Uncertainty budget

An uncertainty budget can be estimated with the following inputs:

  1. reference model parameter configuration string/file

  2. MCMC measurements of model parameters (sensing and actuation) saved in HDF5 files

  3. Time-dependent correction factors (TDCFs) as measured by the calibration lines (queried via NDS2)

  4. Residual unknown systematic error measured by stacked measurements and saved in HDF5 files

  5. systematic error of the photon calibrator

An example configuration file for the uncertainty budget can be found at in the pyDARM git repository.

First we initialize a DARMUncertainty object and then we can draw random samples from the object:

>>> test_unc = pydarm.uncertainty.DARMUncertainty(
... 'H1_20190416.ini', uncertainty_config='H1_20190416_uncertainty.ini')
>>> samples = test_unc.compute_response_uncertainty(
... 1239958818, frequencies, trials=10)

Note

The uncertainty budget configuration file can be combined with the original DARM model configuration file into a single file. For convenience if there is anything provided in the uncertainty_config argument, it will override any key-value in the model configuration.

Plotting