Skip to content

qaqc_tools⚓︎

Quality Assurance of the forecast (QA) and analysis (QC) step.

QAQC ⚓︎

Perform Quality Assurance of the forecast (QA) and analysis (QC) step. Available functions (developed in 4DSEIS project and not available yet):

  • calc_coverage: check forecast data coverage
  • calc_mahalanobis: evaluate "higher-order" data coverage
  • calc_kg: check/write individual gain for parameters;
  • flag data which have conflicting updates
  • calc_da_stat: compute statistics for updated parameters

Copyright © 2019-2022 NORCE, All Rights Reserved. 4DSEIS

calc_coverage(line=None) ⚓︎

Calculate the Data coverage for production and seismic data. For seismic data the plotting is based on the importance-scaled coverage developed by Espen O. Lie from GeoCore.

Parameters:

Name Type Description Default
line array - like

If not None, plot 1D coverage.

 None
Notes
  • Copyright © 2019-2022 NORCE, All Rights Reserved. 4DSEIS
  • Not available in current version of PIPT

calc_da_stat(options=None) ⚓︎

Calculate statistics for the updated parameters. The persentage of parameters that have updates larger than one, two and three standard deviations (calculated from the initial ensemble) are flagged.

Parameters:

Name Type Description Default
options dict

Settings for statistics. write_to_file : bool, optional Whether to write results to a .grdecl file. Defaults to False.

 None
Notes
  • Copyright © 2019-2022 NORCE, All Rights Reserved. 4DSEIS
  • Not available in current version of PIPT

calc_kg(options=None) ⚓︎

Check/write individual gain for parameters. Note form ES gain with an identity Cd... This can be improved

Visualization of the many of these parameters is problem-specific. In reservoir simulation cases, it is necessary to write this to the simulation grid. While for other applications, one might want other visualization. Hence, the method also depends on a simulator specific writer.

Parameters:

Name Type Description Default
options dict

Settings for the Kalman gain computations.

  • 'num_store' : int, optional Number of elements to store. Default is 10.
  • 'unique_time' : bool, optional Calculate for each time instance. Default is False.
  • 'plot_all_kg' : bool, optional Plot all the Kalman gains for the field parameters. If False, plot the num_store. Default is False.
  • 'only_log' : bool, optional Only write to the logger; no plotting. Default is True.
  • 'auto_ada_loc' : bool, optional Use localization in computations. Default is True.

  • 'write_to_resinsight' : bool, optional Pipe results to ResInsight. Default is False.

    requires that ResInsight is open on the computer.

 None
Notes
  • Copyright © 2019-2022 NORCE, All Rights Reserved. 4DSEIS
  • Not available in current version of PIPT

calc_mahalanobis(combi_list=(1, None)) ⚓︎

Calculate the mahalanobis distance as described in "Oliver, D. S. (2020). Diagnosing reservoir model deficiency for model improvement. Journal of Petroleum Science and Engineering, 193(February). https://doi.org/10.1016/j.petrol.2020.107367"

Parameters:

Name Type Description Default
combi_list list

List of levels and possible combinations of datatypes. The list must be given as a tuple with pairs: - level : int Defines which level. Default is 1. - combi_typ : str Defines how data are combined. Default is no combine.

 (1, None)
Notes
  • Copyright © 2019-2022 NORCE, All Rights Reserved. 4DSEIS
  • Not available in current version of PIPT