pyMAISE.datasets.load_loca

pyMAISE.datasets.load_loca(stack_series=False)[source]

Load loss of coolant accident (LOCA) time series data. This data comes from [RPK+20] and consists of 40 time-independent features that describe the initial state of the reactor during the LOCA transient, which are propagated out in time. There are four sequences:

  • Pellet Cladding Temperature: pellet cladding temperature [K],

  • Core Pressure: core pressure [Pa],

  • Water Level: water level [m],

  • Break Flow Rate: break flow rate [kg/s],

with 400 time steps. The original data was randomly sampled for 2000 perturbed data points and the nominal sample, 2001 samples total.

Parameters:

stack_series (bool, default=False) – If false, the data is loaded in 3D with dimensions (samples, time steps, features). If true, the data is loaded in 2D with the sequences stacked like pulse data resulting in dimensions (time steps, features).

Returns:

  • nominal_data (xarray.DataArray) – The 2D or 3D nominal LOCA data. If 2D it will be shaped (400, 44) and if 3D then the shape is (1, 400, 44).

  • perturbed_data (xarray.DataArray) – The 2D or 3D perturbed LOCA data. If 2D it is shape (800000, 44) and if 3D then the shape is (2000, 400, 44).