pyMAISE.datasets.load_chf

pyMAISE.datasets.load_chf(data_path=None)[source]

Load Critical Heat Flux (CHF) synthetic data. This dataset is based on that produced by the NEA in [CDZ24]. The NEA dataset was generated from vertical water-cooled uniformly heated tubes, producing 24576 samples from 59 different sources of measurement. The dataset was collected over experimental measurements spanning 60 years of CHF data collection methods such as visual identification, physical burnout, changes in the test section resistances, and the usage of thermocouples. The input parameters collected include.

D (m): Diameter of the test section ($0.002 - 0.016~m$),
L (m): Heated length ($0.07 - 15.0~m$),
P (kPa): Pressure ($100-20000~kPa$),
G (kg m-2s-1): Mass flux ($17.7-7712.0~\frac{kg}{m^2\cdot s}$),
Tin (C): Inlet temperature length ($9.0-353.62^\circ C$),
Xe (-): Outlet equilibrium quality ($-0.445-0.986$),

with output

CHF (kW m-2): Critical heat flux ($130.0-13345.0~\frac{kW}{m^2}$).

Negative equilibrium quality ($X$) could represent a subcooled fluid. The database was limited to a diameter between ($2 < D < 25~mm$), $L/D$ ratio ($L/D > 50$ for $X > 0$, $L/D > 25$ for $X < 0$), pressure ($100 \le P \le 21000 kPa$), and mass flux ($0 \le G < 8000~\frac{kg}{m^2\cdot s}$). The measured data was also not equally distributed over the whole span along with no data beyond $D=16~mm$ was found in the database.

As the original dataset is confidential, we do not open-source it here on pyMAISE; therefore, what is given here is a synthetic version designed to mimic the behavior of the real data. It consists of 2500 samples, 2000 training samples and 500 testing samples. These were generated by adding random noise to the experimental data.

Parameters:: data_path (None, str, or list of two str(s), default=None) – Path(s) to real NEA data.

If data_path is None or a list of two strings:

Returns:

train_data (xarray.DataArray) – Synthetic CHF training data.
xtrain (xarray.DataArray) – Training inputs.
ytrain (xarray.DataArray) – Training output.
test_data (xarray.DataArray) – Synthetic CHF testing data.
xtest (xarray.DataArray) – Training inputs.
ytest (xarray.DataArray) – Training output.

If data_path is a string:

Returns:

data (xarray.DataArray) – Raw CHF data.
inputs (xarray.DataArray) – Input data.
outputs (xarray.DataArray) – Output data.