Skip to content

Latest commit

 

History

History
107 lines (79 loc) · 4.27 KB

extract_data_example.md

File metadata and controls

107 lines (79 loc) · 4.27 KB
Raw

Example on Data Extraction from *.hdf5 file

The following is an example illustrating how to extract data from an *.hdf5 file produced after running the module extract. For the sake of reproducibility, the *.hdf5 file accompanying the following example is made available under ./docs. A Jupyter Notebook is also provided with the outputs of the run.

Prerequisites

Besides a Python installation, you need to install numpy, matplotlib, and hdf5 to reproduce the results of this example. All of these libraries can be set up via pip by the following command:

pip install numpy && pip install matplotlib && pip install h5py

Example of Data Extraction

In this example, we work with the IEEE14_generators_20220526190314_4eaa02.hdf5 file saved under ./docs. The first step is to import the required dependencies:

import os
import matplotlib.pyplot as plt
import h5py

Then, we open the *.hdf5 file and print a list of the available scenarios, the components extracted, and the available internal measurements (because in this case we worked with signals from a generator):

filepath = os.path.join(os.getcwd(), 'IEEE14_generators_20220526190314_4eaa02.hdf5')

with h5py.File(filepath, "r") as f:

    # Listing the available scenarios
    print("Available Scenarios: %s\n" % f.keys())

    # Getting the list of scenarios (and labels for all of them)
    a_group_key = list(f.keys())[0]

    # Get the data
    print("Available Signals: %s\n" % f[a_group_key].keys())
    b_group_key = list(f[a_group_key].keys())[1]

    # Get an internal measurement from the generator
    print("Available Internal Measurements: %s\n" % f[a_group_key][b_group_key].keys())

The next step is to extract the data from the *.hdf5 file. Here, we first get the time vector and convert it to a NumPy array (see the comment in the code extract below). Likewise, we get the traces of the generator's PSS output and the eigenvalues at scenarios 1 and 16 for the initial and final conditions, respectively. Note that the hierarchy in the *.hdf5 file is kept while reading the data. This eases the manual data extraction after the tool is run.

    # ------------------------------
    # Time vector
    # ------------------------------
    # Scenario 1
    t1 = f['1/time']
    t1 = t1[:] # extracting data from DataSet class

    # Scenario 16
    t16 = f['16/time']
    t16 = t16[:] # extracting data from DataSet class

    # ------------------------------
    # Eigenvalues
    # ------------------------------
    eigs1initial = f['1']['eigs/initial']
    eigs1initial = eigs1initial[:]
    eigs16final = f['16/eigs/final']
    eigs16final = eigs16final[:]

    # ------------------------------
    # PSS Output
    # ------------------------------
    PSSoutput1 = f['1/gen_Bus_1_1/VOTHSG']
    PSSoutput1 = PSSoutput1[:]

    PSSoutput16 = f['16/gen_Bus_1_1/VOTHSG']
    PSSoutput16 = PSSoutput16[:]

Finally, the results are plotted using Matplotlib.

# ------------------------------
# Plotting results
# ------------------------------

fig, ax = plt.subplots(nrows = 2, ncols = 1)
fig.canvas.draw()

ax[0].plot(t1, PSSoutput1, label = 'Scenario 1', color = 'darkblue')
ax[0].plot(t16, PSSoutput16, label = 'Scenario 16', color = 'orange')
ax[0].set_title('PSS Output (VOTHSG)', pad = 20)
ax[0].set_xlabel('Time (s)')
ax[0].legend(loc = 'upper left')

ax[1].scatter(np.real(eigs1initial), np.imag(eigs1initial), color = 'darkblue', marker = 'x', label = 'Scenario 1')
ax[1].scatter(np.real(eigs16final), np.imag(eigs16final), color = 'darkorange', marker = 'x', label = 'Scenario 16')
# ax[1].set_title("Eigenvalues", pad = 20)
# ax[1].legend(loc = 'upper left')
ax[1].set_xlabel("Real Part ($\sigma$)")
ax[1].set_ylabel("$j$-Imaginary Part ($j\omega$)")

 # Getting labels for first plot
_lines, _labels = ax[0].get_legend_handles_labels()

# Other experiments
ax[0].legend(_lines, _labels, bbox_to_anchor = (0, 0.975, 1, 0.2),
            loc = 'lower left', mode = 'expand', ncol = 2, handletextpad = 0.52)

fig.tight_layout()