Using monitors

Monitors in PyFluent allow you to dynamically observe the convergence of your solution by tracking the values of solution variables and residuals. They enable you to visualize the progress of the solver, helping ensure that the solution is progressing as expected and allowing you to diagnose issues early on.

You can integrate PyFluent’s monitor callback mechanism with visualization tools from the Python ecosystem, making it easy to visualize the data of interest.

The following example queries the existing monitors and then uses the monitor callback mechanism to perform a simple tabulation of monitored values per iteration:

>>> # get started with case and data loaded
>>> import ansys.fluent.core as pyfluent
>>> from ansys.fluent.core import examples
>>> import pandas as pd
>>> from tabulate import tabulate
>>> solver = pyfluent.launch_fluent(start_transcript=False)
>>> import_case = examples.download_file(
>>>     file_name="exhaust_system.cas.h5", directory="pyfluent/exhaust_system"
>>> )
>>> import_data = examples.download_file(
>>>     file_name="exhaust_system.dat.h5", directory="pyfluent/exhaust_system"
>>> )
>>> solver.file.read_case_data(file_name=import_case)
>>> # check the active report plot monitors using the settings relevant object
>>> sorted(solver.settings.solution.monitor.report_plots())
>>> [
>>>     "mass-bal-rplot",
>>>     "mass-in-rplot",
>>>     "mass-tot-rplot",
>>>     "point-vel-rplot",
>>> ]
>>> # initialize so that monitors object is usable
>>> solver.solution.initialization.hybrid_initialize()
>>> # check which monitors are available
>>> sorted(solver.monitors.get_monitor_set_names())
>>> [
>>>    "mass-bal-rplot",
>>>    "mass-in-rplot",
>>>    "mass-tot-rplot",
>>>    "point-vel-rplot",
>>>    "residual"
>>> ]
>>> # create and register a callback function that will
>>> def display_monitor_table(monitor_set_name="mass-bal-rplot"):
>>>     def display_table():
>>>         data = solver.monitors.get_monitor_set_data(monitor_set_name=monitor_set_name)
>>>         # extract iteration numbers
>>>         iterations = data[0]
>>>         # filter out additional callbacks
>>>         if len(iterations) > display_table.iter_count:
>>>             display_table.iter_count = len(iterations)
>>>             # extract results
>>>             results = data[1]
>>>             # create a DataFrame
>>>             df = pd.DataFrame(results, index=iterations)
>>>             df.index.name = 'Iteration'
>>>             df.reset_index(inplace=True)
>>>             # The streamed data contains duplicates, so eliminate them
>>>             df = df.drop_duplicates(subset='Iteration')
>>>             print(tabulate(df, headers='keys', tablefmt='psql'))
>>>         display_table.iter_count = 0
>>>     return display_table
>>>
>>>
>>> register_id = solver.monitors.register_callback(display_monitor_table())
>>> # run the solver and see the full tabulated monitor data on each iteration
>>> solver.solution.run_calculation.iterate(iter_count=10)