Monitors and derived quantities

Residuals describe the equations; monitors describe the result being sought. Vanellus evaluates each configured monitor once per SIMPLE iteration and stores its history alongside the residuals in iteration_info.csv.

A monitor records a cell-field statistic, a value at one position, or a conservation balance. Its purpose is to keep the engineering quantity visible while the iterative solution develops.

Field statistics

Minimum, maximum, and mean monitors operate on the cells where their field is solved:

Insulating and compact-thermal-model cells are outside the solved thermal domain, so their display values in solid.vtu do not contribute to temperature reductions. Similarly, placeholder flow values in solid cells do not contribute to flow-field reductions.

These operations use the stored cell values. The mean is volume-averaged, so each included cell contributes in proportion to its physical volume. Refining part of the mesh into smaller cells therefore does not give that region disproportionate influence. Minimum and maximum values are resolved-cell values rather than reconstructed values at mesh nodes or boundaries.

A mean-velocity monitor volume-averages each Cartesian component separately. Minimum- and maximum-velocity monitors operate over the stored components; they do not reduce velocity magnitude. For example, a maximum-velocity monitor reports the greatest stored component rather than the greatest speed.

Point probes

A probe returns the value from the cell containing its [x, y, z] position.

If a probe lies exactly on a cell face, Vanellus takes the value from the cell on the positive side of that face. At the maximum edge of the domain, where there is no cell on the positive side, it takes the value from the final cell inside the domain.

This is a cell-field probe, not an interpolation onto the face, so a value requested exactly on a face is not guaranteed to represent the physical face value. This is particularly important at an interface where the neighboring cell values can differ, such as between two solids with different thermal conductivities.

The position should lie in the region where its field is meaningful. For example, place a pressure probe in fluid and a temperature probe in fluid or a conducting solid. A probe in an insulating solid can return a placeholder rather than a physical prediction.

Conservation balances

Balance monitors integrate fluxes rather than sampling cell values:

Mass flow is the rate at which mass crosses every non-wall fluid boundary. Energy flow includes convective and conductive transfer across the thermal-domain boundary together with volumetric heat sources. A positive absolute balance means more enters than leaves; a negative value means more leaves than enters.

Relative balances make cases with different flow or power levels easier to compare, but they should be interpreted together with the corresponding absolute scale. A small stable imbalance is evidence of conservation quality, not proof that the physical setup is correct.

CTM junction temperatures

A compact thermal model calculates one junction temperature from its power, thermal resistances, and the temperatures on its active faces. This quantity is configured inside the component’s thermal model rather than in the top-level monitors array, and it is kept separate from regular monitor fields in the iteration output. It otherwise behaves like a monitored derived quantity:

• it is evaluated once per SIMPLE iteration;
• its complete history is written to iteration_info.csv as junction_temperature_<tag>; and
• exit_tolerance can allow it to stop the simulation after it settles.

The final junction temperature is also written into that CTM’s cells in solid.vtu, so this value can be trusted as the final junction result. It is set uniformly across the CTM cells for visualization, however, and is not a prediction of the temperature distribution through the package.

Histories and stopping

In iteration_info.csv, regular monitor columns use the monitor_ prefix and CTM junction temperatures use junction_temperature_. Non-finite live values are represented as null in streamed updates and as NaN in the downloaded CSV.

Recording a quantity and allowing it to stop a run are separate choices. A regular monitor or CTM needs exit_tolerance to participate in stopping. Both use the shared convergence_window and test stability rather than an acceptable limit.

Passing the test shows only that the quantity was stable over the most recent window. It does not guarantee that the quantity would remain unchanged if the simulation continued: a history can appear flat for several iterations before drifting or changing again. Choose the window and tolerance for the decision being made, and inspect the complete history.

See also

• How simulations stop explains the convergence test.
• Monitor a result and stop when it settles shows how to configure monitors.
• VTU fields and domains explains how final spatial fields relate to monitor values.