ansys.mapdl.core.Mapdl.cnvtol#

Mapdl.cnvtol(lab='', value='', toler='', norm='', minref='', **kwargs)#

Sets convergence values for nonlinear analyses.

Mechanical APDL Command: CNVTOL

Command default:

For static or transient analysis, check the out-of-balance load for any active degree of freedom using the default VALUE, TOLER, NORM, and MINREF. Also check the translational displacement convergence in most cases. For harmonic magnetic analysis, check the out-of-balance of the degrees of freedom. The energy criterion convergence check is off by default.

Parameters:
labstr

Valid convergence labels.

This command contains some tables and extra information which can be inspected in the original documentation pointed above.

valuestr

Typical reference value for the specified convergence label ( Lab ).

VALUE defaults to the maximum of a program calculated reference or MINREF. For degrees of freedom, the reference is based upon the selected NORM and the current total degree-of- freedom value. For forcing quantities, the reference is based upon the selected NORM and the applied loads.

If VALUE is negative, the convergence criterion based on the specified label is removed, including the default convergence criterion value. The convergence criterion for all other labels remain as they were (either a default value or a previously specified value).

tolerstr

Tolerance value used for the specified Lab convergence label. Default values are described below.

  • If cnvtol is issued with a Lab value specified but no TOLER value, the default tolerance values are:

  • 0.05 (5%) for displacement (U).

  • 1.0E-7 for the joint element constraint check (JOINT). This value rarely needs to be changed. A loose tolerance value may lead to inaccurate or incorrect solutions. When Lab = JOINT, VALUE, NORM, and MINREF are ignored.

  • 1.0E-3 for the volumetric compatibility check (COMP). When Lab = COMP, VALUE, NORM, and MINREF are ignored.

  • 0.05 for energy error (ENGY).

  • For all other Lab labels, the default tolerance value is 0.005 (0.5%).

  • If cnvtol is not issued, the TOLER defaults are as follows:

  • 0.005 (0.5%) for force (F) and moment (M)

  • 1.0E-4 (0.01%) for volume (DVOL)

  • 0.05 (5%) for displacement (U)

  • 0.05 (5%) for hydrostatic pressure (HDSP)

  • 1.0 for temperature (TEMP) when the iterative QUASI solver is used ( thopt,QUASI,,,,,,1)

If you choose to specify a TOLER value, it must be greater than zero and less than 1. This is true for all Lab labels.

The program may adjust the force convergence tolerance if you do not explicitly set a value via cnvtol. See Notes for details.

normint or str

Specifies norm selection:

  • 0 - Infinite norm (check each degree of freedom separately) (default for Lab = U and for Lab = TEMP when the iterative QUASI solver is used ( thopt,QUASI,,,,,,,1).

    The infinite norm is also used for the energy error criterion (ENGY) and is the only option available for ENGY.

  • 1 - L1 norm (check absolute value sum).

  • 2 - L2 norm (check SRSS value) (default, except for Lab = U).

  • 3 - Infinite norm (check each degree of freedom separately). The reference is calculated using the infinite norm of the displacement increment of the substep. Valid only for Lab = U.

minrefstr

The minimum value allowed for the program-calculated reference value. If negative, no minimum is enforced. Used only if VALUE is blank. Default values are as follows:

  • maximum of 0.01 or internally calculated minimum reference value for force (F), moment (M)

  • a small factor times the average element length of the model for displacement (U) convergence

  • 0.01 for volume (DVOL) convergence

  • 1.0E-4 for gradient field residual (GFRS)

  • 1.0E-6 for heat flow (HEAT)

  • 1.0E-6 for diffusion flow (RATE)

  • 1.0E-12 for charge (CHRG)

  • 1.0E-6 for hydrostatic pressure (HDSP)

  • 1.0 for temperature (TEMP) when the iterative QUASI solver is used ( thopt,QUASI,,,,,,,1)

  • 1.0 for energy error convergence (ENGY)

  • 0.0 otherwise

Notes

This command is usually not needed because the default convergence criteria are sufficient for most nonlinear analyses. In rare cases, you may need to use this command to diagnose convergence difficulties.

Values may be set for the degrees of freedom and/or the out-of-balance load for the corresponding forcing quantities.

Issuing cnvtol to set a convergence criterion for a specific convergence label ( Lab ) does not affect the convergence criterion for any other label. All other convergence criteria will remain at their default setting or at the value set by a previous cnvtol command.

If cnvtol is not issued for any force convergence label (F, M, DVOL, and so on as listed under the Lab argument), the default convergence tolerance for a particular force label is increased dynamically during the Newton-Raphson iterations in the range of 1 to 1.66 times the default value. For example, the F label default tolerance is 0.005. Therefore, the maximum convergence tolerance with the adjustment is 0.0083. This adjustment is not activated until the 8th or higher Newton-Raphson iteration. If you do not want the program to adjust the tolerance, issue cnvtol to specify the convergence tolerance for the appropriate force label.

When using the Mechanical APDL graphical user interface (GUI), if a “Delete” operation in a Nonlinear Convergence Criteria dialog box writes this command to a log file ( Jobname.LOG or Jobname.LGW ), you will observe that Lab is blank, VALUE = -1, and TOLER is an integer number. In this case, the GUI has assigned a value of TOLER that corresponds to the location of a chosen convergence label in the dialog box’s list. It is not intended that you type in such a location value for TOLER in an interactive session. However, a file that contains a GUI- generated cnvtol command of this form can be used for batch input or with the input command.

Convergence norms specified with cnvtol may be graphically tracked while the solution is in process using the Graphical Solution Tracking (GST) feature. Issue gst to enable or disable GST. By default, GST is ON for interactive sessions and OFF for batch runs.

The energy convergence check (ENGY) is not available when the arc-length method ( arclen,ON) is used.

For more information on convergence calculations in a nonlinear analysis, see Convergence

This command is also valid in PREP7.