The default slope-based method (CJUMP,CALC,1) calculates the allowable jump such that the predicted slope does not change beyond a user-imposed jump criterion :

(6–1)

The slope-based calculation is susceptible to small or no jumps for small slope values.

Derived using a Taylor-series expansion of the variable evolution, the variable-based method (CJUMP,CALC,2) calculates the allowable jump via a user-imposed jump criterion and the cycle history of the control variables:

(6–2)

Set the minimum number of initial cycles at the beginning of a cyclic-loading analysis before a jump is allowed via Option = INICYC.

To set the minimum number of intermediate cycles before a jump is allowed at any state of the cyclic-loading analysis, use Option = MINCYC (including the initial cycles at the beginning of the cyclic-loading analysis if INICYC is not set).

Defaults are INICYC = 3 and MINCYC = 3, as at least three cycles are needed to establish the global trend.

To set the maximum intermediate number of cycles to jump, use Option = MAXJUMP.

CJUMP,MINCYC,5
CJUMP,INICYC,10
CJUMP,MAXJUMP,20

You can adjust the minimum number of intermediate cycles N _int based on the jump length, thereby limiting jump occurrences in cases of highly nonlinear global trends. Initially, N _int = N _min . The adjustment is calculated as:

are your minimum intermediate cycle-adjustment parameters defining the increase and decrease, respectively, of intermediate cycles. To set your adjustment parameters, enter N _min (Input1), (Input2), and (Input3) via Option = MINCYC.

! Set minimum increase (Δp) to 2 and minimum decrease (Δm) to 1 for intermediate cycles
CJUMP,MINCYC,5,2,1

Defaults:

Cycle-jump calculations occur at the same relative point in time for each cycle.

As shown in Figure 6.8: Analysis with Standard Solutions Preceding the Cyclic-Loading Analysis, points A1 and A2 are at the same relative point in time for the two cycles shown. Similarly, points B1 and B2 are at the same relative point in time.

By default, the program sets the relative time value to correspond to an absolute maximum load value for a defined load cycle. In the figure, the default would be set as B1 (or B2).

Optionally, you can specify a relative time value via Option = RELTIME. The relative time must coincide with a defined time point in the cyclic-loading table.

CJUMP,RELTIME,25

The default control (Option = CONTROL) variable for cycle-jump calculation is stress (S).

Optionally, other control variables can be chosen via Option CONTROL. Multiple control variables are allowed.

Generally, for control variables (such as stress, plastic strain, and degrees of freedom), jump calculations are performed based on equivalent values.

! set stress control var
CJUMP,CONT,S
! set plastic strain control var
CJUMP,CONT,EPPL

For problems involving user-programmable features (such as UserMat), you can set state variables as control variables. You can select specific state variables for jump control. The material ID associated with the state variables must be specified.

For state-variable control, jump calculations are performed based on individual values.

! select material associated with state variable
CJUMP,CONT,SVAR,MAT,2  !material 2
! enter up to 6 state variable entries per line
CJUMP,CONT,SVAR,SEL,2,3,5,6
! enter as many lines as needed
CJUMP,CONT,SVAR,SEL,9,1,100,12,8,19 

Outputting State Variables at Select Time Points

To output state variables at select time points, issue CLOAD,OUTR and OUTRES,ALL before issuing OUTRES,SVAR.

CLOAD,OUTR
OUTRES,ALL,%timvals%
OUTRES,SVAR,%timvals% 

By default, all materials in a body control the jump.

You can select specific materials to control the jump (CJUMP,CNMT). Jump calculations are then performed for the selected materials only. This option applies to both Gauss-point-based (stress, plastic strain, etc.) and degree-of-freedom-based control. For materials not selected, jump calculations are omitted. When a jump occurs, however, extrapolation of values occurs for the entire body.

You can specify up six material IDs on each CJUMP,CNMT command. Issue the command as many times as needed.

CJUMP,CNMT,2,4,6,5  ! up to 6 mat IDs per line (can be less)
CJUMP,CNMT,1,11     ! enter as many lines as needed

You can set a jump criterion to be dependent on the material ID (CJUMP,ADCR,MAT) and/or the control variable (CJUMP,ADCR,CONT).

CJUMP,ADCR,MAT,2,0.5
CJUMP,ADCR,MAT,12,1.0
CJUMP,ADCR,CONT,DOF,0.6
CJUMP,ADCR,CONT,EPPL,0.2

If you define material-ID- and/or control-variable-dependent jump criteria, they override any global jump criteria (defined via CJUMP,CRIT). Global jump criteria (if defined) still apply to any jump-calculation point not affected by a CJUMP,ADCR command. If no global jump criteria are defined, the jump calculation is omitted for any jump-calculation points not having associated jump criteria. Global jump criteria are not required if you define material-ID- or control-variable-dependent jump criteria.

When you define both material-ID- and control-variable-dependent jump criteria for a jump-calculation point, the program uses the smaller value. For example, if your jump criterion is 0.8 for material 2 and your jump criterion is 0.2 for control variable EPPL, a jump criterion of 0.2 will apply to any jump-calculation point where the material ID is 2 and control is EPPL.

For a control-variable-dependent jump criterion, you must still set the appropriate control variable (CJUMP,CONT).

In the case of multiple cycle-jump solutions in a single analysis, each requires a CJUMP,ADCR command to be issued before the associated SOLVE command.

By default, the program uses the slope-based jump-calculation method to calculate jump.

If needed, you can specify the variable-based jump-calculation method instead (CJUMP,CALC,2).

By default, the program uses the minimum value across all jump-calculation points (Gauss point and/or node) as the jump value.

For some problems, the default behavior may be too conservative, as some jump-calculation points lead to smaller-than-desired jumps. In such cases, you can use the statistical jump-calculation method (CJUMP,PERC), which selects the jump value as a percentile of the cumulative frequency of all jump values (Van Paepegem et al., 2001).

The percentile technique relies on the jump-value information from all jump-calculation points. For multiple Gauss-point-based control variables active together (such as stress and plastic strain), the smallest value is stored at a given Gauss point. When both Gauss-point and node-based control are active together (such as stress, plastic strain, and degrees of freedom), the cumulative frequency is determined based on information from both the Gauss points and nodes.

CJUMP,PERC,2.5