TUNSUB

ModelingThe TUNSUB allows you to modify certain solver parameters during a simulation.

Use

<Param_Simulation
     usrsub_param_string = "USER(9998, 0.1)"
     usrsub_dll_name     = "NULL"
     usrsub_fnc_name     = "TUNSUB"
  />

Format

Fortran Calling Syntax
SUBROUTINE TUNSUB (TIME, PAR, NPAR, IFLAG, ISTATE, H, IORDER)
C/C++ Calling Syntax
void TUNSUB (double *time, double *par, int *npar, int *iflag, int *istate, double *h, int *iorder)
Python Calling Syntax
def TUNSUB(time, par, npar, iflag, istate, h, iorder)
MATLAB Calling Syntax
function TUNSUB(time, par, npar, iflag, istate, h, iorder

Attributes

TIME
[double precision]
Current simulation time.
PAR
[double precision]
An array that contains the constant arguments from the list provided in the user-defined statement.
NPAR
[integer]
Number of entries in the PAR array.
IFLAG
[integer]
An integer variable that MotionSolve sets to 1 when the user subroutine is called during MotionSolve initialization. The IFLAG will be set to zero at all other times.
ISTATE
[integer]
An integer variable that MotionSolve sets to 1 when the integrator step is a successful step. If the integrator step is not successful, ISTATE will be set to zero.
H
[double precision]
Current integrator step size (h < h_max, specified in Param_Transient).
IORDER
[integer]
Current integrator order (IORDER < max_order, specified in Param_Transient).

Example

The following TUNSUB example modifies the absolute and relative integrator error tolerance to the value passed in as the second argument in the USER(…) in the model when the angular velocity of the rigid bodies exceed 20 radians/second.

from math import fabs
g_num_rigid_body = 0
g_cg_id_vec = []
g_body_id_vec = []

#define the TUNSUB function
def TUNSUB(time, par, npar, iflag, istate, h, iorder):
    w = 3*[0.0]
    global g_num_rigid_body
    global g_cg_id_vec
    global g_body_id_vec
    
    #set solver parameters
    py_set_dae_error(0.01)
    py_set_dae_hmax(0.01)
    py_set_dae_maxord(5)
    py_set_dae_maxit(4)
    t0 = py_get_starting_time()
    te = py_get_end_time()
    
    #define dependencies during iflag = 1
    if iflag==1:
        tmp_num =0
        tmp_num = py_getnumid("PART")
        tmp_iarray1 = tmp_num*[0]
        tmp_iarray2 = tmp_num*[0]
        [tmp_iarray1, istat] = py_getidlist("PART", tmp_num)
        g_num_rigid_body =0 
        for i in xrange(tmp_num):
            [tmp_str, istat] = py_modfnc("Body_Rigid", tmp_iarray1[i],"cg_id")
            cg_id =int(tmp_str)
            [tmp_str, istat] = py_modfnc("Body_Rigid", tmp_iarray1[i],"IsGround")
            if tmp_str=='TRUE':
                 continue
            tmp_iarray2[g_num_rigid_body] = tmp_iarray1[i]
            tmp_iarray1[g_num_rigid_body] = cg_id
            g_num_rigid_body +=1
        g_cg_id_vec = tmp_iarray1
        g_body_id_vec = tmp_iarray2
        for i in xrange (g_num_rigid_body):
            cg_id = g_cg_id_vec[i];
            [w[0], istat] = py_sysfnc("WX", cg_id)
            [w[1], istat] = py_sysfnc("WY", cg_id)
            [w[2], istat] = py_sysfnc("WZ", cg_id)
        return
    #check if current step is converged i.e. istate = 1
    if istate==1:
        for i in xrange(g_num_rigid_body):
            cg_id = g_cg_id_vec[i];
            #obtain angular velocities for all parts
            [w[0], istat] = py_sysfnc("WX", cg_id)
            [w[1], istat] = py_sysfnc("WY", cg_id)
            [w[2], istat] = py_sysfnc("WZ", cg_id)
          #check if the angular velocities are above some threshold 
          #value
            if (fabs(w[0])>20) | (fabs(w[1])>20) | (fabs(w[2])>20):
             #change tolerance values
                istat = py_set_tol_factor("Body_Rigid", g_body_id_vec[i],"ATOL", par[1])
                istat = py_set_tol_factor("Body_Rigid", g_body_id_vec[i],"RTOL", par[1])

Comments

  1. Only one TUNSUB can be defined in the model and it should be defined in the model section. There is no ID associated with TUNSUB.
  2. Currently TUNSUB is application only for DAE integrator DSTIFF and FIM_D.
  3. TUNSUB is called after every integration step and the solver settings that are modified are applicable for analysis from that time onward.
  4. All the utility subroutines can be called from TUNSUB. In addition there are several special utility subroutines available to modify solver settings.
    set_tol_factor(char*entity_type, intentity_id, char*tol_type, doublefactor, int*errflg)
set_dae_error(doubleerror);
set_dae_hmax(doubleh_max);
set_dae_maxord(intmaxord);
set_dae_maxit(intmaxit);
set_dae_jacobeval(intjacob_eval);
set_dae_jacobinit(intjacob_init);