Modelica.Blocks.Sources.TimeTable does not work (bug in event-handling)

[asodja]

In attachment is a test model consisting of only one block of the Modelica.Blocks.Sources.TimeTable component.

The output (TimeTable.y) is zero all the time.

The problem is that the {{initial()}} and {{time >= pre(nextEvent)}} (nextEvent is initalized to 0) never happens, e.g., {{functionDAE_output2()}} is not called at appropriate time so the if-statment is never evaluated to true:

int functionDAE_output2()
{
  /* ... */
  tmp0 = initial();
  tmp1 = pre($PtimeTable$PnextEvent);
  RELATIONGREATEREQ(tmp2, time, tmp1);
  localData->helpVars[0] = tmp2;
  localData->helpVars[1] = tmp0;
  if (edge(localData->helpVars[0]) || edge(localData->helpVars[1])) {
    real_array_create(&tmp3, &$PtimeTable$Ptable, 2, 7, 2);
    tmp4 = _Modelica_Blocks_Sources_TimeTable_getInterpolationCoefficients(tmp3, $PtimeTable$Poffset, $PtimeTable$PstartTime, time, (modelica_integer)$PtimeTable$Plast, 1e-13);
    $PtimeTable$Pa = tmp4.targ1;
    $PtimeTable$Pb = tmp4.targ2;
    $PtimeTable$PnextEvent = tmp4.targ3;
    $PtimeTable$Plast = tmp4.targ4;
  }

[asodja]

An important note is also that if {{_Modelica_Blocks_Sources_TimeTable_getInterpolationCoefficients()}} would get called, it would result in a segmentation fault. See bug #1262.

[asodja]

Bug in event-handling is also evident when {{CombiTimeTable}} is used instead of {{TimeTable}}.
In that case, parameter {{t_min}} and {{t_max}} never get initialized properly.
They are initialized in initial-equation section:
{{{initial equation

t_min = tableTimeTmin(tableID);
t_max = tableTimeTmax(tableID);

}}}
But equations from inital-equation section get evaluated before the initial() event is handeled.
In this case, in initial() event {{tableTimeInit()}} is called which return valid value of variable {{tableID}}.

[Martin Sjölund]

These models still do not seem to be working. And we use the MA implementation of the external functions, so it is probably initialisation or event handling that is causing problems.

model ttTest
  Modelica.Blocks.Sources.TimeTable timeTable(table=[0,0; 1,0; 1,10; 2,10; 2,5; 3,0; 4,0]);
end ttTest;

model ttTest2
  Modelica.Blocks.Sources.CombiTimeTable combiTimeTable(table=[0,0; 1,0; 1,2; 2, 2; 2,1; 3,0]);
end ttTest2;

[Lennart Ochel]

Both models are working as expected. You have to simulate more than just one time unit before y is getting different to zero.

[Martin Sjölund]

Ah yes, silly examples to not have StopTime in them ;)