Bad value for interpolation in delayImpl

[ni.ding@…]

The following example using delay operator is not properly handled when there is a state event (please see the attached log).

model example_delay

Real x;
Real y;

equation

x = if time < 0.00003 then 0 else 2*Modelica.Math.cos(2

*Modelica.Constants.pi*60*(time));

y = delay(x, 0.00034);

annotation(

OpenModelica_simulationFlags(lv = "LOG_STATS", outputFormat = "csv", s = "trapezoid"),
experiment(StartTime = 0, StopTime = 0.05, Tolerance = 1e-6, Interval = 5e-05));

end example_delay;

In this example, the ring buffer contains values
storeDelayed[0] 0:0 position=1
storeDelayed[0] 5e-005:0 position=2
storeDelayed[0] 3.00001e-005:0 position=3
storeDelayed[0] 5e-005:1.99964 position=4
storeDelayed[0] 0.0001:1.99858 position=5
…
We have
delayImpl: exprNumber = 0, exprValue = 1.99432, time = 0.00035, delayTime = 0.00034
| | | | | | findTime 1.000000e-005
| | | | | | time(0, 9)[4] = 1.000000e-004
| | | | | | time(0, 4)[2] = 3.000010e-005
| | | | | | time(0, 2)[1] = 5.000000e-005
| | | | | | return time[0, 1] = 5.000000e-005
| | | | | | delayImpl: Linear interpolation of 1e-005 between 0 and 5e-005
| | | | | | delayImpl: Linear interpolation of 1e-005 value: 0 and 0 = 0
Here the value to be used in the interpolation at t=5e-005 should be 1.99964 instead of 0, since the state event has occurred at 3.00001e-005. In other words, storeDelayed[0] 5e-005:0 position=2 should be erased in event handling.

[ni.ding@…]

log file

[ni.ding@…]

I close this ticket since the behavior of the delay operator is correct. The interpolation time value 1.000000e-005 is before 3.00001e-005 the event time. The interpolation is between 0 and 0.