Hello Raymond,
Upon reviewing the "InitFcn" you've shared, one of the probable reasons for the varying outcomes between the initial and later simulation runs is the sequence in which the instructions are executed. Specifically, the parameter initialization “Lls = Ls-Lm” and “Llr = Lr-Lm” happens before the value “Lm” is defined and as a result the first run has corrupted or dummy values for “Lls” and “Llr”.
The simulation should in fact error out the first time as well, since the value “Lm” is undefined, but it is not the case. A possible reason is that the value “Lm” had already been initialized in the base workspace. During the compilation phase of the first simulation, the value of “Lm” is correctly updated to 0.64 resulting the subsequent simulations to yield the correct output.
To resolve the issue, you can change the order of instructions in the “InitFcn” as shown below.
Lm = 0.64;
Llr = Lr-Lm;
Lls = Ls-Lm;
I hope this helps.
