Regarding infinite derivative in Simulink ODE5

I am doing an optimization in MATLAB using CMA-ES. My cost function includes the evaluation of a Simulink model which has a dynamic nonlinear system. The parameter optimization is unconstrained and as a result, sometimes it leads to instability of the model (infinite derivative at t_0). There is no analytical way of defining the upper/lower boundaries of the solution space to prevent this instability. Therefore, I need to address this issue by adding a penalty term to the objective function. However, as soon as getting such problem in function evaluation process (Simulink model Run), MATLAB automatically breaks the optimization and stops the simulation. Is there a way to go around this issue by manipulating the 'Diagnostics' of Simulink to produce NaN or sth instead of an error and break of the loop?