You generally get this error when your simulation has an error that causes one of the states to grow exponentially until one of the derivative terms grows so large that it goes out of bounds (like 10^300). Alternatively, it might have turned into a NaN value. The integrator can't handle either condition, so it throws an error.
Typically there's a block which is miswired leading to an unwanted positive feedback loop or an accidental divide by zero condition. In rarer instances there can be an issue created from a fixed step solver trying to handle system dynamics faster than its minimum time resolution. To fix that one you can try reducing the timestep. I'd recommend you look through your state equations and the terms feeding into the integrator for something out of place. If you can't find it, consider attaching a View/Log to a few key signal wires by right-clicking on them and stepping through the simulation. Past those general tips, we'll need a little more information about your simulation to help.
