the error model you suggested looks like a combined error model. But from your equation I don't see whether you expect one or two normally distributed random variable(s). How does the error term look like?
If it is (a+b*|f|)*e, then it's exactly SimBiology's combined error model.
If it is (a*e1 + b*|f|*e2), then unfortunately it can't be implemented in SimBiology as SimBiology does not support custom error models as of release R2019a.