Normal probability plot of what? The residuals?
Many SAS procedures produce normal quantile-quantile plots automatically. The Q-Q plot is very similar to the probability plot, but the Q-Q plot is usually easier to read than the probability plot, and makes deviations from normality more apparent.
A "envelope" usually means that you run an analysis many times and look at the 2.5% and 97.5% percentiles of the results. For a Q-Q plot, you might bootstrap the data and store the data for 1000 Q-Q plots. You then plot the appropriate percentiles as the envelope.
I do not know what the term "simulation envelope from the mixed model" means. Perhaps someone else might be able to provide insight.
