I'm trying to use HOMER and I couldn't find an option to cap the availability of the wind turbines in the system. For instance in the systems in Alaska, the turbines are down about 20% of the time for maintenance. Can you think of a way to have HOMER incorporate this fact in its calculations? Would scaling any output invaliditate the results?

You could scale the power curve (outside of HOMER) but that would underestimate the peak wind power output and the variability of the wind power.  It would be better to pre-process the wind speed data in Excel and set the wind speed to zero 20% of the time.  You could be as careful as you want to be with timing and length of the 'outages'.