You can use QGIS for this analysis but do not confuse line of sight calculations with viewshed analysis. They are related but
r.los returns a raster whose values are null for not-visible or the vertical angle required by the view to see the location. The module
r.viewshed on the other hand returns a raster that simply marks a cell as visible or not (1 or 0). The advantage of
r.viewshed is that, if you have multiple turbines, you can calculate the visibility for each turbine and then simply add all your rasters together and that will give you the number of turbine tips (or hubs depending on the 'observer' height you use) that can be seen at any given location.
BUT be careful! Remember that you are effectively doing things backwards if your 'observer' point is a turbine location. So you must set the target elevation to a normal viewer height (1.75m is a good value, though a value of 2m is often applied). The observer height is then set to the turbine tip height.
You need a decent DEM and I would recommend a resolution of 50m or better (50m is the minimum resolution permissible in the UK - you don't sday where you are based but if it is the UK then I reccomend the OS Terrain50 data set). SRTM is really not good enough away from the equator for serious viewshed analysis of planning-application quality (and therefore credibility).
If you have more than a couple of turbines, using GRASS will quickly become tedious due to all the additional steps, unless you script a solution to add them altogether afterwards. Fortunately there are a number of Windfarm packages out there designed to do just this and OpenWind is a free one which is easy to use. See this thread and this one for more discussion on using GRASS with multiple turbines. Also, here is a thread on how to use OpenWind.