I had similar conditions but I think I got it (see attached). I have images taken 10 mins earlier that show stars so I can get a good reference position for the expected azimuth and elevation and it is in the right place. It is also moving at sidereal rate from images before and after this one.
