I do have a main suspicion, which is that instead of an analytical solution you have a numerical one based on an Excel spreadsheet, and that your resulting change in airspeed is an artifact of accumulations of rounding errors in the increments. I would please like to ask you to do 2 things:
1. I assume the resolution of your spreadsheet is 1 degree, therefore 360 increments. Can you rerun it with more increments, say half a degree and a quarter of a degree? If the airspeed change trends to zero with an increasing resolution, there's our problem.
Vessbot has it.
Brercrow, you did a VERY rough spreadsheet with
10 degree increments (refer section 6 of his
Post), and then claimed a VERY SMALL change in airspeed as a conclusion. Didn’t that ring any alarm bells for you? Especially considering that your conclusions actually negated Newtonian physics, and if true would have earned you a Nobel Prize.