I have no solution for your problem, juts a few remarks:
you have five equations but just two variables you allow to be changed. This can't work.
I guess that you would also like d.a1,da2,alpha.a1 and alpha.a2 to be changed. If thats true, you also have to include them in the "find" command even though you may not be interested in their value.
Now we have six variables for Prime to play with but only five equations. Normally this means that you can expect an infinite number of possible solutions and you could add additional constraints like x.1>0mm or the like.
But in case of your system, Prime can find no solution at all. Thats the first moment you may ant to re-check your variables and equations - maybe the problem really has no solution the way you had set it up.

If a system has no solution its always worth a try to replace "find" with "minerr". That way Prime will try to find the best variable combination where the overall errors are minimal. If we do so in your example, the 'solutions' for all six variables are non-real! This also may explain why "find" fails because "find" will only look for real values if you provide a real valued guess.
But I guess that you are not interested in non-real results anyway and expect real values and maybe also positive values as result.

I can think of two possible reasons for the behaviour:
a.) There exist real and positive solutions, but your guess values are way off and so Primes algorithm can't find them. In this case you would have to provide more realistic guesses for all six variables.
b) Your system really can't be solved because some of your equations/constraints are wrong and demand the impossible. In this case you would have to re-check your setup. your equations and see if they really describe the underlying geometry and if there really is a solution to be expected.