From your post,the geometry and temperature distribution (possibly non-uniform)is not clear. If these are known, and assuming this is not a creep problem, you could almost certainly compute the effects using FEA. For these relatively high temperatures, you will also need to know the relation between the coefficient of expansion and temperature, so a non-linear analysis may be required. If the temperature distribution and geometry are simple enough, you may be able to get close enough with traditional methods. If necessary, you may even be able to calculate the temperature distribution using FEA.