This thesis is centered on the foreseen realization and performance of a proof-of-principle cavity based (hard) X-ray FEL (CBXFEL) demonstrator experiment at the European XFEL facility. A CBXFEL promises to address the prominent issue of longitudinal coherence and (stable) high, narrow bandwidth spectral flux in the hard X-ray regime, in which the usually employed self amplified spontaneous emission (SASE) scheme is severely lacking. In order to study the highly coupled system of FEL production, X-ray propagation and the crystals’ thermal response, affecting the reflection characteristics, a computational framework was set up. It chains the popular Genesis-1.3 FEL program with the self-written, highly optimized parallel X-ray Cavity Propagator (pXCP) wavefront propagation code and a finite element (FE) based modeling of the strongly non-linear thermal diffusion. In order to properly account for low-temperature thermal transport with an increased relevance of phonon boundary scattering, thermal conductivities obtained from first-principles simulation are used. Thorough simulations are carried out, which account for realistic electron bunch distribution, inter RF-pulse bunch fluctuations and various possible errors of the X-ray optics. They reveal that with well inside state of the art optical tolerances, a simplistic two crystal backscattering setup would fulfill the main goal of the demonstrator, which is to proof that seeding and exponential radiation build with spectral narrowing occurs. However, due to the strong heating of the crystals and the following thermoelastic response, stable operation at high peak brilliance will not be feasible. Following the principle, experimental nature of the CBXFEL demonstrator setup, these effects will need to be properly measured. Using this data, counter measures can be developed towards the future realization of a permanent CBXFEL source at the European XFEL facility.