[en] The Gas Cooled Fast Reactor (GCFR)is one of the Generation IV reactor concepts. This concept specifically targets sustainability of nuclear power generation. In nuclear reactors fertile material is converted to fissile fuel. If the neutrons inducing fission are highly energetic, the opportunity exists to convert more than one fertile nucleus per fission, thereby effectively breeding new nuclear fuel. Reactors operating on this principle are called ‘Fast Breeder Reactor’. Since natural uranium contains 99.3%of the fertile isotope ²³⁸U, breeding increases the energy harvested from the nuclear fuel. If nuclear energy is to play an important role as a source of energy in the future, fast breeder reactors are essential for breeding nuclear fuel. Fast neutrons are also more efficient to destruct heavy (Minor Actinide, MA) isotopes, such as Np, Am and Cm isotopes, which dominate the long-term radioactivity of nuclear waste. So the waste life-time can be shortened if the MA nuclei are destroyed. An important prerequisite of sustainable nuclear energy is the closed fuel cycle, where only fission products are discharged to a final repository, and all Heavy Metal (HM) are recycled. The reactor should breed just enough fissile material to allow refueling of the same reactor, adding only fertile material to the recycled material. Other key design choices are highly efficient power conversion using a direct cycle gas turbine, and better safety through the use of helium, a chemically inert coolant which cannot have phase changes in the reactor core. Because the envisaged core temperatures and operating conditions are similar to thermal-spectrum High Temperature Reactor (HTR) concepts, the research for this thesis initially focused on a design based on existing HTR fuel technology: coated particle fuel, assembled into fuel assemblies. It was found that such a fuel concept could not meet the Generation IV criteria set for GCFR: self-breeding is difficult, the temperature gradients within the fuel assemblies would be too high, and fuel economy is poor. Two improved fuel concepts are proposed: (1) a redesign of the classic TRISO coated particle fuel, and (2) an innovative hollow sphere design. Both fuel elements are used in a core design based on direct cooling of the coated particle fuel. To increase the neutronic margins and obtain adequate self-breeding capabilities, the proposed reactor has 2400 MWth power output and a power density of 50 MW/m³. With both types of fuel, it is possible to obtain a closed fuel cycle. Long irradiation intervals (several years) are possible with a low burnup reactivity swing, which reduces the required over-reactivity of the fresh core and reduces control rod requirements during operation. In the closed fuel cycle it is important to be able to predict whether a certain initial fuel composition will in fact yield a new fuel, after irradiation, cool down and reprocessing, with which the reactor can be restarted. A theoretical framework is presented in this thesis which allows calculation of the ‘Breeding Gain’ (BG) of the reactor. The BG quantifies the performance of the fuel for batch i + 1 as a function of the composition of the initial fuel of batch i. If this BG can be made equal to zero, both fuel compositions give the same nuclear performance. To be able to calculate the fuel performance, the reactivity weight, i.e. the contribution of each isotope to the overall reactivity of the reactor, needs to be estimated. It is proposed in this thesis to calculate these reactivity weights using a first-order eigenvalue perturbation calculation. It is shown that this approach yields an expression which reduces to a well-established formula for reactivity weights. All steps in the fuel cycle, i.e. irradiation, cool down and reprocessing, have to be taken into account to calculate the Breeding Gain for the closed fuel cycle. First order nuclide perturbation theory provides an efficient method to calculate the effects of small variations of the initial fuel composition on the performance of the closed fuel cycle. The theory is applied to the closed fuel cycle of a 600MWth Gas Cooled Fast Reactor. The result is that the closed fuel cycle can be obtained if the reprocessing is efficient enough in retrieving the transuranics from the irradiated fuel (> 99%). Calculations were done adding extra MA to the GCFR fuel, to estimate the transmutation potential of the GCFR concept. Extra MA in the fuel improve the Breeding Gain, and reduce the burnup reactivity swing. The GCFR core power density is high in comparison to other gas cooled reactor concepts. Like all nuclear reactors, the GCFR produces decay heat after shut down, which has to be transported out of the reactor under all circumstances. The layout of the primary system therefore focuses on using natural convection Decay Heat Removal (DHR) where possible, with a large coolant fraction in the core to reduce friction losses. However, due to the combination of high power density and low thermal inertia in the core, transients in the GCFR core may lead to high temperatures. To protect the reactor under all circumstances during transients, passive reactivity control devices are researched. These devices control the reactor power under off-nominal conditions when all other control devices fail. The proposed devices use liquid ⁶Li as an absorber, which is passively introduced into the core. Activation of the device is by freeze seals, which melt when the core outlet temperature is too high. These devices can be integrated into the normal control assemblies of the reactor while still keeping enough room available for the regular control elements. The passive devices are shown to adequately limit the power production of the GCFR core. It is also shown that natural circulation DHR is possible under pressurized core conditions.