SUMMARY

Gen IV reactor designs show promise in providing safer cleaner and potentially cheaper options for electricity generation. The Advanced High Temperature Reactor (AHTR) is a Gen IV reactor design that can provide cheaper electricity costs safely due operation at atmospheric pressure and increased thermodynamic efficiency due to higher temperatures. This new design concept is challenging to model, making design optimization more computationally expensive. A new methodology for design optimization of double heterogeneous fuel in the AHTR is evaluated in this dissertation. The approach is to apply valid approximations to the neutronics calculations allowing for a practical analysis of the design space. Then, using advanced sampling techniques and artificial neural networks, surrogate models are created to generate the constrained objective function for optimization. A novel optimization algorithm was developed to efficiently find the optimal design and the region of solutions near the optimal design in a concave nonlinear design space. This approach provides a rigorous design optimization search and characterizes the sensitivity of the solutions near the optimum with regards to the features of the model. This is helps with understanding how optimal costs are affected by changes in external factors.