Adjusting reloading and reactivity mechanisms of CANDU reactors for advanced fuel cycles

This research project main objectives are to set up and apply a methodology that can determine the potential of advanced thorium-based fuel cycles in CANDU reactors and that is able to adjust reactivity devices, in such a way as to maintain their reference efficiency for these new fuels. In order to select these fuel cycles, a large alternative fuel envelope is submitted to several discriminating criteria. A coarse parametric core modeling, that takes into account standard reactivity devices, is first used to highlight candidates presenting the best economical performances and to eliminate non viable options. Then, for the best candidates, the neutronic modeling is optimized before considering reactivity devices adjustment. For every reactivity device managed by the reactor regulating system, innovative generic optimization methods are used to achieve specific objectives for every fuel cycle, all of them being based on the reference natural uranium cycle behavior. Specific optimization objectives are assessed by simulating advanced fuel cycle for specific operating conditions, including : normal on-power refueling period, spurious reactor trip and fueling machine unavailability. Unlike the generalized perturbative approach proposed in the OPTEX code, we have successfully implemented a multi-step method able to maximize both the energy extracted from the fuel using an equilibrium refueling optimization, and the reactivity devices adequacy. We also propose new reactivity device supercell models that provides accurate reactor databases for a fraction of the computing cost usually needed using a full model with a similar spatial discretization. Our approach is verified by comparing our simulation results with results published in the literature for the reference fuel cycle. The methodology developed identified advanced fuel cycles, containing up to 60%v. thorium, thereby increasing resources utilization by more than 50% and multiplying the fuel average exit burn-up by a factor of 4.4 when compared with the reference cycle. The reactivity devices were also retained after our optimization processes, requiring only minor modifications to the original design. We determined that a 10%v. heavy water doping of the light water within liquid zone controllers could increase the average exit burnup of the reference cycle by almost 1%, without any adverse consequence to the reactor control. This method is validated through its systematic application to numerous different cases. It demonstrates its capability to achieve very different objectives related to reactivity devices requirements, thus it can be now used for other similar studies. (author)