A Zero Dimensional Model for Simulation of TRIGA Mark II Dynamic Response

In this paper the development of a model for the nuclear research reactor TRIGA Mark II operating at University of Pavia is presented. Purpose of the modeling is to reproduce the dynamic behavior of the reactor on the entire operative power range, i.e. 0÷250 kW. A zero dimensional approach is accounted for and the coupling between neutronics and thermalhydraulics in natural circulation is considered. The model has been validated through comparison with experimental data, concerning three different power transients. For neutronics, point reactor kinetics model with one energy group and six delayed neutron precursors groups has been adopted. The system reactivity can be modified moving the control rods, which allow the reactor to operate at different power levels. As far as thermalhydraulics is concerned, two regions have been defined, i.e. the fuel and the coolant. Heat exchange (convective and conductive) has been modeled by proper adoption of a global heat transfer coefficient. This has been considered as a function of coolant mass flow rate through the core to introduce the effects of natural circulation, evaluated using Boussinesque approximation for buoyancy effects. Neutronics and thermal-hydraulics are coupled together by means of fuel and moderator temperature feedback coefficients. The large thermal inertia due to the mass of water in the tank containing the reactor core causes temperature variation during transients to be very small. Therefore, moderator temperature feedback coefficient can be neglected. On the contrary, the fuel temperature coefficient strongly influences the dynamic behavior of the system and has been estimated making a best-fit between the model response and the experimental data regarding positive reactivity insertion in the system at three different power levels, i.e. 1 kW, 50 kW and 100 kW. The results obtained show that the fuel temperature coefficient is a monotonically increasing function of fuel temperature and its magnitude is in agreement with the values found in literature. The nonlinear system of 9 coupled ODE has been solved by means of Simulink® (The MathWorks, Inc. 2008a), which represents a reliable tool for dynamic and control analysis. The model reproduces the real behavior of the system in a very satisfying way: error on the power response simulation is less than 7% for the first transient and less than 1% for the other two transients.