Forward sensitivity analysis procedure (FSAP) with improved quasi static method for the estimation of dynamic uncertainty in reactor transients

Abstract

Forward sensitivity analysis procedure (FSAP) with improved quasi static (IQS) method is developed and presented here for the estimation of dynamic uncertainty in reactor transients. The propagation of uncertainty in the macroscopic cross section data with time and its effect on reactor power transient is discussed in this paper. The incorporation of the IQS method in FSAP creates a new kind of point kinetics equations and its solution with the shape function gives the dynamic uncertainty in the flux during the transient. This method is applied to estimate the dynamic uncertainty in the power transients in 3D homogeneous reactor, TWIGL-2D and CANDU-2D benchmark reactors. In TWIGL-2D reactor, a Doppler temperature feedback is introduced and this creates an additional internal uncertainty in the core. The core power, core regional temperature and their uncertainties are estimated for the combination of external and internal uncertainties. It is also shown that as the uncertainty in the macroscopic cross section data is increased, the uncertainty in the core power increases. In CANDU-2D benchmark reactor, the dynamic uncertainty in the core power is estimated with uncertainty in the delayed neutron fraction. The estimated uncertainties in the power transient and core temperature are compared with that of direct solution method. The results are found to agree to a good accuracy. From the comparison of results, it is established that this computational method is capable of estimating the dynamic uncertainty in reactor power transient, caused by uncertainties in macroscopic cross section data. The advantage of this new computational method is that the dynamic uncertainty in the power transient can be estimated in a much faster way as compared to the direct solution method and the dynamic uncertainty estimation is represented as a forward problem in time. It is also shown here that using this computational method, both the reactor transient and its uncertainty can be estimated simultaneously.