عدم اطمینان و آنالیز حساسیت در تولید پارامترهای نئوترونیک برای شبیه سازی های هیدروترمال-نئوترونیک BWR و PWR توأم

The Best Estimate analysis consists of a coupled thermal-hydraulic and neutronic description of the nuclear system's behavior; uncertainties from both aspects should be included and jointly propagated. This paper presents a study of the influence of the uncertainty in the macroscopic neutronic information that describes a three-dimensional core model on the most relevant results of the simulation of a Reactivity Induced Accident (RIA). The analyses of a BWR-RIA and a PWR-RIA have been carried out with a three-dimensional thermal-hydraulic and neutronic model for the coupled system TRACE-PARCS and RELAP-PARCS. The cross section information has been generated by the SIMTAB methodology based on the joint use of CASMO-SIMULATE. The statistically based methodology performs a Monte-Carlo kind of sampling of the uncertainty in the macroscopic cross sections. The size of the sampling is determined by the characteristics of the tolerance intervals by applying the Noether–Wilks formulas. A number of simulations equal to the sample size have been carried out in which the cross sections used by PARCS are directly modified with uncertainty, and non-parametric statistical methods are applied to the resulting sample of the values of the output variables to determine their intervals of tolerance.

Best-estimate computer programs make use of the best physical models and numerical solution methods available to simulate the behavior of nuclear power plants. It is well known that their results are affected by the uncertainty in the methods and the models, and in order to draw proper conclusions from them, it is necessary to apply methodologies for the propagation of uncertainty so that it can be quantified. When the Best Estimate analysis consists of a coupled thermal-hydraulic and neutronic description of the nuclear system's behavior, uncertainties from both aspects should be included and jointly propagated. This paper presents a study of the influence of the uncertainty in the macroscopic neutronic information that describes a three-dimensional core model on the most relevant results of the simulation of a Reactivity Initiated Accident (RIA) which are part of the accident analysis for the licensing basis. In this paper, the uncertainty and sensitivity analysis in a BWR NPP core configuration in a CRDA (Control Rod Drop Accident) and a PWR NPP core configuration in a REA (Rod Ejection Accident) are presented. These accidents are caused by the failure of the driving mechanism of a control rod. As a consequence, a continuous reactivity is inserted in the reactor which has to be compensated for by other reactor feedback mechanisms in order to maintain the values of its safety variables within the regulatory margins. The physical description of the reactor response is based on the coupled neutronic–thermal-hydraulic systems analysis program standard in the industry TRACE (NCR, 2007)/PARCS v2.7 (Downar et al., 2004) and RELAP5 (RELAP5/MOD3.3 Code Manual, 2001)/PARCS v2.7 respectively. The data needed for the complete neutronic description of the core behavior, the cross-sections sets, are obtained by using the CASMO4 (Knott et al., 1995)-SIMULATE3 (Cronin et al., 1995) computer package and processed according to the SIMTAB methodology developed at the Polytechnic University of Valencia (UPV) together with Iberdrola (Rosello, 2004). In order to qualify the results obtained using these cross-sections sets, the model's results for steady-state conditions have been compared to the results of a stand-alone core simulation performed with SIMULATE3, a standard and extensively validated core analysis tool in the industry. Specifically, the core power axial profile and the nuclear integral parameter keff, that indicates its critical state, have been compared yielding very close results, thus validating the model employed in the paper. As in the case of PWRs, the transient is started by the ejection of rod with the maximum reactivity worth until it is completely extracted in 0.1 s. A continuous reactivity insertion drives the transient behavior of the reactor. The Doppler reactivity feedback mechanism resulting from the increase of the fuel temperature counterbalances the positive reactivity insertion due to the extraction of the control rod and finishes the transient by drastically stopping the power increase and then bringing it down to levels below nominal steady state conditions before the fuel enthalpy reaches values close or above the safety limits (170 cal/g or 711.756 kJ/kg) (OECD/NEA, 2001). As a result, the amount of energy deposited in the nuclear fuel stays below the maximum values accepted as safety limits, and assuring that the fuel will not be damaged.

In this paper we have presented the results from an uncertainty and sensitivity analyses of a RIA in a BWR and in a PWR NPP. In the BWR case, the results show that increasing the uncertainty, there is a spread in the time at which the power rises and falls, but the maximum value is not greatly changed, even for uncertainty as large as 10%. The uncertainty analysis in the PWR NPP case has shown that variations about 1% have a higher influence in the output variables of interest, e.g., higher peak power and lower time at which it is reached. Variances of 0.1% have a smaller influence in the output variables but they cannot be neglected from the safety point of view. In both cases, the sensitivity analyses have shown that the most influential uncertainties correspond to the fast diffusion coefficient (1), which determines the leakage, the scattering cross section (3), which determines the moderation, and both fission cross sections, which determine the rate of fission power release. The uncertainty in the absorption cross sections appear to have very little influence in the output variables considered. Comparison of the type of pdf used has shown that a normal pdf is giving results in more “conservative” direction from the point of view of maximum peak value and spread of the time of the power transient, if compared to an uncertainty quantified with a uniform pdf.