Research On The Methods Of Online Doppler Broadening And Nuclear Cross Section Processing

Neutron cross section data is essential to the reactor physics calculations. Thetemperature-dependent ACE data libraries, which are used in continuous-energy MonteCarlo transport codes like RMC and MCNP, are generated by processing code (NJOY,PREPRO, etc.) from ENDF-format neutron data. The treatments include resonancereconstruction, linearization and Doppler broadening. The typical Kernel Broadeningmethod has a high time cost practically unacceptable in schemes likeNeutronics-Themal Hydraulics coupling calculation that needs to generatetemperature-dependent ACE libraries frequently for a large number of materials, whichhampers the improvement of efficiency of Monte Carlo transport calculation.Methodology of online Doppler broadening and cross section processing isdiscussed and fulfilled in this thesis for the purpose of finding an online cross sectiongenerating method with excellent accurary and speed and developing an independentcross section data processing system for Monte Carlo transport calculations. The maincontent is as follow.Firstly, methods of resonance reconstruction, linearization and treatment ofenergy/angular distributions of secondary particles are studied in order to producezero-temperature ACE libraries which are necessary in the following research aboutonline Doppler broadening. A processing code CSP is also developed to finish the job.Secondly, the Online Doppler broadening method based on random sampling andthe FDB fast broadening method based on Gauss integration algorithm and parallelcomputing were both studied. The corresponding codes MCNP5t and FDOPPLER arealso developed. Interpolation of thermal scattering data is implemented to generate ACEthermal libraries at desired temperature in a much faster speed than that of NJOY. Theinterpolating algorithm is optimized to gain better accuracy.Finally, a cross section data processing and Monte Carlo transport couplingmethod is developed to fulfill the online production of temperature-dependent ACE datalibraries. A coupling code system is also developed. A series of benchmarks and modelsare chosen to validate the cross section processing method and coupling system whichclearly improve the efficiency of data processing with good accuracy. According to the above research, the method to generate zero-temperature ACEdata libraries is proved to be accurate. The Online Doppler broadening method based onrandom sampling has given the ability to get ride of special Doppler broadening code,which provides a valuable research direction. The FDB fast broadening method basedon Gauss integration algorithm and parallel computing shows the same level ofaccuracy with Kernel Broadening method and a much higher speed, which meets therequirements of the online production. The cross section data processing and MonteCarlo transport coupling system improves the efficiency of continuous-energy MonteCarlo transport calculation tremendously and achieved the desired target of this thesis.