Central worth discrepancy

The paramount fundamental problem to resolve is that of the central worth discrepancy. It is important not only for its academic interest or for reassuring the reactor physicist in the adequacy of the calculational methods and nuclear data he u.ses, but also for practical reasons. Safety factors that account for the CWD must be included in the design of fast reactors an economic penalty is associated with these safety factors. Related to the question of the CWD is the more general problem of the multigroup calculation of bilinear functionals. The questions arising here are (I) for which applications, and under what conditions, are spectral fine structure effects non-negligible (2) how can these fine structure effects be taken into account in the multigroup formulation. 

All of the anatyses were based on two-dimensioBal diffusion theory, and used ENDF/B-68 cross sections. One may conclude from comparison of experiment and analysis, that the two-zone system can be calculated as accurately, if not as simply, as a reactor with a single core region. The primary difttculties (e.g., central worth discrepancies) are the same as those encountered in the simpler cores. 

A definite trend appears in the U-23S and Pu-239 central worth figures. For hard spectrum cores the calculated values fall below the measured, while for soft spectrum cores the calculated worth Is considerably higher. The discrepancy is consistent for U-235 and Pu-239. 

In general, the calculated ENDF/B and measured reactivity worths, as well as the neutron spectra, compare reasonably well. However, central worths for several ENDF/B materials exhibit substantial deviation from measured results and a reevaluation of the cross-section data may be justified. Recently, detailed comparisons of calculated ENDF/B and GAM data with measured results in the Thermionic Critical Experiment corroborate many of the large discrepancies found in the present study. 

Hie principal results of this study are given in Table 1. Especially evident in these results is the improvement over previous ENDF/B versicms of the and central in-cell reaction rate ratios. The ceidral worth discrepancy is still present, however, although the agreement with experiment shows a slight inqirovement. Table 1 also provides new lh/% Ak/k conversicm factors for the assemblies vdiich are necessary for central worth comparisons. These values are based cm the Keepin delayed-neutron data and shoidd be taken to sq iersede previously quoted values. 

Sodium whs progressively removed from a central zone in Assembly-51. The zone radii and heights were selected to allow an approximate identification bf the zone over which the sodium worth is positive. Calculations of the worths of sodium removed are less negative or more positive than measured, as is shown in the table. While the discrepancies between calculations and experiments are large, these results and, hence, early conceptual design values for loss of sodium, are conservaUve. 

Central perturbation worths of Pu and U are overcalculated by 30% here, as has been the case in analysis, of many earlier experiments. While no explanation of these discrepancies between perturbation calculations and experiments can be given yet, it is noted that significantly better agreement is obtained for larger core samples whose worths arb primarily due to Pu and U. 

---