Consolidation of spent fuel is under active consideration as the U.S. Department of Energy plans to dispose of spent fuel as required by the Nuclear Waste Policy Act of 1982. During consolidation, the fuel pins are removed from an intact fuel assembly and repackaged into a more compact configuration. After repackaging, approximately 30 kg of residual spent fuel assembly hardware per assembly remains that is also radioactive and requires disposal. Understanding the nature of this secondary waste

more »

... stream is critical to designing a system that will properly handle, package, store, and dispose of the waste. This report presents a methodology for estimating the radionuclide inventory in irradiated spent fuel hardware. Ratios are developed that allow the use of 0RIGEN2 computer code calculations to be applied to regions that are outside the fueled region. The ratios are based on the analysis of samples of irradiated hardware from spent fuel assemblies. The results of this research are presented in three volumes. In Volume 1, the development of scaling factors that can be used with 0RIGEN2 calculations to estimate activation of spent fuel assembly hardware is documented. The results from laboratory analysis of irradiated spent-fuel hardware samples are also presented in Volume 1. In Volumes 2 and 3, the calculated flux profiles of spent nuclear fuel assemblies are presented for pressurized water reactors and boiling water reactors, respectively. The results presented in Volumes 2 and 3 were used to develop the scaling factors documented in Volume 1. iii^,M SUMMARY Under the NWPA, the DOE will be accepting spent nuclear fuel for disposal. Included with the fuel will be a significant quantity of activated metal in the form of fuel assembly end fittings, grid spacers, guide tubes, etc. that must be characterized for disposal. If spent fuel assemblies are proposed for disposal by consolidation, spent fuel hardware will then represent a separate waste form. This report presents a method for characterizing these activated metal components for classification within the Federal Waste Management System. The radionuclide inventory in materials irradiated in a reactor is dependent upon several parameters. The major variables are initial material composition, irradiation history, and location in the reactor. In the case of spent fuel assembly hardware, the structural materials are composed of various alloys of stainless steel, Inconel, and Zircaloy. The irradiation history is a function of the burnup of the fuel assembly; the location is governed by the fuel assembly structure. For a specific fuel assembly, the 0RIGEN2 computer code (Croff 1980) can be used to estimate the radionuclide inventory of materials irradiated in the fueled region of the reactor. Input parameters include the composition of the material to be irradiated and the irradiation history. The results of the calculation are applicable to materials irradiated in the core's fueled region. Outside the fueled region, the results are not applicable due to changes in the absolute magnitude of the neutron flux and shifts in the neutron spectrum. For spent fuel assembly hardware, much of the material of interest is located at the end fittings, outside of the fueled region. To estimate the radionuclide inventory in these components. Scaling factors are applied to 0RIGEN2 calculations to compensate for the changed neutron flux outside the fueled region. A total of 38 samples was obtained from three spent fuel assemblies. Each sample was individually analyzed for both elemental composition and for radionuclide content. Based on the results of the analysis, scaling factors were developed that relate the activation rate in the reactor's fueled region V to those outside. These factors are presented in Table S .l. These factors are applied to 0RIGEN2 results by multiplying the fueled-region radionuclide inventories, as calculated in 0RIGEN2, by the appropriate scaling factor for the region in which the material is located. These global factors are only useful for average radionuclide inventory estimates, and should not be applied to small sections due to significant variations in the neutron flux with respect to position. Based on the sample analysis described in this report, it was found that the activation rate of small regions within the top end fitting of a PWR fuel assembly can vary by an order of magnitude. For example, a sample taken from the lower portion of the top end fitting will have 10 times more radioactivity than an identical sample at the upper portion of the top end fitting. The bottom end fitting showed much less variation. It is therefore important to note that these are average scaling factors to be applied to general regions. vi the Westinghouse fuel rods. This difference places the Combustion Engineering top end fitting further away from the fueled region. Hence, it is in a lower flux region than the Westinghouse top fitting. However, once the flux reaches the top end fitting, the flux has an order of magnitude reduction over the end fittings length for both assembly types. The General Electric bottom end fitting has a lower scaling factor than either of the pressurized water reactors. This is apparently due to the greater length of the end fitting and a greater reduction of flux over its length. Calculations were also done in order to estimate these scaling factors. The calculations included using the one-dimensional neutronics code, ANISN. The fuel assemblies were individually modeled, in order to produce assemblyspecific scaling factors. The calculated scaling factors were compared against those that were empirically derived. Results of a comparison of calculations to laboratory analysis indicated that the analyzed samples had higher radionuclide inventories than predicted , the indications ar that the use of one-dimensional neutronics calculations may not be appropriate to determine scaling factors, and as a result, the recommended scaling factors are empirically derived from the sample analyses rather than calculations. The calculations, as shown in Volume 2, indicate that the neutron flux changes significantly outside the fueled region. This is true both in the magnitude of the flux (which decreases rapidly) and its spectrum (as seen in the rapidly changing one-group spectrum averaged cross sections). The presence of a significantly changing flux is shown by the order-of-magnitude reduction in the radionuclide inventory in the pressurized water reactor (PWR) top end fittings from the results of the laboratory analysis. The discrepancies between measured and calculated end fitting results are probably due to the one-dimensional representation of the assembly geometry in the calculations. Calculations with more accurate geometric representation should lessen the discrepancy. It was found that even in the fueled region, where the 0RIGEN2 calculations apply directly, comparing calculational results with laboratory analysis indicated higher concentrations than the calculation predicted, while for others the concentrations were lower. For ^^Co, there was evidence vii of both underprediction and overprediction from different sets of samples. More detailed analysis is needed to resolve the discrepancies. In the interim, there are discernable patterns in how activation rates vary in spent fuel hardware, these patterns, within appropriate uncertainty bounds, can be used to estimate radionuclide inventories for activated hardware and are the basis for the scaling factors presented in Table 1 .1. An interesting fact that was noted in the laboratory measurements was the variation in the minor constituents of the base metals. The level of cobalt and niobium impurity varied significantly in samples of the same alloy. The uncertainty in the values in Table 1.1 does not allow for any uncertainty in the initial composition of the material being irradiated. For cobalt, the level of impurity was found to vary by a factor of two between different samples of Inconel 718. In the stainless steel samples, the same variation was observed. Several of the samples had cobalt levels in excess of 0.1%, an upper limit for many specifications. This would lead to an underestimation of the amount of ^^Co, even if the correct scaling factors were used. For niobium, the variation range was much larger. In Zircaloy, elemental analyses reflected niobium values from below detectable limits up to several hundred ppm. Curiously, ^^Nb was detectable in most of the samples. This is an important result since most items made of Zircaloy are disposed of as low-level waste with no consideration of ^^Nb content. Our analyses showed that levels of ^^Nb were a significant fraction of 10 CFR 61 limits (17% -97%). viii CONTENTS ABSTRACT iii SUMMARY V