W. Müller, D. Maric, and W. Wurtinger
Institut für Sicherheitstechnologie (ISTec) GmbH
Köln, Germany

Scaling factors or comparable methods to correlate hard to measure nuclides with easy measurable key nuclides are widely accepted as a means to characterise activity inventories in low and intermediate level waste streams from different sources. An overview about this practice in western Europe is given. The data base available at ISTec and its sources, the kind of evaluation and the results are described.

Beyond the common application of the scaling factors they may, however, be used to determine the effects of modifications in plant design and operation or waste management on the transport behaviour of radionuclides on their way from build-up into waste streams and on their activity concentrations in wastes. Examples are given for the removal of Co bearing alloys in PWRs and for the influence of the use of MOX fuel on transuranics concentration in NPP wastes.

Finally, information is provided about the practical implementation of the results in routine waste characterisation and documentation procedures.

Declaration of activity inventories for individual waste packages today requires nuclide specific information for a number of radionuclides, which are not easily measurable from outside a package. To provide the necessary information it is well established to use mathematical methods. These methods are either based on activation or burn-up calculations or on the relation of the nuclides to be determined to easily measurable so called key nuclides. The last mentioned method offers the greatest potential for activity determination of hard to measure nuclides. It is most commonly known as the scaling factor method although different mathematical procedure are used in application of the method.

The method, although designed for activity determination, today offers some more possibilities. It may be e.g. a valuable tool to analyse the effects of changes in design and operation of NPP on the radionuclide composition in waste streams. Inversely, it may be used as an indicator for hidden developments in a NPP due to observed deviations in nuclide composition of wastes.

Finally, all these features and possibilities can be implemented in a waste characterisation and waste flow control system for routine use by plant personnel as well as for sophisticated investigations on physico-chemical effects. The development of the method along this line will be highlighted in the following sections.

The scaling factor describes a relationship between the activity concentrations of two radionuclides occurring in the same waste product or waste sample. The simplest form of this relationship which is commonly associated with the term "scaling factor" is an average ratio between the activity concentrations. Mathematically this can be expressed in the following way:

with

Instead the arithmetic averaging as in eq. (2) other methods like the geometric mean can be used in eq. (1).

A first improvement of the scaling factor method is the evaluation of measurement results by linear regression analysis. This methods provides two fitting parameters instead of one for the scaling factor and a correlation coefficient R as a measure of the quality of the correlation between the radionuclides considered. Both methods rely on a linear relationship between two radionuclides which are correlated. In practice it has become apparent, that this assumption is valid only for a few nuclides over the whole range of measured activity concentrations or over a variety of waste streams.

Since most measurement data are lognormally distributed [1] the geometric mean should be more appropriate than the arithmetic mean. It is therefore used by some authors for activity determination [2]. A more illustrative mathematical equivalent of the geometric mean is a linear regression of the logarithms of the measured activity concentrations. Therefore this method is routinely used at ISTec and has been established as a common method throughout Europe especially in the frame of a project funded by the European commission [3]. Table I lists the mathematical expressions used for activity determination for the methods mentioned so far.

For routine application the available measurement results from different waste samples are evaluated in the manner described above. Obviously, an increasing number of results improves statistical accuracy. At ISTec more than 900 samples can be evaluated coming from different countries and waste streams as Table II shows for European LWR data.

From these data, up to now correlations with acceptable quality can be derived for 30 radionuclides as Table III demonstrates. The results of these correlations are used to calculate the activity concentration of the resp. radionuclide for waste packages, whose key nuclide activity is known. Thus activity inventories can be provided for the majority of radionuclides which have to be declared according to transport regulations or acceptance requirements.

Table IV gives an impression how these correlations and their parameters may look like for a number of radionuclides. The evaluations were based on data from all PWR waste streams from different countries and the accumulated data of all European countries involved.

Differences in the results mainly come from the different number of data available from the individual countries, which cause an accordingly bigger scatter the lower the number of available data. Some of the differences, however, can well be explained by different materials and designs or different operation histories.

With increasing data and knowledge about the plant-specific physico-chemical background of the correlations evaluations can be performed not only for activity determination alone, but also for the analysis of effects of modifications in NPP design and operation. Thus empirical evidence can be provided for effects of e.g.:

• modification of purification or cleaning systems on activity inventories in concerned waste streams
• material changes at primary coolant bearing systems on coolant and waste inventories of dose-rate dominating radionuclides
• changed fuel design or composition on a emitters in coolant and waste streams
• modifications in primary coolant chemistry on dose-rate dominating radionuclides.

Examples will be given in the following sections.

Co 60 is the dominating radionuclide with regard to dose-rate build-up on primary coolant system components in NPP. For German PWRs Co bearing alloys have been identified as a major source of this radionuclides. Therefore these alloys have been avoided in later designs and replaced in some older plants with corresponding reductions of Co 60 concentrations and dose-rate development. In parallel, the primary coolant chemistry has been changed to generally reduce the formation of corrosion products and thus prevent their activation.

The effect of these measures is reflected in Fig. 1, which compares activity concentrations of Fe 55 and Co 60 in waste streams from PWRs of different age. Fe 55 in this illustration represents general corrosion, which obviously has been lowered markedly from the second to the third generation but increased to some extent in the latest so called convoi design. Co 60 was reduced to a comparable extent in the first step, however, even more from the third generation to the convoi plants. Since Co transport is significantly influenced by the available solid fraction of the corrosion products in the coolant the first Co 60 reduction step should mainly be due to the primary coolant chemistry change which lowered this fraction. Thus it is mainly the second reduction step that can be attributed to material changes to avoid Co alloys.

If these observations are transferred to activity determination in waste streams it is obvious that scaling factors or correlations based on Co 60 as key nuclide should change accordingly. Fig. 2 illustrates this for the two radionuclides compared before. The same Co 60 activity is correlated to a higher Fe 55 activity for the latest plant generations since Co 60 concentrations have decreased more than Fe 55 concentrations. Thus not only the effect of the individual measures can be tracked but also related consequences for activity determination can be derived from this experience.

Mixed oxide (MOX) fuel elements have been used in routine operation in German NPPs since 1986. Experience is available mainly from PWRs since in BWRs the use of MOX fuel has been tested only in the last few years. As the MOX fuel elements contain transuranic elements from the very beginning a rise in the radionuclide inventory of higher transuranic elements (e.g. Am, Cm) and the heavier Plutonium isotopes (e.g. Pu 242, Pu 244) in wastes from NPPs which use MOX fuel might be expected. This rise should have influence on the correlation of the mentioned elements to the key nuclides.

The influence of MOX fuel might be superimposed by other effects like fuel defects or outer contamination of fuel elements. Primary information about these parameters is hardly available. They may, however, be related to the plant age. To avoid age-dependent effects in the evaluation again four generations of NPPs were investigated.

As Fig. 3 shows, the activity of transuranic elements is lower in waste from younger nuclear power plants than from older ones. This holds true irrespective of the type of fuel used. The primary explanation of this effect is an obviously decreasing contamination of the outer surface of the fuel cladding for NPP's of later generations. In some cases fuel defects in early years of operation may be an additional reason. The comparison of Fig. 3 and 4 shows that this fact superimposes the effect of MOX fuel elements. Since the difference between operation with and without MOX is in the range of the inherent scatter of data the changes in activity concentration of transuranic elements due to MOX fuel are hardly above the detection limit.

As the activity concentration of Co 60 decreases from older to younger nuclear power plants to nearly the same extent as the activity concentration of the transuranics but because of completely different reasons (see above), the correlations of the transuranic elements to Co 60 remain essentially unchanged.

Today continuous updating of the existing data base enables an increasing differentiation in evaluation and more complex determinations. An example for this is the activity determination by use of the dose-rate. Since most of the dose rate is caused by the key nuclides Co 60 and Cs 137 the dose-rate can be considered as an equivalent to their activity concentration in the waste. For this purpose a simple relation between g emitting activity, dose-rate and the waste mass was developed and qualified by comparison with measurements.

Up to now dose rate calculations for 40 package types, 8 radionuclides and 3 distances have been carried out. Thus, all packages and g emitting radionuclides relevant for radioactive wastes of the German nuclear power plants have been taken into consideration. Based on the g -emitting radionuclides, especially Co-60 and Cs-137, all other radionuclides can be determined using the approach of the previous section.

In addition further methods exist for the determination of special radionuclides or groups of nuclides in certain waste streams: e.g. for ashes, plant specific ratios for Co60 to a emitters are determined. In practice, generally, the various procedures for determining activity inventories are used in combination.

All the methods described have been incorporated as separate modules in waste flow control systems like AVK, AVK-ELA and ReVK developed by ISTec [4]. These systems are PC based and require only key nuclide activities or dose rates, mass and package type for activity determination. Thus the analytical tools in combination with extensive measurements for waste characterization are nowadays used for routine waste activity declaration at the highest possible level with the lowest possible efforts.

Scaling factors and comparable methods are used today on a routine base for activity determination of waste packages. They are in general part of integrated waste characterisation and documentation systems which are applied by waste producers. Beyond that scaling factors meanwhile have a considerable potential to trace modifications in plant design and operation and to detect hidden developments which influence nuclide composition in NPP coolant and waste streams.

1. W. Müller et al., "Calculation of Nuclide Inventories by Using Key Nuclides, 3^rd Intern. Seminar on Radioactive Waste Products, Würzburg, 1997
2. D.W. James et al., "Scaling Factor Derivation and Use for 10CFR61 Radionuclides", EPRI-RP-2414-53, Palo Alto, 1991
3. W. Müller et al., "Development of Methods to Provide an Inventory of important Radionuclides: Analytical Methods and Correlation Data", Proc. of the 4^th Europ. Conf. on Management and Disposal of Radioactive Waste, p. 272, Luxemburg, 1996
4. D. Gründler et al., "Data Processing System for Tracking and Documentation Radioactive Wastes and Residues", ICEM ´97, Singapore, 1997

BACK