Ch. Cosemans and A. De Goeyse
NIRAS/ONDRAF

The 'Belgian Nuclear Waste Inventory' gathers all relevant information on radioactive waste, produced or to be produced by the Belgian nuclear industry, into a main database. This database is an indispensable tool for Belgium's radioactive waste management (safety assessment of transport-, interim storage- and final disposal projects, development and planning of future repository installations, ...).

The Inventory provides data on the radioactive waste produced by the seven operational Belgian nuclear power plants, nuclear fuel plants, waste outcoming from the abroad-reprocessing of irradiated fuel, research and development and medical and industrial activities. It distinguishes operational waste from decommissioning waste and non-conditioned waste from conditioned waste. This paper describes the methodology developed by the Nationale Instelling voor Radioactief Afval en Verrijkte Splijtstoffen/Organisme National des Déchets Radioactifs et des Matières Fissiles Enrichies (NIRAS/ONDRAF, the Belgian National Institute for the Management of Radioactive Waste and Enriched Fissile Materials) to achieve this Inventory as well as the main results of this exercise.

As the collection of these data requires a systematic and uniform approach, waste streams have been identified for every waste producer. For each of those waste streams, a questionnaire has to be completed. The structure of this questionnaire defined by NIRAS/ONDRAF reflects the layout of the Inventory which consists of three sections.

The first section covers the radioactive waste volumes. The data included in this segment of the Inventory detail the existing stockpiles of radioactive waste on Belgian territory at a given time and also contain a forecast of radioactive waste arisings in the future. By combining the existing stockpiles of radioactive waste on Belgian territory and the assumptions used in forecasting future waste arisings, the report results into provisional allocation of the radioactive waste to possible repository sites.

The second section describes the chemical properties, i.e., absolute density, material content and chemical composition. The chemical composition specifies the chemical constituents present in the radioactive waste. Data on the quantities of chemical components is provided by way of their respective masses.

The third and last section contains the radiological properties ,i.e., dose rates, activities and isotopic content of the average waste package and the methods used to determine them. The characterisation of this isotopic content focuses on the nuclides that are considered as being relevant for the Belgian final repository projects. Furthermore, the report shows the relative contribution of High Level Waste (HLW), Intermediate Level Waste (ILW) and Low Level Waste (LLW) to the alpha and beta activity in the total waste volume.

Belgium's nuclear industry consists of:

• seven nuclear power plants of the Pressurized Water Reactor-type (PWR) (Doel 1, 2, 3 and 4 and Tihange 1, 2 and 3), generating 5.6 GW(e) of electricity and covering nearly 60% of Belgium's electricity needs
• two nuclear fuel fabrication plants, namely Franco-Belge de Fabrication de Combustible International and Belgonucléaire, producing UO₂- and MOX-fuel respectively
• a central waste treatment facility, namely Belgoprocess
• a reprocessing facility, namely Eurochemic, shut down in 1974, now being decommissioned
• a nuclear research institute (Studie Centrum voor Kernenergie-Centre d'Etudes Nulcéaires), operating a graphite moderated reactor (BR1), a Material Testing Reactor (BR2) and a PWR (BR3), the latter shut down in 1986, now being decommissioned
• a major radioisotope producing facility (Institut des RadioElements);
• a nuclear facility operated by the European Community (Instituut voor ReferentieMaterialen en BMetingen)
• a number of universities and other research facilities
• a medical and pharmaceutical branch and
• other non-nuclear related plants, using mainly sources of radioisotopes.

With the exception of the reprocessing of spent fuel, Belgium is self sufficient as far as treatment and conditioning of radioactive waste is concerned. Reprocessing of spent fuel is carried out by Cogéma in France. All radioactive waste produced during this abroad reprocessing is also due to return to Belgium.

As mentioned earlier, the seven active nuclear power plants are situated at Doel and Tihange. Each nuclear power plant site operates a facility which processes and conditions waste generated by the PWR's primary circuits (i.e., concentrates, ion exchange resins and primary filter cartridges). All other radioactive wastes from all sources, including low level waste produced by the nuclear power plants, are sent to Belgoprocess, which undertakes all further treatment and conditioning.

Current radioactive waste management policies in Belgium envisage two repositories, namely a shallow land burial facility and a deep repository. By consequence, a knowledge of the number of primary waste packages already produced and expected to arise and their chemical and radiological content is vital if any judgement is to be made on the possibilities of final disposal. The Inventory aims to provide this information.

As the Inventory uses a certain waste classification, the following text starts with an outline of this categorizing and describes the methodology which leads to the Inventory. The remainder of this paper features a review of the main results.

After being submitted to a designated process and conditioning method, the resulting conditioned waste is allocated to one of two following groups^{1, 2}:

• conditioned waste characterized by radiological properties, i.e., activity concentrations of the nuclides present and their respective radioactive half lives, which, within a certain time period, are reduced by radioactive decay of the activity concentrations involved, in such a way that the resulting levels are insignificant from a radiological point of view. Furthermore, this time period is compatible with means of controlling this evolution, allowing shallow land burial as a disposal option. Waste volumes that fit this description are designated as category A-wastes.
• conditioned waste characterized by radiological properties, i.e., activity concentrations of the nuclides present and their respective radioactive half lives, which require isolation from the biosphere. At present, this isolation is considered feasible by way of deep geological disposal in stable, virtually impermeable geological layers. Waste volumes which require such a disposal route are designated as category B- or category C-wastes; the heat output generated by the primary waste package determines the category.

A NIRAS/ONDRAF report³ which deals with a shallow land burial facility features a list of twenty nuclides (see Table I). These nuclides are considered relevant for any shallow land burial facility. Each of these nuclides is linked to a volumetric activity concentration. This concentration is based on a safety analysis.

In order to identify the possible final repository of a primary waste package (i.e. a shallow land burial facility or a deep geological repository), the presence of each relevant nuclide has to be quantified. This determination also leads to an activity concentration. The ratio between this concentration in the primary waste package and the maximum admissible concentration is calculated for each relevant nuclide. If the sum of these ratios is less than 1, the primary waste package belongs to category A. As such, it is qualified for a shallow land burial facility.

Category A waste originates from nuclear electricity production, production and utilization of nuclides for pharmaceutical and industrial means, nuclear research activities and decommissioning of nuclear facilities. This variety of waste sources leads to a diversity of non-conditioned waste, varying from ion exchange resins arising from the treatment of liquids passing through the primary circuit of a nuclear power plant to protective clothing and paper towels generally used by all branches of the nuclear industry.

If the sum of the ratios calculated earlier is greater than 1, the primary waste package is allocated to category B or C: it no longer qualifies for a shallow land burial facility. On the basis of the heat output produced by the primary waste package, the latter is assigned to either category B or category C. The criteria used is based on the maximum permissible heat output per unit length of repository gallery.

If the heat output is less than 20 Watt per cubic meter (W/m;), the associated category is B. Furthermore, primary waste packages designated as category B do not impose special requirements as far as stacking is concerned. Primary waste packages generating more than 20 W/m; are allotted to category C.

Category B waste is produced during the fabrication and reprocessing of nuclear fuel, decommissioning of nuclear facilities in general and nuclear power plants in particular. Category C waste comprises vitrified waste and waste containing hulls and endpieces. This waste is produced during the reprocessing of spent fuel. If this reprocessing is not carried out, the spent fuel itself also belongs to category C. Waste of this categorie already in stock consists mainly of vitrified waste which originates from the reprocessing facility Eurochemic. This factory, whose activities ceased in 1974, is situated at the site of Belgoprocess.

Finally, NIRAS/ONDRAF distinguishes a fourth waste category, namely category R. This category contains various Radium (Ra) contaminated wastes produced during uranium mill tailings. According to their nature, these wastes are accumulated in bunkers and immobilized. These bunkers are covered by a multi-barrier system which represents a radiological protection and a confinement for Radon (Rn) gasses produced during the radioactive decay of Ra. These bunkers are situated at the site of the Union Minière factory. Although this storage facility is built according to the prevailing regulations, it is not an authorized shallow land burial facility.

The Inventory⁴ covers the existing stockpiles of conditioned waste and includes forecasts of future waste arisings. Existing stockpiles of conditioned waste include, on the one hand, volumes which have already been taken over by NIRAS/ONDRAF. They are stored in purpose-build facilities on the Belgoprocess site. On the other hand, the existing stockpiles also comprise volumes which, up to now, haven't been taken over by NIRAS/ONDRAF. They are stockpiled either at the Belgoprocess site or on the site of the Doel or Tihange nuclear power plants.

Future waste arisings not only consist of conditioned waste, but also include non-conditioned waste. Non-conditioned waste quantities are either produced during normal operation or during decommissioning of nuclear facilities. They are transformed into a conditioned form assuming a process and conditioning scenario.

Data on existing stockpiles of conditioned waste are provided by a central NIRAS/ONDRAF database. Existing records in this database hold data which cover :

• surface dose rate and dose rate at 1 meter from the package (mSv/h) and date of measurement
• total beta- and alpha activity (Bq) and date of measurement
• masse (kg)
• nuclide content, specifying:

As the collection of data requires a systematic and uniform approach, this operation is planned in five stages :

• Identification of waste streams by the waste producer.
• Completion of a questionnaire for each and every waste stream by the waste producer.
• Evaluation of the information by ONDRAF/NIRAS.
• Transformation of data concerning non-conditioned waste streams into results applicable to corresponding conditioned waste streams by ONDRAF/NIRAS.
• Transfer of resulting data into a main ONDRAF/NIRAS-database.

A waste stream is defined as a homogeneous waste entity, distinguished by its nuclide content and, to a lesser extent, by its physical and chemical properties. As such, a waste stream is defined by the type of raw waste, its activity spectrum and level and its material composition. These parameters impose a particular process and conditioning method which, at the end, determines the type of waste package and conditioning material required. Consequently, a waste stream is univocally linked to one type of waste package and one type of conditioning material.

During this identification phase, a unique code is assigned to each waste stream. This code consists of three elements :

• a code of up to six letters, identifying the waste producer
• a three-digit number, representing the waste stream's serial number and
• a letter, indicating the waste's state (non-conditioned or conditioned).

As a waste stream is built up from waste volumes sharing homogeneous radiological characteristics, each non-conditioned waste stream is univocally linked to a specific processing and conditioning method, leading to a specific conditioned waste package and a distinct type of repository, i.e., shallow land burial facility or deep geological disposal.

Whereas the waste producer represents the source of data and provides all necessary information, NIRAS/ONDRAF developed the questionnaire. The structure of this questionnaire mirrors the layout of the Inventory, covering following aspects :

• General information and waste quantities, including waste type, type of package, physical components and a forecast of waste arisings.
• Physical and chemical characteristics of the waste, including density, material content and chemical composition.
• Radiological characteristics of the waste, focusing on dose rates and nuclide activities.

The waste producers are requested to complete one questionnaire for each waste streams. This questionnaire is available as a computer program, functioning in a Windows environment and providing a detailed explanation of the required answers.

All data provided by the waste producer in the questionnaire is subject to an evaluation by NIRAS/ONDRAF. This evaluation focuses on:

• internal consistency and completeness;
• comparison of the information with data for related waste streams; and
• validity of the information.

During this evaluation process, NIRAS/ONDRAF is assisted by external experts. As the waste producer has to take full responsibility for the answers given in the questionnaire, any inconsistency resulting from this survey requires the waste producer's assistance. This feedback results in a completed questionnaire upon which both parties agree.

Using a mathematical model of the designated process and conditioning method, NIRAS/ONDRAF transforms all data concerning non-conditioned waste streams into information applicable to conditioned waste streams. This model incorporates the following parameters that relate to a distinguished process and conditioning method:

• conditioned waste quantities are calculated on the basis of the corresponding non-conditioned waste quantities using appropriate volume reduction factors;
• chemical and/or radiological characteristics are

As mentioned earlier, the central NIRAS/ONDRAF database contains data concerning existing conditioned waste volumes. All information relative to waste arisings is transferred into a purpose-designed relational database. This database consists of a central database and several "satellite-databases". As was the case for the questionnaire, the structure of this relational database reflects the scheme of the Inventory, which is:

• a central database that holds all general data regarding the waste stream
• a first "satellite" comprising the forecasted waste volumes
• a second "satellite" including the chemical composition of the waste stream and
• a third "satellite" containing the nuclide radioactivities.

The code allocated to a waste stream during the identification phase acts as the link between the central "unit" and the "satellites-databases".

All calculations regarding forecasted waste arising are based on a number of assumptions, giving rise to the "reference scenario". The assumptions consider the following aspects:

• the installations of waste producers, comprising existing facilities as well as future installations
• the associated duration of operations and
• the environmental circumstances.

As far as existing installations are concerned, the reference scenario considers all nuclear facilities mentioned earlier. Moreover, it assumes that no further facilities are built with the following exceptions :

• plants designed to process and condition existing non-conditioned waste volumes
• extensions of storage facilities currently in operation on the Belgoprocess site and
• a new cyclotron on the Institut des RadioElements-site.

The Ra contaminated waste quantities, at present stored at the Union Minière site, are also taken into account.

Finally, the reference case looks at three repositories:

• a shallow land burial facility for category A-wastes
• a deep repository for category B- and C-wastes and
• a pending repository for category R-wastes.

The operational life of all nuclear power plants in Belgium is assumed to be 40 years. This operational time span is followed by a stand-by period of 5 years, after which the decommissioning will take place. Decommissioning activities span a period of 10 years. According to this assumption, the nuclear power plants first put into service (Doel 1 and Tihange 1) in 1975 will be shut down in 2015; the last nuclear power plants put into service (Doel 4 and Tihange 3) in 1985 will be shut down in 2025. Decommissioning activities of all other nuclear installations will be carried out without stand-by periods, leading to a cessation of all nuclear activities in 2085.

All mathematical methods used during calculations are based on present technology regarding process and conditioning methods. It is assumed that all irradiated fuel will be reprocessed. However, present reprocessing contracts only cover 16% of the spent fuel expected to arise. The reference scenario looks at the exemption criteria currently proposed by the International Atomic Energy Agency (I.A.E.A.). It should be noted that, at present, the corresponding release levels are not accepted by the Belgian legislator.

Calculations based on the assumptions used in developing the "reference scenario", lead to a cumulative volume of 100,500 m; ⁵. This amount includes waste volumes in stock on 1^st January 1996 and those predicted to arise after this date. As summarized below, this amount includes waste volumes of categories A, B, C and R. Total waste volumes are given in m;, when conditioned. The arisings of these wastes are illustrated in Figure 1 and Table II.

About 70% of the total waste volume is produced during decommissioning work; the remaining 30% originates from operational activities.

The sensitivity of the calculated total waste volume is investigated by way of alternative scenarios for possible deviations from the "reference scenario". These different investigated alternatives include:

• the cessation of reprocessing of spent fuel, i.e., only the spent fuel already under contract is reprocessed and the remaining irradiated fuel is conditioned and allocated to a repository
• a moratorium on nuclear electricity production within 10 years, i.e., a final shut down of all nuclear power plants in 2007
• a 10-year prolongation of nuclear electricity production
• the construction of a new nuclear power plant after the shut down of an existing one
• the incorporation of waste volumes which are exempted by the "reference scenario"
• a reduction of operational waste production by 25% and
• an improvement of volume reduction factors by 30% due to more advanced conditioning technologies.

All but two of these alternative scenarios lead to a total volume which deviates more than 10% from the "reference volume". The two alternatives resulting into a change of the "reference volume" by more than 10% are:

• the construction of a new nuclear power plant after the shut down of an existing one. This scenario increases the "reference volume" by 45%, highlighting the fact that 45% of the total waste volume produced originates from nuclear power plants; and
• the incorporation of waste volumes which are exempted by the "reference scenario". This scenario increases the "reference volume" by 300%, showing the importance of exemption criteria.

The reference volume consists of primary waste packages. As a general rule, a primary waste package contains following components:

• the raw waste, including any possible inner package and an appropriate lid
• a matrix and
• a primary waste containment, including inner shielding.

The chemical composition of this entirety, shown in Figure 2, constitutes the chemical composition of the reference volume.

Fig. 2. Composition of a primary waste package.

The assessment of the chemical composition of the primary waste packages is carried out by way of a chemical analysis of its components⁶, i.e., the raw waste, the matrix and the primary waste containment. It therefore uses the relative composition of the primary waste package, which specifies the mass of the primary waste package and the fractional masses of the raw waste and the matrix.

Whereas the chemical composition of matrices and primary waste containments are readily available in NIRAS/ONDRAF databases, the raw waste necessitates the method to distinguish between conditioned and non-conditioned waste. As mentioned earlier, the non-conditioned waste requires the use of a model that transforms all data related to this non-conditioned waste into corresponding data for conditioned waste. All calculations are based on data provided by the subsequent questionnaires.

Results are available on a molecular level as well as on an element level. On the molecular level, the chemical composition is specified by the material name and the corresponding mass. A material is designated by either an explicit name (for instance polyvinyl chloride), an ambiguous name (for instance concrete), or a chemical formula (for instance SiO₂). On the element level, all materials are broken down into their element compositions.

Each material also resides one of the following categories:

• cellulosics, this category only contains cellulose, but this material gives rise to complexing agents which can increase nuclide mobility, but also leads to gas formation due to radiological degradation
• metals, a category containing all identified "metallic" components, which are of potential importance to corrosion studies. Corrosion, in an anaerobic environment, leads to gas formation
• oxides, comprising all identified oxides that can influence the sorption of nuclides in the near field
• plastics, comprising all organic mono and polymers supporting studies focusing on possible degradation of these molecules. These degradation products may lead to gas formation or may influence nuclide mobility, although their contribution is small compared to that of cellulose
• salts, including all identified organic and inorganic ions because they can play an important role in several organic and biological processes and
• unknowns, a category that reflects the incomplete or unknown chemical composition of certain waste streams.

Samples are shown in Table III.

As far as the shallow and deep repositories are concerned, the following noteworthy points emerge from this Table:

• the category "oxides" emanates as the most important one from a quantitative point of view, 74% of the total waste mass belongs to this category. Materials like concrete and cement constitute the bulk of this mass. Their overwhelming presence can be accounted for by the following facts:

• the category "metals" represents 19% of the total mass. Metallic materials are not only present as a construction material for primary waste containments, but can also be found in raw waste.

• the category "oxides" emerges as an important one, 35% of the total waste mass belongs to this category. Again, concrete and cement are major materials but, SiO₂ also appears as an important oxide, due to the fact that it forms a significant oxide of vitrified waste streams
• the category "metals" represents 45% of the total waste mass. A major metallic component is Zircalloy-4, which mirrors the fact that waste streams linked to reprocessing activities are of major importance for a deep geological repository and
• the category "plastics" comprises 11% of the total waste mass. An explanation can be found in the fact that bitumen are often used as a matrix in waste streams destined for a deep geological repository.

Total activities in wastes are derived by multiplying the specific activity of the nuclides present in each waste stream by the respective waste volumes, followed by summating the individual nuclide activities. The results of this calculation are shown in Table IV.

For the existing wastes on 1^st January 1996, the specific activities relate to the activities of the waste at that date. For the predicted waste arisings, the specific activities are those estimated to exist at the time of arising. As such, the derived total activities presented in Table 4 do not represent the total activity at any given time, but depict an upper limit. It is not likely that this limit will be exceeded by the actual waste produced at the end of the Belgian nuclear program.

Beta activities clearly dominate the total activity content, representing 97.4% of the total activity. Bearing in mind that only category A wastes are compatible with a shallow land burial facility, 99.5% of the total activity of this repository consists of beta emitters. In a deep geological repository, the latter contribute 97.4%.

Figure 3 shows the evolution of the activity of category A, B and C wastes. This evolution takes the radioactive decay into account and also looks at the waste production scenario in terms of annually produced volumes ^7,8.

As category A wastes originate from various sources, its nuclide content contains a variety of activation products, fission products and actinides.

In current waste arisings, Plutonium-238 (Pu-238) accounts for 32.6% of the total alpha activity. Other major contributors are Plutonium-239 (Pu-239) and Plutonium-240 (Pu-240) (both 13.1%) and Americium-241 (Am-241) (16.3%). Initially, the beta activity is dominated by Iron-55 (Fe-55) (40.2%), Cobalt-60 (Co-60) (20.5%) and Nickel-63 (Ni-63) (25.8%). Due to radioactive decay, the dominance of relatively short-lived nuclides like Fe-55 and Co-60 decreases and the relative contribution of Ni-63 increases to 85.7% after a period of approximately 300 years.

Existing stockpiles of category B wastes emanate essentially from treatment of secondary fluids produced during reprocessing of spent fuel which took place at the Eurochemic factory. This is clearly reflected by the relative contribution of certain nuclides to the alpha and beta activity. The alpha activity consists mostly of Pu-238, Pu-239, Pu-240 and Am-241: as a whole, they contribute 98.0% to the total. The fission products Strontium-90 (Sr-90) and Cesium-137 (Cs-137) and their respective daughters Yttrium-90 (Y-90) and Barium-137m (Ba-137m) dominate the beta/gamma activity, representing 83.1% of the total.

Predicted waste arisings originate mostly from decommissioning activities and contains highly activated components. The dominant presence of Ni-63 clearly mirrors this fact.

Category C wastes mainly originate from reprocessing of spent fuel, the main objective of which is to separate Uranium and Plutonium from other actinides and fission products formed during irradiation. Inherently, from a radiological point of view, the nuclide content is dominated by fission products and actinides.

For predicted waste arisings, Am-241 and Curium-244 (Cm-244) dominate the alpha activity as their relative contribution is around 97.4%. For existing waste, the Cm-244 contribution has strongly diminished as a result of radioactive decay. The Am-241 nuclide governs the alpha activity.

The beta activity in forecasted waste arisings is dominated by Sr-90, Ruthenium-106 (Ru-106), Cs-134, Cs-137 and Cerium-144 (Ce-144). As a whole, their relative contribution is approximately 83.6%. In waste already produced, radioactive decay, diminishes the relative contribution of Ru-106, Cs-134 and Ce-144 in favor of the longer-lived species Sr-90 and Cs-137.

1. L. FROMENT, E. GOVAERT and Ch. COSEMANS, "Classification générale des déchets conditionnés", ONDRAF/NIRAS report 97-3446 (1997).
2. ONDRAF/NIRAS report NIROND 97-04, "Comparaison des diverses options pour la gestion à long terme des déchets radioactifs de faible activité et de courte durée de vie" (1997)
3. ONDRAF/NIRAS report NIROND 94-04, "Le dépot définitf en surface, sur le territoire Belge, des déchets radioactifs de faible activité et de courte durée de vie-synthèse et recommandations" (1994).
4. Ch. COSEMANS, "Inventaris 1996 van het radioactief afval in België-Verklarende nota bij de vragenlijst", ONDRAF/NIRAS report 96-2278 (1996).
5. Ch. COSEMANS and R. VERBEKE, "Inventaire des volumes des déchets radioactifs", ONDRAF/NIRAS report 96-4777 (1996).
6. Ch. COSEMANS, "Chemische samenstelling van het referentievolume geconditioneerd radioactief afval", ONDRAF/NIRAS report 97-2867 (1997).
7. Ch. COSEMANS, "Berekening van de activiteitsevolutie van de afvalinventaris : concept", ONDRAF/NIRAS report 97-3406 (1997).
8. L. FROMENT, "Estimation du spectre d'un site de surface de 60 000 m;", ONDRAF/NIRAS report 97-1342 (1997).