Philippe Fournier
Deputy Director, UP3 Plant
COGEMA, La Hague

Martial Huver
Technical Director
Eurisys Mesures

As part of the new waste management programme launched in La Hague, COGEMA operates a sorting unit in the AD2 workshop which is aimed at optimizing waste conditioning according to activity content. The ultimate goal is to select the type of conditioning most suited to the actual activity of the waste and to reduce the volume of LLW to be stored in surface repositories.

Technological waste arisen from process facilities throughout the site are measured in the AD2 sorting unit. According to measured activity, waste is encapsulated in a concrete container or simply immobilized in a steel drum.

Measurement of a activity is performed by combining passive neutron coincident counting and g scanning tuned for transuranic isotopes. This technology allows setting detection thresholds much lower than 185 MBq of a activity per metric ton of waste, corresponding to less than 1 mg per drum. If the waste activity is below the set limit, then the waste is suitable for conditioning in a steel drum.

Measurement of bg activity is performed by segmented g scanning. bg activity limits below which waste can be conditioned in a steel drum is defined by ANDRA^* specifications depending upon the type of radionuclides contained in waste.

After almost two years of continuous operation over 30,000 drums containing technological waste arisen from the plant workshops have been measured in AD2. The results have been spectacular: 80% of waste previously conditioned in concrete containers have been conditioned in conventional steel drums.

COGEMA La Hague has launched a programme called NGD (Nouvelle Gestion des Déchets) to improve management of solid LLW. It is aimed at providing Quality Assurance tools and procedures for tracing waste throughout the plants and reducing the volume of low-level, short lived technological waste. The techniques and processes developed for the NGD should halve the ultimate volume of conditioned waste. By the year 2000, the volume of short lived technological waste should be less than 0.8 m³, down from 1.4 m³ per MTU when the programme started at the end of 1995.

NGD is based on four basic principles: 1) controlling and eventually minimizing consumable entering controlled areas; 2) zoning facilities according to the type of waste that might be generated in nuclear facilities - two zones are defined, one zone comprising areas generating "nuclear waste" i.e. active waste or waste that can not be classified as inactive, and another zone comprising areas generating "conventional waste," i.e. inactive waste; 3) sorting waste before it exits the facility for treatment and conditioning according to the area of origin and the nature of the waste (aluminum, other metals, objects suitable to incineration, etc.); and 4) controlling and sorting waste on the basis of measured activity contained in the waste.

NGD will be implemented in stages: a dedicated unit is operating since 1996 in the AD2 workshop for centralized waste activity control and sorting; facility zoning according to waste management criteria, waste sorting before it exits the facility of origin and waste incineration will be available by the year 1999.

The stakes of final waste sorting according to measured activity is to optimize conditioning of waste that used to be systematically encapsulated in concrete containers. Within AD2 waste is measured by activity. If the waste is below thresholds values it will be simply conditioned in conventional 213 l steel drums and immobilized by grouting.

The purpose of this paper is to describe the sorting unit installed in the AD2 workshop for waste activity measurement. It comprises two lines, one dedicated to measuring a activity, the other one dedicated to measuring bg activity.

Solid active waste arising in La Hague process facilities, comprising clothes, gloves, tools, etc. are placed in 120 l standard steel drums and sent to AD2 for characterization, sorting and conditioning.

AD2 sorting unit performs waste characterization and sorting. It operates as shown in Fig. 1.

Upon receipt in AD2, drums are identified, controlled and weighed. Each drum is characterized by: the workshop of origin; the nature of waste (vinyl, metal); and the expected isotopic composition (standard spectra have been established for workshops). Drums are identified by reading a bar code. Depending on the workshop of origin and the expected isotopic composition, drums are routed to the a counting line or the bg counting line.

At this stage, before entering counting lines, drums are controlled and weighed. X ray control allows detection of prohibited objects that might have been introduced in the drums such as metals if the waste is to be incinerated, and liquids if the waste is to be compacted.

a activity measurement of drums routed to the a counting line is performed by combining specific g scanning tuned for transuranic isotopes (station 11) and passive neutron coincident counting (station 6). g scanning is used to measure uneven isotopes of plutonium (Pu²³⁹, Pu²⁴¹), Am²⁴¹ and U²³⁵. Passive neutron counting measures even isotopes of plutonium and other actinides as Cm²⁴⁴ that are neutron emitters. This combination allows measuring less than 1 mg of plutonium per drum. Depending on measured a activity content, drums will eventually be conditioned in fiber concrete containers or steel drums.

Results from station 11 also allows detection of potential bg activity above thresholds. In this case the drum is routed to the bg counting line for complementary bg counting.

bg activity measurement of drums routed to the bg counting line is performed by segmented g scanning (station 5). This method measures signal attenuation by sector thus providing very fine measurements down to hundreds of bequerels per drum. Waste with measured activity below the thresholds set by ANDRA will eventually be conditioned in steel drums, others will be conditioned in concrete containers. If results indicate that there might be actinides in a drum, it is routed to the a counting line for complementary a counting.

For final conditioning, 120 l standard drums are compacted, exepted those to be compacted. Five compacted drums are placed in 213 l steel drums (called C0 drums) and immobilized by pouring concrete. If waste requires encapsulation in a concrete container, the 213 l steel drum is placed in a fiber concrete overpack called CBF-C1 container (Fig.2) and fiber concrete is poured for encapsulation.

AD2 sorting unit is fully automated as shown in Fig. 3. Automation allows treating up to 30,000 drums a year with a high level of accuracy. It also ensures traceability of waste according to stringent QA requirements. More than 98% of drums are treated with no operator intervention.

A main in-process computer collects and treats all data for waste characterization: workshop of origin, nature of waste, and expected isotopic composition. As waste is being controlled and measured in the AD2 sorting unit more data are collected: weight, a and bg activity.

The in-process computer controls programmable logical controllers of conveyors, lag storage areas, and conditioning lines. Bar codes reading stations are disposed on the various lines of conveyors for double check. This computer routes the drums to a or bg counting line according to the expected isotopic composition.

Counting stations operate independently. Final activity is computed by an in-process computer (PC40) which collects data from interpreters of counting station (PC11, PC6 and PC5). On the basis of measurement results PC40 evaluates the need for additional counting. Final results are sent to the main in-process computer.

AD2 counting stations are capable of treating up to 30,000 drums a year with a high level of accuracy. This is made possible by achieving very short counting sequences (including measurement interpretation) from less than five minutes (station 5) to no more than thirty minutes (station 11).

AD2 sorting unit was developed by COGEMA in cooperation with CEA and Eurisys Mesures.

Station 11 determines the isotopic composition of plutonium contained in waste, and detects quantities of plutonium as little as 0,02 mg of Pu²⁴¹ representing approximately 0.2 mg of plutonium per drum. The measurement method uses matrix correction by g interrogation and source localization by segmented data acquisition.

Station 11 is equipped with three Ge-HP planar detectors for determining plutonium isotopic composition, and three Ge-HP coaxial detectors for detecting small plutonium quantities. The lower limit of plutonium content for determining accurately isotopic composition is typically 10 mg of plutonium per drum.

The counting sequence consists of:

• preliminary measurement (120 seconds) for evaluating the waste density and performing required QA control;
• main measurement (20 minutes) for spectrum acquisition;
• data processing and filing (around 5 minutes) which include establishing spectrum, calculating activity and isotopic composition, and performing required QA controls.

QA controls are aimed at identifying problems that may have affected counting such as counting time which is not in accordance with specifications.

Thanks to a detection efficiency of 33% of total emitted neutrons, the passive neutron counting of AD2 allows detecting less than 1 mg of plutonium in a drum.

Station 6 is installed in a shielded compartment to eliminate ambient background radiation. It is equipped with 46 He³ detectors surrounding the source to the top (5 detectors), bottom (5 detectors) and circumference (36 detectors), which are connected to amplifiers.

The interpreter (PC6) analyses signals from groups of detectors which permits source localization. It also detects coincident neutrons.

The whole counting sequence lasts about 30 minutes.

Station 5 works according to two dynamic ranges: one tuned for very low activity waste (hundreds of becquerels), the other one tuned for low activity waste (some megabecquerels up to some gigabecquerels). It systematically looks for the 28 g emitters that might be present in the waste (activation products, fission products, Pu²³⁹, Pu²⁴¹, Am²⁴¹, U²³⁵).

It is equipped with three Ge-HP planar detectors for the low activity waste, and three Ge-HP coaxial detectors for the very low activity waste.

The interpreter (PC5) computes the measured activity and the uncertainty for each radionuclide identified in the waste. For those of the 28 radionuclides searched that were not found the detection threshold is provided.

The counting sequence consists of:

• preliminary measurement (30 seconds) for selecting the set of detectors according to the activity of the drum (low or very low);
• main measurement (3 minutes) for spectrum acquisition ;
• data processing and filing (seconds) which include establishing spectrum, calculating activity, and performing required QA controls.

In 1996, 24,984 120 l standard drums were processed in the AD2 sorting unit. Close to 80% of solid waste previously conditioned in fiber concrete drums (CBF-C1) have been conditioned in steel drums (CO) thus resulting in a volume reduction of 3. Table I gives the respective numbers of C0 drums and CBF-C1 containers produced in AD2 in 1994 and 1996. Fig. 4 illustrates 1996 production showing the percentage of C0 drums and CBF-C1 container.

Table II gives the mean activity of conditioned waste produced in AD2 in 1996.

AD2 sorting unit is a unique industrial counting complex for LLW non destructive activity measurement which can treat up to 30,000 drums a year. After nearly two years of successful operation a significant volume reduction has been achieved for technological waste generated in process facilities throughout the plants of La Hague site.