GNS HAS DEVELOPED A QUIVER system to remove damaged fuel rods from PWRs and BWRs (see NEI January 2018 p28) and establish a disposal concept meeting the requirements of German utilities.

The first hot campaign took place in October 2018 at Unterweser (a 1410MW PWR) in Lower Saxony. The campaign comprised handling and dispatch of three Quivers, to complete removal of the remaining spent fuel from the spent fuel pool.

The first campaign

The first hot campaign was originally scheduled for Germany’s Biblis A. This campaign was postponed and the campaign at Unterweser was brought forward for regulatory reasons. Before the first dispatch campaign at Unterweser could start in October 2018, an extended work programme had to be successfully completed. This comprised loading the damaged fuel rods into the Quivers, as well as installation and site acceptance testing of the complete dispatch equipment1.

The handling of the Quivers takes place at two different levels inside the containment. A loading station is positioned in the spent fuel pool, while a service station is at the reactor floor outside the pool.

The dispatch of the first Quiver started at Unterweser on 12 October 2018 and was completed on 21 October. The drying process lasted about six days. The maximum dose rate at the service station was less than 70μSv/h.

The major results of the first three dispatch cycles are:

  • The qualified processes for handling, drying and welding are robust and reliable.
  • The ‘out of pool’-handling results in very low radiation exposures for the service personnel.
  • It is feasible to dry damaged fuel in an industrial process on-site.

Drying damaged fuel

One of the major challenges during the encapsulation of damaged fuel is drying fuel that might be entirely soaked with water, due to cladding failure during reactor operation or during storage in the spent fuel pool. During qualification and certification of the Quiver system, the drying of damaged fuel rods was simulated. It used test rods with Al2O3 powder representing the fuel and with test leaks of different sizes representing different cladding failures (pin hole, hairline crack, etc).

In parallel, some experiments in drying sections of real damaged fuel were performed in a hot cell at SCK•CEN, Mol, Belgium. However, the first hot campaign had to prove that drying damaged fuel is feasible with the planned setup, and determine to what extent it behaves similarly to the hot experiments at SCK•CEN, Mol and to cold experiments earlier carried out at the Höfer & Bechtel test site for qualifying and certifying the drying procedures and equipment.

Drying the real damaged fuel turned out to be comparable to drying the test rods that were used for qualification and certification, as regards the time required and the pressure curves that monitored the drying process. The main differences were radiological aspects, which could not be tested with the original equipment beforehand and only to a limited extent at SCK•CEN, Mol. 

The drying process is heat-assisted vacuum drying (shown right) and it is monitored by different pressure gauges. The pressure inside the Quiver is monitored by a very robust pressure gauge. Between this first pressure gauge and other more sophisticated pressure gauges, a very tight filter of sintered metal is placed to absorb potential radioactive aerosols evaporating from the Quiver. Next to that filter is a tunable laser absorption spectroscopy monitor (TLSM) that monitors the partial pressure of water in the gas stream. There is a high precision capacitive pressure gauge in parallel.

This is followed by a measuring aperture and another high-precision capacitive pressure gauge.

The mass flow rate of water (the drying rate) is calculated by combining the pressure difference of the two pressure gauges, before and after the measuring aperture, with the partial pressure of water.

By integrating the mass flow over the drying time, the amount of water removed from the Quiver and the fuel inside is calculated. The calculated values were in accordance with the amount of water drained from the condensers of the vacuum pumps.

Figure 1 shows the course of the most relevant parameters during drying of the first Quiver at Unterweser.

  • Yellow: the temperature of the heating air entering the heating path of the Quiver.
  • Green: shows the temperature at the outlet side.
  • Red: the pressure measured by the capacitive pressure gauge before the aperture.
  • Blue: the pressure, measured by the TLSM.
  • Purple: the calculated water-mass flow.
  • Black: integration of the mass flow, i.e. the removed amount of water.

As can be seen from the graph, the first stage of the drying was performed at temperatures of about 100°C. In this stage, free water that was taken on during loading (under water in the spent fuel pool) and remains in the tubes containing the fuel rods is removed. This first stage lasted about two days and about 11 liters of water was removed. The pressure curves of the TLSM and of the capacitive pressure gauge, although located at the same place in the system, show quite different values during this first stage of drying. The difference between the two readouts is related to the fact that the TLSM was optimised for performance at pressures below 300Pa (3mbar). This can be seen in the diagram for the later stage of the drying process.

After removing the free water from the tubes of the Quiver, the pressure and mass flow decreases sharply (see end of 14 October). After reaching a pressure of 50Pa, the heating temperature is increased to >150°C at the outlet side. The increase in temperature is followed by a short pressure increase. When the pressure is back down to 50Pa, the measuring aperture is switched to a smaller one to enhance the resolution, in order to calculate the mass flow.

According to the theoretical models and also according to the results from the Mol experiments, at the second stage of the drying, at a higher temperature, the wet fuel itself is dried. As can be seen from the diagram, the amount of water removed during that second stage of the drying is around 100ml.

According to the certified drying process, dryness is reached when a pressure buildup test shows a value for the water partial pressure, measured by the TLSM, below a defined threshold. The pressure build-up test is performed with heating in full operation and without vacuum pumping for 30 minutes. The threshold is derived from an internationally accepted value of 400Pa for a 30-minute pressure build-up test. The threshold is further reduced with respect to the volume of the Quiver and the vacuum system integrated. The precondition to enter the pressure build-up test is that the mass flow of water is below 1g/d.

Figure 1 shows two pressure build-up tests that were performed on 17 and 19 October 2018. The first was performed with the mass flow slightly above 1g/d, and the second when the mass flow was below 1g/d. The total pressure, monitored by the capacitive pressure gauge, rises higher than the water partial pressure, measured by the TLSM. This is as expected, since the total pressure is also influenced by small leakages in the vacuum system, while the TLSM really shows the pressure build-up by evaporation of residual moisture. The final pressure buildup test is shown in more detail in Figure 2.

The difference in the two pressure curves at very low pressures in the diagram above is due to a slight offset of the zero point. (The difference is about 3Pa or 0.03mbar.)

Radiological aspects

The radiological aspects of drying damaged fuel could not be fully tested with the Mol experiments. Some information was also collected from other hot cell experiments with nuclear fuel.

The accessible radioactive gases inside the failed fuel rods were extracted by applying vacuum to these rods. The amount of radioactive gases that would be released from the failed fuel rods was calculated. As the worst case, it was assumed that all accessible radioactive gases would be released from the failed fuel rods in very short time after applying vacuum to the Quiver. Of main interest were the gaseous nuclides H-3 and Kr-85, which have long half-lives and are still in existence after the typical decay time of the damaged fuel rods. While H-3 typically evaporates, almost all of it subsequently condenses in the condensers of the vacuum pumps. Kr-85, as a noble gas, will not be retained and will leave the vacuum pumps via the exhaust gas line. A monitor measuring radioactive gases is included integrated into the vacuum pump unit, at the exhaust of the last vacuum pump. The plant also installed a noble gas monitor, taking samples from its exhaust gas line. Finally, the vent stack noble gas monitor was also monitoring how much of each radioactive gas was released.

Figure 3 shows the measured count rate of the monitor of the vacuum pump unit during drying of the first Quiver at Unterweser. Radioactive gases were released during the whole drying period, with the maximum rate during the first stage in the first two days. So the assumption that the total amount of accessible radioactive gases will be released during a very short period after applying vacuum to the Quiver turned out to be conservative. The total amount of released radioactive gases was below the calculated values but within the expected order of magnitude.

Another radiological question was whether other volatile radioactive substances would evaporate from the damaged fuel during the drying process. Quite a number of studies have been published on the release rates of fission products from spent nuclear fuel in aqueous conditions3-6.

It is generally understood that the release rates are highest for fission gases, followed by iodine and cesium. Other fission products that are mobile in the in-reactor stage and accumulate on grain boundaries or in the fuel cladding gap (e.g. Sr, Rb, Mo) are released somewhat faster than the uranyl ions that are dissolved on corrosion of the UO2 matrix. Lanthanides and actinides dissolve congruently with the UO2 matrix. To avoid possible contamination of the vacuum equipment, a filter of sintered metal is placed in the vacuum line behind the Quiver, in front of the high precision pressure gauges. It was expected that all radioactive aerosols escaping from the Quiver would be deposited there so this filter was continuously monitored using a dose rate probe.

Figure 4 shows the dose rate at the filter over the drying time. The dose rate at the filter increases mostly during the first stage of drying and is steady when there is almost no moisture passing the filter. Further investigations of the filter after the drying showed that the contamination was almost all Cs-137. It seems that a small amount of the water-soluble Cs-137 is evaporating and contaminating the vacuum system ahead of the filter. Behind the filter, almost no contamination was detectable.

Most of the contamination is trapped in the suction head, directly on top of the Quiver. The suction head was decontaminated after every drying cycle and the filter was replaced after each Quiver.

The radiological aspects were not fully known before the first dispatch of the Quiver at Unterweser but in the end, they did not cause any real problems. All precautions taken during development of the dispatch equipment were effective.

Tests at SCK•CEN

In parallel to developing the quiver described above, a series of tests were performed on samples from two fuel rods in Tihange 1, Belgium (irradiated to a burnup of about 50MWd/kgU) to study hydraulic resistance and vapour flow through the samples. One rod failed during operation, with perforated cladding. From both the failed and intact rods samples of different length (50cm, 17cm, 10cm) were prepared, by cutting the fuel attached to the cladding by bonding. The samples were tested in a device comprising two instrumented containers, with the fuel rod sample between them and the connection between the two sealed in such a way that any gas or vapour flow had to pass through the clamped fuel rod segment (see Figures 5&6).

The test setup allows the following tests:

  • Hydraulic resistance for dry gas flow and vapour flow
  • Wetting – drying sequence
  • Water-pocket drying.

Before embarking on real spent-fuel rod segments, mock-up tests were performed using a 50cm solid bar with a machined central channel, and a segment of identical length filled with fine Al2O3 powder and sealed at both ends with a sintered metallic filter plug. The hydraulic resistance of these test samples was low enough to establish a measurable gas flow. Hydraulic resistance on these mock-up samples was measured by establishing a flow of Argon by constant pressure difference between the upper and lower container, resulting in 640μm and 195μm hydraulic radius. The hydraulic resistance for the irradiated rod samples was much higher than that of the mock-up samples and the gas flow was below the measurement range for the flow meters.

Instead of using flow measurements, the hydraulic parameters were determined by measuring the rate of pressure change in either of the two vessels as a function of pressure difference over the sample. Under ideal gas behaviour and considering a situation with the pressure in vessel 1 (top) being higher than in vessel 2 (bottom), a mass flow will be established from vessel 1 to vessel 2 which is expressed as: 



Under conditions of laminar flow, the molar flow rate Qm(t) can also be expressed as: 



In these expressions, Qm(t) is the instantaneous mass flow rate (expressed in g.s-1) through the segment, P1(t) and P2(t) are the bottom and top pressures as a function of time, V1 is the volume of the top vessel, r is the radius for an effective capillary for the gas flow path, η(T) is the dynamic viscosity of the gas at temperature T (eg Ar, air or H2O), M is the molar mass of the considered gas, L is the flow path length, R is the universal gas constant. From (2), the effective hydraulic radius can be readily calculated: 



The hydraulic radii of the different irradiated rod samples were between 85μm and 110μm.

A complete wetting and drying sequence consisted of inserting an excess amount of water in the lower vessel such that the lower part of the fuel rod segment would be completely immersed (see the dashed line in Figure 6). The gas cushion above the water was then pressurised such that the sample segment was progressively filled with water until the moisture readout in the top vessel indicated the presence of liquid water i.e. full percolation occurred. The system was then soaked for a minimum period of 2 hours to allow finer cracks and gaps to be wetted as well.

The lower vessel was then drained and top and bottom vessels were heated to a preset temperature while being pumped. During the pumping sequence, the pressure was monitored, as well as the moisture content in the exhaust line. After reaching pressures below 100Pa (1mbar) in both top and bottom vessel, a pressure rebound test was performed. To this end, the exhaust lines were shut and the pressure increment was monitored for 30 minutes. If the pressure did not exceed 400Pa (4mbar), the test was considered complete.

The drying sequence, plotted in Figure 7, clearly showed several phases: in the first phase the pressure rapidly dropped to ~10mbar, at which point the pressure stabilised while liquid water was slowly removed from the fuel column. The humidity in the exhaust lines remained elevated (dew point 10-20°C). Once the liquid water was removed from the segment, the pressure and humidity dropped gradually.

Given the performance of the pumping system, the vacuum was expected to asymptotically approach ~0.5mbar. In the example shown in Figure 7, a first near-successful dryness demonstration test was reached after around 6 hours. Upon further drying, the pressure and humidity gradually evolved to 0.3-0.4mbar and 40 °C, respectively. A successful dryness test was performed after 24 hours. Further drying did not result in any significant changes in vessel pressure or relative humidity of the exhaust gas. The test was concluded after 96 hours with a third dryness test, which was again successful.

The wetting and drying sequence yielded a successful demonstration of the feasibility of the drying principle but was difficult to quantify. Water removal rate was quantified using two methods. First the hydraulic resistance of a fuel rod segment was assessed under dry conditions (see above), and in a second stage, “water pocket tests” were performed at different temperatures. To this end, 10ml of water was inserted in the top vessel which was sealed, the whole system was heated and pumping was performed from the bottom vessel. During such tests, the bottom vessel pressure and relative humidity were continuously monitored (Figure 8). The top vessel “water pocket” dried out after several hours (in Figure 7 there was no water- pocket). Depending on the drying temperature, the drying time was shorter or longer and correspondingly, the lower vessel pressure was at a higher or lower equilibrium during the drying process (~4mbar for 3h when drying at 130°C and ~2.5mbar for more than twelve hours when drying at 110 °C). 

Figure 8 also illustrates that after reaching the high temperature, the bottom vessel pressure experiences a stable level for several hours, after which the pressure gradually drops to a low level, asymptotically evolving to 0.3mbar. The relative humidity level approaches -40 °C. The solid lines are the times at which the pressure dropped to half of its initial value and was used to determine the removal rate. The dashed lines denote the times consistent with the theory according to equation 1.

From the same water pocket drying experiments, one can determine vapour flow rates by shortly closing the valves of the bottom chamber and monitoring the instantaneous pressure increment.

The mass flow rates for water vapour for four different rod segments are shown in Figure 9. They are determined by interpreting the time needed for the pressure to drop to half of its initial value to be the time required to remove 10g of water. Mass flow rates are calculated from the hydraulic radius as derived from the dry hydraulic resistance measurements (Figure 9). There is excellent agreement between the two approaches. The slight offset is attributed to the above described method to determine the drying time (solid vs. dashed lines).

This shows that the Equation (3) is applicable to vapour-removal and the full-scale situation can be assessed. The plenum of a PWR fuel rod holds typically 25 cm3. The worst case situation is a cladding breach on one end of the fuel rod where the filled plenum has to be emptied through the full fuel stack. Hydraulic radii depend on irradiation history, as well as fuel-chemical reactions during storage with fuel-water contact. Once liquid water is removed from a water pocket, its pressure will further gradually decrease until the pressure difference is no longer sufficient to drive the gas through the gap and cracks (ending laminar flow).

In the quiver 32 fuel rods are dried simultaneously. At the beginning each rod with a cladding breach has a water pocket at one end with saturation pressure corresponding to 155 °C and vacuum at the perforation at distance L with a hydraulic radius r. The rod with the worst connectivity will take longest to dry.

In Figure 1 155°C is applied, starting at 3:00am on 15 October 2018 after all the water outside the fuel rods in the quiver has evaporated, the sharp rise in exhaust-line pressure at 6:00am indicating that the fuel has reached 155°C. Around twelve hours later, the pressure steeply drops and settles at around 50Pa (0.5mbar), with a water removal rate around 20g/day, and these values stay stable for nearly 12 hours after which they drop further to asymptotically reach P ~ 20Pa (0.2mbar) and q ~ 1 g/d.

The pressure drop observed after 24h can be interpreted as the water pocket in the worst connecting fuel rod being emptied of liquid water. According to Eq. (2) this could be a rod similar to the one that is indicated by the red line in Figure 5 (r=80μm, L=4m) for which removal of 25 cm3 of water at 155 °C would need about 30h.

The IAEA criterion on dryness is an isolation test showing a pressure increase of less than 4mbar = 400Pa within 30 min. The first isolation test was performed after 6:00am on 17 October, displaying a pressure increase of 250Pa, i.e. dryness according to IAEA. Dryness can in any case be assumed from the pressure level and evolution after 6:00am on 16 October. 

The six-day drying time comprises two days to empty the quiver, one day drying the fuel itself and three days to reach the specific criterion of 1g/d vapour-removal rate, after which an isolation test is deemed credible.


After the first successful hot handling of PWR-Quivers inside a nuclear plant the next milestone will be the cold trial and afterwards hot handling of BWR-Quivers. Licence approval for the BWR-cask including the BWR-Quiver has already been issued. Hot handling of the first BWR-Quivers started in May 2019 and at the end of 2019 we expect to load a BWR cask with BWR-Quivers.  


1  S. Bechtel, W. Faber, H. Hofer, F. Juttemann, M. Kaplik, M. Kobl, B. Kuhne, M. Verwerft, The German Quiver Project – Quivers for Damaged and Non-Standard Fuel Rods, ATW-Int. J. Nucl. Power 64(3) (2019) 151-159.

2  IAEA, Management of Damaged Spent Nuclear Fuel, in: IAEA Nuclear Energy Series, IAEA, Vienna, report NF-T-3.6 (2009).

3  L. Johnson, C. Ferry, C. Poinssot, P. Lovera, Spent fuel radionuclide source-term model for assessing spent fuel performance in geological disposal. Part I: Assessment of the instant release fraction, 346(1) (2005) 56-65.

4  D. Serrano-Purroy, F. Clarens, E. González-Robles, J.P. Glatz, D.H. Wegen, J. de Pablo, I. Casas, J. Giménez, A. Martínez-Esparza, Instant release fraction and matrix release of high burn-up UO2 spent nuclear fuel: Effect of high burn-up structure and leaching solution composition, J. Nucl. Mater. 427(1) (2012) 249-258.

5  A. Martínez-Torrents, D. Serrano-Purroy, R. Sureda, I. Casas, J. de Pablo, Instant release fraction corrosion studies of commercial UO2 BWR spent nuclear fuel, J. Nucl. Mater. 488 (2017) 302-313.

6  K. Lemmens, E. González-Robles, B. Kienzler, E. Curti, D. Serrano-Purroy, R. Sureda, A. Martínez-Torrents, O. Roth, E. Slonszki, T. Mennecart, I. Günther-Leopold, Z. Hózer, Instant release of fission products in leaching experiments with high burn-up nuclear fuels in the framework of the Euratom project FIRST- Nuclides, J. Nucl. Mater. 484(Supplement C) (2017) 307-323. 

Author information: Christoph Rirschl, Head of Cask Loading Services, GNS Gesellschaft für Nuklear-Service mbH; Hagen Höfer, Managing Director Höfer & Bechtel GmbH; Marc Verwerft, Head of fuel materials group at Belgian Nuclear Research Centre (SCK•CEN); Wolfgang Faber, Head of fuel assembly operation and disposal at PreussenElektra GmbH