Between 2001 and 2005, the fresh fuel for one of Fortum’s two VVER-440 units at Loviisa will be supplied by BNFL – the first time a Western fuel supplier has provided reload quantities of fuel for a VVER-440.

The contract, covering around 100 assemblies per year, was signed in December 1999, but its roots were back in March 1996. That is when BNFL signed a contract with Imatran Voima Oy (IVO) and Paks in Hungary to deliver five lead test assemblies (LTAs) to Loviisa. Under the contract, BNFL had to obtain regulatory approval for loading these LTAs at Loviisa, and approval in principal for the possible future full reloads at Paks.

The new fuel design was known as the new optimised VVER assembly, enhanced version 3 (Nova E-3). The assemblies – four fixed and one follower – were loaded into Loviisa 2 in September 1998.


Compared to traditional VVER fuel, Nova E-3 has a reduced fuel rod diameter, which provides a better fuel-to-moderator ratio and reduces fuel costs. Despite the reduced rod diameter, there is a net increase in uranium mass because the new design uses solid fuel pellets. This helps to keep burnups low as required by the Finnish regulator. Other new features include a removable shroud which is 1.6mm thick, and spacer grids, shroud tubes and fuel tubes made of Zircaloy-4. The fuel rods are removable. The design (one of six assessed) optimised fuel cycle costs, uranium mass, surface heat flux and mechanical strength.

All the materials used in the Nova E-3 have been successfully used in LWRs and the fuel rod design is identical in terms of materials and construction to standard PWR fuel rods. Out-of-reactor testing suggests that the primary coolant water chemistry in VVER-440s is less aggressive than that in PWRs, where optimised Zircaloy-4 cladding has operated successfully for many years.

Material specifications were written

to limit the cobalt content of all major


The Nova E-3 design was supported

by thermal hydraulic and mechanical test programmes.

The thermal hydraulic programme included:

•Pressure drop tests on full size prototypes. Pressure drop tests carried out in a flow loop at Loviisa confirmed that the pressure drops of the fixed and follower LTAs were almost identical to those of the corresponding reference assemblies.

•Critical heat flux tests, to determine the burnout characteristics of the assemblies.

•Subchannel mixing tests, to assess fluid exchange and thermal mixing between adjacent subchannels in the fuel assembly.

•Flow stability tests on a full size prototype fixed fuel assembly, which concluded that there was no evidence of resonance.

•A flow endurance test on the full size prototype used in the flow stability tests.

The mechanical test programme included tests on:

•Shroud strength, nozzle strength and shroud to nozzle joint.

•Grid strength.

•Instrument tube to grid sleeve joint strength.

•Fuel rod/instrumentation tube to bottom nozzle joint strength.

•Grid spring stiffness.

•Holddown coil spring stiffness.

•Fuel assembly vibration.

•Fuel rod vibration.

•Fuel follower drop.

The aim of the drop test was to demonstrate that the follower fuel assembly is strong enough to withstand the impact forces arising from a reactor scram and breakage of the intermediate rod, which supports the weight of the follower and absorber assemblies in normal operation. Controlled and free drops simulating each of these events showed that the follower fuel assembly remained intact after a free drop.

For Paks, all the design variants would require the use of burnable poisons to maintain acceptable subcriticality in the Paks refuelling ponds. It is envisaged that gadolinia-doped rods will be included in any fuel supplied to Paks


The Nova E-3 design has increased thermal hydraulic margins, giving the operators greater flexibility in core management.

Fuel rod thermomechanical design calculations were carried out using BNFL’s Enigma-B code. The code predicts the behaviour of fuel rods during normal operation and anticipated operational occurrences (AOOs). The validation database includes 500 fuel rod irradiations.

The thermo-mechanical performance of the Nova E-3 has been evaluated against several design criteria for fuel failures. They include limits on clad stress and strain, fuel temperature, rod internal pressure and clad corrosion. The bounding power histories included peak linear fuel ratings of 35kW/m and maximum assembly burnups of 60 MWd/kgU – well beyond the current VVER-440 licensing limits.

Mechanical design calculations used a combination of finite element analysis and hand calculations. The mechanical performance of the Nova E-3 has been assessed against the stress and fatigue criteria of the ASME Boiler and Pressure Vessel Code and the ‘LWR Fuel Assembly Mechanical Design and Evaluation’ requirements of the American Nuclear Society’s Guide 57.5. Fuel assembly nuclear design calculations were carried out by Fortum Engineering and VTT Energy using CASMO-HEX, HEXBU-3D and ELSI-1440 – the standard tools used for fuel management at Loviisa.

The principal nuclear design requirement for the LTAs was they should be capable of being loaded in any fresh assembly location in Loviisa 2 in Cycle 19 without any restrictions. In practice, this meant selecting an enrichment to give equal or lower assembly powers and equal or lower fuel rod powers relative to the reference fuel. Due to improvements in the fuel-to-moderator ratio in the Nova E-3 design this was achieved with a lower enrichment than in the reference fuel.

Further thermal hydraulic design calculations confirmed that bulk boiling does not occur in any sub-channel in the reactor. In practice, because the reactors at Loviisa have been uprated by 9% to 1500MWt, the thermal margin to bulk boiling is the most limiting factor in licensing. For potential future reloads Paks has used the transport code HELIOS 2D to develop a dual enrichment fuel assembly design.


In the loading plan for the LTAs at Loviisa two of the four fixed assemblies were placed in high power symmetric core locations in the central region, and two were placed in lowly rated symmetric locations on the periphery. Each of the fixed assemblies was positioned under one of the 210 assembly outlet thermocouples. Due to the shrouded nature of the VVER-440 fuel design, these measurements can be used to determine the power of the assembly. In practice, since there is 30° symmetry in the core loading pattern, there are several additional reference temperature measurements available. Transitional loading patterns are being planned for Loviisa to give a smooth and fast transition to an equilibrium cycle containing a full Nova E-3 core.

Potential reloads at Paks were simulated using the KARATE code for loading pattern searches, starting from a full core of reference fuel until equilibrium conditions were achieved. Seventy-eight fixed assemblies and 12 follower assemblies were loaded annually, maintaining one-sixth core symmetry. Follower and fixed assemblies were irradiated for three and four years, respectively. The highest burnup assemblies were placed on the core periphery consistent with a low leakage loading pattern. The calculations showed that safety parameters remained within the licensed limits during all cycles.


Whole-plant fault analyses were carried out to demonstrate that neither the loading of a full core of Nova E-3 fuel nor the loading of five LTAs would adversely affect the safety of the reactor. The data generated by these analyses were used as input to ‘hot rod’ transient analyses and to structural analyses.

To consider the effect on safety of a core made up entirely of a new fuel design would normally require a full scope safety analysis. However, in 1992 the AGNES (Advanced General and New Evaluation of Safety) project was launched at Paks and this full scope safety analysis examined all the initiating events considered worldwide to affect plant plus some cases specific to VVER-440s. The limiting initiating faults chosen for re-analysis in the Project were:

•Analysis of the LOCA with double ended cold leg break.

•Asymmetric control rod ejection transient.

•Withdrawal of control assembly group at 100 % power without scram (ATWS).

•Withdrawal of control assembly group (with scram).

All were analysed by the Hungarian Atomic Energy Research Institute, using codes already validated and licensed for VVER-440 applications.

For the loading of the LTAs, it was considered that the inclusion of five new assemblies would not affect the course of the plant transient at Loviisa. Fortum Engineering was able to provide the results of plant analyses carried out for a core made up of reference fuel assemblies.

For large break LOCA the thermal hydraulic and mechanical behaviour of an array of fuel rods was calculated during the reflood phase of the fault. For the other two faults, which are intact circuit faults, the thermal hydraulic behaviour of the ‘hot rod’ was calculated. The ‘hot rod’ analyses met licensing criteria generally accepted by regulators in Europe and the US – maximum clad temperatures of 1200°C, maximum clad oxidation of 17% and maximum radial averaged fuel enthalpy of 963J/gUO2. It also showed that the maximum clad strain due to dynamic fuel rod swelling (clad ballooning) during the LOCA did not cause contact between adjacent rods or rod rupture. Maximum clad temperature during dynamic ballooning was actually less than during a case with no ballooning. It was concluded that a coolable geometry would be maintained during this fault.

The fourth of the limiting faults analysed, withdrawal of control assembly group (with scram), was an AOO. For this case it was shown that the Bezrukov correlation limit was met and therefore that departure from nucleate boiling (DNB) failures would not occur as a result of this initiating event.

A structural analysis of the effects of a LOCA was carried out. The displacement versus time histories of the reactor internals due to a LOCA were calculated by Paks and the dynamic response of the fuel assembly was calculated by BNFL. The LOCA event alone leads to high, but acceptable, stresses in the shroud tube and the fuel rod cladding and high, but also acceptable, impact grid-to-shroud impact forces. Since the shroud and cladding stresses remain below the maximum allowable stresses as defined by the ASME code of practice, the integrity of the fuel assembly is assured and the coolability of the fuel is unimpaired.

Nevertheless yield strength of the shroud and the cladding is exceeded locally leading to the possibility of some permanent local deformation. An elastic/plastic finite elements analysis confirmed that the small residual bow in the shroud tube due to this local yielding woiuld be insufficient to prevent insertion of the assemblies in a large break LOCA.

Considering the combined effect of a postulated earthquake and a LOCA at Paks showed that the fuel rods would remain coolable and that the reactor can be safely shutdown.


The fuel was manufactured at BNFL Springfield’s Oxide Fuel Complex (OFC) using the automated PWR fuel rod production line with minor modifications. With further modifications to take account for the hexagonal geometry of VVER-440 fuel it was also possible to assemble the LTAs on the Complex’s PWR production line.

Five LTAs were loaded in Loviisa 2 in Cycle 19 and have nearly completed two full cycles of irradiation. The fuel has continued to operate as expected; the outlet temperatures of the four fixed LTAs are very similar to those of the reference fuel. This confirms that the powers of the LTAs are the same, within measurement uncertainty, as those of the reference fuel in symmetric core locations.

One of the LTAs was removed from the core at Loviisa 2 in September 1999; the shroud was removed and the fuel inspected. Visual inspection revealed the fuel to be undistorted and showing little evidence of corrosion, as expected. Three weeks of detailed measurements confirmed the dimensional stability of the fuel and its low corrosion. The fuel assembly will be reloaded at the next outage, when another of the LTAs will be removed for routine inspection.