© C. Giovedi et al., published by EDP Sciences, 2016

1 Introduction

The available data shows that the steady state performance of steel cladding in the first PWR was considered excellent [1,2]. The material used in the early PWR was mainly AISI 304 (12% cold worked). Nonetheless, some reactors operated using annealed AISI 348, which presents a better corrosion resistance due to the addition of niobium and tantalum in its composition.

The substitution of stainless steel by zircaloy as cladding material was due to the lower absorption for thermal neutrons of the zirconium-based alloys which enables to operate with lower enrichment cost. Despite the stainless steel economics penalty, the main advantage of using this material as cladding comes from the reduction of the probability of the violent oxidation reaction that occurs with zirconium-based alloys at high temperatures, as it has occurred in the Fukushima Daiichi accident [3]. As a consequence of this, iron-based alloys once again can be considered as a good option to replace zirconium-based alloys as cladding material improving the safety under accident scenarios [4]. Considering the previous good experience of AISI 348 as cladding, this material could be again applied to replace zirconium-based alloys as PWR fuel cladding.

In order to evaluate the fuel performance of fuel rods using AISI 348 as cladding, it is necessary to modify the current fuel performance codes to insert correlations and properties of this material. In this sense, TRANSURANUS code appears as a good option due to its flexibility for different fuel rod designs and reactor types, time range of the problems to be treated and materials data bank, which includes AISI 316 (both 20% cold worked and annealed correlations are programmed into the code) [5,6].

The adapted version of the TRANSURANUS code to evaluate the AISI 348 performance under irradiation was assessed using Yankee Rowe available data from open literature. The reason why Yankee Rowe fuel rod was selected is because it was the unique PWR (for which information was available to the authors) in which AISI 348 was used as cladding material. The aim of this paper is to present the obtained results in the framework of this activity.

1.1 TRANSURANUS code

TRANSURANUS is a computer code for the thermal and mechanical analysis of fuel rods in nuclear reactors developed at the European Institute for Transuranium Elements (ITU). The code consists of a clearly defined mechanical-mathematical framework into which physical models can easily be incorporated [5].

In order to introduce the AISI 348 data in the TRANSURANUS code, a set of references has been searched and collected. A selection has been made in order to use reliable data, when necessary data are not available either values coming from similar stainless steel (AISI 347) or typical values (i.e. applicable for a variety of stainless steel) were used. A comparison of the main properties for non-irradiated annealed AISI 304, 316 and 348 is presented in Table 1. The data show that the properties for AISI 316 and AISI 348 are very close which enable to expect a similar performance for both materials under irradiation.

Based on the literature research, the following properties related to the annealed AISI 348 were introduced in the TRANSURANUS code to obtain the adapted version: elasticity constant, Poisson's ratio, strain due to swelling, thermal strain, thermal conductivity, creep strain (thermal and irradiation creep rate), yield stress, rupture strain, burst stress, specific heat, density and melting temperature.

It was assumed that correlations already programmed in TRANSURANUS for the AISI 316 are acceptable and validated enough being the TRANSUNARUS originally developed to deal with fast breeder reactor fuel and considering its validation program [6]. In addition, the new correlations related to the AISI 348 properties somewhat reflect the same structure of the equivalent formula already programmed for the AISI 316. These correlations similarities should (at least partially) ensure that code numerical stability issue is not to be expected.

The AISI 348 behavior predicted by the modified code version has been compared against AISI 316 behavior which is part of the original (hence validated) code version. In general, the two steels present, as expected, similar trends. AISI 316 has shown a bit more conservative results in respect to AISI 348.

1.2 Description of Yankee Rowe NPP features

The Yankee Rowe PWR has been owned and operated since startup in 1960 by the Yankee Atomic Electric Co. at Rowe, Massachusetts. The reactor and its initial core and stainless steel reloads were designed and built by Westinghouse. Yankee Rowe was the first fully commercial PWR of 250 MWe, which started up in 1960 and operated to 1992 [10]. Yankee Rowe produced 44 billion kilowatt-hours of electricity from 1961–1992 when it was permanently shutdown for economic reasons. The plant was successfully decommissioned between 1992–2007 with structures removed and the site restored to stringent federal and state remediation standards [11].

Starting from its 7th cycle of operation, the reactor began to change to zircaloy cladding, the transition was completed with cycle 12. The stainless steel clad reactor core consisted of 76 assemblies and 24 cruciform control rods. A typical stainless steel assembly was made up of 9 subassemblies each arranged in a 6 × 6 array, to make up an 18 × 18 fuel rod array. The subassemblies were tied together along their length to form a complete integral fuel assembly.

The clad material was both seamless and welded annealed AISI 348 and represents the only large scale fuel experience with this steel in a PWR. The chemical composition of the adopted AISI 348 is identical to the niobium stabilized AISI 347, with the exception of a 0.10% limit on tantalum to reduce the neutron absorption cross-section. The fuel rod was also unique in that 6 physically separated fuel stacks spaced by equally spaced stainless steel discs. Each segment contains about 25 pellets. The objective of such design was to minimize differential thermal expansion between fuel and clad. There were no reported stainless steel clad fuel failures. The average fuel rod heat generation rate was 114 W cm^–1, the design rate was 353 W cm^–1 (with a peak as high as 410 W cm^–1). The maximum cladding surface temperature was 343 °C. A total of 16 assemblies were examined, all the assemblies were in excellent conditions with a minor amount of crud deposited [1].

2 Methodology

2.1 Yankee Rowe general data and assumptions

In order to prepare the input deck to perform the simulation considering the Yankee Rowe reactor design and operational parameters, it was collected in the literature all the available data, which are presented in Table 2 as well as the necessary assumptions.

2.2 TRANSURANUS model and assumptions

The simulations were carried out adopting the recommended TRANSURANUS models for PWR. The geometric characteristics, thermal-hydraulic parameters and power profile obtained from the literature for the Yankee Rowe fuel rod were implemented in the TRANSURANUS input deck according to the data presented in Table 2.

Considering that in TRANSURANUS code the analysis is performed slice per slice, it was necessary to assume a discretization for the Yankee Rowe fuel rod, which is presented in Figure 1. In order to prepare this model, it was considered the following information: the fuel rod had six physically separated fuel stacks with a perforated stainless steel disk between them localized at equally spaced axial locations, each segment contains about 25 UO₂ pellets, the active fuel length is 229.9 cm and the height of the fuel pellet is 1.46 cm [1,12].

The plenum length is not presented in the literature. Then, for calculation was assumed a value of 14 cm, which represents a conservative value for a PWR fuel rod with an active length of 229.9 cm.

The cladding and pellet roughness are also not presented in the literature for the Yankee Rowe fuel rod, and then it was assumed typical values for PWR. The same was considered for grain diameter, open porosity and plenum spring characteristics.

The simulation to assess the behavior of the Yankee Rowe fuel rod was carried out considering the information related with the rod E6-C-f6 as described in reference [12]. The selected rod target of the present simulation was irradiated into three core cycles identified as Core I, Core II and Core IV. Boundary conditions and axial power profile have been derived from reference [12]. Noticeably the axial power profile has been derived considering the average core power reported in Table 2. Data related with E6-C-f6 fuel rod are available in four axial positions, which have been interpreted as average values along the related length. Thus, constant piecewise trend has been adopted into TRANSURANUS simulation (Fig. 2). The calculated peaking factors have been imposed both to the linear power and to the neutron flux. The resulting profile is bottom skewed for the cycles Core I and II where the power was controlled by control rods, rather in Core IV boron was introduced as chemical shim resulting in a flatter axial profile [13].

The irradiation period is consistent with the information available in reference [12], adding 24 h for the power rise, 12 h for the power decrease, in addition 48 h has been set as shutdown period between two core cycles.

Finally, an average neutron flux equal to 6.3 × 10¹³ n cm^–2 s^–1 has been set in order to achieve a fluence level close to the value available in the literature, i.e. 6.0 × 10²¹ n cm^–2. Regarding the fuel-cladding contact model, the perfect slip model has been adopted.

Fig. 1 Yankee Rowe fuel rod assumed discretization based on the literature data [1]. 1, 2, 3, 4, 5, 6: 36.5 cm (25 fuel pellets) divided in 4 segments each one of 91 mm (apart the first two meshes of 85 mm); a, b, c, d, e: 40 mm (stainless steel disk); Plenum: 140 mm.

Fig. 2 E6-C-f6 fuel rod axial peaking factor for Core I (a), Core II (b) and Core IV (c), available data (blue dots [12]) and related interpolation (red curve).

3 Results and discussion

The results obtained from the Yankee Rowe fuel model are shown hereafter. Table 3, Table 4 and Table 5 list, respectively, the outcomes of the simulation at the end of the Core I, Core II and Core IV cycle, compared with available information taken from reference [12]. It should be noted that no tuning has been done for carrying out the simulation.

The parameters attaining to the Core I cycle are reasonably reproduced by the TRANSURANUS code (Tab. 3), noticeably the burnup matches fairly good in all four locations.

Regarding the fuel temperature calculated by the code, centerline and surface values are provided since for the reference data no specification about the radial position is provided. It can be seen that reference fuel temperature is within the code prediction for all the four axial positions.

The same considerations apply for the clad temperature (apart for the top position which is slightly underpredicted due to the underestimation of the coolant temperature) in relation with both reference data radial position and calculated values.

Additional calculated data are provided in Table 3 regarding fission gas release which remains very low; fuel and clad axial elongation, both are lower than 0.5%; maximum fluence value and plenum pressure which is double of its starting value.

Table 4 compares reference and calculated data related with the Core II cycle. Also for this irradiation step the code gives reasonable results, showing the same (as in Core I) good compliance regarding the burnup data.

Calculated values of fuel and clad temperature include the corresponding reference data. Notwithstanding the accumulation of the burnup, the fission gas release is still low (0.12%); fuel and clad axial elongation do not change so much from the previous cycle (both slightly increased); the gap is reducing but still open; the plenum pressure is slightly increased from the previous cycle.

Table 5 reports the comparison discussed above but at the end of the Core IV cycle. Also at this stage of the simulation, the code shows the same capabilities in relation with the burnup, fuel and clad temperature. Coolant temperature is also reasonably predicted as well.

At the end of the whole simulation, the fission gas release is below 0.2%; fuel and clad elongation are well below 1%; the gap kept open with a minimum value of about 16 μm (about 1/4 of its initial value) and the plenum pressure is less than the triple of its initial value.

In relation with the fuel and clad relative elongation it can be seen that the code is able to reproduce one of the objective of the particular Yankee Rowe rod design, namely to minimize the differential thermal expansion between fuel and clad.

In general, the TRANSURANUS code performed reasonably well even facing with a rod design which is quite far from the typical (current) PWR technology (e.g. clad material, filling gas type, lack of gap pressurization, presence of different segments within the fuel rod). Any predicted parameters for the simulated fuel rod are of no concern regarding their corresponding design data.

4 Conclusion

The assessment of the modified TRANSURANUS code benefits of the availability in the open literature of data related with Yankee Rowe NPP, which was one of the few plants in which AISI 348 has been used as cladding material. A specific Yankee Rowe fuel model has been set up, fully considering the available information and doing some assumptions for covering some lacks (e.g. fuel rod upper plenum height). When such assumptions had to be made, conservative values have been adopted (considering Yankee Rowe and typical PWR rod design).

The carried out calculations show reasonably agreement with available data confirming the modified code capabilities. This constitutes an indication of the modified TRANSURANUS code capabilities to perform fuel rod investigation of fuel rod manufactured with AISI 348 cladding material.

Acknowledgments

The authors are grateful to the technical support of USP, IPEN-CNEN/SP and to the financial support of IAEA to attend the TopFuel 2015 meeting.

References

All Tables

All Figures

	Fig. 1 Yankee Rowe fuel rod assumed discretization based on the literature data [1]. 1, 2, 3, 4, 5, 6: 36.5 cm (25 fuel pellets) divided in 4 segments each one of 91 mm (apart the first two meshes of 85 mm); a, b, c, d, e: 40 mm (stainless steel disk); Plenum: 140 mm.
In the text

	Fig. 2 E6-C-f6 fuel rod axial peaking factor for Core I (a), Core II (b) and Core IV (c), available data (blue dots [12]) and related interpolation (red curve).
In the text