Beyond RPV design life1 October 2003
An overview of the vessel surveillance programmes carried out in Spain, life predictions based on the existing surveillance data and research that will provide more information. By Antonio Ballesteros, Ignacio Marcelles and Juan Bros
The reactor pressure vessel (RPV) is rated a highest priority key (Category 1) component in all national categorisation exercises because the RPV is considered irreplaceable or prohibitively expensive to replace. This in turn means that if it degrades sufficiently it could be the feature that limits the nuclear plant's operational life. The RPV houses the reactor core and because of its function it has a direct safety significance.
The majority of early Westinghouse designed plants had a 'design basis life' (as distinct from physical life) of 30-40 years. The design basis life was based mostly on fatigue usage factors, not general technical studies of material degradation. Now that light water reactors are being operated for longer than originally intended, this has become
relevant, as radiation embrittlement is the most important degradation mechanism limiting RPV life. Neutron irradiation degrades the mechanical properties of RPV steels and the extent of the degradation is determined by a number of factors such as neutron fluence, irradiation temperature, neutron flux and the concentration of deleterious elements in the steel.
Technical and economic considerations mean the operating life of a nuclear plant could be easily 50 or 60 years.
This might require the upgrading of RPV surveillance
programmes and the use of modern techniques and approaches, such as Charpy reconstitution and master curve testing, to cover the extended operation time.
RPV Surveillance DATA
Changes in material properties due to neutron irradiation are monitored by means of surveillance programmes. Every PWR and BWR pressure vessel has an ongoing radiation surveillance programme for the vessel material. To date several hundred surveillance capsules have been removed from their host pressure vessel and tested. The results from these surveillance capsules have been used to develop heatup and cooldown curves and to analyse potential or postulated accident or transient conditions.
The structural integrity evaluation of Spanish reactors vessels follows the regulations, guidelines, codes and standards developed in the USA, since the reactors were designed by Westinghouse and General Electric and
are similar in design and operation to US reactors. An exception is Trillo I, supplied by Siemens KWU (now Framatome ANP).
In general, the weld is the most limiting material of the beltline region for the older reactors (see Table 1).
It is well known that certain residual elements, such as copper and phosphorus, favour embrittlement. At levels higher than 1%, nickel also causes embrittlement, although it has a beneficial effect since it produces a lower initial value of the reference temperature RTNDT. The role played by other residual elements, such as tin, antimony and arsenic, is not clear. Copper produces fine precipitates which cause embrittlement. Copper and nickel are also believed to have a synergy effect in hardening. Phosphorus embrittles as a result of two mechanisms: fine precipitates are formed in a manner analogous to the case of copper; and phosphorus precipitation is segregated at the grain boundaries, causing them to be weakened (non-hardening embrittlement) and leading to the possibility of unstable crack growth. Table 2 shows the range of concentration of Cu, Ni and P measured in the RPV surveillance steels of Spanish LWRs.
Because phosphorus can also contribute to hardening, it is not always clear whether its effect on embrittlement is due to hardening or segregation or both. A predominant role for hardening would be consistent with the observed characteristically low susceptibility of Spanish PWR vessel steels
to embrittlement phenomena caused by grain boundary segregation of impurities.
The ratio of the yield stress change to the increase in T41J, measured by Charpy testing, is frequently a good measure of whether non-hardening embrittlement is occurring (The change in the reference temperature, RTNDT, is usually
evaluated as the difference in 41 J index temperature from average Charpy curves measured before and after irradiation). Figure 1 shows this ratio for Spanish PWR vessel steels. The scatter observed at low fluence values is not relevant since small increases in yield strength can produce high DT41J/Dys ratios. The conclusion from Figure 1 is that there is not a non-hardening embrittlement component in the Spanish pressure vessel (PV) steels under consideration.
A total of 26 surveillance capsules have been tested and analysed in Spain. The surveillance data makes it possible to verify the theoretical embrittlement trend curve and to detect any anomaly in the irradiation conditions. Figure 2 shows a comparison of measured and predicted values using the Eason model. The predictions are reasonably good at low fluence levels, but become more dispersed at higher
Prediction to 40 years
The steels used for the reactor vessel become progressively more brittle throughout the service life of the component, because of the neutron radiation to which they are exposed. This progressive degradation must be understood in order to guarantee the structural integrity of the vessel throughout its service life. There are two key parameters which make it possible to evaluate and quantify vessel degradation: upper shelf energy (USE), defined as the average energy value for all Charpy specimens whose test temperature is above the upper end of the transition region, and RTNDT (reference temperature). The initial values of RTNDT are obtained by means of Charpy and drop-weight tests, while increases in RTNDT and the value of USE are measured exclusively by means of Charpy tests. RTNDT tends to increase throughout the service life of the reactor, while USE tends to decrease.
The pressurised thermal shock (PTS) US rule 10CFR50.61, establishes requirements on the ability of a PWR vessel to withstand events in which the vessel is both rapidly overcooled (thermally shocked) and pressurised (or repressurised). The code requires the calculation of the
projected values of RTNDT at the end of life (EOL) and comparison with given limit values. The PTS screening criterion is 270ºF (132ºC) for plates, forging and axial weld materials, and 300ºF (149ºC) for circumferential weld materials.
On the other hand, Regulatory Guide 1.99 revision 2 requires that the reference temperature RTNDT at the quarter thickness (1/4T) position in the vessel wall at EOL be less than 200 ºF (93ºC) for RPV beltline materials of new plants. This could also be considered a good recommendation for existing plants.
Table 3 shows the projected values of RTNDT at 32 effective full power years (EFPY) for the most limiting beltline material of the Spanish vessels at the inside location of the reactor vessel wall. In the past, 32 EFPY was associated with 40 calendar years of operation, representing an average availability factor of 80%. For PWRs operating in Spain the projected RTNDT values are lower than the PTS limit values established in 10CFR50.61.
As established in Appendix G of 10CFR50, 102J is this initial minimum USE for reactor vessel beltline materials in the transverse direction for base material and along the weld for weld material. Through the vessel life USE must be at least 68J, unless it is demonstrated that lower values can provide margins of safety against fracture equivalent to those required by the ASME code. Regulatory Guide 1.161 was developed by the US NRC to provide comprehensive guidance for evaluating vessels when the USE falls below this limit.
All the USE values at 32 EFPY listed in Table 3 are over 68J. In several cases it was difficult to get an accurate projected value of the USE at 32 EFPY, since the impact tests of the unirradiated material were not performed to a high-enough temperature in the upper shelf region. In these cases, the projected USE values were based on the surveillance data of capsules which accumulate a neutron fluence greater than the projected neutron fluence inside the reactor vessel wall at 32 EFPY.
It is worth mentioning two specific types of activities performed in Spain: analysing the impact of power uprating on the RPV structural integrity assessment, and implementing ASME code cases N-640 and N-588.
Several Spanish reactors have undertaken a major
initiative to increase the economic value of their plants, by uprating. The Spanish nuclear regulatory body, CSN, has reviewed applications for power uprates at Almaraz, Ascó, Vandellós and Cofrentes, to determine whether adequate safety margins exist at the higher power and to ensure that regulatory limits are not exceeded. The power uprating studies performed by the utilities include re-evaluation of the projected neutron fluence at end of life and, consequently, of the projected RTNDT and USE values and the impact in the pressure temperature limit curves.
Cofrentes used ASME code cases N-588 and N-640 to update the pressure temperature limit curves. N-588 allows an alternative procedure for calculating the applied stress intensity factors of Appendix G of ASME XI for axial and circumferential welds. N-640 allows the use of KIc rather than KIa to determine pressure temperature limit curves. The use of KIc reduces one of the conservatisms used to
generate these curves.
Towards 60 years
Section XI.M31 of NUREG 1801 (the GALL report)
recommends actions and methods to evaluate the embrittlement status of the vessel for a period of 60 years. For instance, if a plant has a surveillance programme that
consists of capsules with a projected fluence of less than the 60-year fluence at the end of 40 years, at least one capsule is to remain in the reactor vessel and should be tested during the period of extended operation. This is of application to the Spanish BWRs since the lead factor is close to one.
One important remark in NUREG 1801 is a recommendation to remove the standby capsules if the lead factors are relatively high. For example, in a reactor with a lead factor of three, after 20 years the capsule test specimens would have received a neutron exposure equivalent to what the reactor vessel would see in 60 years. The capsule is to be removed and placed in storage since further exposure would not provide meaningful metallurgical data. The standby capsules would be available for reinsertion into the reactor if additional licence renewals are sought (for example, for 80 years of operation). Although up to now Spanish plants are not contemplating any life extension beyond 40 years, Almaraz will be the first Spanish plant to follow this NUREG recommendation. Two standby capsules from each unit will be removed from the vessel in the next outages as they will have accumulated a neutron fluence slightly greater than the
projected neutron fluence on the vessel at 60 years.
When all the surveillance capsules have been removed, some means must be established to ensure that the on-going exposure of the reactor vessel is consistent with the basis used to project the effects of embrittlement to the end of life. It is not possible to say now what the operating philosophies will be in the future. The general recommendation is that, if possible, ex-vessel neutron dosimetry be installed one or more fuel cycles prior to withdrawing the last (typically the 60-year) surveillance capsule. The dosimetry is then removed and analysed at the same time as the surveillance capsule, and replacement ex-vessel dosimetry is installed. The simultaneous measurement inside and outside the
reactor vessel provides information to characterise the reactor vessel exposure.
Facing 60-year operation the most promising techniques are the reconstitution of surveillance specimens and master curve testing. The former solves the limitation in amount of surveillance material available for irradiation and testing, and the latter provides a technically sound approach for defining a unique fracture toughness transition temperature T0 for ferritic steels clearly superior to the old RTNDT.
In January 2002 CSN and the utilities represented by UNESA started a 3-year project (CUPRIVA) focused on compact test (CT) and Charpy specimen reconstitution and master curve testing. Two pilot plants are participating in the project, Santa María de Garoña (BWR) and Ascó II (PWR), which provided surveillance material for investigation. The available irradiated broken specimens base material from Ascó II are being used to machine reconstituted pre-cracked Charpy V-notch (PC-CVN) specimens. The master curve testing results of the Ascó II specimens will be compared with fracture toughness results obtained by conventional testing of available irradiated compact test 1/2T CT specimens. The base and weld surveillance material of the Garoña RPV is being investigated. Reconstituted CT and PC-CVN specimens are being tested according to the
master curve approach. For the Garoña reactor it will be possible to compare master curve testing results from unirradiated and irradiated specimens. Preliminary results will be available by the end of 2003. The test results from Ascó and Garoña will be used to determine with greater accuracy than previous evaluations the remaining life of these reactors. The possible application of the ASME code cases
N-629 and N-631 will be considered. The study will include a comparison of RTNDT and RTT0 values. This last indexing temperature, as defined in the ASME code cases, is based on the master curve concept, and ensures that the KIc curve
will bound the actual material fracture toughness data by functional equivalency of RTNDT and RTT0.
As is usual for BWRs designed by General Electric, the surveillance practice of the Cofrentes and Garoña includes the reinsertion in the vessel of capsules with reconstituted specimens. The surveillance data provided by these non-mandatory capsules allows a better definition of the
embrittlement trend curves for these reactors, and a better prediction of the projected RTNDT and USE values at end of life. In addition to the reinsertion of surveillance capsules, Garoña carried out a modification of the surveillance holder to improve the lead factor of the surveillance capsules, which is now slightly greater than one. For Cofrentes the last capsule being manufactured, to be reinserted in the
vessel in October 2003, will include, in addition to the limiting beltline base and weld materials, a reference steel or correlation monitor material, JRQ, that will make it possible to detect and analyse any anomaly or change in the irradiation conditions.