The Atucha II nuclear power plant is a unique pressurized heavy water reactor being constructed in Argentina. The original plant design was by KWU in the 1970’s using the then German methodology of break preclusion, which assumed that the largest break-opening area would be 10-percent of the cross-sectional area of the largest pipe diameter. That philosophy was used for the design of the emergency core cooling system in the 1970’s. The plant construction was halted for several decades, but a recent need for power was the driver for restarting the construction. The construction is progressing with initial start-up in 2011. Since the 10-percent of the cross-sectional area is a smaller ECCS design requirement than the normally assumed double-ended-guillotine break, the safety evaluation of the plant for beyond design basis seismic loading of the nuclear plant was a regulatory requirement. This overview paper describes a Robust LBB Evaluation that was conducted in great detail to assess the safety aspects of the piping system under beyond design basis seismic loading and the implications to the ECCS. Key aspects involved: • Static and dynamic material property testing, • Determination of weld residual stresses, • Determination of crack sizes that might evolve by worst case SCC growth rates under weld residual stresses and normal operating stresses, • Determination of leakage rates as a function of time with the upper-bounding crack growth rates, • Development of seismic hazard curves for the site, • Development of FE models of the containment building and primary NSSS system within the building, • Determination of normal operating stresses, SSE stresses and 10−6 seismic stresses using worst case soil foundation assumptions, • Evaluation of flaw behavior for circumferential cracks using the shapes from the natural crack growth. • Evaluation of margins on the critical flaw size and times to leakage, and • Standard LBB analyses, as well as Transition Break Size evaluations. The key result from this effort was that even with all the normal operating plus 10−6 seismic event loading, the pipe system behaved more like it was displacement-controlled than load-controlled. The displacement-controlled behavior made the pipe much more flaw tolerant, and it was found that a DEGB was not possible because the flaw could never reach the critical flaw size without greatly surpassing the leakage and water make-up capacity of the plant. Since there are many details in this multi-year effort, only the key points will be summarized in this paper while other details will be the topics of other papers.

British Energy (BE) has funded a large work programme to assess the possible impact of primary water stress corrosion cracking on dissimilar metal welds in the primary circuit of the Sizewell ‘B’ pressurised water reactor. This effort has included the design and manufacture of representative pressuriser safety/relief valve (SRV) nozzle welds both with and without a full structural weld overlay, multiple residual stress measurements on both mock-ups using the deep hole and incremental deep hole methods, and a number of finite element weld residual stress simulations of both the mock-ups and equivalent plant welds. Three organisations have performed simulations of the safety/relief valve nozzle configuration: Westinghouse, Engineering Mechanics Corporation of Columbus (EMC2) and the Australian Nuclear Science and Technology Organisation (ANSTO). The simulations employ different welding heat input idealisations, make different assumptions about manufacturing history, and use a variety of different material constitutive models, ranging from simple bilinear kinematic hardening to a full mixed isotropic-kinematic formulation. The availability of both high quality measurements from well characterised mock-ups, and a large matrix of simulations, offers the opportunity for a “mini-round-robin” examining both the accuracy and key solution variables of dissimilar metal weld finite element simulations. This paper is one of a series at this conference that examine various aspects of the BE work programme. It draws together residual stress measurement results and the results of all three simulation campaigns (described in detail in other papers at this conference) to examine the impact of manufacturing history, thermal modelling assumptions, material constitutive models and other key solution variables on the accuracy of residual stress predictions in this dissimilar metal weld geometry.

Welding is a commonly used and one of the most important material-joining processes in industry. The incidences of defects had been located by ultrasonic testing (UT) in various pressurizer nozzle dissimilar metal welds (DMW) at nuclear power plants. In order to evaluate the crack propagation, it is required to calculate the stress distribution including weld residual stress and operational stress through the wall thickness in the weld region. The analysis procedure in this paper included not only the pass-by-pass welding steps, but also other essential fabrication steps of surge, safety/relief and spray nozzles. In this paper, detailed welding simulation analyses have been conducted to predict the magnitude of these stresses in the weld material. To prevent primary water stress corrosion cracking (PWSCC) in pressurized water reactors (PWR) on susceptible welded pipes with dissimilar metal welds, the weld overlay process has been applied to repair nuclear reactor pipe joints in plants. The objectives of such repairs are to induce compressive axial residual stresses on the pipe inside surface, as well as increase the pipe thickness with a weld material that is not susceptible to stress-corrosion cracking. Hence, understanding the residual stress distribution is important to evaluate the reliability of pipe joints with weld overlay repairs. The finite element results in this paper showed that, after deposition of the DMW nozzle and stainless steel welds, tensile weld residual stresses still exist at regions of the DMW through the thickness. This tensile weld residual stress region was significantly reduced after welding the overlay. The overlay weld also provides a more uniform and large compressive region through the thickness which has a beneficial effect on the structural integrity of the DMW nozzle welds in the plant.

Cracks have been discovered in Inconel 82/182 bimetal welds at the hot leg to reactor pressure vessel (RPV) nozzle welds at the V. C. Summer nuclear power plant in the United States and at the Ringhals plant in Sweden. One question raised by these incidents is what impact will cracks such as these have on the possible acceptability of leak-before-break for these large-diameter piping systems. In order to address this question, fracture toughness tests in both the L-C and L-R orientation with the crack centered in both the Inconel 82/182 weld as well as in the buttered region of the bimetal weld were conducted on specimens machined from a bimetal weld obtained from a cold leg from a cancelled CE plant. In addition, weld residual stress analyses have been conducted to assess the propensity for primary-water stress-corrosion cracking (PWSCC). As part of this paper, the results from both the fracture toughness tests and weld residual stress analyses will be discussed, along with their potential impact on LBB.

This paper examines the inherent conservatisms of alternative girth weld defect acceptance criteria from the 2007 API 1104 Appendix A, CSA Z662 Appendix K, and the proposed EPRG Tier 2 criteria. The API and CSA codes have the same empirical limit-load criteria, where it has previously been shown that the conservatism on the failure stress is ∼30 to 50 percent compared to pipe test data prior to applying any safety factors. In terms of flaw length, it was found that the API/CSA limit-load equation might allow a flaw of 5% of the pipe circumference, where the properly validated limit-load equation would allow a flaw of 75% of the circumference, i.e., a safety factor of 30 percent on load corresponded to a safety factor of 15 on flaw length for that example case. Similarly there are conservatisms in a proposed EPRG Tier 2 girth weld defect acceptance criterion. This proposed criterion was directly based on curved-wide-plate data to assure that toughness was sufficient to meet limit-load conditions for a curved-wide plate. However, the curved-wide plates are really an intermediate-scale test, and still require proper scaling to pipes of different diameters. The proposed Tier 2 EPRG allowable flaw length is 7T from a large database of curved-wide-plate tests with the a/t value of less than 0.5 (or a < 3mm), and the failure stress being equal to the yield strength of the base metal (also requires the weld metal overmatch the base metal strength, and the Charpy energy at the defect location have a minimum > 30 J and average > 40 J). However, the widths of those curved-wide-plate tests are typically a factor 5 to 12 times less than typical large-diameter pipes. The proper limit-load/fracture mechanics scaling solution would have the flaw length proportioned to the plate width, not the specimen thickness. Additionally, the proper limit-load solution for a pipe in bending gives a much larger tolerable flaw size at the yield stress loading than a plate or pipe under pure tension. Example calculations showed that the EPRG Tier 2 approach is conservative on the flaw lengths by approximately 9 for pure axial tension loading, and between 34 to 79 for a pipe under bending. Suggestions are presented for an improved procedure that accounts for proper limit-load solutions for pipe tests, effects of pipe diameter, effects of internal pressure, and also a much simpler approach to incorporate the material toughness than the 2007 API 1104 Appendix A Option 2 FAD-curve approach. The fracture analyses could evoke SENB, SENT testing, or have relatively simple Charpy test data to assess the transition temperatures to ensure ductile initiation will occur.