Scaling, Passive Systems, and the AP-1000

Abstract

The development of the AP-1000 design and of its precursor the AP-600 started in the aftermath of the Chernobyl event (1986) when the need came from the scientific and technological community for a resilient system against deliberate threats by humans. The “passive system” design concept became relevant. The first AP-1000 entered into operation around 3 decades after that event. This paper discusses the issue of how much the progress in nuclear science and technology since the end of the 1980s has affected the AP-1000 design. Five interconnected areas are identified: (1) reliability of passive systems, (2) scaling and uncertainty, (3) coupling between three-dimensional neutron physics and thermal hydraulics, (4) consideration of large-break loss-of-coolant accidents, and (5) simulation of instrumentation and control systems. All these areas are relevant for the AP-1000 and standard pressurized water reactors; however, the areas (1) and (2) have specific applicability for the AP-1000 and constitute the main concerns of this paper. The conclusion from qualitative investigation is that the safety demonstration of the AP-1000 did not take full benefit from progress in these areas, namely, inadequacies characterize the scaling database and the processes for determining the reliability of thermal-hydraulic passive systems did not receive proper attention.

Keywords:

• AP-1000
• scaling in licensing
• reliability of passive systems
• accident analysis

Acknowledgments

The contributions of all coauthors of publications in the list of references are gratefully acknowledged; a much larger number of scientists (several dozens, impossible to list all of them) have contributed to the findings in this paper. Thanks to the anonymous reviewers who provided comments to clarify the paper and to the journal editor in charge of the review process.

Disclosure Statement

No potential conflict of interest was reported by the author.

Notes

^a Thousands of scientists have pushed the boundary of knowledge in these areas. Apologies for identifying here only papers where the current author contributed (a comprehensive list would involve hundreds of citations, well beyond the purposes, here; see e.g., the list of references in Ref. 11. Selected papers in the present list of references connect with pillar documents from the international science and technology communities.

^b Interested readers may consult the paper by Aksan et al., “Thermal-Hydraulic Phenomena for Water-Cooled Nuclear Reactors, J. Nuclear Engineering and Design, 330, pp. 166–186 (2018), where 116 phenomena are listed, and the paper by D’Auria and Bestion, “Nuclear Thermal-Hydraulic Phenomena: Bases and Challenges,” J. Nuclear Technology, 208, 6, pp. 990–1011 (2022), where a deep definition for the word phenomenon is provided, as well as guidance for the use of phenomena in the development and understanding of nuclear thermal hydraulics.

^c Phenomena are characterized by (typically) several parameters, by parameter ranges, and by combinations of these parameter ranges. The combination of (a large number of intervals of) parameter ranges intuitively brings an extremely large number of expected conditions, a tiny fraction of which (one may postulate <1%) can be covered by experiments. Therefore, looking at experiments only is not sufficient from the viewpoint of nuclear reactor safety. Experiments are needed to qualify the codes, then the codes are used to explore the “extremely large number of expected conditions.”

^d For instance, the solution of the Fourier equation in a cylindrical bar with a 1-cm diameter is the same as the solution in a 1-m diameter bar; this is not necessarily true for two-phase flow solution in a 1-cm and a 1-m pipe because of flow regime maps depending on diameter.

^e In different terms, if scaling capabilities of (empirical) constitutive models, part of the balance of equations, are not proven for two-phase flows (this can be the case for interfacial drag or added mass), these terms shall not be used to design experiments aimed at demonstrating scaling capabilities.

^f A comprehensive analysis (not an exhaustive list) to characterize these weaknesses implies at the least, not considering the origins of the concerned postulated initiating event (PIE) and the assumptions for the availability of systems, the following:

1. Identification and characterization of phenomena for each DBA.2. Characterization of parameter ranges.3. Check that any set of parameter ranges (at least a few relevant ones) is part of experiments in differently scaled facilities (typically, at least three facilities).4. Check that the concerned code (i.e., the code adopted for the best-estimate analyses of a DBA) is capable (without tuning or adjustment of input parameters) of reproducing experiments at the previouls step, ending up with a suitable qualitative and quantitative accuracy evaluation.

An example of a lack of experimental (and consequently code-scaling capability demonstration) scaling bases is the draining of the IRWST following a large-break LOCA, including the interaction between containment and the primary circuit. One may identify many more relevant situations (i.e., additional examples) if one looks at the condensation at the liquid-steam interface in the CMT at a reactor-scale heat transfer area when data from experiments are limited to a much smaller liquid-steam interface area. (It is consensus knowledge that condensation in the CMT may stop delivery of liquid from the CMT to the primary circuit.) This justifies the red triangle entering the green zone for the AP-600 in Fig. 2.

^g Examples of differences between the AP-600 and AP-1000, relevant to scaling and related to ADS 1, 2, and 3, are the detailed geometry of discharge lines (including spargers at the bottom) and the energy flows of steam in those lines, i.e., connected with decay core power. Condensation modalities and consequent pressure oscillations originating in the IRWST pool are of concern.

^h In different terms, curves DP versus L (where DP starts from the hot leg connection with the pressure vessel, and L is the abscissa along the flow direction in nominal conditions) should have been defined, matching to the best prototype values (unknown, but derivable at the time when concerned facilities where put into operation). The location of geometric discontinuities should have been the same as in the reactor design and not optimized, e.g., to facilitate discharge from the CMT, avoiding instabilities, undue condensation, etc.

ⁱ This paragraph constitutes a comment on the application of scaling procedures to experiments in models of the AP-600 primarily, and the AP-1000 (i.e., dealing with the meaning of SF as outlined in Sec. III.A), rather than an evaluation of the quality of the experimental database.

Additional information

Funding

This work was supported by the University of Pisa (2022-Institutional).