Beyond-design-basis screening by a three-bound critical excitation envelope for base-isolated nuclear power plants

In this paper, a resonance-focused Critical Excitation (CE) overlay is developed as a secondary check for base-isolated nuclear power plants. The goal is to capture worst-case, yet physically plausible, motions that align with the isolation period and key equipment periods, while remaining compatible with ASCE 4–16 and ASCE 43–19 practices. The method generates three CE bounds (lower, mean, upper) under explicit Arias intensity, PGA, and PGV constraints. At the same time, routine suite-mean results are read against these bounds to flag under-targeting, proximity, or design-extension concerns. Well-established isolated-reactor models, including two-degree-of-freedom with low-damping rubber, lead rubber bearing, friction pendulum system, and six-degree-of-freedom configurations, are used. As a case study, ground motions are selected for the Diablo Canyon site and matched to the SDC-5 Design Response Spectrum. A code-consistent suite of ground motion records forms the design baseline for averaging. A screening study compares acceleration- and displacement-targeted CE objectives. The displacement objective produces higher peaks in isolator displacement, isolation-plane base shear, and floor acceleration for six of seven seed ground motions; therefore, it is adopted for design-level evaluation. With the displacement objective, suite-mean responses are placed within the CE three-bound for key metrics, indicating conservative and stable estimates without missing resonance. The overlay provides clear decision triggers: (a) below the CE lower bound, supplementation or retuning is indicated; (b) near the CE mean, demand capture is adequate and remaining margins are checked; and (c) trending toward the CE upper bound signals a Beyond-Design-Basis Earthquake (BDBE) condition and prompts targeted checks. The CE overlay thus serves as a transparent, code-compatible safety gate and supports BDBE reasoning without any arbitrary multipliers.