Structural Coherence Analysis of Nuclear Reactor Anomalies in Small Modular Reactors via LESE-EIA Framework

We apply the LESE-EIA framework to the detection and structural classification of anomalies in small modular reactors (SMRs), using the PWR-SMR-2026-01 dataset (1,518 high-fidelity OpenMC statepoints, three anomaly classes). The LESE coherence metric is derived from the single postulate dW/dτ = κ(1+log W), yielding the universal fragility threshold W* = e⁻¹ ≈ 0.3679 without calibration. Analysis at volumetric level (70×70×50 mesh) and scalar level (all 1,518 statepoints) achieves 100% recall with zero calibration parameters. LESE captures structural deviation for fuel temperature perturbations below the Monte Carlo noise floor that single-frame statistical methods cannot detect.