Phase-Resolved Reorganization of Seismic Load Paths in a Full-Scale Mass-Timber Rocking-Wall Building

Self-centering rocking systems redistribute seismic demand as contact conditions at structural interfaces change during excitation. Conventional peak-response measures, global-response model calibration, and system-identification summaries can obscure the transient redistribution among diaphragms, wall couplings, and restoring components. This study addresses that gap with a phase-resolved analysis of heterogeneous measurements from fourteen full-scale shake-table tests of a two-story mass-timber building with post-tensioned cross-laminated timber rocking walls. Acceleration, strain, post-tensioning-force, and wall-uplift records were aligned, screened, and placed on a common analysis grid; uplift-defined operating phases were then evaluated using normalized subsystem activity, entropy, effective participation, conditional directed predictability, and dimensional force-uplift work. Median effective participation increased from 2.7 subsystems in the closed state to approximately 5.8 during rocking. The median closed-to-rocking Jensen–Shannon divergence was 0.195, the median post-recontact recovery index was 0.813, and restoring work was 4.6–10.7% greater on the south wall line. Rocking-phase floor-diaphragm activity co-varied with south restoring work (Spearman ρ = 0.873, p < 0.001). The new scientific result is that wall uplift reorganizes the composition of measured subsystem activity rather than simply scaling a fixed response pattern. The proposed indicators therefore provide complementary, phase-conditioned targets for experimental comparison and nonlinear-model validation, while remaining distinct from equilibrium force fractions, damage indices, or unrestricted causal measures.