Enhancing structural coupling: A frequency-based methodology for optimal interface expansion

This paper presents a novel methodology for optimal interface expansion in dynamic substructuring, addressing inaccessible and continuous interfaces. Unlike modal parameter-based approaches, the proposed frequency domain method directly targets interface Degrees of Freedom (DoFs), avoiding modal identification errors. The System Equivalent Model Mixing (SEMM) technique is employed to merge numerical and experimental models into a hybrid model, with the Global Coherence Correlation Metric (GCCM) used as a correlation metric incorporating both phase and magnitude of Frequency Response Functions (FRFs). Optimal Sensor Placement (OSP) is formulated as a metaheuristic search problem, solved using the Mountain Gazelle Optimizer (MGO) to efficiently navigate large solution spaces. Methodology verification and validation are performed on three case studies: two identical cantilever beams with 17 accessible and 3 inaccessible interface DoFs, a square free-free plate with 36 accessible DoFs, and L-shaped free-free plate experimentally tested with 32 internal and 17 interface DoFs. Results show optimal correlation over energy-based placement methods across all cases. A Monte Carlo noise study with 1,134,000 simulations indicates that under low-noise conditions (Γ > 0.95), optimal configurations favour spatially distributed sensors with aggregation indices ≥ 2.5, whereas higher noise levels promote clustered arrangements (1.5≤A≤2.0). Experimental validation of the L-plate yields maximum mode errors of 6.72% between expanded and measured responses, confirming the method’s capability to accurately predict interface dynamics from optimal accessible measurements.