Numerical and Experimental Validation of a Novel Force Correction Method for Hybrid Simulation Applied to the Partitioned G-α Algorithm

This paper presents the development of a novel force correction methodology integrated into the partitioned Gα algorithm for hybrid simulation (HS). The method was applied within an experimental campaign conducted at the University of Trento on steel frames equipped with energy dissipative devices, as part of the European project DISSIPABLE. The hybrid simulation technique adopted in this campaign relies on dynamic substructuring, allowing for the simulation of the nonlinear response of complex mechanical systems by coupling physical and numerical substructures. A key challenge in the implementation of HS lies in discrepancies between actuatorimposed displacements and the actual displacements of the physical substructure, which can compromise both the stability and the accuracy of the simulation. Thus, the proposed correction allows to perform tests, maintaining the internal control of actuators even in presence of axial deformability of the specimens and bolt-hole clearance. The paper details the algorithmic formulation of the correction scheme and evaluates its performance through convergence and error propagation analyses on benchmark problems. Experimental validation demonstrates that the correction strategy yields conditionally stable and first-order accurate results. The applied force corrections remained below 30% of the total force feedback, with the required correction decreasing as seismic intensity increased, reflecting the fact that bolt-hole clearance effects are more prominent at low loading levels.