Shaking table real‐time iterative control using online system matrix correction

Offline iterative control (OIC) is a widely employed technique in shaking table tests for accurately reproducing earthquake waveforms. However, repeated offline iterations can cause cumulative damage to fragile specimens, while the continuously changing dynamic characteristics of nonlinear specimens reduce the control accuracy of OIC. To overcome these limitations, real-time iterative control (RIC) has been introduced and applied to eliminate the need for multiple iterations. To further improve the stability and accuracy of RIC, this study introduced RIC with online system matrix correction (RICSC) method, discussed the control performance of the RICSC method. The RICSC method evaluates the accuracy of the identified system matrix using the following indices: the coherence function (CF) weighted sum, the CF, and the autocorrelation power density spectrum (AS). Based on these evaluations, the system matrix is corrected via frame correction (FC) or frequency point (FP) correction algorithms, thereby enhancing waveform reproduction accuracy and control stability. The performance of the RICSC method was verified via numerical simulations and shaking table tests under 20 test conditions. The results show that the FP correction algorithm (RICSC-FP) achieves the fastest convergence of absolute error, and its reproduction accuracy is higher than those of the traditional RIC and FC (RICSC-FC) algorithms. Both numerical and experimental results demonstrate that the RICSC method provides higher reproduction accuracy than OIC after just one iteration.