Multi-input multi-output random vibration control of a multi-axis electro-hydraulic shaking table

An online multi-input multi-output (MIMO) random vibration control method is presented for a multi-axis hydraulic shaking table system to replicate the reference acceleration power spectral density (PSD) by the reproduction of the acceleration time domain histories that are proportional to the reference. The offline iterations are implemented using conventional random vibration control to compensate for the changes in the frequency response function (FRF) of the system. The successful test mainly depended on the experience of the operator. In online MIMO random vibration control, the reference PSD is transformed into time domain histories by filtering white noise. The impedance matrix of the shaking table system, which is the inverse of the FRF, is calculated and updated online. The time histories are compensated based on the system impedance matrix and then used as the drive signals to the shaking table. The overlap-save method and fast Fourier transformation are proposed to implement the transformation and compensation online. Two-axis random vibration experimental results obtained with the proposed control method show that both the reference PSD and the time domain histories can be replicated with high precision in the output of the shaking table without operator intervention.