The shaking table is an essential testing tool in the development of earthquake resistant buildings and infrastructure, so improving its performance is an important contribution to saving lives. Currently the bandwidth and accuracy of shaking tables is such that earthquake motion often cannot be replicated with the desired fidelity. A new model-based motion control method is presented for multi-axis shaking tables. The ability of this method to decouple the control axes is demonstrated. A linear parameter varying modal control approach is used – i.e. the modes of vibration of the system are controlled individually, with the modal decomposition repeated at each time step to account for parameter variations. For each mode, a partial non-linear dynamic inversion is performed in the control loop. Feedback is based on a combination of position and acceleration measurements. A command feedforward method is proposed to increase the tracking bandwidth, thus the controller has a two degree-of-freedom structure. Experimental and simulation results are presented for a large (43 tonne total) six degree-of-freedom shaking table. The simulation results are based on a detailed, validated model of the table. Experimental results show that the controller gives exceptional performance compared a conventional proportional controller: for example the horizontal acceleration bandwidth is six-times higher at over 100Hz, which is also many times higher than the hydraulic resonant frequency. These results will allow a step change in earthquake simulation accuracy.
- shaking table, electrohydraulic servosystem, earthquake simulation, multi-axis control, parallel kinematic mechanism, modal control
FingerprintDive into the research topics of 'Model-based motion control for multi-axis servohydraulic shaking tables'. Together they form a unique fingerprint.
- Department of Mechanical Engineering - Head of Department
- Centre for Power Transmission and Motion Control
- Centre for Autonomous Robotics (CENTAUR)
- Institute for Mathematical Innovation (IMI)
- EPSRC Centre for Doctoral Training in Statistical Applied Mathematics (SAMBa)
Person: Research & Teaching