Author ORCID Identifier

https://orcid.org/0000-0002-8469-8391

Date of Award

17-8-2025

Document Type

Thesis

School

School of Mechanical Engineering

Programme

Ph.D.-Doctoral of Philosophy

First Advisor

Dr.M.Ganesh

Keywords

Robotics, Parallel Manipulator, Hybrid Shake Table, Mathematical Modelling, System Identification and Control, Parameter Estimation, Load-Balancing Mechanism

Abstract

Catastrophic earthquakes endanger human lives by severely damaging critical infrastructure such as nuclear plants, bio-laboratories, and bridges. Developing earthquake-resistant structures is vital to mitigate these risks. Shake tables replicate recorded earthquake ground motions to assess structural performance. This research presents a novel three-legged, six degrees of freedom (6-DOF) hybrid shake table with a cost-effective design and decoupled motion control, offering a viable alternative to conventional six-legged models.

This architecture effectively decouples actuator motions by employing hydraulic actuators for 3-DOF spatial movement and electromechanical actuators for 3-DOF planar motion. Servo-hydraulic actuators support the static payload against gravity, while lightweight electromechanical actuators enable planar motion.

In the hybrid shake table, actuator assemblies and planar links cause inaccuracy while positioning the platform with a heavy payload due to the cantilever effect. While precise fabrication reduces inaccuracy, sustained loading weakens stiffness and necessitates an effective load-balancing mechanism. A passive load-balancing mechanism is preferred to eliminate actuator redundancy. However, its integration introduces complex coupled ordinary differential equations (ODEs), requiring system identification for precise motion control.

A fusion-based system identification approach was devised to select a suitable passive load-balancing mechanism from multiple alternatives using minimal experimental data. This approach integrates experimental data for identifying the shake table's parameters with simulation data for characterising passive load-balancing mechanisms. A regression-based nonlinear least-squares (NLS) technique enhanced by the trust-region-reflective (TRR) algorithm is proposed for system identification.

Among the passive mechanisms analysed, the hydraulic damper demonstrated superior load-balancing performance. The use of a passive hydraulic damper to counteract the load-balancing effect is an innovative approach in parallel manipulator research, improving stiffness and damping characteristics. Additionally, its impact on motion control and overall system performance is analysed.

The electromechanical actuator assembly of the hybrid shake table consists of a servo motor, gearbox assembly, electric cylinder with a ball screw drive, and a recirculating ball linear motion guide block, requiring detailed modelling that accounts for frictional effects. Consequently, a state feedback controller is designed to ensure precise motion control.

Download

Share

COinS