The feedforward minimal control synthesis algorithm and its application to the control of shaking-tables

Abstract

Since the early 1990s, research in Europe, the Far East, and America on the control of shaking-tables for seismic investigations has been made all the more urgent, following a series of catastrophic earthquakes across the globe. The fidelity of the results from shaking-table tests, on which many designs of earthquake-proof systems depend, is largely governed by the accuracy of the controlled motion of the table and the test structure.

Coincident with these events has been the development of more advanced, adaptive, control systems that provide an opportunity for high-precision motion control in the face of dynamic interaction, non-linearity, and parameter variation within the test system. This paper discusses a new adaptive control algorithm, called the feedforward minimal control synthesis (FFMCS) algorithm, which offers improved shaking-table control, when used in conjunction with a more conventional fixed-gain controller. FFMCS has been developed as part of a joint Japan-UK initiative, motivated by prior testing of shaking-tables located in both Japan and Europe.

This paper describes the dynamic modelling of a typical shaking-table, in order to emphasize the key features required of its controller. Also detailed are the three-variable control (TVC) strategy, a conventional algorithm that is widely used within shaking-table test facilities, together with the development and stability proof of the FFMCS algorithm. A motivating and comparative implementation study of FFMCS on the control of a simple single-axis servohydraulic rig is also included, as well as comparative simulation studies on the previously modelled shaking-table.