Developments in the Design of Multivariable Controllers for the Structural Testing of Systems with Unknown Dynamics

Multi-actuator structural testing has traditionally been regarded, for control purposes, as a multi-loop, single-input single-output problem. This approach does not take interaction into account, which could be considerable. The result of this is often poor laboratory reproduction of actual service data or the employment of complicated and usually time-consuming data analysis techniques in order to produce the desirable results. This paper shows how this problem can be eliminated by the use of multivariable frequency-domain methods in the design of the servo-controllers. In particular, the method used is the Nyquist array, extended with the use of Perron-Frobenius theory of positive matrices. The controllers produced by the proposed method are simple and easy to implement, and the process does not involve mathematical modelling of the structure and the test equipment. Although the results presented in this paper are produced by simulation they are expected to be accurate since the primary frequency response data is obtained experimentally. Interaction has been reduced to 10% and flat frequency responses achieved. Simulation confirmed that field data can be accurately reproduced in the laboratory.