Identification of simplified structural models from input-output time histories and application to controller design

A technique, based on methods taken from adaptive control, to identify simplified models of complex building structures using measurements of the input excitation and forces and the output displacement or acceleration of the structure, as well as some basic assumptions about the structure of the simplified model to be identified, has been developed. Examples are presented in which the input and output to relatively complex plane frame models are used to identify simplified models with significantly fewer degrees of freedom. One of the complex frames is taken from a widely regarded benchmark for controller design. The simplified models identified using the proposed technique are successfully used in controller design applications. One of these applications involves the design of multi-objective optimal controllers. Another application places the identified model in a building control and sensor monitoring system. In both of these applications, testing of the designed control system is performed using an evaluation model of the benchmark building provided by the authors of the benchmark study.