A Decentralized Control Algorithm for Large-Scale Building Structures

Decentralized control strategy is more suitable for structural control of large-scale structural systems as it increases in the feasibility of control implementation and decreases the risk on the failure of the control system compared with the conventional centralized control approach. In this article, a decentralized control algorithm is proposed for large-scale linear building structures. A large-scale building structure is divided into a set of smaller substructures based on its finite element model. Interconnections between adjacent substructures are treated as disturbances to the individual substructure. Each substructure is controlled by its own local controller using linear quadratic Gaussian control scheme with acceleration measurements as feedback signals. A computational procedure is developed for the recursive estimation of the unknown disturbances to each substructure. Two cases, with substructure interface measurement and without substructure interface measurement respectively, are considered. A numerical example of the decentralized control of the 20-story Structural Engineers Association of California (SAC) benchmark linear building under seismic excitation is studied to evaluate the performance of the proposed algorithm. Simulation results demonstrate that the decentralized control algorithm has quite good control performance compared with the conventional centralized control approach. Therefore, the proposed decentralized control algorithm is viable for structural control of large-scale linear structural systems.