Decentralized LQG Control for Adaptive High-Rise Structures

Abstract

Abstract Adaptivity of buildings introduces new challenges and opportunities for both architects and engineers. With the possibility of active load compensation, new types of lightweight structures can be realized. However, those demand suitable control engineering methods to ensure safe and robust control. In this contribution, we introduce an approach for decentralized linear quadratic Gaussian (LQG) control of adaptive structures. Many state of the art methods in decentralized structural control focus on damping the response of substructures that are either derived by decomposition of a global finite element (Fe) model or later assembled to form a complete structure. In contrast, we derive local models by means of model order reduction techniques which allows for fully decentralized control without the need to communicate states or estimate interaction forces. We demonstrate the decentralized control of local subsystems for an adaptive structures demonstrator building in simulations. Performance and energy demand are found to be comparable to a centralized controller which makes the presented approach well suitable for application. Monte Carlo simulations with both varying model parameters and system eigenvalues were conducted to analyze the robustness of the decentralized LQG controllers.