Precise integration strategy for aseismic LQG control of structures

Abstract

SUMMARYAlthough LQG regulators have been used in many engineering fields, the relevant control technique still requires further development, particularly when the disturbances are random and non-stationary. Therefore this paper presents a new method for non-stationary seismic random excitations in which LQG control is applied to adjacent tall buildings which are connected together by active hydraulic actuators. The ground motion is induced by non-stationary uniformly modulated white noise excitation of the bedrock which passes through a single layered soil which is regarded as a single-peak filter. This represents an advance on other papers in the literature, which use stationary random white noise or filtered white noise as the ground excitation. The method uses the pseudo excitation method in conjunction with two different precise integration schemes to analyse structural non-stationary seismic random responses via solution of the differential Riccati equations of this transient LQG control process. Unlike previous work in the literature, the performance index used has a finite time region, rather than an infinite one. Reasonable control effects are demonstrated via numerical simulation for two adjacent 20-storey buildings which are connected by hydraulic activators.

Robust H ∞ control for aseismic structures with uncertainties in model parameters

“…To evaluate the responses of buildings under random seismic excitation, the pseudo-excitation method (PEM) was used (Lin et al, 2001;Lin and Zhang, 2005;Song et al, 2006), which has been proven to be a rigorous, exact and highly effi cient method for random vibration analysis. Figure 4 depicts the time-varying standard deviations of q 1 and , for perturbations of ΔK= +20% and ΔC= +20%.…”

“…masts and long bridges, to reduce damage caused by earthquakes, wind or other environmental forces. Bakioglu and Aldemir (2001) proposed a numerical algorithm for the sub-optimal solution of optimal closedopen loop control based on the prediction of on-coming earthquake excitations, in which Taylor series and the Kalman fi ltering technique were both used; Akhiev et al (2002) proposed a multipoint instantaneous optimal control of structures; Wu and Yang (2000) applied the LQG control strategy to reduce the acceleration responses of a 310m TV tower subjected to strong winds with an active mass driver installed on the upper observation deck; and Song et al (2006) presented a precise integration strategy when using the LQG control strategy to avoid collision between adjacent tall buildings during earthquakes.…”

Comprehensive Assessment of the Seismic Performance of an Innovative Hybrid Semiactive and Passive State Control System for a Low‐Degree‐of‐Freedom Structure Using Real‐Time Hybrid Simulation

“…In recent years, structural engineering has focused on developing and implementing techniques to reduce seismic vulnerability in structural systems, while encouraging resilient designs that prioritize occupant safety [1][2][3]. Structural control systems have been extensively investigated and implemented in response to these performance requirements, improving the seismic response of civil structures using external devices [4], such as passive control systems [5][6][7][8][9][10][11][12][13][14][15][16][17], active control systems [18][19][20][21][22], semiactive control systems [23][24][25][26][27], and more commonly hybrid control systems [28][29][30][31][32][33][34][35][36]. Passive control devices have been the most widely adopted in buildings because of their lower installation and maintenance costs, lack of energy requirements, and low capacity to cause structural damage.…”

Earthquake Response Reduction of Building Structures Using Learning-Based Lattice Pattern Active Controller

“…The physical properties of a structure need to be known to control the building structure by general control methods such as linear-quadratic Gaussian (LQG), linear quadratic regulation (LQR), and proportional-integral-derivative (PID) control [Song et al, 2006;Alavinasab and Khajepour, 2006]. If they are undefined, an artificial neural network (ANN) method can be useful.…”