RECENT DEVELOPMENT OF EQUIVALENT FORCE CONTROL METHOD FOR HYBRID TESTING
B. Wu,G. Xu,L. Deng,Z. Chen,Z. Wang,H. Jiang,H. D. Wang,F. L. Wang
2009-01-01
Abstract:The equivalent force control (EFC) method replaces numerical iteration with a feedback control strategy to solve the nonlinear equations of motion in pseudodynamic and real-time substructure tests (RST) using an implicit integration method. The EFC method is briefly introduced first in this paper. Then the effectiveness and accuracy of the EFC method are reported with the full-scale quasi-static pseudo-dynamic tests of reinforced concrete (RC) shear wall structures and a reinforced masonry structure, and the RSTs of an offshore platform with MR damper specimen. It is shown that the EFC method can deliver excellent performance in all cases. Introduction In recent years, different approaches have been developed for real-time structural testing using servo-hydraulic actuators and reaction-wall or reaction-frame facilities. These include the effective force test method (Zhao et al. 2006) and real-time substructure test (RST) methods (Nakashima et al. 1992; Wu et al. 2005, 2006; Jung and Shing 2006). The latter is a hybrid experimental technique that combines numerical simulation with physical testing. While the effective force method is conceptually simple and does not require real-time numerical computation during a test, it is not as versatile as RST methods. While many different numerical algorithms are available for RST (Darby et al. 2001; Wu et al. 2005, 2006), for structures with multiple degree of freedom (MDOF), an integration method with unconditional stability is highly desirable. Many implicit integration schemes are unconditionally stable but they require an iterative solution strategy for nonlinear systems, which is a challenge for RST. Implicit schemes have been implemented with different solution strategies to handle structural nonlinearity for RST, see, e.g., Shing et al. (2004), Bayer et al. (2005), Mosqueda and Ahmadizadeh (2007). Bayer et al. (2005) and Jung and Shing (2006) have used an unconditionally stable implicit time integration method with a specially designed nonlinear solution strategy that combines a Newton-type iterative method with subincrementation. To avoid spurious Professor, School of Civil Engineering, Harbin Institute of Technology, Harbin 150090 PhD students, School of Civil Engineering, Harbin Institute of Technology, Harbin 150090 Lecture, School of Civil Engineering, Harbin Institute of Technology at Weihai, Weihai 264209 Associate Professor, School of Civil Engineering, Harbin Institute of Technology, Harbin 150090 Proceedings of the 9th U.S. National and 10th Canadian Conference on Earthquake Engineering Compte Rendu de la 9ième Conférence Nationale Américaine et 10ième Conférence Canadienne de Génie Parasismique July 25-29, 2010, Toronto, Ontario, Canada • Paper No 1451 loading and unloading of a specimen, the commands for the actuators are generated by a quadratic (Jung and Shing 2006) or linear (Bayer et al. 2005) interpolation based on iterative trial quantities. To avoid the numerical iteration process associated with implicit integration, Wu et al. (2007) have proposed the equivalent force control (EFC) method for RST. This paper describes basic concept of the EFC method, and recent applications to quasi-static pseudo-dynamic tests (PDT) and RST with various experimental substructures. These include a reinforced concrete (RC) shear wall, a three-story prefabricated RC shear wall structure, a three-story frame-supported reinforced masonry structure, and a MR damper specimen. Overview of EFC Method EFC with constant-average-acceleration method The concept of the EFC method can be explained by expressing the numerical solution of the equations of motion with the constant-average-acceleration (CAA) method in the following form (Wu et al. 2007) N 1 PD 1 E 1 1 1 EQ, 1 ( ) ( , , ) i i i i i i + + + + + + + + = R d K d R a v d F (1) where N N PD 2 4 2 t t = + Δ Δ M C K (2) N N N EQ, 1 1 N N 2 4 4 2 ( ) ( ) i i i i i t t t + + = + + + + + Δ Δ Δ M M C F F M a C v d (3) In above equations, d, v, and a are the displacement, velocity, and acceleration vectors; MN, CN, RN are the mass matrix, damping matrix, and static restoring force vectors of the numerical substructure; RE is the restoring force vector of the experimental substructure including static, damping, and inertia forces; F is the external excitation force vector; and Δt the integration time interval; KPD is called the pseudo-dynamic stiffness, and FEQ the equivalent force (EF). The solution of Eq.1 can be interpreted as finding the response of a hybrid system, which consists of numerical and experimental substructures with real and pseudo-dynamic forces, to an explicit equivalent load. This response can be obtained by directly applying the equivalent force to the hybrid system, in which the numerical substructure and pseudo-dynamic forces are evaluated in a computer, using a force-feedback control strategy. The block diagram representing an EFC system for a linear single-degree-of-freedom (SDOF) structure is shown in Fig. 1, in which KN and KE are the stiffness of the numerical and experimental substructures, respectively; CE is the damping coefficient of the experimental substructure; and TA(s) is the transfer function of the dynamics of the actuator-specimen system. The equivalent force controller shown in Fig. 1 is a proportional-integral-derivative (PID) controller, in which KP, KI, and KD are the proportional, integral and derivative gains, respectively. The factor KF after the equivalent force command FEQ,i+1 is used to eliminate the steady-state error. After being processed by a EF controller, the force error eEQ between the equivalent force command modified by KF and the equivalent force feedback F′EQ,i+1 is converted to a displacement command di+1 by a conversion factor CF. Then, the actuator is controlled with a displacement control mode. The term d′i+1 represents the displacement response of the experimental substructure subjected to the command di+1. At the end of the (i+1)th step, the measured reaction force RE and the calculated RN is fed back to Eq.1 to calculate the displacement at this time step, and the corresponding velocity and acceleration responses of the structure are calculated according to the CAA method. Figure 1. Block diagram of EFC with PID controller for a SDOF structure. Based on Laplace’s Terminal-Value Theorem (Ogata 2005), one can obtain the steadystate error of the whole system subjected to an EF input FEQ,i+1. For the case of KI=0, which means a P or PD controller is used, the steady-state error of the system subjected to a unit step EF command is 1/(1+KP). Therefore, KF should be equal to (1+KP)/KP to eliminate the control error (Wu et al. 2007). For the case of KI≠0, which means a PI or PID controller is used, the corresponding steady-state error is zero, requiring that KF be equal to 1. EFC with Implicit mid-point method The EFC is not limited to be used with the CAA algorithm. The implicit mid-point method (IMM) is another unconditionally stable integration algorithm. When the IMM was employed in PDT with EFC, the equation of motion can be rearranged as ( ) ( ) PD 1 N 1 E 1 EQ, 1 i i i i + + + + × + + = K x R x R x F (4) in which N N PD 2 4 2 t t ⎛ ⎞ = + ⎜ ⎟ Δ Δ ⎝ ⎠ M C K (5) N N N EQ, 1 1 2 2 4 2 2 i i i i d v t t t + + ⎛ ⎞ = + + × + ⎜ ⎟ Δ Δ Δ ⎝ ⎠ M C M F F (6) The block diagram of IMM-EFC is the same as Fig.1, but d should be replaced by x. The displacement and velocity at step (i+1) are determined with 1 1 2 i i i + + = − d x d (7) 1 1 2 2 i i i i t + + − = − Δ d d v v (8) d′i+1 eEQ + +