Damage detection algorithm based on identified system Markov parameters (DDA/ISMP) in building structures with limited sensors

Abstract

In this study, a new approach for global/local damage detection in a finite element model of structures, with limited sensors, is proposed using identified system Markov parameters. The proposed damage detection is directly related to the Markov parameters, locations of actuators and sensors. Also there is no explicit relation between DDA/ISMP and mode shapes. So, it is unnecessary to install a sensor at each DOF for measuring output to identify mode shapes unlike the other damage detection techniques in which mode shapes play an important role in damage identification. The stiffness of all elements of a structure is identified using the proposed DDA/ISMP. The effects of noise, numbers and locations of sensors on the identification precision are investigated. The results demonstrate that, with the limited sensors and the noise contamination in the measured responses, the DDA/ISMP can effectively identify the locations, types and quantities of damages, both locally and globally. To illustrate the efficiency of DDA/ISMP, a four-storey steel moment frame structure and a five-storey shear building are used. Our numerical results show that the DDA/ISMP technique in damage detection is more effective than the scheme proposed by Xu et al. Also, the time consumed in DDA/ISMP is considerably less than the method introduced by Xu.

Developing a Smart Structure Using Integrated Subspace‐Based Damage Detection and Semi‐Active Control

“…In this article, structural characteristics of the reduced structure with stories are identified based on measured acceleration data of the master DOFs, considered as the system outputs, by using the damage detection algorithm based on identified system Markov parameters (DDA/ISMP) proposed by Amini and Karami (). Therefore, the motion equation of the condensed structure in Equation is rewritten in the continuous state space form as follows: in which, the vector …”

“…The relative identification error in story stiffness, , is defined as follows: where the vector is the identified stories stiffness of the primary structure. In this study, the noise signals are added to the measured responses according to a specified given noise intensity ratio, R , using Amini and Karami () noise contamination equation. Two noise intensities, including and , are considered in the measured response of the used structure to assess the effects of noise contamination on the identification precision.…”

“…Therefore, in this 20‐story structure, it is preferred to use at least four sensors under the high‐level noise condition. Here, the proposed method is compared with the DDA/ISMP method proposed by Amini and Karami () to demonstrate the reliability and feasibility of the obtained results. As shown in Figure a, case (I) with two sensors has the maximum relative identification error compared to the other cases.…”

“…Here, the proposed method is compared with the DDA/ISMP method proposed by Amini and Karami () to demonstrate the reliability and feasibility of the obtained results. In the DDA/ISMP method, it is necessary to install a sensor at each story to identify whole structure; thus, 20 sensors are employed in the used structure, which is close to case (IV).…”

Torsional control of asymmetric buildings using online 3‐D damage detection and adaptive stiffness devices

“…Therefore, to overcome these shortcomings, in this study, a new damage detection method is presented for identifying torsionally coupled building structures that can simultaneously detect the lateral and torsional stiffness of the each story during excitations in almost real time. Amini and Karami 53 proposed a local-global damage detection method in the finite element model of 2-D frame structures based on the identified Markov parameters of the system. In their method, the stiffness matrix of the structure has a direct relationship to the Markov parameters (impulse responses), the location matrices of the inputs and outputs of the system, which, unlike the other methods, it is not a function of mode shapes of the system.…”