Data-driven damage tracking and hysteresis evaluation of earthquake-excited structures with test validation. (15th March 2020)

• Record Type: Journal Article
• Authors: Shan, Jiazeng; Zhang, Hanqing; Ouyang, Yuting; Shi, Weixing
• Is Part Of: Engineering structures. Volume 207(2020)
• Publication Date: 2020-03-15
• Subjects: Damage evaluation -- Nonlinear behavior -- Data-driven -- Earthquake -- Model reference; Structural engineering -- Periodicals; Structural analysis (Engineering) -- Periodicals; Construction, Technique de la -- Périodiques; Génie parasismique -- Périodiques; Pression du vent -- Périodiques; Earthquake engineering; Structural engineering; Wind-pressure; Periodicals; 624.105
• DOI: 10.1016/j.engstruct.2020.110214
• Languages: English
• ISSNs: 0141-0296

Abstract

Highlights: A data-driven system-level damage tracking and hysteresis evaluation methodology, regarding interstory hysteretic degradation and global damping properties, is presented by adopting the model-reference concept. The behavior deviation due to structural damage and hysteresis can be tracked and quantified by the response-oriented damage index with three sub-level variables, which can account for the peak, cumulative and residual effects. The interstory hysteresis may be satisfactorily tracked and favorably quantified as compared to the modal frequency change and the engineering demand parameters. Abstract: A data-driven damage tracking and hysteresis evaluation method for earthquake-excited structures that adopts the model-reference concept is presented in this paper. A series of target substructures with local interstory hysteresis is first decentralized from the global structure system adopting the Rayleigh damping model. A linear reference model, designed to be identified from nondestructive seismic events, is used to represent the desired healthy behaviors. The tracking error of structural responses with a substructure-level loading reconstruction is subsequently introduced to quantify the potential behavior deviation during a disastrous earthquake. Such a structural behavior deviation, in terms of a normalized form, has been proposed in a novel data-driven damage model; accordingly, a logistic framework has been presented for the quantification of structural damage of interests. Two illustrative examples, including a numerical simulation of a three degree-of-freedom degradation model and an experimental investigation of a large-scale reinforced concrete frame through a shaking table test, have been employed. Comprehensive results and detailed discussion are presented to illustrate the damage tracking and quantification performances of the proposed data-driven evaluation method.