Structural identification of super high arch dams using Gaussian process regression with improved salp swarm algorithm. (1st July 2023)

• Record Type: Journal Article
• Authors: Kang, Fei; Wu, Yingrui; Ma, Jianting; Li, Junjie
• Is Part Of: Engineering structures. Volume 286(2023)
• Publication Date: 2023-07-01
• Subjects: Arch dams -- Structural identification -- Inverse/back analysis -- Gaussian process regression -- Improved salp swarm algorithm; 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.2023.116150
• Languages: English
• ISSNs: 0141-0296

Abstract

Highlights: A back analysis framework is proposed for structural identification of concrete dams. The prototype measured data of dam are used to build displacement monitoring model. The Gaussian processes surrogate model is used to replace finite element calculation. An improved salp swarm algorithm is proposed to identify the material parameters. The proposed method is verified on a super high arch dam and shows good performance. Abstract: Structural identification is critical for evaluating the operation of high arch dams, and has become an important topic. This paper proposes a novel inverse analysis framework combining an improved salp swarm algorithm and Gaussian process regression to identify the material parameters of concrete dams. First, the hydrostatic component is separated from the prototype measured data using the multiple linear regression method. Subsequently, the Gaussian process regression is adopted to build the mapping relationship between elastic modulus and the calculated relative hydrostatic component of the dam. Finally, the improved salp swarm algorithm is used to estimate the real elastic modulus of the dam by minimizing the discrepancy between the measured and calculated relative hydrostatic components. The performance of the proposed method is verified on a real concrete arch dam with sufficient monitoring data. Results show that the proposed inverse analysis method can identify the material parameters accurately and efficiently.