Structural identification of super high arch dams using Gaussian process regression with improved salp swarm algorithm

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. Subse-quently, 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.