Deterministic combination prediction model of concrete arch dam displacement based on residual correction. (October 2022)

• Record Type: Journal Article
• Authors: Xiong, Fangjin; Wei, Bowen; Xu, Fugang; Zhou, Lingkai
• Is Part Of: Structures. Volume 44(2022)
• Page Start: 1011
• Page End: 1024
• Publication Date: 2022-10
• Subjects: Concrete arch dam -- Residual sequence multiscale permutation entropy -- Long-term and short-term memory network -- Gated recurrent unit -- Auto-regressive integrated moving average model; Structural engineering -- Periodicals; 624.1
• DOI: 10.1016/j.istruc.2022.08.040
• Languages: English
• ISSNs: 2352-0124

Abstract

Abstract: The traditional deterministic model of arch dam displacement can separate the component effects of hydraulic, temperature, and aging. However, the prediction accuracy is not high, the residual sequence contains chaotic elements, and the dam prototype observation data have not been fully explored. Thus, a deterministic combination prediction model of arch dam displacement considering residual correction is proposed. According to the chaotic characteristics of the residual sequence of the deterministic model, multiscale permutation entropy is introduced to screen the intrinsic modal components with obvious noise, and the residual sequence is decomposed and reconstructed by the combined denoising method of adaptive empirical mode decomposition and wavelet packet. The residual sequence after reconstruction is predicted by identifying the high-and low-frequency characteristics of the signal and by using the long short-term memory network model, gated recurrent unit model, and auto-regressive integrated moving average model. Based on the superposition of residual prediction value and deterministic model prediction value, a deterministic combined prediction model of concrete arch dam displacement considering residual correction is established. The engineering case shows that compared with the fitting accuracy of the traditional model, the fitting accuracy of the combined model is greatly improved. This result can provide a reference for the displacement prediction of similar hydraulic structures.