Modified dam deformation monitoring model considering periodic component contained in residual sequence
Dongyang Yuan, Bowen Wei, Bin Xie, Zimeng Zhong
Abstract
Considering that dam deformation and its influencing factors present complex nonlinearities, a modified deformation monitoring model is proposed in this study. On the basis of the systematic analysis of the effects of environmental influencing factors on dam deformation, this study attempts to explore the functional relation between dam deformation and sediment deposition based on the theory of classical dam deformation monitoring model, and the theory of dam deformation monitoring is developed. Meanwhile, considering that the residual sequence of monitoring model contains certain periodic components, singular spectrum analysis is adopted to denoise and to reconstruct residual by extracting the trend and periodic components, and then the reconstructed residual sequence is trained and forecasted by autoregressive integrated moving average model. By superimposing the residual forecast value with forecast value of classical statistical model, a modified deformation monitoring modeling method is established. Examples show that, compared with conventional models, the forecast capacity of the proposed method is improved to a large extent, which effectively corroborates the rationality and effectiveness of the modeling method. A new technical support for guaranteeing the safe operation of dams is provided.