Combination forecast model for concrete dam displacement considering residual correction

Abstract

In conventional dam displacement monitoring models, forecast precision is below the standard, the fitting residual sequence contains chaotic components, and information mining of dam prototype observation data is limited. In consideration of the chaotic characteristics of the fitting residual sequence in regression model, the multi-scale wavelet analysis is used to decompose and reconstruct the residual sequence in this study; back propagation neural network and autoregressive integrated moving average model are used to forecast the reconstructed residual sequence by identifying the high-frequency and low-frequency characteristics of signals. By superimposing the residual forecast value with the forecast value of regression model, the combination forecast model for concrete dam displacement considering residual correction is proposed. Examples show that, compared with conventional models, the proposed combination model is better in fitting precision and convergence speed. Forecast capability is significantly improved for dam displacement forecast when effective components contained in residual sequence are considered. A new method of displacement forecast for high slope and other hydraulic structures is presented.

Modified hybrid forecast model considering chaotic residual errors for dam deformation

“…Chaos is a kind of irregular random behavior with the characteristics of initial value sensitivity and ergodicity. Related researches [21][22][23][24]47] show that the chaos phenomenon exists both in dam deformation prototype observation data and residual of dam deformation monitoring model. In practical engineering, dam deformation is not only effected by the water pressure, temperature change, but also effected by many other internal and external environments, for example, uplift pressure, sediment pressure, seepage, structural damage, material aging, and monitoring error.…”

“…The existence of chaotic effect components in dam prototype observation data and fitting residual sequence has been effectively validated. [21][22][23][24] Effectively increasing the information mining scale and considering the chaotic components contained in residual sequence will present great significance to improve the capability of forecast model.…”

Displacement monitoring model of concrete dams using the shape feature clustering‐based temperature principal component factor

“…Under the water pressure of a reservoir, the displacement of a concrete dam is composed of three parts: the deformations of the dam body and foundation, and the rotation of the reservoir basin. 5 The hydraulic component can be represented by a polynomial function of the reservoir water depth as follows:…”

“…For normally operated dams, the time effect displacements rapidly increase in the initial operation stage and tend to be stable later on. Thus, the time effect component can be represented by a combination of the linear and logarithmic function as follows 2,5,28 :…”

“…The former mainly includes statistical models and finite element method (FEM)-based hybrid models, in which specific functions are predefined for the three most significant deformation components: hydraulic, temperature, and time effect. [4][5][6] These explicit mathematical expression-represented models can be used not only for a prediction but also for causal mechanism interpretations of the deformation behavior of concrete dams. Furthermore, some advanced machine learning tools, including the neural network, 7,8 the support vector machine, 9,10 the random forests, 11 the extreme learning machine, 12 and the boosted regression tree, 13 have been indicated to possess strong data mining abilities and have also been employed to establish prediction models.…”