Two spatial association–considered mathematical models for diagnosing the long-term balanced relationship and short-term fluctuation of the deformation behaviour of high concrete arch dams

Abstract

The safety of a high concrete arch dam should be rapidly diagnosed from different angles. Displacement is an actual comprehensive reflection of the arch dam, and it is very important to diagnose the overall deformation behaviour by displacement-based mathematical monitoring models. In this article, based on the spatial association validation of the measured displacement of two high arch dams by the empirical orthogonal function decomposition and the Pearson correlation analysis, two spatial association–considered mathematical models were proposed for the dam displacement of multimonitoring points: one model for the long-term balanced relationship and one model for the short-term fluctuation. To diagnose the abnormality of the dam long-term spatial association, each displacement time series of the multimonitoring points on the dam body with strong spatial associations was decomposed by wavelet multiresolution analysis, and the decomposed high-frequency components, which had the same periodicity as the causal factors of the reservoir water level or air temperature, were determined to establish the cointegration monitoring model. The second model was a combination prediction model, with two sub-models established from the modelling angles of the hydraulic, seasonal and time causal factors and the adjacent point displacement factors, and this second model was mainly used for identifying dam short-term local abnormal deformation behaviour. Engineering examples show that the deformation behaviour of an arch dam under normal conditions has strong spatial associations. The two proposed models have high accuracy and interpreting ability and can effectively reduce the number of needed monitoring models.

Hybrid artificial intelligence-based inference models for accurately predicting dam body displacements: A case study of the Fei Tsui dam

“…The influential factors for the dam body displacement at different elevations were determined as follows: For EL172.5, the influential factors were water level (1), air temperature (3), and dam body temperature (17); for EL150.0, they were water level (1), air temperature (3), and dam body temperature (17); for EL115.0, water level (1) and dam body temperature (15); for EL90.0, water level (1) and dam body temperature (10); and for EL57.5, water level (1) and dam body temperature (7). The factors and subfactors for these observation points (Table 5) were employed as parameters for the AI inference models in this study.…”

“…Then, the three-dimensional finite element analysis was conducted to establish a function mapping the correlation between the actual and measured displacements through a regression method. Wang, Xu 17 established two displacement-based mathematical models to diagnose the overall deformation behavior of a high arch dam.…”

Hybrid model for concrete dam deformation in consideration of residual correction by frequency division

“…11,12 Therefore, many scholars focus on the set collective innovation of intelligent algorithms and decomposition tools. Wang et al 13 proposed a multipoint hybrid model based on principal component analysis to predict the deformation information of the entire dam. Mata 14 showed that the artificial neural network (ANN) model is an effective method to represent the performance of concrete dam behavior under environment loads.…”

A spatial association-coupled double objective support vector machine prediction model for diagnosing the deformation behaviour of high arch dams

“…However, mathematical models are mostly established for single monitoring point and from the single modelling angle of causality, whereas spatial associations of displacements at multiple monitoring points are ignored. 32,33 To solve this problem, data fusion technologies, including the PCA, projection pursuit analysis and entropy, have been used to extract useful information from measured displacements of multiple monitoring points, and special mathematical models or monitoring indexes have been established to determine the overall variation of dam displacement field. [34][35][36][37] To consider the similarity of displacements among multiple monitoring points, Chen et al 38 proposed a spatio-temporal clustering and diagnosis method for concrete arch dams.…”