To ensure the safety of concrete dams, a large number of monitoring instruments are embedded in the bodies and foundations of the dams. However, monitoring data are often missing due to failure of monitoring equipment, human error and other factors that cause difficulties in diagnosis of dam safety and failure to precisely predict their deformation. In this paper, a new method for imputing missing deformation data is proposed. First, since the traditional deformation increment speed distance index of the deformation similarity index does not take into account the fact that there is little change in deformations occurring in two consecutive days, the denominator of the index tends to be equal to zero. In this paper, an improved index for solving this problem is proposed. A combined weighting method for calculating the deformation similarity comprehensive index and the k-means clustering method is then proposed and used to classify deformation monitoring points. Subsequently, a panel data model that imputes different types of missing data is established. The method proposed in this paper can impute missing concrete dam deformation data more accurately; therefore, it can effectively solve the missing deformation monitoring data problem.

Deformation is the most intuitive indicator of the actual working status of a concrete dam. Zoning the variation regulation of dam deformation is one of the key parts of dam safety evaluation and risk assessment. However, the sample points reflecting deformation and variation characteristic information are non-uniformly distributed, thus it is difficult to cluster the data samples by traditional clustering methods. To solve this problem, a spatio-temporal zoning method of dam deformation considering non-uniform distribution of monitoring information is proposed. Firstly, the preprocessed deformation data are utilized to establish the similarity-distance zoning indicators using the absolute deformation, the deformation increase and the relative deformation increase respectively; then the deformation data are transferred into the Cartesian coordinate system, known as sample points. Secondly, utilize the improved Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm to cluster the points. The clustering parameters $M$ and $\delta $ are determined by an optimization algorithm with an evaluation index as the objective function, then the sample points representing time sections or spatial monitoring points are clustered through dynamically updating the neighborhood radius value $\varepsilon $ . Moreover, several artificial data sets are selected to demonstrate that the improved DBSCAN algorithm is with more obvious superiority in non-uniform clustering compared to traditional algorithms. Deformation data of an existing concrete dam are presented and discussed to validate the established zoning method.

In view of the problem that traditional dam outlier identification methods mostly rely on single-monitoring-point models and do not fully consider the spatio-temporal correlation characteristics of deformation between monitoring points, which can easily lead to the misdiagnosis of outliers, this paper proposes a novel Ward-VMD-BiLSTM-Iforest method for identifying gross errors in dam deformation monitoring. By integrating spatio-temporal clustering, variational mode decomposition (VMD), and BiLSTM neural networks, the method effectively identifies outliers while avoiding the misclassification of data mutations caused by environmental changes. Compared to traditional models (GRU, LSTM, and BiLSTM), the HHO-BiLSTM model demonstrates superior performance, achieving an R² of 0.97775 at TCN08, with a reduced MAE and better accuracy. In comparison with the Raida and Romanovsky criteria, the proposed method achieves 100% precision and 100% recall, significantly improving detection accuracy and reducing misjudgment. This method provides an effective and reliable solution for dam deformation outlier detection.

The safe operation of cascade dam groups is critical to water resource management and disaster prevention. However, existing methods are ill‐suited for real‐time structural health monitoring (SHM) and system‐level health assessment due to their static nature and inability to model dynamic interdependencies. To bridge this gap, this study proposes ordered entropy as a novel, systems‐theoretic metric that quantifies the dynamic health state of an entire dam group. We develop an integrated, monitoring‐data‐to‐health‐assessment framework that systematically transforms raw monitoring data through robust preprocessing (Apriori–DBSCAN), spatial field reconstruction (Co‐Kriging), and multisource information fusion (order degree with PCA). By integrating the maximum entropy principle and extreme value theory with the derived entropy sequence, we establish a novel method for estimating the real‐time failure probability as a health state quantification output that serves SHM of hydraulic systems. Applied to the Y River cascade dam group, the framework yields a service‐state failure probability of 1.33 × 10 ⁻⁶ –1.47 × 10 ⁻⁶ per year, demonstrating its rationality and practicality, with the temporal variation of failure probability showing a dominant correlation with reservoir water level—consistent with engineering mechanics principles and operational observations. This work provides both a new theoretical paradigm and a practical toolchain for advancing data‐driven SHM, prognostic health management, and operation state awareness of hydraulic infrastructures.