Dam safety is an important basic part of national water network security. Building a dam deformation prediction model based on monitoring data is crucial to ensure dam safety. However, traditional statistical regression methods have shortcomings, such as a weak nonlinear fitting ability when constructing dam deformation monitoring and prediction models. The residual part of the statistical regression results usually contains parts that cannot be effectively explained by the linear regression method, that is usually highly variable and noisy. In this study, the phase space reconstruction method is used to smooth the residual term of the statistical regression model to eliminate noise interference. On this basis, an improved long short-term memory (LSTM) neural network is used to learn the nonlinearity contained in the residual term of the linear regression. Considering the impact of parameter selection on model performance, the gray wolf optimization (GWO) algorithm is used to determine the optimal parameters of the model for better performance. A high arch dam is used as a case study, with multiple measuring points used as research objects. The experimental results show that the phase space reconstruction can effectively smooth the high-frequency components in the residual term and remove noise interference. In addition, the GWO algorithm can effectively determine the hyperparameters of the LSTM network, thereby constructing a residual prediction model with high prediction accuracy. The combination of statistical models and deep learning prediction methods can effectively improve the model prediction performance while preserving the model interpretability and transparency.

The multi-level modeling technology of Building Information Modeling (BIM), combined with Three-dimensional Geographic Information System (3DGIS) macro-scene visualization technology and location information, can realize the transmission of decentralized information from various disciplines to multi-disciplinary collaborative information sharing services. It can be applied independently for the whole life cycle, which plays a positive role in reducing the cost and improving the efficiency of engineering planning, design, construction, operation, and maintenance. In this paper, the data integration and function integration methods of 3DGIS and BIM are designed. In order to avoid the breaking problems caused by attribute information loss and excessive simplification in the process of BIM data integration, the attribute mapping between 3DGIS and BIM based on Industry Foundation Classes (IFC) and City Geography Markup Language (CityGML) and the data simplification method considering the geometric characteristics of BIM are designed. By setting the relevant preconditions and thresholds of patch merging, on the premise of maintaining the structural characteristics of BIM data surface, reduce the amount of model data to improve the efficiency of BIM data loading, rendering and display effect in 3D geospatial scene. Through the data and function integration of 3DGIS and BIM, we can effectively manage the data of large-scale model, and calculate and obtain the geospatial location and direction of key parts of buildings through the coordinate transformation of BIM, which can effectively assist the rapid and accurate positioning of BIM in virtual 3D scene and expand the visualization ability of 3DGIS. By effectively integrating 3DGIS and BIM, this paper gives full play to the spatial management advantages of 3DGIS and the component management advantages of BIM. The rationality and operability of the method are verified by its application in the operation and maintenance management project of concealed facilities in actual buildings.

To address the disconnection between carbon emission accounting and building information modeling (BIM), inconsistent accounting boundaries, the lack of systematic carbon optimization methods, and the absence of construction-stage tracking in hydraulic engineering projects, a BIM-based life-cycle carbon emission accounting and optimization method is proposed. A BIM carbon information model is established based on the IFC standard, with a three-level “component-material-emission factor” mapping that enables automatic extraction of material quantities and automatic updating upon design changes. An emission factor database covering four stages (material production, construction, operation and maintenance, demolition and recycling) is constructed and a life-cycle carbon accounting model with itemized, traceable calculation sheets is developed. Three-dimensional heat maps are used to visualize component-level carbon distribution. Taking life-cycle carbon emission and total project cost as objectives, the NSGA-II algorithm is employed to derive the carbon-cost Pareto front under structural safety, schedule and cost constraints. IoT-based real-time data enables dynamic carbon tracking during construction. A case study of a concrete gravity dam shows that the total life-cycle emission is 0.3945 MtCO2, with the material production stage accounting for 84.1% and cement production accounting for 72.4% of the total, making it the largest emission source. Substituting slag or fly-ash cement reduces cement-related emissions by 24.7%-28.8% (equivalent to 21.3%-24.8% of the material production stage). Rail-road combined transport and waterway-road combined transport cut aggregate transport emissions by 61.9% and 57.1%, respectively. NSGA-II yields 15 Pareto-optimal solutions; within the near-zero-cost interval, emissions are reduced by 19.8% at only 0.2% additional cost (unit abatement cost about 26 CNY/t), while the deep-decarbonization scenario reduces emissions by 33.5% with the marginal abatement cost rising to 559 CNY/t, revealing an increasing marginal abatement cost pattern. Dynamic tracking corrects the overrun deviation from +8.2% to +3.5%. The proposed method provides systematic technical support for low-carbon construction of hydraulic engineering projects.