The measured displacement-driven structural health monitoring (SHM) model is a significant approach to interpreting the influencing mechanism of environmental loads on dam displacement and to predicting the future operating behavior. Considering the large ambient temperature difference in cold area and the insufficient interpretive performance of the hydrostatic-seasonal-time (HST) model on thermal displacement behavior caused by the short-term dynamic fluctuations of ambient temperature, a hydrostatic-thermal-time (HTT) model considering hysteresis effect of boundary temperature is proposed to quantitatively analyze the influencing mechanism of environmental loads on displacement behavior of concrete gravity dams in cold area. Meanwhile, to improve the predictive performance of the proposed HTT model, whale optimization algorithm (WOA)-optimized gated recurrent unit (GRU) network is employed to effectively excavate the complex nonlinear relationship between dam displacement and its explanatory variables. A roller compacted concrete (RCC) gravity dam in cold area is taken as an example, and the cause of the disharmonious horizontal displacement behavior of an overflow dam crest is explained via the proposed HTT model and numerical simulation method. The conclusion is conducive for better understanding the displacement behavior of concrete dams in cold area. Meanwhile, engineering examples show that the proposed HTT model can reasonably interpret the thermal displacement behavior caused by the ambient temperature variations, and the WOA-optimized GRU network-based HTT model exhibits excellent fitting and predicting capabilities. A novel technique is provided for the accurate prediction of dam displacement.

The mean square deviation (MSD) of an arch dam dynamic displacement response during a flood discharge period is an important indicator used to evaluate the vibration intensity. The whole distribution of the MSD of the dynamic displacement response reflects the whole dynamic displacement field. Inversion of the whole dynamic displacement field of the arch dam based on limited measuring points is of significance. In this study, the Ertan arch dam was used to propose an inversion algorithm for the whole prototype dynamic displacement field of the arch dam. First, inversion theory of flow-induced vibration response for the arch dam is introduced. Second, the arch dam prototype vibration test under flood discharge excitation is conducted to determine the dynamic displacement response of limited measuring points. Third, the full-scale finite element model of the Ertan arch dam is set up to conduct modal analysis, and the first nine modes are cutoff to extract the modal parameters for the inversion of equivalent excitation source load spectra. Finally, the whole dynamic displacement field of the arch dam is calculated by flow-induced vibration response positive analysis based on the equivalent excitation load spectra. Inversion results show that the whole arch dam dynamic displacement field is obtained through only seven dynamic displacement response measuring points arranged on the crest of the dam. Simultaneously, the inversion and measured values are in good agreement. This method provides a novel technique to reasonably evaluate the high arch dam during the flood discharge period.