In water conservation projects, reservoirs play a key role in controlling and regulating floods. In recent years, extreme rainfall events have occurred more frequently. Urban reservoirs in danger under extreme rainfall happen occasionally. The breaching of urban reservoirs seriously threatens the lives and property of the population downstream. A lack of clarity regarding the characterization of the flooding response remains owing to urban reservoir overtopping under different rainfall conditions. This study used observed data from the heavy rainfall event in Zhengzhou on July 20, 2021, with Jiangang Reservoir in Zhengzhou as the study site. The study designed a storm-flood process with different rainfall scenarios that combines the long-duration rainstorm formula and Chicago rain pattern method, establishing a response relationship between extreme rainfall and flooding processes. Flood evolution simulation analyses were carried out for different scenarios of rainfall and flooding processes under three working conditions: rainfall only, dam failure only and dam failure coupled with rainfall, to characterize the flood response due to dam failure under different rainfall conditions. The results showed that the peak of the rainfall occurs earlier in the total rainfall calendar time, the faster the dam failure rate of the reservoir. As the rain peak occurs at a time that is later in the total rainfall calendar, and the recurrence period and rainfall duration increase, the flood volume increases and urban inundation becomes more severe. The rainfall type had less of an effect on the difference in the extent of inundation between different working conditions. The extent of flood inundation for the dam failure-coupled rainfall scenario is much greater than that for the dam failure only scenario. The increase in inundation extent in the dam failure coupled rainfall scenario at the higher levels of flood severity is smaller. Study results provide hydraulic element support for flood hazard assessment.

Current research trends in significant hydraulic engineering projects focus on investigating the seismological properties of intensity and frequency content of pulse-type near-field earthquakes on the structural response. Conversely, the duration impact is not expressly addressed in the seismic design code for underground buildings. Currently, various duration indicators of as-recorded strong ground motions mainly consider the effective duration of the initial acceleration component record. In contrast, the duration indicators for the effective velocity duration (EVD) of the original velocity time-history component record have rarely been addressed. Specifically, there is a gap between the effective velocity duration and the structural response. To illustrate the impact on the structural response, an EVD of pulse-type NFGM duration was used. This EVD can be calculated for seismic excitations with set threshold values that enable a quantitative examination of the duration effects. A fluid-hydraulic tunnel-rock interaction system was built and used to estimate the seismic response characteristics induced by different duration NFGMs. The investigation’s findings highlight that the inelastic dynamic response and damage degree are strongly affected by the EVD. Additionally, the fixed threshold value of 5–95% showed an excellent correlation coefficient with the structural response. The significant duration was also found to be the most suitable alternative indicator to replace the EVD index. In addition, the reduced time-history methodology of near-fault earthquake records is presented and validated, with this method being used to improve the efficiency of the dynamic time-history analysis of hydraulic arched tunnels.

Rapidly developed deep learning methods, widely used in various fields of civil engineering, have provided an efficient option to reduce the computational costs and improve the predictive capabilities. However, it should be acknowledged that the application of deep learning methods to develop prediction models that efficiently assess the nonlinear dynamic responses of cross-fault hydraulic tunnels (CFHTs) is lacking. Thus, the objective of this study is to construct a rational artificial neural network (ANN) prediction model to generate the mass data and fragility curves of CFHTs. Firstly, an analysis of 1080 complete nonlinear dynamic time histories via incremental dynamic analysis (IDA) is conducted to obtain the mass data of the drift ratio of the CFHT. Then, the hyper-parameters of the ANN model are discussed to determine the optimal parameters based on four examined approaches to improve the prediction capacity and accuracy. Meanwhile, the traditional probabilistic seismic demand models of the predicted values obtained by the ANN model and the numerical results are compared with the statistical parameters. Eventually, the maximum likelihood estimation couping IDA method is applied to assess the seismic safety of CFHTs under different damage states. The results show that two hidden layers, ten neurons, and the ReLU activation function for the ANN model with Bayesian optimization can improve the reliability and decrease the uncertainty in evaluating the structural performance. Moreover, the amplitude of the seismology features can be used as the neurons to build the input layers of the ANN model. It is found through vulnerability analysis that the traditional seismic fragility analysis method may overestimate the earthquake resistance capacity of CFHTs compared with maximum likelihood estimation. In practical engineering, ANN methods can be regarded as an alternative approach for the seismic design and performance improvement of CFHTs.

Machine learning (ML) approaches, widely used in civil engineering, have the potential to reduce computing costs and enhance predictive capabilities. However, many ML methods have yet to be applied to develop models that accurately analyze the nonlinear dynamic response of cross-fault hydraulic tunnels (CFHTs). To predict CFHT models and fragility curves effectively, we identify the most effective ML techniques and improve prediction capacity and accuracy by initially creating an integrated multivariate earthquake intensity measure (IM) from nine univariate earthquake IMs using principal component analysis. Structural reactions are then performed using incremental dynamic analysis by a multimedium-coupled interaction system. Four techniques are used to test ML–principal component analysis (PCA) feasibility. Meanwhile, mathematical statistical parameters are compared to standard probabilistic seismic demand models of expected and computed values using ML-PCA. Eventually, multiple stripe analysis–maximum likelihood estimation (MSA-MLE) is applied to assess the seismic performance of CFHTs. This study highlights that the Gaussian process regression and integrated IM can improve reliable probability and reduce uncertainties in evaluating the structural response. Thorough numerical analysis, using the suggested methodology, one can efficiently assess the seismic fragilities of the tunnel by the predicted model. ML-PCA techniques can be viewed as an alternate strategy for seismic design and CFHT performance enhancement in real-world engineering.