The landslide caused a huge disaster to the living environment and seriously threatened the lives and property safety of nearby residents. Assess or predict landslide-susceptible the landslide displacement through monitoring are great beneficial to guide landslide control and mitigate these hazards by taking appropriate preparatory measures. In this paper, a new water cycle algorithm optimization BP neural network (BPNN) dynamic prediction model (WCA-BPNN) was established to make up for the shortcoming of BPNN convergence speed. A typical step-wise landslide——Langshuwan Landslide happened in the Three Gorges Reservoir area of China is taken as a case, and the displacement monitoring data of 4 years was used for time series analysis and modeling. The long-term creep effect of the landslide and the short-term acceleration effect of the climate are considered in the model, and the accumulative displacement is divided into two kinds of trend displacement and periodic displacement. The key influencing factors of landslide periodic displacement were screened by gray relational grade analysis method, and then used as learning data. In addition to comparing the predicted value of the model with the measured value, it also compares the accuracy of the three models of BPNN, support vector machine, extreme learning machine under the training conditions of the same learning data set. The results show that the WAC-BPNN model has faster convergence speed and higher prediction accuracy than the traditional BPNN model, and it is also the most accurate of the four models.

Application of an enhanced BP neural network model with water cycle algorithm on landslide prediction

Landslides are natural phenomena, causing serious fatalities and negative impacts on socioeconomic. The Three Gorges Reservoir (TGR) area of China is characterized by more prone to landslides for the rainfall and variation of reservoir level. Prediction of landslide displacement is favorable for the establishment of early geohazard warning system. Conventional machine learning methods as forecasting models often suffer gradient disappearance and explosion, or training is slow. Hence, a dynamic method for displacement prediction of the step-wise landslide is provided, which is based on gated recurrent unit (GRU) model with time series analysis. The establishment process of this method is interpreted and applied to Erdaohe landslide induced by multi-factors in TGR area: the accumulative displacements of landslide are obtained by the global positioning system; the measured accumulative displacements is decomposed into the trend and periodic displacements by moving average method; the predictive trend displacement is fitted by a cubic polynomial; and the periodic displacement is obtained by the GRU model training. And the support vector machine (SVM) model and GRU model are used as comparisons. It is verified that the proposed method can quite accurately predict the displacement of the landslide, which benefits for effective early geological hazards warning system. Moreover, the proposed method has higher prediction accuracy than the SVM model.

A novel displacement prediction method using gated recurrent unit model with time series analysis in the Erdaohe landslide

Early-warning of rock burst in coal mine by low-frequency electromagnetic radiation

For water inrush from coal floor, due to different kinds of controlling factors and their internal correlations, the accuracy of prediction model is mostly below expectation. In this paper, it studies on which controlling factors should be selected for water inrush prediction model because all these factors have different influence on water inrush incidents based on the analysis of in situ data. Some factors are proved having limited impacts on water inrush, it is no necessary to collect in situ data of those factors from coal mining work face. Therefore, the workload and expense will decrease. In this paper, an index system of factors influencing water inrush from coal floor is established based on the current water inrush controlling theory and detailed analysis of in situ data obtained from mining regions. Following the Wrapper method in feature selection, 10 main controlling factors were selected from 14 existing indicators which were thought could affect water inrush. After training on dynamic GRU model which is made for water inrush prediction, a comparison among dynamic GRU model and stable SVM and BPMN models turns out the advantages of the previous with a higher accuracy in train, validation and test set against the latter. It is believed GRU model is able to predict water inrush water inrush from coal floor with high accuracy and hence enhances mining safety.

A novel dynamic predictive method of water inrush from coal floor based on gated recurrent unit model

In order to study the influence of water level fluctuation on the stability of landslide in the Three Gorges Reservoir (TGR), Xigouwan landslide was selected to be a research object to analyze its deformation mechanism as the fluctuation of the reservoir water level. Based on the long-term field monitoring of landslide, the evolution law of the surface has been studied. Based on the numerical simulation used by FLAC3D, five different water levels of TGR operating conditions are set, and the plastic zone, the factor of safety, change of displacement, and the seepage characteristics of the landslide were obtained to analyze the influence of water level fluctuation on the stability of landslide.

The influence of water level fluctuation on the stability of landslide in the Three Gorges Reservoir

Water level change upstream of a reservoir highlights the risk of a landslide-prone area on the banks of a reservoir. This paper conducted a study on the deformation mechanism of a selected landslide that occurred in the Three Gorges Reservoir (TGR) after the water level of the reservoir changed. The long-monitored surface deformation of the slide mass revealed that the deformation of the landslide was related to the water level changes in the reservoir, especially of the change between flood and floodless seasons. The measured internal lateral displacements in the landslide showed that such a landslide was characterized by a trail-mode. FLAC3D was adopted to model the landslide by examining the plastic zone, factor of safety, and the displacement in the x-direction in consideration of four conditions: the natural state of a landslide in the TGR, the initial impoundment, the subsequent rise of water level, and the drawdown of water level. The numerical results indicated that the landslide mass tended to be unstable during the initial impoundment; the subsequent rise of water level had a limited effect on the landslide happening, but the drawdown of water level directly triggered the landslide. The landslide changed from push-mode to trail-mode. It is strongly recommended that drawdown of the water level in the reservoir be carefully controlled to mitigate the effect on landslide mass.

Stability analysis of a typical landslide mass in the Three Gorges Reservoir under varying reservoir water levels

Establishing a prediction model of soil liquefaction is an effective way to evaluate the site's quality and prevent the relevant loss caused by the earthquake. Considering the complexity of the liquefaction mechanism and the disadvantage of shear wave not being able to test the type of soil, the standard penetration test (SPT) data and the grey wolf optimization (GWO) algorithm were applied to try to improve the prediction accuracy of the SVM model in this paper. First, the optimal value of C and g of SVM was calculated and selected by iterating the GWO; then, the selected parameters were submitted into the SVM to train the prediction model with the training set; finally, the initial parameter of GWO was judged and updated by testing the test set and evaluating whether the performance of trained model until the goal of accuracy was meet. Besides, the GWO-SVM based on the dataset without the parameter of the shear wave velocity was also trained and tested to prove the advantage of combining the SPT data and shear wave data. It was indicated that the GWO algorithm could not only improve the accuracy of SVM fitting and optimize the performance of the prediction but also can fasten the operation; combining the SPT data and shear wave data was able to improve the prediction accuracy.

Yonggang Zhang

Papers

Stability analysis of a typical landslide mass in the Three Gorges Reservoir under varying reservoir water levels

The adoption of a support vector machine optimized by GWO to the prediction of soil liquefaction

Study on Excavation Sequence of Pilot Tunnels for a Rectangular Tunnel Using Numerical Simulation and Field Monitoring Method