Slope displacement prediction based on multisource domain transfer learning for insufficient sample data

Accurate displacement prediction is critical for the early warning of landslides.The complexity of the coupling relationship between multiple influencing factors and displacement makes the accurate prediction of displacement difficult.Moreover,in engineering practice,insufficient monitoring data limit the performance of prediction models.To alleviate this problem,a displacement prediction method based on multisource domain transfer learning,which helps accurately predict data in the target domain through the knowledge of one or more source domains,is proposed.First,an optimized variational mode decomposition model based on the minimum sample entropy is used to decompose the cumulative displacement into the trend,periodic,and stochastic components.The trend component is predicted by an autoregressive model,and the periodic component is predicted by the long short-term memory.For the stochastic component,because it is affected by uncertainties,it is predicted by a combination of a Wasserstein generative adversarial network and multisource domain transfer learning for improved prediction accuracy.Considering a real mine slope as a case study,the proposed prediction method was validated.Therefore,this study provides new insights that can be applied to scenarios lacking sample data.

Slope displacement prediction based on morphological filtering

Combining mathematical morphology (MM),nonparametric and nonlinear model,a novel approach for predicting slope displacement was developed to improve the prediction accuracy.A parallel-composed morphological filter with multiple structure elements was designed to process measured displacement time series with adaptive multi-scale decoupling.Whereafter,functional-coefficient auto regressive (FAR) models were established for the random subsequences.Meanwhile,the trend subsequence was processed by least squares support vector machine (LSSVM) algorithm.Finally,extrapolation results obtained were superposed to get the ultimate prediction result.Case study and comparative analysis demonstrate that the presented method can optimize training samples and show a good nonlinear predicting performance with low risk of choosing wrong algorithms.Mean absolute percentage error (MAPE) and root mean square error (RMSE) of the MM-FAR&LSSVM predicting results are as low as 1.670% and 0.172 mm,respectively,which means that the prediction accuracy are improved significantly.

Short-term displacement prediction for newly established monitoring slopes based on transfer learning

This study makes a significant progress in addressing the challenges of short-term slope displacement prediction in the Universal Landslide Monitoring Program,an unprecedented disaster mitigation program in China,where lots of newly established monitoring slopes lack sufficient historical deformation data,making it difficult to extract deformation patterns and provide effective predictions which plays a crucial role in the early warning and forecasting of landslide hazards.A slope displacement prediction method based on transfer learning is therefore proposed.Initially,the method transfers the deformation patterns learned from slopes with relatively rich deformation data by a pre-trained model based on a multi-slope integrated dataset to newly established monitoring slopes with limited or even no useful data,thus enabling rapid and efficient predictions for these slopes.Subsequently,as time goes on and monitoring data accumulates,fine-tuning of the pre-trained model for individual slopes can further improve prediction accuracy,enabling continuous optimization of prediction results.A case study indicates that,after being trained on a multi-slope integrated dataset,the TCN-Transformer model can efficiently serve as a pretrained model for displacement prediction at newly established monitoring slopes.The three-day average RMSE is significantly reduced by 34.6%compared to models trained only on individual slope data,and it also successfully predicts the majority of deformation peaks.The fine-tuned model based on accumulated data on the target newly established monitoring slope further reduced the three-day RMSE by 37.2%,demonstrating a considerable predictive accuracy.In conclusion,taking advantage of transfer learning,the proposed slope displacement prediction method effectively utilizes the available data,which enables the rapid deployment and continual refinement of displacement predictions on newly established monitoring slopes.

Efficiency of scalar and vector intensity measures for seismic slope displacements

Ground motion intensity measures are usually used to predict the earthquake-induced displacements in earth dams,soil slopes and soil structures.In this study,the efficiency of various single ground motion intensity measures(scalar IMs)or a combination of them(vector IMs)are investigated using the PEER-NGA strong motion database and an equivalent-linear sliding-mass model.Although no single intensity measure is efficient enough for all slope conditions,the spectral acceleration at 1.5 times of the initial slope period and Arias intensity of the input motion are found to be the most efficient scalar IMs for flexible slopes and stiff slopes respectively.Vector IMs can incorporate different characteristics of the ground motion and thus significantly improve the efficiency over a wide range of slope conditions.Among various vector IMs considered,the spectral accelerations at multiple spectral periods achieve high efficiency for a wide range of slope conditions.This study provides usefiul guidance to the development of more efficient empirical prediction models as well as the ground motion selection criteria for time domain analysis of seismic slope displacements.

Failure mode and dynamic response of a double-sided slope with high water content of soil

A double-sided slope with high water content in sandy clay was considered under the action of seismic load. Its failure mode and dynamic response were investigated using a hydraulic servo shaking table test. The typical characteristic of failure mode and dynamic responses of the double-sided slope were analyzed. Experimental results show that slope failure undergoes a process of progressive deformation. The slope failure mode can be explained as creep sliding landslide. AFA(Amplification Factor of Acceleration) at the surface and inner parts of the slope shows an increasing trend with the increase of relative elevation. The relationship between AFA and EAA(Excitation Amplitude of Acceleration) is nonlinear. An empirical formula is proposed to describe preferably the relationship between AFA,relative elevation and dimensionless EAA. The AFA at the middle and upper parts of the slope increases apparently with increasing EFA(Excitation Frequency of Acceleration).

On the role of topographic ampli fi cation in seismic slope instabilities

Surface wave generation due to body wave propagation near ground surface has been discussed in the literature.This phenomenon,typically occurring in topographic changing areas,along with its interaction with body waves(SV),decreases precision of formulas for evaluation of slope displacement.This signi fi cant fact caused the researchers not only to investigate the combined surface and SV waves motion pattern,but also to consider its effect on structures built on the slopes.In order to reveal the phenomenon,several fi nite element numerical studies have been performed by ABAQUS programme.Besides,two physical model slopes simulating the landslide occurrence have been constructed and tested by shaking table device.The results of induced and calculated accelerations obtained by two approaches have been compared and Rayleigh wave generation has been proved.Furthermore,the slope displacements have been calculated by various empirical methods and the results were compared with numerical ones.The results proved that in order to increase the precision of empirical formulas for displacement prediction,surface wave effect should be taken into account.Finally,a concept of'effective depth of sur fi cial ampli fi cation'is introduced and its effect on dynamic slope stability is analysed.

Effectiveness of Fiber Bragg Grating monitoring in the centrifugal model test of soil slope under rainfall conditions

Centrifugal model testsare playing an increasingly importantrolein investigating slope characteristics under rainfall conditions. However, conventional electronic transducers usually fail during centrifugal model tests because of the impacts of limitedtest space, high centrifugal force, and presence of water, with the result that limited valid data is obtained. In this study, Fiber Bragg Grating(FBG) sensing technology is employed in the design and development of displacement gauge, an anchor force gauge and an anti-slide pile moment gauge for use on centrifugal model slopes with and without a retaining structure. The two model slopes were installed and monitored at a centrifugal acceleration of 100 g. The test results show that the sensors developed succeed in capturing the deformation and retaining structure mechanical response of the model slopes during and after rainfall. The deformation curvefor the slope without retaining structure shows a steepresponse that turns gradualfor the slope with retaining structure. Importantly, for the slope with the retaining structure, results suggest that more attention be paid to increase of anchor force and antislide pile moment during rainfall. This study verifies the effectiveness of FBG sensing technology in centrifuge research and presents a new and innovative method for slope model testing under rainfall conditions.

Automatic prediction of time to failure of open pit mine slopes based on radar monitoring and inverse velocity method

Radar slope monitoring is now widely used across the world, for example, the slope stability radar(SSR)and the movement and surveying radar(MSR) are currently in use in many mines around the world.However, to fully realize the effectiveness of this radar in notifying mine personnel of an impending slope failure, a method that can confidently predict the time of failure is necessary. The model developed in this study is based on the inverse velocity method pioneered by Fukuzono in 1985. The model named the slope failure prediction model(SFPM) was validated with the displacement data from two slope failures monitored with the MSR. The model was found to be very effective in predicting the time to failure while providing adequate evacuation time once the progressive displacement stage is reached.

New multiplexed system for synchronous measurement of out-of-plane deformation and two orthogonal slopes

We propose a novel system for synchronous measurement of out-of-plane deformation and two orthogonal slopes using a single camera. The linearly polarized reference beam introduced by an optical fiber interferes with the unpolarized object beam to measure the out-of-plane deformation. A modified Mach–Zehnder interferometer is used to measure the two orthogonal slopes of the out-of-plane deformation. One of the object beams of the Mach–Zehnder interferometer is an unpolarized beam, and the other object beam is split into two orthogonal linearly polarized object beams by a polarizing prism. The two beams are orthogonally polarized. Hence, they will not interfere with each other. The two polarized beams respectively interfere with the unpolarized beam to simultaneously measure the two orthogonal slopes of the out-of-plane deformation. In addition, the imaging lens and apertures are respectively placed in three optical paths to independently control the carrier frequencies and shearing amounts. The effectiveness of this method can be proved by measuring two pressure-loaded circular plates.