Step type landslide displacement prediction method based on VMD-Mama algorithm

Abstract

The invention discloses a step-type landslide displacement prediction method based on a VMD-Mama algorithm. The step-type landslide displacement prediction method comprises the steps of constructing a data set in the same region, performing variational mode decomposition, performing variational mode decomposition, performing periodic term displacement component prediction and performing prediction component superposition and prediction precision evaluation. According to the invention, the unit prediction model and the multivariate prediction model are respectively constructed for different displacement components based on the Mama model, and the prediction precision of the accumulated displacement of the landslide is improved.

Description

Technical Field

[0001] The invention relates to a landslide displacement prediction method, in particular to a step-type landslide displacement prediction method based on a VMD-Mamba algorithm, and belongs to the technical field of geological disaster prevention and control. Background Art

[0002] Landslides are a complex and common natural geological disaster. They are one of the most widespread and serious geological disasters in the world, causing serious harm and damage to humans and their living environment. Especially in China, due to the large number of mountainous areas and complex terrain, landslides are frequent and have far-reaching impacts, making them one of the most serious geological disasters in China. In recent years, with the acceleration of urbanization and the impact of climate change, the frequency and destructive power of landslides have increased, becoming a major threat to the safety of people's lives and property and economic development.

[0003] However, due to the complexity of landslide formation mechanism and inducing factors, landslide displacement prediction is still a difficult task. Taking the Bajimen landslide in the Three Gorges Reservoir area as an example, due to the combined influence of its own internal control conditions and multiple external inducing factors, the landslide deformation monitoring curve shows a step-like characteristic over time. This deformation curve increases sharply in the rainy season, which increases the difficulty of landslide deformation prediction. When directly predicting this stage-by-stage sudden increase in displacement curve, the model fitting is difficult and the prediction effect of the landslide step stage is poor. Therefore, when dealing with step-type landslides with many inducing factors such as the Three Gorges area, how to use the existing monitoring data to achieve accurate prediction of landslide displacement is a thorny problem that needs to be solved today.

[0004] In the prior art, a landslide displacement prediction method based on GRU is disclosed in the patent with announcement number CN113111573B. The landslide surface displacement deformation is monitored, and the replacement value of the abnormal value in the obtained landslide displacement data is calculated using a constructed polynomial model; all the landslide displacement data replaced by the replacement value are mean normalized, and a GRU model and a training set, a validation set, and a test set are constructed; the GRU model is iterated cyclically using the training set, and the parameters of the trained GRU model are adjusted using the validation set; the GRU model after the parameter adjustment is tested using the test set, and the obtained model prediction results are reversed and normalized to complete the landslide displacement prediction and improve the prediction accuracy. However, in the prior art, the prediction model is constructed based on the recurrent neural network, and the gated recurrent unit (GRU) and the long short-term memory network (LSTM) are the most widely used. However, the gradient disappearance or gradient explosion phenomenon may occur, which makes it difficult for the model to capture long-distance dependencies. At the same time, although LSTM and GRU can handle nonlinear relationships, their modeling capabilities are still limited in complex and highly nonlinear sequence data, resulting in poor prediction results. In addition, in the field of step-type landslide prediction, in order to improve the accuracy of the landslide prediction model, a suitable landslide displacement decomposition algorithm is needed. However, traditional landslide displacement decomposition algorithms, such as sliding average, HP filtering, and empirical mode decomposition, have more or less defects. Summary of the invention

[0005] The purpose of the present invention is to provide a step-type landslide displacement prediction method based on the VMD-Mamba algorithm in order to solve at least one of the above technical problems. The method is based on the combination of the variational mode decomposition algorithm and the Mamba model framework. The prediction result is more accurate and reliable than the current mainstream landslide prediction model. The model structure is simple and the calculation complexity is low. It has good practical value and can improve the accuracy and timeliness of landslide displacement prediction.

[0006] The present invention achieves the above-mentioned object through the following technical scheme: a step-type landslide displacement prediction method based on VMD-Mamba algorithm, the step-type landslide displacement prediction method comprising the following steps: Step 1: constructing a data set in the same area: the data set includes a landslide displacement monitoring data set and a landslide influencing factor data set, and preprocessing the landslide displacement monitoring data set and the landslide influencing factor data set to obtain a landslide prediction data set in the area; Step 2, variational mode decomposition: selecting a landslide displacement monitoring data set in the landslide prediction data set, and decomposing the accumulated landslide displacement into a trend term displacement component and a period term displacement component by using a variational mode decomposition algorithm; Step 3: Prediction of displacement component of trend item: According to the trend characteristics of displacement component of trend item obtained by variational mode decomposition, a Mamba-trend item prediction model is constructed, wherein the Mamba-trend item prediction model belongs to a unit prediction model, and the predicted value of trend item in the next three days is predicted based on the displacement component of trend item in the previous five days, so as to obtain the predicted value of displacement component of trend item; Step 4, prediction of periodic term displacement components: select the landslide influencing factor data set in the landslide prediction data set, screen out the most relevant factor set affecting the landslide displacement through grey correlation analysis, and obtain the trend characteristics of the periodic term displacement components according to variational mode decomposition, and construct a Mamba-periodic term prediction model; wherein, the Mamba-periodic term prediction model is a multivariate model, and based on the most relevant factor set and the periodic term displacement components of the previous five days, the periodic term displacement components of the next three days are predicted to obtain the predicted value of the periodic term displacement components; Step 5: Prediction component superposition and prediction accuracy evaluation: The predicted value of the displacement component of the trend term is superimposed with the predicted value of the displacement component of the period term to obtain the cumulative landslide displacement prediction value, and the comprehensive prediction ability of the landslide displacement prediction model is evaluated by calculating the root mean square error, mean absolute error, mean absolute percentage error and determination coefficient, and the accuracy of the prediction results is analyzed and evaluated.

[0007] As a further technical solution of the present invention: in the step 1, the landslide displacement monitoring data set is constructed in days for training and verification of the landslide displacement prediction model; The landslide influencing factor dataset is all relevant inducing factors in the region, including but not limited to topography, hydrogeological conditions, rainfall, reservoir water level changes, surface pore changes and human engineering activities.

[0008] As a further technical solution of the present invention: in the step 1, the landslide displacement monitoring data set is preprocessed, including monitoring outliers and processing missing values ​​on the data set and normalizing the data set, including the following steps: A. Preprocess the data set by dealing with outliers and filling in missing values, and then normalize it to between 0 and 1 to eliminate the impact of different dimensions on model training; B. The landslide displacement monitoring data set is divided into 7:2:1, that is, the first 70% of the data set is used as the training set of the model, the subsequent 20% is used as the validation set of the model, and finally 10% of the data set is retained as the test set of the model.

[0009] As a further technical solution of the present invention: in the step 2, the variational mode decomposition algorithm specifically includes: 1) Decompose the landslide displacement time series into multiple IMFs, each of which represents an oscillation mode with different frequency components, where IMF is the intrinsic mode function; 2) Through the recursive iterative algorithm, each IMF is gradually optimized to achieve input signal decomposition; 3) Combine the decomposed IMFs to extract the components representing trend displacement T and periodic displacement P, thereby achieving data denoising and feature extraction.

[0010] As a further technical solution of the present invention: decomposing the input signal into a plurality of IMFs, each IMF corresponding to a center frequency; The center frequency calculation is to update the center frequency and bandwidth of each IMF in each iteration step so that each IMF is concentrated in the frequency domain.

[0011] As a further technical solution of the present invention: the Mamba-trend item prediction model in step three and the Mamba-period item prediction model in step four together constitute a Mamba model, and the Mamba model includes a state space model, a dynamic selection mechanism and a hardware perception architecture; The state space model is used to capture and model long-term dependencies in time series data. The state space model represents the evolution of the time series as a sequence of hidden states through state transition and observation mechanisms. The dynamic selection mechanism implements the selective replication and sensor head synthesis tasks by inputting parameterized state-space model parameters, so that the state-space model can filter out irrelevant information and remember relevant information indefinitely; The hardware-aware architecture uses scans to loop through the computational model but does not implement extended states to avoid IO accesses between different levels of the GPU memory hierarchy, thereby achieving faster inference and linear scaling to longer sequence lengths.

[0012] As a further technical solution of the present invention: in the step 4, the grey relational analysis specifically includes: The influence of influencing factors on landslides is determined by grey correlation analysis. The larger the correlation factor, the more important the influencing factor is for predicting landslide displacement; the smaller the correlation factor, the less important the influencing factor is for predicting landslide displacement.

[0013] The beneficial effects of the present invention are: 1) The variational mode decomposition algorithm adopted in the present invention is a new signal decomposition method proposed on the basis of empirical mode decomposition (EMD), which can process nonlinear and non-stationary signals. Compared with the traditional EMD, it avoids the problem of mode aliasing and the decomposition result is more accurate; the variational mode decomposition algorithm ensures the concentration of the frequency distribution of each mode through the sub-signal decomposition with limited frequency band, and enhances the ability to capture the local characteristics of the signal; at the same time, the variational mode decomposition algorithm can effectively deal with noise interference, and introduce constraints by Lagrange multipliers to ensure the stability and robustness of signal decomposition; compared with linear filtering methods such as sliding average and HP filtering, VMD can decompose multiple intrinsic mode functions (IMFs) in the signal, revealing richer signal characteristics; 2) The Mamba model used in the present invention is a structured state space sequence model with a selection mechanism and a scanning module. It has stronger data feature extraction and memory capabilities, and can fully learn and describe landslide displacement components with different characteristics. Compared with models such as LSTM and Transformer, the Mamba model shows higher fitting degree and prediction performance in learning short-term landslide data, especially when dealing with local complex change characteristics of landslides, and is superior to traditional LSTM and GRU models in terms of computational efficiency and prediction accuracy. 3) Since the influencing factors of different periodic displacement components and trend displacement components are different, the present invention constructs a unit prediction model and a multivariate prediction model for different displacement components based on the Mamba model, thereby further improving the prediction accuracy of the accumulated displacement of the landslide. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a schematic diagram of the process structure of the present invention. DETAILED DESCRIPTION

[0015] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0016] Embodiment 1, as Figure 1 As shown, this embodiment provides a step-type landslide displacement prediction method based on the VMD-Mamba algorithm, and the step-type landslide displacement prediction method specifically includes the following steps: S1: Construct a dataset within the same region.

[0017] The data set includes the landslide displacement monitoring data set and the landslide influencing factor data set in the same area; among them, the landslide displacement monitoring data set is the most representative and typical displacement monitoring point data selected in the area (i.e. the target area); the landslide influencing factor data set is to collect data on factors affecting landslides such as reservoir water level and rainfall in the target area.

[0018] The landslide displacement monitoring dataset is constructed on a daily basis and is mainly used for the training and verification of landslide displacement prediction models. The landslide influencing factor dataset needs to collect all relevant inducing factors in the target area, including but not limited to topography, hydrogeology, rainfall, reservoir water level changes, surface pore changes and human engineering activities.

[0019] The landslide displacement monitoring dataset and the landslide influencing factor dataset are preprocessed to obtain the landslide prediction dataset in the target area.

[0020] The landslide displacement monitoring dataset is divided into a target dataset and a label dataset. The preprocessing of the landslide displacement monitoring dataset includes outlier monitoring, missing value processing, and normalization of the target dataset and the label dataset. Specifically, the following steps are included: 1) Preprocess the target dataset and the label dataset by handling outliers and filling missing values, and then normalize the target dataset and the label dataset to between 0 and 1 to eliminate the impact of different dimensions on model training; 2) The landslide displacement monitoring dataset is divided into 7:2:1, where the first 70% of the dataset is used as the training set of the model, the middle 20% is used as the validation set of the model, and the last 10% of the dataset is reserved as the test set of the model.

[0021] It should be noted that, specifically, the dataset used is from the Bazimen landslide in the Three Gorges Reservoir area, where the target dataset is the landslide displacement monitoring dataset of the monitoring point GSCX3, and the label dataset is the reservoir water level and rainfall daily record dataset.

[0022] S2: Variational Mode Decomposition (VMD).

[0023] The landslide displacement monitoring dataset in the landslide prediction dataset is selected, and the cumulative landslide displacement is decomposed into trend term displacement component and periodic term displacement component by using variational mode decomposition (VMD) algorithm.

[0024] The variational mode decomposition algorithm specifically includes the following steps: 1) Due to the complexity of the landslide displacement time series, the landslide displacement time series is decomposed into multiple intrinsic mode functions (IMFs), each of which represents an oscillation mode with different frequency components; 2) Through the recursive iterative algorithm, each intrinsic mode function is gradually optimized to make it physically meaningful, realize the decomposition of the input signal, and then achieve the optimal decomposition of the input signal; 3) Combine the decomposed IMFs to extract the components representing trend displacement T and periodic displacement P, thereby achieving data denoising and feature extraction.

[0025] It should be noted here that the decomposition of the input signal is optimized through a recursive iterative algorithm, that is, the input signal is decomposed into several intrinsic mode functions (IMFs), each IMF corresponds to a center frequency; since each part of the input signal has a clear separation in the frequency domain, the different frequency components in the input signal can be effectively decomposed; the calculation of the center frequency is to update the center frequency and bandwidth of each IMF in each iteration step, so that each IMF is as concentrated as possible in the frequency domain.

[0026] Variational mode decomposition is modal reconstruction. Through optimization, each IMF has good interpretability in the time domain and frequency domain, so that the reconstructed mode can truly reflect the characteristics of the original signal (input signal). In addition, the parameter selection of variational mode decomposition is appropriately selected according to the specific situation of the cumulative displacement curve and the optuna parameter optimization framework.

[0027] S3: Prediction of displacement component of trend term.

[0028] According to the trend characteristics of the displacement component of the trend item obtained by variational mode decomposition, a Mamba-trend item prediction model is constructed; among them, the Mamba-trend item prediction model belongs to a unit prediction model, which predicts the trend item prediction value of the next three days based on the displacement component of the trend item of the previous five days to obtain the prediction value of the displacement component of the trend item.

[0029] S4: Prediction of displacement components of periodic terms.

[0030] A data set of landslide influencing factors in the landslide prediction data set was selected, and the most relevant factor set affecting the landslide displacement was screened out through grey correlation analysis (the purpose is to exclude unnecessary factors and reduce the training difficulty of the model). The trend characteristics of the periodic term displacement component were obtained according to the variational mode decomposition, and the Mamba-periodic term prediction model was constructed. Among them, the Mamba-periodic term prediction model is a multivariate model. Based on the most relevant factor set and the periodic term displacement component of the previous five days, the periodic term displacement component of the next three days is predicted to obtain the predicted value of the periodic term displacement component.

[0031] The influence of the influencing factors in the landslide influencing factor data set on the landslide is determined by grey correlation analysis. The larger the correlation influence factor, the more important the influencing factor is for predicting the landslide displacement, and the smaller the correlation influence factor, the less important the influencing factor is for predicting the landslide displacement. In this embodiment, the influencing factors with a correlation less than 0.6 are excluded, and the remaining influencing factors are constructed as the most relevant factor set.

[0032] The Mamba-trend item prediction model in step S3 and the Mamba-period item prediction model in step S4 together constitute a Mamba model, which includes a state space model, a dynamic selection mechanism, and a hardware-aware architecture.

[0033] Among them, the state space model (SSM) is used to capture and model long-term dependencies in time series data. The SSM represents the evolution of the time series as a sequence of hidden states through state transition and observation mechanisms. The state space model can not only effectively process long sequence data, but also has significant advantages in capturing dynamic changes in the data. Specifically, the state space model maps the input sequence to a series of state vectors and makes predictions based on the state transition matrix and output matrix. This enables the state space model to accurately predict future sequences while maintaining long-term memory.

[0034] The dynamic selection mechanism enables the state-space model to filter out irrelevant information and remember relevant information indefinitely through SSM parameters parameterized according to the input, achieving important synthetic tasks such as selective copying and sensing heads.

[0035] The hardware-aware architecture uses scans instead of convolutions to loop through the computational model, but does not implement extended state to avoid IO accesses between different levels of the GPU memory hierarchy, thereby achieving faster inference and linear scaling to longer sequence lengths.

[0036] S5: Prediction component superposition and prediction accuracy evaluation.

[0037] The predicted value of the displacement component of the trend term is superimposed with the predicted value of the displacement component of the period term to obtain the predicted value of the cumulative landslide displacement, and the root mean square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE) and determination coefficient (R 2 ), evaluate the comprehensive prediction ability of the prediction model, and analyze and evaluate the accuracy of the prediction results.

[0038] Embodiment 2: This embodiment provides two comparative examples (an LSTM method based on a recurrent neural network and a state-of-the-art Transformer method based on a self-attention mechanism) for comparison with the step-type landslide displacement prediction method based on the VMD-Mamba algorithm of the present invention. At the same time, in order to verify the generalization performance of the method, the GSCX5 monitoring data set of the Bazimen landslide monitoring point in the Three Gorges Reservoir area is added as a generalization performance verification data set.

[0039]

[0040]

[0041] It can be seen from Table 1 and Table 2 that the experimental results of the present invention and other landslide prediction methods on the monitoring points GSCX3 and GSCX5 of the Bajimen landslide in the Three Gorges Reservoir area are compared, and the performance of the method in predicting step-type landslides and stage-by-stage sudden displacement differences is evaluated. In general, the R2 of the displacement prediction results of the two monitoring points of the present invention reached above 0.97, and its average RMSE and R2 were improved by 112% and 2.43% compared with the other two prediction models. At the same time, the generalization performance of the model is also better than the other two models, and the average RMSE and R2 are 35.44% and 2.79% higher than the optimal advanced method "Transformer". It is worth noting that the present invention has achieved the best accuracy in the three-day prediction, which is significantly better than other advanced methods. The results show that the method proposed by the present invention is an effective improvement of the existing step-type landslide prediction method, and verify that the method proposed by the present invention has stronger feature learning ability and prediction ability.

[0042] Working principle: In view of the extremely complex characteristics of the influencing factors and displacement process of step-type landslides, the variational mode decomposition (VMD) algorithm is used to decompose the cumulative landslide displacement into trend term landslide displacement and period term landslide displacement. The Mamaba-trend term prediction model and the Mamba-period term prediction model are respectively constructed to fit and predict the displacement components. Among them, the gray correlation analysis is used to screen the influencing factors, and the appropriate influencing factors are selected as the influencing factor set of the periodic term model to reduce the complexity of model fitting and the difficulty of calculation. A step-type landslide displacement prediction model combining the VMD algorithm and the Mamba model is comprehensively constructed, and the accuracy of the prediction results is analyzed and evaluated.

[0043] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

[0044] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. A step-type landslide displacement prediction method based on VMD-Mamba algorithm, characterized in that: The step-type landslide displacement prediction method comprises the following steps: Step 1: constructing a data set in the same area: the data set includes a landslide displacement monitoring data set and a landslide influencing factor data set, and preprocessing the landslide displacement monitoring data set and the landslide influencing factor data set to obtain a landslide prediction data set in the area; Step 2, variational mode decomposition: selecting a landslide displacement monitoring data set in the landslide prediction data set, and decomposing the accumulated landslide displacement into a trend term displacement component and a period term displacement component by using a variational mode decomposition algorithm; Step 3: Prediction of displacement component of trend item: According to the trend characteristics of displacement component of trend item obtained by variational mode decomposition, a Mamba-trend item prediction model is constructed, wherein the Mamba-trend item prediction model belongs to a multivariate prediction model, and the displacement component of trend item of the previous five days is used to predict the predicted value of trend item of the next three days, so as to obtain the predicted value of displacement component of trend item; Step 4, prediction of periodic term displacement component: select the landslide influencing factor data set in the landslide prediction data set, screen out the most relevant factor set affecting the landslide displacement through grey correlation analysis, and obtain the trend characteristics of the periodic term displacement component according to variational mode decomposition, and construct a Mamba-periodic term prediction model; wherein, the Mamba-periodic term prediction model belongs to a unit model, and based on the most relevant factor set and the periodic term displacement component of the previous five days, the periodic term displacement component of the next three days is predicted to obtain the predicted value of the periodic term displacement component; Step 5: Prediction component superposition and prediction accuracy evaluation: The predicted value of the displacement component of the trend term is superimposed with the predicted value of the displacement component of the period term to obtain the cumulative landslide displacement prediction value, and the comprehensive prediction ability of the landslide displacement prediction model is evaluated by calculating the root mean square error, mean absolute error, mean absolute percentage error and determination coefficient, and the accuracy of the prediction results is analyzed and evaluated.

2. The method for predicting displacement of a step-type landslide according to claim 1, characterized in that: In the step 1, the landslide displacement monitoring data set is constructed in days and is used for training and verification of the landslide displacement prediction model; The landslide influencing factor dataset is all relevant inducing factors in the region, including but not limited to topography, hydrogeological conditions, rainfall, reservoir water level changes, surface pore changes and human engineering activities.

3. The method for predicting displacement of a step-type landslide according to claim 1, characterized in that: In the step 1, the landslide displacement monitoring data set is preprocessed, including monitoring outliers and processing missing values ​​on the data set and normalizing the data set, including the following steps: A. Preprocess the data set by dealing with outliers and filling in missing values, and then normalize it to between 0 and 1 to eliminate the impact of different dimensions on model training; B. The landslide displacement monitoring data set is divided into 7:2:1, that is, the first 70% of the data set is used as the training set of the model, the subsequent 20% is used as the validation set of the model, and finally 10% of the data set is retained as the test set of the model.

4. The method for predicting displacement of a step-type landslide according to claim 1, characterized in that: In the step 2, the variational mode decomposition algorithm specifically includes: 1) Decompose the landslide displacement time series into multiple IMFs, each of which represents an oscillation mode with different frequency components, where IMF is the intrinsic mode function; 2) Through the recursive iterative algorithm, each IMF is gradually optimized to achieve input signal decomposition; 3) Combine the decomposed IMFs to extract the components representing trend displacement T and periodic displacement P, thereby achieving data denoising and feature extraction.

5. The method for predicting displacement of a step-type landslide according to claim 4 is characterized in that: Decompose the input signal into several IMFs, each of which corresponds to a center frequency; The center frequency calculation is to update the center frequency and bandwidth of each IMF in each iteration step so that each IMF is concentrated in the frequency domain.

6. The method for predicting displacement of a step-type landslide according to claim 1, characterized in that: The Mamba-trend item prediction model in step 3 and the Mamba-period item prediction model in step 4 together constitute a Mamba model, and the Mamba model includes a state space model, a dynamic selection mechanism, and a hardware-aware architecture; The state space model is used to capture and model long-term dependencies in time series data. The state space model represents the evolution of the time series as a sequence of hidden states through state transition and observation mechanisms. The dynamic selection mechanism implements the selective replication and sensor head synthesis tasks by inputting parameterized state-space model parameters, so that the state-space model can filter out irrelevant information and remember relevant information indefinitely; The hardware-aware architecture uses scans to loop through the computational model but does not implement extended states to avoid IO accesses between different levels of the GPU memory hierarchy, thereby achieving faster inference and linear scaling to longer sequence lengths.

7. The method for predicting displacement of a step-type landslide according to claim 1, characterized in that: In the step 4, the grey correlation analysis specifically includes: The influence of influencing factors on landslides is determined by grey correlation analysis. The larger the correlation influence factor, the more important the influencing factor is for predicting landslide displacement; the smaller the correlation influence factor, the less important the influencing factor is for predicting landslide displacement.

Images