Landslide pore water pressure prediction method based on IoT

Abstract

The invention discloses a landslide pore water pressure prediction method based on IoT. The method comprises the steps of sensing deployment, data collection, data processing, pore water pressure prediction and self-adaptive early warning. The invention relates to the technical field of landslide pore water pressure prediction, in particular to an IoT (Internet of Things)-based landslide pore water pressure prediction method, which comprises the following steps of: deploying various sensors to collect various data indexes, and comprehensively performing multi-modal characteristic decomposition and processing; the method for predicting the pore water pressure by adopting an improved long and short term neural network combined with adaptive noise and variational mode decomposition improves the prediction accuracy and prediction depth of the landslide pore water pressure, and improves the usability of the method. And a self-adaptive early warning mechanism is constructed, and early warning design is carried out on the predicted values of the multiple modes and indexes obtained through prediction.

Description

Technical Field

[0001] The present invention relates to the technical field of landslide pore water pressure prediction, and in particular to a landslide pore water pressure prediction method based on IoT. Background Art

[0002] The landslide pore water pressure prediction method based on the Internet of Things (IoT) is an advanced geological disaster monitoring and early warning method. It deploys a high-precision sensor network in landslide-prone areas to collect key data such as pore water pressure and moisture content in the soil in real time. These data are transmitted to the central data processing platform through wireless communication technology. Using big data analysis and machine learning algorithms, the probability of landslide occurrence is dynamically evaluated and predicted. Once the pore water pressure is detected to be abnormal or reaches a critical value, the system can quickly issue a warning signal to notify relevant departments and residents to take emergency response measures, thereby effectively reducing the risks and losses caused by landslides and significantly improving the intelligence and scientific level of geological disaster management.

[0003] However, among the existing landslide pore water pressure prediction methods, there are traditional methods for predicting landslide pore water pressure that only focus on collecting water pressure data itself, or only predict other environmental data as auxiliary features, thus ignoring various complex potential factors that affect the landslide pore water pressure prediction process, resulting in insufficient prediction results and poor interpretability. Technical problems; Among the existing landslide pore water pressure prediction methods, there are traditional landslide pore water pressure prediction due to the complex environment and susceptibility to multiple influences such as extreme weather, soil corrosion and sensor interference, resulting in poor data quality and difficulty in supporting subsequent prediction tasks. Technical problems; Among the existing landslide pore water pressure prediction methods, there is a technical problem that the subsequent use of landslide pore water pressure prediction is difficult to quantify and implement, resulting in poor availability of landslide pore water pressure prediction results. Summary of the invention

[0004] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a landslide pore water pressure prediction method based on IoT. (1) In the existing landslide pore water pressure prediction methods, the traditional methods for predicting landslide pore water pressure only focus on collecting water pressure data itself, or only make predictions based on auxiliary features of other environmental data, thereby ignoring various complex potential factors affecting the landslide pore water pressure prediction process, resulting in insufficient prediction results and poor interpretability. The present solution creatively adopts the deployment of multiple sensors to collect multiple data indicators, and comprehensively decomposes and processes multi-modal features, and then adopts an improved long-term and short-term neural network combined with adaptive noise and variational mode decomposition to predict pore water pressure. The method improves the prediction accuracy and prediction depth of landslide pore water pressure and improves the usability of the method. In the existing landslide pore water pressure prediction methods, There is a technical problem that the traditional landslide pore water pressure prediction is difficult to support subsequent prediction tasks due to the complex environment and susceptibility to extreme weather, soil corrosion and sensor interference, which leads to poor data quality. This solution creatively adopts multi-directional redundant sensor deployment and data integration, and cloud storage through the Internet of Things platform, which improves the availability of data from multiple aspects and angles; in the existing landslide pore water pressure prediction methods, there is a technical problem that the subsequent use of landslide pore water pressure prediction is difficult to quantify and implement, which leads to poor availability of landslide pore water pressure prediction results. This solution creatively constructs an adaptive early warning mechanism, and designs early warnings for the predicted values ​​of multiple modes and indicators obtained by prediction, which improves the overall practicality and scalability of the method, and also provides a theoretical and practical basis for landslide pore water pressure prediction and deeper Internet of Things applications.

[0005] The technical solution adopted by the present invention is as follows: The present invention provides a landslide pore water pressure prediction method based on IoT, which includes the following steps:

[0006] Step S1: sensor deployment;

[0007] Step S2: data collection;

[0008] Step S3: data processing;

[0009] Step S4: pore water pressure prediction;

[0010] Step S5: Adaptive early warning.

[0011] Further, in step S1, the sensor deployment is used to design a sensor layout plan for landslide-prone areas and perform sensor layout, specifically to build a sensor cluster, and to arrange sensors in the landslide-prone areas according to the soil characteristics and terrain information according to the sensor cluster, and to collect data;

[0012] The sensor cluster specifically includes a pore water pressure sensor, a soil moisture sensor, a rainfall sensor, a temperature sensor, a humidity sensor and a landslide acceleration sensor.

[0013] Further, in step S2, the data collection is used to collect the raw data required for landslide pore water pressure prediction, specifically, based on the sensor cluster, multi-sensor redundant deployment is performed, and data redundant collection is performed, and sensor data processing is performed using an integrated algorithm to obtain raw data for pore water pressure prediction, and the real-time raw data for pore water pressure prediction is sent to the Internet of Things cloud platform through wireless communication technology for data storage;

[0014] The multi-sensor redundant deployment specifically refers to deploying three sensors of the same type at the location where the sensors are arranged to perform redundant data collection;

[0015] The redundant data collection specifically refers to collecting repeated types of sensor data to reduce the situation of data inaccuracy or failure;

[0016] The integration algorithm specifically adopts a weighted average algorithm;

[0017] The raw data for pore water pressure prediction specifically include pore water pressure data, soil moisture data, rainfall data, temperature data, humidity data and soil displacement data.

[0018] Further, in step S3, the data processing is used to optimize the original data, specifically to optimize the original pore water pressure prediction data to obtain pore water pressure prediction optimized data, which specifically includes the following steps:

[0019] Step S31: data cleaning, specifically, performing outlier detection and processing on the sensor data in the raw data for pore water pressure prediction, and supplementing missing and abnormal data by a mean filling method to obtain a cleaned data set;

[0020] Step S32: timestamp synchronization, specifically performing a timestamp synchronization operation on the cleaned data set to obtain a synchronized time series data set;

[0021] Step S33: data standardization, specifically, performing Z-score standardization on the synchronous time series data set to obtain a standardized sensor data set;

[0022] Step S34: data optimization processing, specifically, optimizing the raw data for pore water pressure prediction through data cleaning, timestamp synchronization and data standardization to obtain optimized data for pore water pressure prediction;

[0023] The pore water pressure prediction optimization data specifically includes standardized pore water pressure data, standardized soil moisture data, standardized rainfall data, standardized temperature data, standardized moisture data and standardized soil displacement data.

[0024] Further, in step S4, the pore water pressure prediction is used to predict the pore water pressure according to the optimized data, specifically, according to the pore water pressure prediction optimization data, an improved long-term and short-term neural network combining adaptive noise and variational mode decomposition is used to predict the pore water pressure, and pore water pressure prediction reference data is obtained, which specifically includes the following steps:

[0025] Step S41: decomposing the sensor data, specifically using a complete set empirical mode decomposition method to decompose the pore water pressure prediction optimization data to obtain an intrinsic mode function set;

[0026] Step S42: variational mode decomposition, specifically applying a variational mode decomposition method, constructing an optimization objective function, performing variational mode decomposition, and obtaining high-frequency decomposition mode data;

[0027] Step S43: extracting decomposition modal features, specifically using a statistical feature extraction method to extract multimodal features from the inherent modal function set and the high-frequency decomposition modal data to obtain a multimodal feature data set of pore water pressure;

[0028] Step S44: constructing an improved long short-term memory network, specifically, based on the multimodal feature data set of pore water pressure, by connecting the modal feature data sets in the multimodal feature data set of pore water pressure, constructing a feature-improved long short-term memory neural network, performing multimodal prediction of pore water pressure, and merging the prediction results of each mode to obtain a model prediction output, and the calculation formula is:

[0029] ;

[0030] In the formula, f t is the unit identifier of the feature-modified long short-term memory neural network, used to represent the input unit, output unit and forgetting unit, t is the time index, sig(·) is the S-type activation function, W f is the LSTM network unit weight, h t-1 is the output hidden state at the previous time t-1, (F1:F n ) is the multimodal feature connection data, which is used to represent the connection of the modal feature data sets in the pore water pressure multimodal feature data set. F1 is the modal feature data of the first mode. M is the modal characteristic data of the Mth mode, where M is specifically set to 4, which is used to represent the total number of characteristic modes in the multimodal characteristic data set of pore water pressure, and bf is the unit bias term;

[0031] W m is the classification prediction result of the mth mode, m is the mode index, is the classification weight vector, is the importance weight of the mth mode;

[0032] P(t) is the final predicted pore water pressure multimodal model prediction output, h t is the output hidden state of the feature-modified LSTM neural network;

[0033] Step S45: constructing an adaptive optimization algorithm, specifically, by adding adaptive weights, constructing an eagle hunting adaptive weight search algorithm, updating the model hyperparameter combination, and obtaining the optimal model parameter combination;

[0034] The calculation formula for constructing the hawk hunting adaptive weight search algorithm is:

[0035] ;

[0036] Where W new is the adaptive weight value of the updated eagle hunting search algorithm, W old is the adaptive weight value of the eagle hunting search algorithm before the update, rand is a random factor with a value range of [0,1], is the adaptive update factor weight, X target is the optimal solution of the current iterative search, which is used to represent the current optimal hyperparameter combination of the model. is the current number of iterations, T is the maximum number of iterations;

[0037] Step S46: training the pore water pressure prediction model, specifically, training the model through the sensor data decomposition, the variational mode decomposition, the decomposition mode feature extraction, the improved long short-term memory network and the adaptive optimization algorithm to obtain the pore water pressure prediction model Model WP ;

[0038] Step S47: pore water pressure prediction, specifically using the pore water pressure prediction model Model WP , performing pore water pressure prediction based on the pore water pressure prediction optimization data to obtain pore water pressure prediction reference data;

[0039] The pore water pressure prediction reference data specifically includes pore water pressure prediction value, soil moisture prediction value, rainfall prediction value and soil displacement prediction value.

[0040] Further, in step S5, the adaptive warning is used to construct an adaptive warning mechanism according to the predicted pore water pressure result, specifically, to construct an adaptive warning mechanism according to the pore water pressure prediction optimization data and the pore water pressure prediction reference data, obtain adaptive warning reference data, and perform adaptive warning;

[0041] The adaptive early warning mechanism specifically constructs a four-dimensional adaptive early warning threshold group, and reads the predicted values ​​in the pore water pressure prediction reference data in real time and performs data comparison based on the four-dimensional adaptive early warning threshold group, and performs adaptive early warning when the early warning conditions are met;

[0042] The four-dimensional adaptive warning threshold group specifically includes a warning threshold group and a warning category group;

[0043] The warning threshold group specifically includes a pore water pressure threshold, a soil moisture threshold, a rainfall threshold and a soil displacement threshold;

[0044] The warning category group specifically includes normal warning, warning warning and critical warning;

[0045] The construction of the adaptive early warning mechanism specifically adopts an adaptive fuzzy logic control method, including the following steps:

[0046] Step S51: input initialization, specifically, initializing input variables according to the four-dimensional adaptive warning threshold group to obtain output warning detection data;

[0047] Step S52: designing a fuzzy membership function, specifically designing a fuzzy membership function for each warning threshold group in the four-dimensional adaptive warning threshold group;

[0048] Step S53: constructing a fuzzy rule base, specifically, constructing a fuzzy rule base based on the pore water pressure prediction reference data and the manual warning history records, so as to determine the comprehensive warning type;

[0049] Step S54: fuzzy reasoning, specifically using Mamdani reasoning to build a fuzzy reasoning engine to obtain fuzzy reasoning output;

[0050] Step S55: defuzzification conversion, specifically using a centroid defuzzification method to convert the fuzzy reasoning output into a warning value to obtain a warning output;

[0051] Step S56: Real-time adaptive warning, specifically, performing real-time prediction warning according to the warning output, and dynamically adjusting the fuzzy membership function during the warning process, and performing adaptive warning update and maintenance.

[0052] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0053] (1) In the existing landslide pore water pressure prediction methods, there are traditional methods for predicting landslide pore water pressure, which only focus on collecting water pressure data itself, or only predict other environmental data as auxiliary features, thus ignoring various complex potential factors that affect the landslide pore water pressure prediction process, resulting in insufficient prediction results and poor interpretability. This scheme creatively deploys multiple sensors to collect multiple data indicators, and comprehensively decomposes and processes multi-modal features, and then uses an improved long-term and short-term neural network combined with adaptive noise and variational mode decomposition to predict pore water pressure. This method improves the prediction accuracy and prediction depth of landslide pore water pressure and improves the usability of the method.

[0054] (2) In view of the technical problem that the existing landslide pore water pressure prediction methods are complex and susceptible to extreme weather, soil corrosion and sensor interference, resulting in poor data quality and difficulty in supporting subsequent prediction tasks, this solution creatively adopts multi-directional redundant sensor deployment and data integration, and uses the Internet of Things platform for cloud storage, thereby improving the availability of data from multiple perspectives.

[0055] (3) In view of the technical problem that the subsequent use of landslide pore water pressure prediction is difficult to quantify and implement in the existing landslide pore water pressure prediction methods, which leads to poor usability of landslide pore water pressure prediction results, this scheme creatively constructs an adaptive early warning mechanism. By designing early warnings for the predicted values ​​of multiple modes and indicators, the overall practicality and scalability of the method are improved, and it also provides a theoretical and practical basis for landslide pore water pressure prediction and deeper Internet of Things applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 A schematic diagram of a flow chart of a landslide pore water pressure prediction method based on IoT provided by the present invention;

[0057] Figure 2 A schematic diagram of a landslide pore water pressure prediction system based on IoT provided by the present invention;

[0058] Figure 3 This is a schematic diagram of the data processing flow in step S3;

[0059] Figure 4 Schematic diagram of the process of pore water pressure prediction in step S4.

[0060] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0061] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0062] In the description of the present invention, it should be understood that terms such as “upper”, “lower”, “front”, “back”, “left”, “right”, “top”, “bottom”, “inside” and “outside” indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as limiting the present invention.

[0063] Example 1, see Figure 1 The present invention provides a landslide pore water pressure prediction method based on IoT, which includes the following steps:

[0064] Step S1: sensor deployment;

[0065] Step S2: data collection;

[0066] Step S3: data processing;

[0067] Step S4: pore water pressure prediction;

[0068] Step S5: Adaptive early warning.

[0069] By performing the above operations, in the existing landslide pore water pressure prediction methods, there is a traditional method for predicting landslide pore water pressure, which only focuses on collecting water pressure data itself, or only predicts auxiliary features of other environmental data, thereby ignoring various complex potential factors affecting the landslide pore water pressure prediction process, resulting in insufficient prediction results and poor interpretability. Technical problems, this scheme creatively adopts the deployment of multiple sensors to collect multiple data indicators, and comprehensively decomposes and processes multi-modal features, and then adopts an improved long-term and short-term neural network combined with adaptive noise and variational mode decomposition to predict pore water pressure, which improves the prediction accuracy and prediction depth of landslide pore water pressure and improves the usability of the method.

[0070] Example 2, see Figure 1 and Figure 2In step S1, the sensor deployment is used to design a sensor layout plan for landslide-prone areas and perform sensor layout, specifically to build a sensor cluster, and to arrange sensors in the landslide-prone areas according to the soil characteristics and terrain information according to the sensor cluster, and to collect data;

[0071] The sensor cluster specifically includes a pore water pressure sensor, a soil moisture sensor, a rainfall sensor, a temperature sensor, a humidity sensor and a landslide acceleration sensor;

[0072] The pore water pressure sensor is used to monitor the pore water pressure, water saturation and soil stability in the soil in real time;

[0073] The soil moisture sensor is used to measure the soil moisture content to assist in determining the moisture accumulation situation;

[0074] The rainfall sensor is used to measure rainfall;

[0075] The temperature sensor is used to monitor the ambient temperature and assist in characterizing the change in pore water pressure;

[0076] The humidity sensor is used to monitor the environmental humidity and assist in characterizing the change in pore water pressure;

[0077] The landslide acceleration sensor is used to monitor the displacement and vibration of soil or rock and provide landslide early warning information support.

[0078] Example 3, see Figure 1 , Figure 2 and Figure 3 , this embodiment is based on the above embodiment. In step S2, the data collection is used to collect the original data required for landslide pore water pressure prediction, specifically, based on the sensor cluster, multi-sensor redundant deployment is performed, and data redundant collection is performed, and sensor data processing is performed using an integrated algorithm to obtain pore water pressure prediction original data, and the real-time pore water pressure prediction original data is sent to the Internet of Things cloud platform through wireless communication technology for data storage;

[0079] The multi-sensor redundant deployment specifically refers to deploying three sensors of the same type at the location where the sensors are arranged to perform redundant data collection;

[0080] The redundant data collection specifically refers to collecting repeated types of sensor data to reduce the situation of data inaccuracy or failure;

[0081] The integration algorithm specifically adopts a weighted average algorithm;

[0082] The raw data for pore water pressure prediction specifically include pore water pressure data, soil moisture data, rainfall data, temperature data, humidity data and soil displacement data.

[0083] By performing the above operations, in view of the technical problem that in the existing landslide pore water pressure prediction methods, the traditional landslide pore water pressure prediction has poor data quality due to the complex environment and is susceptible to multiple factors such as extreme weather, soil corrosion and sensor interference, and is difficult to support subsequent prediction tasks, this solution creatively adopts multi-directional redundant sensor deployment and data integration, and performs cloud storage through the Internet of Things platform, thereby improving the availability of data from multiple aspects and angles.

[0084] Example 4, see Figure 1 , Figure 2 and Figure 4 This embodiment is based on the above embodiment. In step S3, the data processing is used to optimize the original data, specifically to optimize the original pore water pressure prediction data to obtain the pore water pressure prediction optimization data, which specifically includes the following steps:

[0085] Step S31: data cleaning, specifically, performing outlier detection and processing on the sensor data in the raw data for pore water pressure prediction, and supplementing missing and abnormal data by a mean filling method to obtain a cleaned data set;

[0086] Step S32: timestamp synchronization, specifically performing a timestamp synchronization operation on the cleaned data set to obtain a synchronized time series data set;

[0087] Step S33: data standardization, specifically, performing Z-score standardization on the synchronous time series data set to obtain a standardized sensor data set;

[0088] Step S34: data optimization processing, specifically, optimizing the raw data for pore water pressure prediction through data cleaning, timestamp synchronization and data standardization to obtain optimized data for pore water pressure prediction;

[0089] The pore water pressure prediction optimization data specifically includes standardized pore water pressure data, standardized soil moisture data, standardized rainfall data, standardized temperature data, standardized moisture data and standardized soil displacement data.

[0090] Example 5, see Figure 1 and Figure 2This embodiment is based on the above embodiment. In step S4, the pore water pressure prediction is used to predict the pore water pressure according to the optimized data. Specifically, based on the pore water pressure prediction optimized data, an improved long-term and short-term neural network combining adaptive noise and variational mode decomposition is used to predict the pore water pressure to obtain pore water pressure prediction reference data. Specifically, the following steps are included:

[0091] Step S41: sensor data decomposition, specifically using a complete set empirical mode decomposition method to decompose the pore water pressure prediction optimization data to obtain an intrinsic mode function set, and the calculation formula is:

[0092] ;

[0093] Where X(t) is the original sensor signal, which is used to represent the signal data in the pore water pressure prediction optimization data, I is the total number of modes after decomposition, i is the decomposition mode index, and IMF is i (t) is the intrinsic mode function corresponding to the i-th decomposed mode, t is the time index, and r(t) is the residual signal after decomposition;

[0094] Step S42: variational mode decomposition, specifically applying a variational mode decomposition method, constructing an optimization objective function, performing variational mode decomposition, and obtaining high-frequency decomposition mode data;

[0095] The calculation formula of the optimization objective function is:

[0096] ;

[0097] Where F is the variational mode decomposition objective function, min is the minimum function, and u k is the value of the kth variational high-frequency mode function, w k is the central frequency value of the kth variational high-frequency mode, K is the total number of high-frequency modes obtained by variational mode decomposition, k is the high-frequency mode index, X(t) is the original sensor signal, is the penalty factor, j is the imaginary unit;

[0098] Step S43: extracting decomposition modal features, specifically using a statistical feature extraction method to extract multimodal features from the inherent modal function set and the high-frequency decomposition modal data to obtain a multimodal feature data set of pore water pressure;

[0099] The pore water pressure multi-modal characteristic data set specifically includes a pore water pressure modal characteristic data set, a soil moisture modal characteristic data set, a rainfall modal characteristic data set and a soil displacement modal characteristic data set;

[0100] The statistical feature extraction specifically extracts mean features, standard deviation features, maximum features, skewness features and kurtosis features;

[0101] Step S44: constructing an improved long short-term memory network, specifically, based on the multimodal feature data set of pore water pressure, by connecting the modal feature data sets in the multimodal feature data set of pore water pressure, constructing a feature-improved long short-term memory neural network, performing multimodal prediction of pore water pressure, and merging the prediction results of each mode to obtain a model prediction output, and the calculation formula is:

[0102] ;

[0103] In the formula, f t is the unit identifier of the feature-modified long short-term memory neural network, used to represent the input unit, output unit and forgetting unit, t is the time index, sig(·) is the S-type activation function, W f is the LSTM network unit weight, h t-1 is the output hidden state at the previous time t-1, (F1:F n ) is the multimodal feature connection data, which is used to represent the connection of the modal feature data sets in the pore water pressure multimodal feature data set. F1 is the modal feature data of the first mode. M is the modal characteristic data of the Mth mode, where M is specifically set to 4, which is used to represent the total number of characteristic modes in the multimodal characteristic data set of pore water pressure, and b f is the unit bias term;

[0104] W m is the classification prediction result of the mth mode, m is the mode index, is the classification weight vector, is the importance weight of the mth mode;

[0105] P(t) is the final predicted pore water pressure multimodal model prediction output, h t is the output hidden state of the feature-modified LSTM neural network;

[0106] Step S45: constructing an adaptive optimization algorithm, specifically, by adding adaptive weights, constructing an eagle hunting adaptive weight search algorithm, updating the model hyperparameter combination, and obtaining the optimal model parameter combination;

[0107] The calculation formula for constructing the hawk hunting adaptive weight search algorithm is:

[0108] ;

[0109] Where W new is the adaptive weight value of the updated eagle hunting search algorithm, Wold is the adaptive weight value of the eagle hunting search algorithm before the update, rand is a random factor with a value range of [0,1], is the adaptive update factor weight, X target is the optimal solution of the current iterative search, which is used to represent the current optimal hyperparameter combination of the model. is the current number of iterations, T is the maximum number of iterations;

[0110] Step S46: training the pore water pressure prediction model, specifically, training the model through the sensor data decomposition, the variational mode decomposition, the decomposition mode feature extraction, the improved long short-term memory network and the adaptive optimization algorithm to obtain the pore water pressure prediction model Model WP ;

[0111] Step S47: pore water pressure prediction, specifically using the pore water pressure prediction model Model WP , performing pore water pressure prediction based on the pore water pressure prediction optimization data to obtain pore water pressure prediction reference data;

[0112] The pore water pressure prediction reference data specifically includes pore water pressure prediction value, soil moisture prediction value, rainfall prediction value and soil displacement prediction value.

[0113] By performing the above operations, in the existing landslide pore water pressure prediction methods, there is a traditional method for predicting landslide pore water pressure, which only focuses on collecting water pressure data itself, or only predicts auxiliary features of other environmental data, thereby ignoring various complex potential factors affecting the landslide pore water pressure prediction process, resulting in insufficient prediction results and poor interpretability. Technical problems, this scheme creatively adopts the deployment of multiple sensors to collect multiple data indicators, and comprehensively decomposes and processes multi-modal features, and then adopts an improved long-term and short-term neural network combined with adaptive noise and variational mode decomposition to predict pore water pressure, which improves the prediction accuracy and prediction depth of landslide pore water pressure and improves the usability of the method.

[0114] Embodiment 7, see Figure 1 and Figure 2 This embodiment is based on the above embodiment. In step S5, the adaptive warning is used to construct an adaptive warning mechanism according to the predicted pore water pressure result. Specifically, an adaptive warning mechanism is constructed according to the pore water pressure prediction optimization data and the pore water pressure prediction reference data, and adaptive warning reference data is obtained, and adaptive warning is performed;

[0115] The adaptive early warning mechanism specifically constructs a four-dimensional adaptive early warning threshold group, and reads the predicted values ​​in the pore water pressure prediction reference data in real time and performs data comparison based on the four-dimensional adaptive early warning threshold group, and performs adaptive early warning when the early warning conditions are met;

[0116] The four-dimensional adaptive warning threshold group specifically includes a warning threshold group and a warning category group;

[0117] The warning threshold group specifically includes a pore water pressure threshold, a soil moisture threshold, a rainfall threshold and a soil displacement threshold;

[0118] The warning category group specifically includes normal warning, warning warning and critical warning;

[0119] The construction of the adaptive early warning mechanism specifically adopts an adaptive fuzzy logic control method, including the following steps:

[0120] Step S51: input initialization, specifically, initializing input variables according to the four-dimensional adaptive warning threshold group to obtain output warning detection data;

[0121] Step S52: designing a fuzzy membership function, specifically designing a fuzzy membership function for each warning threshold group in the four-dimensional adaptive warning threshold group, and the calculation formula is:

[0122] ;

[0123] In the formula, f N (·) is the normal warning fuzzy membership function, f W (·) is the warning fuzzy membership function, f D (·) is the critical warning fuzzy membership function. When the fuzzy membership function takes the value of 0, it means no warning is given. When the fuzzy membership function takes the value of 1, it means a warning is given. x is the output warning detection data. is the lower limit threshold of normal warning, is the normal warning upper threshold, is the lower limit of the warning warning threshold, is the upper threshold of the warning alert. is the lower threshold of the emergency warning, is the upper threshold of the emergency warning;

[0124] Step S53: constructing a fuzzy rule base, specifically, constructing a fuzzy rule base based on the pore water pressure prediction reference data and the manual warning history records, so as to determine the comprehensive warning type;

[0125] Step S54: fuzzy reasoning, specifically using Mamdani reasoning to build a fuzzy reasoning engine to obtain fuzzy reasoning output;

[0126] Step S55: defuzzification conversion, specifically using a centroid defuzzification method to convert the fuzzy reasoning output into a warning value to obtain a warning output;

[0127] Step S56: Real-time adaptive warning, specifically, performing real-time prediction warning according to the warning output, and dynamically adjusting the fuzzy membership function during the warning process, and performing adaptive warning update and maintenance.

[0128] Table 1 is a table showing the specific contents of the four-dimensional adaptive warning threshold group, as shown in the table, pore water pressure threshold (unit: kPa), including normal warning, warning warning and critical warning;

[0129] Soil moisture thresholds, including normal warning, warning warning, and critical warning;

[0130] Rainfall threshold (unit: mm / h), including normal warning, warning warning and critical warning;

[0131] Soil displacement threshold (unit: mm), including normal warning, warning warning and critical warning.

[0132] Table 1 Specific contents of four-dimensional adaptive warning threshold group

[0133]

[0134] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0135] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

[0136] The present invention and its embodiments are described above, which is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it and design structural methods and embodiments similar to the technical solution without creativity without departing from the purpose of the invention, they should all fall within the protection scope of the present invention.

Claims

1. A landslide pore water pressure prediction method based on IoT, characterized by: The method comprises the following steps: Step S1: sensor deployment, building a sensor cluster and arranging sensors; Step S2: data collection, based on the sensor cluster, multiple sensors are deployed redundantly, and data is collected redundantly to obtain raw data for pore water pressure prediction; Step S3: data processing to obtain pore water pressure prediction optimization data; Step S4: pore water pressure prediction, based on the pore water pressure prediction optimization data, adopt an improved long short-term neural network combined with adaptive noise and variational mode decomposition to predict pore water pressure, and obtain pore water pressure prediction reference data, specifically including the following steps: step S41: sensor data decomposition; step S42: variational mode decomposition; step S43: decomposition mode feature extraction; step S44: constructing an improved long short-term memory network; step S45: constructing an adaptive optimization algorithm; step S46: pore water pressure prediction model training; step S47: pore water pressure prediction; Step S5: Adaptive warning, building an adaptive warning mechanism, obtaining adaptive warning reference data, and performing adaptive warning.

2. The method for predicting landslide pore water pressure based on IoT according to claim 1, characterized in that: In step S1, the sensor deployment is used to design a sensor layout plan for a landslide-prone area and to arrange the sensors, specifically to build a sensor cluster, and to arrange the sensors in a landslide-prone area according to the soil characteristics and terrain information according to the sensor cluster, and to collect data; The sensor cluster specifically includes a pore water pressure sensor, a soil moisture sensor, a rainfall sensor, a temperature sensor, a humidity sensor and a landslide acceleration sensor.

3. The method for predicting landslide pore water pressure based on IoT according to claim 2 is characterized in that: The data collection is used to collect the raw data required for landslide pore water pressure prediction, specifically, based on the sensor cluster, multi-sensor redundant deployment, and data redundant collection, and sensor data processing using an integrated algorithm to obtain pore water pressure prediction raw data, and send the real-time pore water pressure prediction raw data to the Internet of Things cloud platform through wireless communication technology for data storage; The multi-sensor redundant deployment specifically refers to deploying three sensors of the same type at the location where the sensors are arranged to perform redundant data collection; The redundant data collection specifically refers to collecting repeated types of sensor data to reduce the situation of data inaccuracy or failure; The raw data for pore water pressure prediction specifically include pore water pressure data, soil moisture data, rainfall data, temperature data, humidity data and soil displacement data.

4. The method for predicting landslide pore water pressure based on IoT according to claim 3 is characterized by: In step S3, the data processing is used to optimize the original data, specifically to optimize the original pore water pressure prediction data to obtain pore water pressure prediction optimized data, which specifically includes the following steps: Step S31: data cleaning, specifically, performing outlier detection and processing on the sensor data in the raw data for pore water pressure prediction, and supplementing missing and abnormal data by a mean filling method to obtain a cleaned data set; Step S32: timestamp synchronization, specifically performing a timestamp synchronization operation on the cleaned data set to obtain a synchronized time series data set; Step S33: data standardization, specifically, performing Z-score standardization on the synchronous time series data set to obtain a standardized sensor data set; Step S34: data optimization processing, specifically, optimizing the raw data for pore water pressure prediction through data cleaning, timestamp synchronization and data standardization to obtain optimized data for pore water pressure prediction; The pore water pressure prediction optimization data specifically includes standardized pore water pressure data, standardized soil moisture data, standardized rainfall data, standardized temperature data, standardized moisture data and standardized soil displacement data.

5. The method for predicting landslide pore water pressure based on IoT according to claim 4 is characterized by: In step S4, the following steps are specifically included: Step S41: decomposing the sensor data, specifically using a complete set empirical mode decomposition method to decompose the pore water pressure prediction optimization data to obtain an intrinsic mode function set; Step S42: variational mode decomposition, specifically applying a variational mode decomposition method to perform variational mode decomposition by constructing an optimization objective function; Step S43: extracting decomposition modal features, specifically using a statistical feature extraction method to extract multimodal features from the inherent modal function set and the high-frequency decomposition modal data to obtain a multimodal feature data set of pore water pressure; Step S44: constructing an improved long short-term memory network, specifically, based on the multimodal feature data set of pore water pressure, by connecting the modal feature data sets in the multimodal feature data set of pore water pressure, constructing a feature-improved long short-term memory neural network, performing multimodal prediction of pore water pressure, and merging the prediction results of each mode to obtain a model prediction output; Step S45: constructing an adaptive optimization algorithm, specifically, by adding adaptive weights, constructing an eagle hunting adaptive weight search algorithm, updating the model hyperparameter combination, and obtaining the optimal model parameter combination; Step S46: training the pore water pressure prediction model, specifically, training the model through the sensor data decomposition, the variational mode decomposition, the decomposition mode feature extraction, the improved long short-term memory network and the adaptive optimization algorithm to obtain the pore water pressure prediction model Model WP ; Step S47: pore water pressure prediction, specifically using the pore water pressure prediction model Model WP , performing pore water pressure prediction based on the pore water pressure prediction optimization data to obtain pore water pressure prediction reference data; The pore water pressure prediction reference data specifically includes pore water pressure prediction value, soil moisture prediction value, rainfall prediction value and soil displacement prediction value.

6. The method for predicting landslide pore water pressure based on IoT according to claim 5, characterized in that: In step S44, the improved long short-term memory network is constructed, and the calculation formula is: ； In the formula, f t is the unit identifier of the feature-modified long short-term memory neural network, used to represent the input unit, output unit and forgetting unit, t is the time index, sig(·) is the S-type activation function, W f is the LSTM network unit weight, h t-1 is the output hidden state at the previous time t-1, (F1:F n ) is the multimodal feature connection data, which is used to represent the connection of the modal feature data sets in the pore water pressure multimodal feature data set. F1 is the modal feature data of the first mode. M is the modal characteristic data of the Mth mode, where M is specifically set to 4, which is used to represent the total number of characteristic modes in the multimodal characteristic data set of pore water pressure, and b f is the unit bias term; W m is the classification prediction result of the mth mode, m is the mode index, is the classification weight vector, is the importance weight of the mth mode; P(t) is the final predicted pore water pressure multimodal model prediction output, h t is the output hidden state of the feature-modified LSTM neural network; In step S45, the calculation formula for constructing the hawk hunting adaptive weight search algorithm is: ； Where W new is the adaptive weight value of the updated eagle hunting search algorithm, W old is the adaptive weight value of the eagle hunting search algorithm before the update, rand is a random factor with a value range of [0,1], is the adaptive update factor weight, X target is the optimal solution of the current iterative search, which is used to represent the current optimal hyperparameter combination of the model. is the current iteration number, and T is the maximum iteration number.

7. The method for predicting landslide pore water pressure based on IoT according to claim 6, characterized in that: In step S5, in step S5, the adaptive warning is used to construct an adaptive warning mechanism according to the predicted pore water pressure results, specifically, to construct an adaptive warning mechanism based on the pore water pressure prediction optimization data and the pore water pressure prediction reference data, obtain adaptive warning reference data, and perform adaptive warning.

8. The method for predicting landslide pore water pressure based on IoT according to claim 7, characterized in that: In step S5, the adaptive early warning mechanism specifically constructs a four-dimensional adaptive early warning threshold group, and reads the predicted values ​​in the pore water pressure prediction reference data in real time and performs data comparison based on the four-dimensional adaptive early warning threshold group, and performs adaptive early warning when the early warning conditions are met; The four-dimensional adaptive warning threshold group specifically includes a warning threshold group and a warning category group; The warning threshold group specifically includes a pore water pressure threshold, a soil moisture threshold, a rainfall threshold and a soil displacement threshold; The warning category group specifically includes normal warning, warning warning and critical warning; The adaptive early warning mechanism is constructed by adopting an adaptive fuzzy logic control method, which includes the following steps: step S51: input initialization; step S52: designing fuzzy membership function; step S53: constructing a fuzzy rule base; step S54: fuzzy reasoning; step S55: defuzzification transformation; step S56: real-time adaptive early warning.

Images