Geotechnical engineering investigation safety monitoring system and method based on artificial intelligence

Abstract

The invention discloses a geotechnical engineering investigation safety monitoring system and method based on artificial intelligence, and relates to the technical field of rock-soil body deformation safety monitoring, and the method specifically comprises the following steps: obtaining multi-source data including deformation, stress and environmental data of a rock-soil body, and building a model input data source; performing data fusion processing; a data fusion processing result is used as input, and a rock-soil body stress and deformation incidence relation is obtained through data simulation; the method comprises the following steps: respectively constructing a short-term recursive calculation model and a long-term trend analysis model through a time sequence prediction method based on deep learning by taking a rock-soil body stress and deformation incidence relation as a substrate; and obtaining a real-time multi-source data value, obtaining a displacement prediction vector of the rock-soil body according to the short-term recursive calculation model, comparing the displacement prediction vector with a preset threshold value, performing real-time early warning when the displacement prediction vector exceeds the preset threshold value, and visualizing a long-term trend analysis result. According to the invention, the engineering safety is enhanced, and the long-term stability and reliability of infrastructures are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of rock and soil deformation safety monitoring technology, and in particular to a rock and soil engineering investigation safety monitoring system and method based on artificial intelligence. Background Art

[0002] Geotechnical engineering investigation is a crucial component of infrastructure construction, involving the assessment of geotechnical stability and safety. Traditional monitoring methods, such as total station monitoring, inclinometer measurements, and pore water pressure testing, primarily rely on field measurements and empirical judgment. However, these methods suffer from data update lag, limited measurement points, and accuracy affected by human factors, making them unable to meet the high-precision, real-time, and intelligent requirements of modern engineering. Furthermore, due to the complex forces acting on geotechnical structures and the nonlinear and time-varying deformation characteristics, traditional monitoring methods struggle to accurately predict future stability states, impacting the accuracy of engineering safety assessments. In recent years, the development of the Internet of Things (IoT) and artificial intelligence (AI) technologies has driven the intelligentization of geotechnical monitoring. The IoT enables large-scale data collection, while AI improves prediction accuracy through big data analysis and deep learning. However, existing monitoring methods still face challenges in multi-source data integration, traditional time series models fail to fully account for the nonlinear characteristics of geotechnical structures, and geotechnical investigations rely too heavily on human resources. Consequently, even subtle geotechnical deformations are often difficult to detect and fail to balance short-term sudden changes with long-term trend analysis. Therefore, more efficient, accurate, and intelligent methods for geotechnical deformation safety monitoring are urgently needed. Summary of the Invention

[0003] The purpose of the present invention is to provide a geotechnical engineering investigation safety monitoring system and method based on artificial intelligence to solve the problems raised in the prior art.

[0004] To achieve the above-mentioned object, the present invention provides the following technical solution: a geotechnical engineering investigation safety monitoring method based on artificial intelligence, wherein the geotechnical engineering investigation safety monitoring method specifically comprises the following steps:

[0005] Step S100: Acquire multi-source data, including deformation, stress and environmental data of rock and soil, and establish a model input data source;

[0006] Step S200: performing data fusion processing on the acquired multi-source data;

[0007] Step S300: using the result of data fusion processing as input, solving the stress and deformation of the continuous rock and soil mass by finite element method through data simulation to obtain the correlation relationship between the stress and deformation of the rock and soil mass;

[0008] Step S400: Based on deep learning and taking the relationship between rock and soil stress and deformation as the basis, a short-term recursive calculation model and a long-term trend analysis model are constructed respectively through a time series prediction method;

[0009] Step S500: Acquire real-time multi-source data values, obtain displacement prediction vectors of the geotechnical mass according to the short-term recursive calculation model, compare them with preset threshold values, issue real-time warnings when the threshold values are exceeded, and visualize long-term trend analysis results.

[0010] In step S100, multi-source data is acquired, including deformation, stress and environmental data of the rock mass, and a model input data source is established, specifically:

[0011] Equipped with IoT devices to conduct real-time monitoring, obtain deformation, stress and environmental data of rock and soil, and establish model input data source;

[0012] Including surface monitoring, underground monitoring, vibration monitoring and environmental monitoring;

[0013] The surface monitoring includes laser ranging and inclination measurement;

[0014] The underground monitoring is used to monitor the pore water pressure of rock and soil;

[0015] The vibration monitoring equipment includes a MEMS accelerometer and a seismic array;

[0016] The environmental monitoring includes temperature, humidity, rainfall and remote sensing images.

[0017] In step S200, data fusion processing is performed on the acquired multi-source data, specifically:

[0018] The data fusion process specifically includes the following steps:

[0019] Step S201: Replace outliers in the acquired multi-source data;

[0020] Step S202: Set timestamps, and align multi-source data with different time and space based on the time scale between the timestamps to achieve data fusion of the multi-source data;

[0021] Step S203: different data are kept at the same scale by a standardization transformation method. The specific formula is: X'=(X-μ) / a; wherein X' represents the data after standardization transformation; X represents the data after data fusion; μ represents the average value of data X; a represents the standard deviation of data X.

[0022] In step S201, outlier replacement is performed on the acquired multi-source data, specifically:

[0023] Step S211: sort the different types of multi-source data in descending order to obtain a descending sequence;

[0024] Step S212: Calculate the upper quartile, median, and lower quartile of the data according to the descending sequence; the upper quartile, median, and lower quartile are the data corresponding to the upper quartile point Q1, the median point Q2, and the lower quartile point Q3 in the descending sequence, respectively; Among them, m represents the number of data in the descending sequence, and m is a positive integer; Q1 and Q2 are rounded up; Q3 is rounded down;

[0025] Step S213: Calculate the difference between the upper quartile and the lower quartile, and take the absolute value of the difference as IQR;

[0026] Step S214: using the upper quartile and lower quartile as upper and lower limits, respectively, and determining data greater than the upper quartile plus 1.5 times the IQR and data less than the lower quartile minus 1.5 times the IQR as outliers;

[0027] Step S215: replace the outlier with the value of the upper quartile plus 1.5 times the IQR, and replace the outlier that is smaller than the lower quartile minus 1.5 times the IQR with the value of the lower quartile minus 1.5 times the IQR.

[0028] In step S300, the result of data fusion processing is used as input, and the stress and deformation of the continuous rock and soil body are solved by finite element method through data simulation to obtain the correlation relationship between the stress and deformation of the rock and soil body, which is specifically:

[0029] Step S301: construct an equilibrium equation for the rock-soil body according to the equilibrium relationship between the stress tensor of the rock-soil body and the external body force of the rock-soil body under static conditions. The equilibrium equation is represented as follows: Where δ represents the stress tensor; f represents the external body force; represents stress divergence;

[0030] Wherein, the stress tensor δ is composed of normal stress and shear stress;

[0031]

[0032] δ xx represents the normal stress along the x direction (acting on the yz plane);

[0033] δ yy represents the normal stress along the y direction (acting on the xz plane);

[0034] δ zz represents the normal stress along the z direction (acting on the xy plane);

[0035] τ xy represents the shear stress in the xy plane (acting along the x direction on the y axis);

[0036] τ yz represents the shear stress in the yz plane (acting along the y direction toward the z axis);

[0037] τ zx represents the shear stress on the zx plane (acting along the z direction toward the x-axis);

[0038] Due to static equilibrium, the shear stress satisfies symmetry;

[0039] Preferably, the external body force f includes gravity and pore water pressure of the rock and soil mass;

[0040] The gravity f g =ρg; where ρ represents the density of the rock mass; g represents the acceleration of gravity;

[0041] The pore water pressure of the rock and soil Among them, ρ w Indicates the density of pore water in rock and soil; It represents the gradient of the pore water head height of the rock mass. The negative sign indicates that the direction of the pore water pressure in the rock mass is opposite to the direction of the water head gradient and depends on the direction of water flow.

[0042] Step S302: Solve the stress and strain relationship of the rock and soil mass to obtain the displacement vector of the rock and soil mass.

[0043] In step S303, the stress and strain relationship of the rock and soil mass is solved to obtain the displacement vector of the rock and soil mass, which specifically includes the following steps:

[0044] Step S311, discretize the rock and soil into finite triangular units;

[0045] Step S312: For each triangular unit, use the shape function to approximate the deformation displacement to obtain the displacement interpolation relationship of each point in the triangular unit;

[0046] Preferably, the shape function is:

[0047]

[0048] Among them, u(x,y,z) represents the deformation displacement of the rock mass at the spatial coordinate (x,y,z), which is expressed by the shape function N i (x, y, z) is obtained by interpolating the node displacement;

[0049] u i is the displacement value of node i, n is the total number of nodes involved in the interpolation;

[0050] The displacement of any point in the unit is obtained by summing the product of the shape function of all nodes and the displacement of the corresponding nodes;

[0051] Step S313: Calculate the stiffness matrix based on the deformation and displacement of each triangular unit; the stiffness matrix is used to describe the relationship between the deformation and force of the rock mass when subjected to force;

[0052] The stiffness matrix is specifically represented by:

[0053] K=∫ V B T DBdV;

[0054] Where K represents the stiffness matrix; B represents the shape function derivative matrix; B T The transposed matrix representing the derivative moments of the shape function;

[0055] The shape function derivative matrix B is composed of the partial derivatives of the shape function with respect to the coordinates, and is used to represent the relationship between the stress and strain in the triangular element and the node displacement;

[0056] D represents the elasticity matrix;

[0057] Preferably, the elastic matrix is determined by elastic modulus and Poisson's ratio, and is used to convert strain into stress;

[0058] V represents the volume of the triangular unit;

[0059] Step S314: Calculate the displacement vector of the rock and soil mass according to the stiffness matrix.

[0060] The calculation formula of the rock and soil displacement vector is:

[0061] KU=f;

[0062] Where U represents the displacement vector, which is a column vector consisting of the displacement values of all nodes.

[0063] In step S400, based on deep learning and taking the relationship between rock and soil stress and deformation as the basis, a short-term recursive calculation model and a long-term trend analysis model are constructed respectively through the time series prediction method. Specifically,

[0064] Step S401: construct a short-term recursive calculation model by a time series prediction method, wherein the short-term recursive calculation model includes an input gate, a forget gate, an output gate, a cell state, and a hidden state part;

[0065] Step S402: Use the random forest method to construct multiple subsets by randomly sampling with replacement from the original training data, perform secondary training on each subset using the method of step S401, take the average value of the secondary training of all subsets as the final prediction result, and construct a long-term trend analysis model based on the secondary training results over a certain period of time.

[0066] In step S500, real-time multi-source data values are obtained, and the displacement prediction vector of the geotechnical mass is obtained according to the short-term recursive calculation model. The displacement prediction vector is compared with the preset threshold value, and a real-time warning is issued when the threshold value is exceeded. The long-term trend analysis results are visualized, specifically:

[0067] Step S501: acquiring real-time multi-source data values, and obtaining a displacement prediction vector of the rock and soil mass according to the short-term recursive calculation model;

[0068] Step S502: Setting a displacement threshold value of the rock and soil mass, comparing it with the preset threshold value, and issuing a real-time warning when the threshold value is exceeded;

[0069] Step S503: Visualize the long-term trend analysis results.

[0070] An artificial intelligence-based geotechnical engineering investigation and safety monitoring system, comprising a data acquisition module, a data fusion module, a correlation analysis module, a deep learning module, and an intelligent early warning and visualization module;

[0071] The data acquisition module is used to equip IoT devices to obtain deformation, stress and environmental data of the rock and soil mass and establish a model input data source;

[0072] The data fusion module is used to perform fusion processing on the collected multi-source data, including the processing steps of outlier replacement, time alignment and data standardization, matching data from different sources through timestamps, and using standardization methods to ensure data scale consistency;

[0073] The correlation analysis module is used to establish the correlation relationship between rock and soil stress and deformation based on the finite element method;

[0074] The deep learning module is used to build a short-term recursive calculation model and a long-term trend analysis model based on the time series prediction method, and introduces an attention mechanism to improve prediction accuracy to fully consider the nonlinear and time-varying characteristics of the rock and soil mass;

[0075] The intelligent early warning and visualization module is used to perform real-time early warning and visualization analysis based on the prediction results;

[0076] Compared with the existing technology, the beneficial effects of the present invention are: by introducing the Internet of Things and deep learning technology, the present method has significantly improved the intelligence level of geotechnical engineering monitoring, which not only enables large-scale, real-time data collection, but also improves monitoring accuracy through multi-source data fusion, avoiding the errors caused by limited measuring points and data lag in traditional methods; at the same time, the intelligent prediction model based on deep learning can fully consider the nonlinear and time-varying characteristics of the rock and soil body, greatly improving the accuracy of deformation trend and stability prediction; in addition, combined with the intelligent early warning mechanism, it can achieve real-time response to geological environment changes, discover potential risks in advance and take effective measures, enhance engineering safety, improve the long-term stability and reliability of infrastructure, and reduce the waste of human resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 This is a schematic diagram of the steps of a geotechnical engineering investigation and safety monitoring method based on artificial intelligence of the present invention;

[0078] Figure 2 This is a structural schematic diagram of an artificial intelligence-based geotechnical engineering investigation safety monitoring system of the present invention. DETAILED DESCRIPTION

[0079] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making creative efforts are within the scope of protection of the present invention.

[0080] Example: Figure 1-Figure 2 As shown, the present invention provides a technical solution, a geotechnical engineering investigation safety monitoring method based on artificial intelligence, such as Figure 1 As shown, the geotechnical engineering investigation safety monitoring method specifically includes the following steps:

[0081] Step S100: Acquire multi-source data, including deformation, stress and environmental data of rock and soil, and establish a model input data source;

[0082] Step S200: performing data fusion processing on the acquired multi-source data;

[0083] Step S300: using the result of data fusion processing as input, solving the stress and deformation of the continuous rock and soil mass by finite element method through data simulation to obtain the correlation relationship between the stress and deformation of the rock and soil mass;

[0084] Step S400: Based on deep learning and taking the relationship between rock and soil stress and deformation as the basis, a short-term recursive calculation model and a long-term trend analysis model are constructed respectively through a time series prediction method;

[0085] Step S500: Acquire real-time multi-source data values, obtain displacement prediction vectors of the geotechnical mass according to the short-term recursive calculation model, compare them with preset threshold values, issue real-time warnings when the threshold values are exceeded, and visualize long-term trend analysis results.

[0086] In step S100, multi-source data is acquired, including deformation, stress and environmental data of the rock mass, and a model input data source is established, specifically:

[0087] Equipped with IoT devices to conduct real-time monitoring, obtain deformation, stress and environmental data of rock and soil, and establish model input data source;

[0088] Including surface monitoring, underground monitoring, vibration monitoring and environmental monitoring;

[0089] The surface monitoring includes laser ranging and inclination measurement;

[0090] The underground monitoring is used to monitor the pore water pressure of rock and soil;

[0091] The vibration monitoring equipment includes a MEMS accelerometer and a seismic array;

[0092] The environmental monitoring includes temperature, humidity, rainfall and remote sensing images.

[0093] In step S200, data fusion processing is performed on the acquired multi-source data, specifically:

[0094] The data fusion process specifically includes the following steps:

[0095] Step S201: Replace outliers in the acquired multi-source data;

[0096] Step S202: Set timestamps, and align multi-source data with different time and space based on the time scale between timestamps to achieve data fusion of multi-source data;

[0097] Step S203: different data are kept at the same scale by a standardization transformation method. The specific formula is: X'=(X-μ) / a; wherein X' represents the data after standardization transformation; X represents the data after data fusion; μ represents the average value of data X; a represents the standard deviation of data X.

[0098] In step S201, outlier replacement is performed on the acquired multi-source data, specifically:

[0099] Step S211: sort the different types of multi-source data in descending order to obtain a descending sequence;

[0100] Step S212: Calculate the upper quartile, median, and lower quartile of the data according to the descending sequence; the upper quartile, median, and lower quartile are the data corresponding to the upper quartile point Q1, the median point Q2, and the lower quartile point Q3 in the descending sequence, respectively; Among them, m represents the number of data in the descending sequence, and m is a positive integer; Q1 and Q2 are rounded up; Q3 is rounded down;

[0101] Step S213: Calculate the difference between the upper quartile and the lower quartile, and take the absolute value of the difference as IQR;

[0102] Step S214: using the upper quartile and lower quartile as upper and lower limits, respectively, and determining data greater than the upper quartile plus 1.5 times the IQR and data less than the lower quartile minus 1.5 times the IQR as outliers;

[0103] Step S215: replace the outlier with the value of the upper quartile plus 1.5 times the IQR, and replace the outlier that is smaller than the lower quartile minus 1.5 times the IQR with the value of the lower quartile minus 1.5 times the IQR.

[0104] In step S300, the result of data fusion processing is used as input, and the stress and deformation of the continuous rock and soil body are solved by finite element method through data simulation to obtain the correlation relationship between the stress and deformation of the rock and soil body, which is specifically:

[0105] Step S301: construct an equilibrium equation for the rock-soil body according to the equilibrium relationship between the stress tensor of the rock-soil body and the external body force of the rock-soil body under static conditions. The equilibrium equation is represented as follows: Where δ represents the stress tensor; f represents the external body force; represents stress divergence;

[0106] Wherein, the stress tensor δ is composed of normal stress and shear stress;

[0107]

[0108] δ xx represents the normal stress along the x direction (acting on the yz plane);

[0109] δ yy represents the normal stress along the y direction (acting on the xz plane);

[0110] δ zz represents the normal stress along the z direction (acting on the xy plane);

[0111] τ xy represents the shear stress in the xy plane (acting along the x direction on the y axis);

[0112] τ yz represents the shear stress in the yz plane (acting along the y direction toward the z axis);

[0113] τ zx represents the shear stress on the zx plane (acting along the z direction toward the x-axis);

[0114] Due to static equilibrium, the shear stress satisfies symmetry;

[0115] Preferably, the external body force f includes gravity and pore water pressure of the rock and soil mass;

[0116] The gravity f g =ρg; where ρ represents the density of the rock mass; g represents the acceleration of gravity;

[0117] The pore water pressure of the rock and soil Among them, ρ w Indicates the density of pore water in rock and soil; It represents the gradient of the pore water head height of the rock mass. The negative sign indicates that the direction of the pore water pressure in the rock mass is opposite to the direction of the water head gradient and depends on the direction of water flow.

[0118] Step S302: Solve the stress and strain relationship of the rock and soil mass to obtain the displacement vector of the rock and soil mass.

[0119] In step S303, the stress and strain relationship of the rock and soil mass is solved to obtain the displacement vector of the rock and soil mass, which specifically includes the following steps:

[0120] Step S311, discretize the rock and soil into finite triangular units;

[0121] Step S312: For each triangular unit, use the shape function to approximate the deformation displacement to obtain the displacement interpolation relationship of each point in the triangular unit;

[0122] Preferably, the shape function is:

[0123]

[0124] Among them, u(x,y,z) represents the deformation displacement of the rock mass at the spatial coordinate (x,y,z), which is expressed by the shape function N i (x, y, z) is obtained by interpolating the node displacement;

[0125] u i is the displacement value of node i, n is the total number of nodes involved in the interpolation;

[0126] The displacement of any point in the unit is obtained by summing the product of the shape function of all nodes and the displacement of the corresponding nodes;

[0127] Step S313: Calculate the stiffness matrix based on the deformation and displacement of each triangular unit; the stiffness matrix is used to describe the relationship between the deformation and force of the rock mass when subjected to force;

[0128] The stiffness matrix is specifically represented by:

[0129] K=∫ V B T DBdV;

[0130] Where K represents the stiffness matrix; B represents the shape function derivative matrix; B T The transposed matrix representing the derivative moments of the shape function;

[0131] The shape function derivative matrix B is composed of the partial derivatives of the shape function with respect to the coordinates, and is used to represent the relationship between the stress and strain in the triangular element and the node displacement;

[0132] D represents the elasticity matrix;

[0133] Preferably, the elastic matrix is determined by elastic modulus and Poisson's ratio, and is used to convert strain into stress;

[0134] V represents the volume of the triangular unit;

[0135] Step S314: Calculate the displacement vector of the rock and soil mass according to the stiffness matrix.

[0136] The calculation formula of the rock and soil displacement vector is:

[0137] KU=f;

[0138] Where U represents the displacement vector, which is a column vector consisting of the displacement values of all nodes.

[0139] In step S400, based on deep learning and taking the relationship between rock and soil stress and deformation as the basis, a short-term recursive calculation model and a long-term trend analysis model are constructed respectively through the time series prediction method. Specifically,

[0140] Step S401: construct a short-term recursive calculation model by a time series prediction method, wherein the short-term recursive calculation model includes an input gate, a forget gate, an output gate, a cell state, and a hidden state part;

[0141] in,

[0142] The forget gate is characterized as: g t =σ(W g ·[h t-1 ,H t ]+b g );g tRepresents the output vector of the forget gate, whose element value range is between [0,1], and is used to determine the cell state C at the previous moment t-1 What information needs to be forgotten; σ represents the Sigmoid activation function, which maps the input value to the [0,1] interval; W g Represents the weight matrix of the forget gate; [h t-1 ,H t ] means to change the hidden state h of the previous moment t-1 and the current input H t Splicing; b g Represents the bias vector of the forget gate;

[0143] The input gate is characterized as follows: t =σ(W i ·[h t-1 ,H t ]+b I );I t Represents the output vector of the input gate, which is used to determine the current input H t How much new information is added to the cell state? I represents the weight matrix of the input gate; b I represents the bias vector of the input gate;

[0144] The output gate is characterized as follows: t =σ(W o ·[h t-1 ,H t ]+b o );

[0145] o t Represents the output vector of the output gate, which is used to determine the current cell state C t How much information should be output to the hidden state h t W o represents the weight matrix of the output gate; b o Represents the bias vector of the output gate;

[0146] The cell state is characterized by: C t =g t ⊙C t-1 +I t ⊙tanh(W c ·[h t-1 ,H t ]+b c );C t Represents the cell state at the current moment, used to store and transmit information; ⊙ represents the element-by-element multiplication operation; W c represents the weight matrix used to update the cell state; b cRepresents the bias vector used to update the cell state; tanh represents the hyperbolic tangent activation function, which maps the input value to the [0,1] interval;

[0147] The hidden state is characterized as: h t =o t ⊙tanh(C t );h t Represents the hidden state at the current moment, which is the output of the short-term recursive computation model and will be passed to the next time step. It can also be used as the final output result for prediction;

[0148] As a preference, an attention mechanism is added, specifically:

[0149] P1=W p1 H; P2 = W p2 H; P3 = W p3 H; P1 represents the query matrix, which is used to find information related to itself in the attention mechanism;

[0150] P2 represents the key matrix, which is used to match with the query matrix and calculate the correlation;

[0151] P3 represents the value matrix, which contains the information in the input sequence;

[0152] W p1 、W p2 and W p3 are the weight matrices used to generate the query matrix, key matrix, and value matrix respectively;

[0153] H represents the input time series data;

[0154]

[0155] Attention(P1, P2, P3) represents the output of the attention mechanism, which is the result of weighted summation of the input sequence;

[0156] Softmax represents a normalization function that obtains the attention weight of each position so that the sum of the weights is 1;

[0157] d K Represents the dimension of the key vector. The square root is taken to prevent the dot product result from being too large, thus avoiding the gradient vanishing or gradient exploding problem.

[0158] Step S402: Use the random forest method to construct multiple subsets by randomly sampling with replacement from the original training data. Perform secondary training on each subset using the method of step S401. The average value of the secondary training of all subsets is used as the final prediction result. A long-term trend analysis model is constructed based on the secondary training results over a certain period of time.

[0159] Preferably, the formula of the long-term trend analysis model is characterized as follows:

[0160]

[0161] in, represents the average value of the predicted deformation at time t+t', where t is the current time and t' is the time span;

[0162] l represents the number of subsets constructed in the random forest;

[0163] w j represents the weight of the j-th subset;

[0164] f j (H) is the prediction result of the jth subset for input H;

[0165] In step S500, real-time multi-source data values are obtained, and the displacement prediction vector of the geotechnical mass is obtained according to the short-term recursive calculation model. The displacement prediction vector is compared with the preset threshold value, and a real-time warning is issued when the threshold value is exceeded. The long-term trend analysis results are visualized, specifically:

[0166] Step S501: acquiring real-time multi-source data values, and obtaining a displacement prediction vector of the rock and soil mass according to the short-term recursive calculation model;

[0167] Step S502: Setting a displacement threshold value of the rock and soil mass, comparing it with the preset threshold value, and issuing a real-time warning when the threshold value is exceeded;

[0168] Step S503: Visualize the long-term trend analysis results.

[0169] Example,

[0170] The section between K12+300 and K12+700 on a certain highway is a deep-cut slope with a maximum excavation depth of 35m. The strata are highly weathered sandstone interbedded with mudstone, with well-developed joints, making shallow landslides prone to occur during the rainy season. The design utilizes a "prestressed anchor cable + lattice beam" support system, requiring real-time monitoring of slope deformation, stress, and environmental factors to ensure construction and operational safety.

[0171] Monitoring system deployment and data collection:

[0172] Sensor selection and arrangement:

[0173] Surface deformation monitoring:

[0174] Laser rangefinder (Leica TM50, accuracy ±0.3 mm): three measuring lines were arranged along the slope mouth, slope foot and platform, with monitoring points set at intervals of 20 m along each measuring line;

[0175] BeiDou GNSS receiver (Trimble R10, positioning accuracy ±2mm+1ppm): A base station and rover are set up at the top of the slope and in the surrounding stable areas to monitor three-dimensional displacement;

[0176] Subsurface stress monitoring:

[0177] Vibrating wire strain gauge (VWP-1000, accuracy ±0.01% FS): placed in the anchorage section of the anchor cable to monitor prestress loss;

[0178] Pore water pressure meter (Dynagage PT-500, range 0-5 MPa, accuracy 0.1% FS): drilled and buried to the sliding zone depth (15 m) to monitor groundwater pressure;

[0179] Environmental monitoring:

[0180] Rain gauge (tipping bucket type, accuracy ±0.5mm): placed on the top of the slope; temperature and humidity sensor (SHT30, accuracy ±0.5℃, ±2%RH);

[0181] Video monitoring (2 million pixels): real-time recording of slope crack development;

[0182] Construction period: high-frequency data collection (1 time / 10 minutes); operation period: low-frequency data collection (1 time / hour), which automatically switches to 1 time / 5 minutes during rainfall;

[0183] Finite element modeling and stress-deformation analysis:

[0184] Geometric model and parameters:

[0185] The slope is simplified into a two-dimensional plane strain model, with meshing of triangular elements (minimum side length 1m) and 5000+ nodes.

[0186] Geotechnical parameters: elastic modulus E = 50 MPa, Poisson's ratio ν = 0.3, density ρ = 2.2 g / cm 3 , cohesion c=20kPa, internal friction angle

[0187] Solving balanced equations:

[0188] Stress tensor (principal stress): σ1 = 0.8 MPa (compressive stress), σ3 = 0.3 MPa (compressive stress);

[0189] Volume force: gravity f g =22kN / m 3 , pore water pressure ( When fp=-100kPa);

[0190] Displacement calculation:

[0191] The shape function uses linear interpolation: u = N1u1 + N2u2 + N3u3;

[0192] Stiffness matrix K = ∫B T DBdV, the maximum displacement of the slope top is U = 12mm (the finite element calculation takes about 15 minutes, which is reduced to 5 minutes after parallel calculation optimization);

[0193] Short-Term Recurrent Model (LSTM):

[0194] Input characteristics: displacement rate, anchor stress, pore water pressure, rainfall (time window = 72h, step size = 1h);

[0195] Model structure: 2-layer LSTM (64 units per layer) + attention mechanism, training data is one year of construction data;

[0196] Long-term trend model (random forest):

[0197] Feature engineering: integrating 10 features including annual rainfall, seasonal temperature difference, and stress relaxation rate of support structures;

[0198] An artificial intelligence-based geotechnical engineering investigation safety monitoring system, such as Figure 2 As shown, the geotechnical engineering investigation safety monitoring system includes a data acquisition module, a data fusion module, a correlation analysis module, a deep learning module and an intelligent early warning and visualization module;

[0199] The data acquisition module is used to equip IoT devices to obtain deformation, stress and environmental data of the rock and soil mass and establish a model input data source;

[0200] The data fusion module is used to perform fusion processing on the collected multi-source data, including the processing steps of outlier replacement, time alignment and data standardization, matching data from different sources through timestamps, and using standardization methods to ensure data scale consistency;

[0201] The correlation analysis module is used to establish the correlation relationship between rock and soil stress and deformation based on the finite element method;

[0202] The deep learning module is used to build a short-term recursive calculation model and a long-term trend analysis model based on the time series prediction method, and introduces an attention mechanism to improve prediction accuracy to fully consider the nonlinear and time-varying characteristics of the rock and soil mass;

[0203] The intelligent early warning and visualization module is used to perform real-time early warning and visualization analysis based on the prediction results;

[0204] This module acquires real-time data, uses short-term prediction models to calculate geotechnical displacement prediction vectors, and compares these to preset thresholds, triggering an alert when these thresholds are exceeded. The module also provides a visual interface for long-term trend analysis, allowing project managers to intuitively view monitoring results, optimize decision-making, and improve the safety and stability of geotechnical projects.

[0205] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. An artificial intelligence-based geotechnical engineering investigation safety monitoring method, characterized by: The geotechnical engineering investigation safety monitoring method specifically comprises the following steps: Step S100: Acquire multi-source data, including deformation, stress and environmental data of rock and soil, and establish a model input data source; Step S200: performing data fusion processing on the acquired multi-source data; Step S300: using the result of data fusion processing as input, solving the stress and deformation of the continuous rock and soil mass by finite element method through data simulation to obtain the correlation relationship between the stress and deformation of the rock and soil mass; Step S400: Based on deep learning and taking the relationship between rock and soil stress and deformation as the basis, a short-term recursive calculation model and a long-term trend analysis model are constructed respectively through a time series prediction method; Step S500: Acquire real-time multi-source data values, obtain displacement prediction vectors of the geotechnical mass according to the short-term recursive calculation model, compare them with preset threshold values, issue real-time warnings when the threshold values are exceeded, and visualize long-term trend analysis results.

2. The artificial intelligence-based geotechnical engineering investigation safety monitoring method according to claim 1, characterized in that: In step S100, multi-source data is acquired, including deformation, stress and environmental data of the rock mass, and a model input data source is established, specifically: Equipped with IoT devices to conduct real-time monitoring, obtain deformation, stress and environmental data of rock and soil, and establish model input data source; Including surface monitoring, underground monitoring, vibration monitoring and environmental monitoring; The surface monitoring includes laser ranging and inclination measurement; The underground monitoring is used to monitor the pore water pressure of rock and soil; The vibration monitoring equipment includes a MEMS accelerometer and a seismic array; The environmental monitoring includes temperature, humidity, rainfall and remote sensing images.

3. The artificial intelligence-based geotechnical engineering investigation safety monitoring method according to claim 2, characterized in that: In step S200, data fusion processing is performed on the acquired multi-source data, specifically: The data fusion process specifically includes the following steps: Step S201: Replace outliers in the acquired multi-source data; Step S202: Set timestamps, and align multi-source data with different time and space based on the time scale between timestamps to achieve data fusion of multi-source data; Step S203: different data are kept at the same scale by a standardization transformation method. The specific formula is: X'=(X-μ) / a; wherein X' represents the data after standardization transformation; X represents the data after data fusion; μ represents the average value of data X; a represents the standard deviation of data X.

4. The artificial intelligence-based geotechnical engineering investigation safety monitoring method according to claim 3, characterized in that: In step S201, outlier replacement is performed on the acquired multi-source data, specifically: Step S211: sort the different types of multi-source data in descending order to obtain a descending sequence; Step S212: Calculate the upper quartile, median, and lower quartile of the data according to the descending sequence; the upper quartile, median, and lower quartile are the data corresponding to the upper quartile point Q1, the median point Q2, and the lower quartile point Q3 in the descending sequence, respectively; Among them, m represents the number of data in the descending sequence, and m is a positive integer; Q1 and Q2 are rounded up; Q3 is rounded down; Step S213: Calculate the difference between the upper quartile and the lower quartile, and take the absolute value of the difference as IQR; Step S214: using the upper quartile and lower quartile as upper and lower limits, respectively, and determining data greater than the upper quartile plus 1.5 times the IQR and data less than the lower quartile minus 1.5 times the IQR as outliers; Step S215: replace the outlier with the value of the upper quartile plus 1.5 times the IQR, and replace the outlier that is smaller than the lower quartile minus 1.5 times the IQR with the value of the lower quartile minus 1.5 times the IQR.

5. The artificial intelligence-based geotechnical engineering investigation safety monitoring method according to claim 4, characterized in that: In step S300, the result of data fusion processing is used as input, and the stress and deformation of the continuous rock and soil body are solved by finite element method through data simulation to obtain the correlation relationship between the stress and deformation of the rock and soil body, which is specifically: Step S301: construct an equilibrium equation for the rock-soil body according to the equilibrium relationship between the stress tensor of the rock-soil body and the external body force of the rock-soil body under static conditions. The equilibrium equation is represented as follows: Where δ represents the stress tensor; f represents the external body force; represents stress divergence; Step S302: Solve the stress and strain relationship of the rock and soil mass to obtain the displacement vector of the rock and soil mass.

6. The artificial intelligence-based geotechnical engineering investigation safety monitoring method according to claim 5, characterized in that: In step S303, the stress and strain relationship of the rock and soil mass is solved to obtain the displacement vector of the rock and soil mass, which specifically includes the following steps: Step S311, discretize the rock and soil into finite triangular units; Step S312: For each triangular unit, use the shape function to approximate the deformation displacement to obtain the displacement interpolation relationship of each point in the triangular unit; Step S313: Calculate the stiffness matrix based on the deformation and displacement of each triangular unit; the stiffness matrix is used to describe the relationship between the deformation and force of the rock mass when subjected to force; Step S314: Calculate the displacement vector of the rock and soil mass according to the stiffness matrix.

7. The artificial intelligence-based geotechnical engineering investigation safety monitoring method according to claim 6, characterized in that: In step S400, based on deep learning and taking the relationship between rock and soil stress and deformation as the basis, a short-term recursive calculation model and a long-term trend analysis model are constructed respectively through the time series prediction method. Specifically, Step S401: construct a short-term recursive calculation model by a time series prediction method, wherein the short-term recursive calculation model includes an input gate, a forget gate, an output gate, a cell state, and a hidden state part; Step S402: Use the random forest method to construct multiple subsets by randomly sampling with replacement from the original training data, perform secondary training on each subset using the method of step S401, take the average value of the secondary training of all subsets as the final prediction result, and construct a long-term trend analysis model based on the secondary training results over a certain period of time.

8. The artificial intelligence-based geotechnical engineering investigation safety monitoring method according to claim 7, characterized in that: In step S500, real-time multi-source data values are obtained, and the displacement prediction vector of the geotechnical mass is obtained according to the short-term recursive calculation model. The displacement prediction vector is compared with the preset threshold value, and a real-time warning is issued when the threshold value is exceeded. The long-term trend analysis results are visualized, specifically: Step S501: acquiring real-time multi-source data values, and obtaining a displacement prediction vector of the rock and soil mass according to the short-term recursive calculation model; Step S502: Setting a displacement threshold value of the rock and soil mass, comparing it with the preset threshold value, and issuing a real-time warning when the threshold value is exceeded; Step S503: Visualize the long-term trend analysis results.

9. An artificial intelligence-based geotechnical engineering investigation safety monitoring system, applying the artificial intelligence-based geotechnical engineering investigation safety monitoring method according to any one of claims 1 to 8, characterized in that: The geotechnical engineering investigation safety monitoring system includes a data acquisition module, a data fusion module, a correlation analysis module, a deep learning module and an intelligent early warning and visualization module; The data acquisition module is used to equip IoT devices to obtain deformation, stress and environmental data of the rock and soil mass and establish a model input data source; The data fusion module is used to perform fusion processing on the collected multi-source data, including the processing steps of outlier replacement, time alignment and data standardization, matching data from different sources through timestamps, and using standardization methods to ensure data scale consistency; The correlation analysis module is used to establish the correlation relationship between rock and soil stress and deformation based on the finite element method; The deep learning module is used to build a short-term recursive calculation model and a long-term trend analysis model based on the time series prediction method, and introduces an attention mechanism to improve prediction accuracy to fully consider the nonlinear and time-varying characteristics of the rock and soil mass; The intelligent early warning and visualization module is used to perform real-time early warning and visualization analysis based on the prediction results.

Images