Hydropower station slope three-dimensional deformation early warning method and system based on measurement robot

Abstract

This invention provides a three-dimensional deformation early warning method and system for hydropower station slopes based on a measurement robot, relating to the field of hydropower station safety monitoring technology. First, a measurement robot is deployed to continuously observe the hydropower station slope, thereby obtaining detailed three-dimensional observation data of each observation point on the slope surface. Then, based on the relevant information of each observation point in the three-dimensional observation data, a deformation transmission link is constructed between all observation points of the hydropower station slope. The cross-regional deformation transmission characteristics of the hydropower station slope are explored, and the deformation transmission sequence and associated state data of observation points in different observation areas are obtained. The cross-regional deformation transmission characteristics and the spatial position change trajectory of each observation point are integrated to construct a three-dimensional deformation evolution path. Finally, based on the three-dimensional deformation evolution path, a three-dimensional deformation early warning information containing a description of the deformation evolution trend and an early warning trigger identifier is generated and promptly sent to the hydropower station safety monitoring terminal, realizing real-time early warning of hydropower station slope deformation.

Description

Technical Field

[0001] This invention relates to the field of hydropower station safety monitoring technology, and more specifically, to a method and system for early warning of three-dimensional deformation of hydropower station slopes based on a measurement robot. Background Technology

[0002] In the construction and operation of hydropower stations, slope stability is crucial, and its three-dimensional deformation directly affects the safe operation of the hydropower station. Traditional methods for monitoring slope deformation in hydropower stations mostly rely on manual periodic measurements or single-sensor monitoring. Manual periodic measurements suffer from long measurement intervals and untimely data acquisition, making it difficult to capture the dynamic deformation process of the slope in real time and failing to provide timely warnings for sudden deformation. Single-sensor monitoring methods, such as using traditional displacement sensors, typically only acquire deformation information at specific points on the slope, failing to comprehensively reflect the overall deformation status of the slope and making it difficult to discover the transmission relationship of deformation between different areas of the slope. Furthermore, existing monitoring methods often only perform simple analysis of data from a single monitoring point in terms of data processing and analysis, without fully considering the transmission and correlation of deformation between various observation points on the slope, making it impossible to accurately predict the evolution trend of slope deformation, leading to problems such as deformation at different points on the slope. Summary of the Invention

[0003] In view of the aforementioned problems, and in conjunction with the first aspect of the present invention, embodiments of the present invention provide a method for early warning of three-dimensional deformation of hydropower station slopes based on a measurement robot, the method comprising: A measurement robot is deployed to continuously observe the slope of the hydropower station and obtain three-dimensional observation data of the slope. The three-dimensional observation data includes the spatial location information and observation time information of each observation point on the slope surface. Based on the spatial location information and observation time information of each observation point in the three-dimensional observation data, a deformation transmission link is established between all observation points of the hydropower station slope. The deformation transmission link includes the deformation transmission direction and transmission association information between the observation points. Based on the deformation transmission link, the cross-regional deformation transmission characteristics of the hydropower station slope are extracted. The cross-regional deformation transmission characteristics include the deformation transmission sequence and transmission association status data of observation points in different observation areas. By integrating the cross-regional deformation transmission characteristics and the spatial position change trajectory of each observation point in the three-dimensional observation data, a three-dimensional deformation evolution path of the hydropower station slope is constructed. The three-dimensional deformation evolution path includes the continuous transmission trend of spatial position change of the observation points and the deformation diffusion path between regions. Based on the three-dimensional deformation evolution path, three-dimensional deformation early warning information of the hydropower station slope is generated. The three-dimensional deformation early warning information includes a description of the deformation evolution trend and an early warning trigger identifier. The three-dimensional deformation early warning information is sent to the hydropower station safety monitoring terminal.

[0004] Furthermore, embodiments of the present invention also provide a three-dimensional deformation early warning system for hydropower station slopes based on a measurement robot, comprising: A processor; a machine-readable storage medium for storing machine-executable instructions of the processor; wherein the processor is configured to execute the above-described method for early warning of three-dimensional deformation of hydropower station slopes based on a measurement robot by executing the machine-executable instructions.

[0005] Based on the above, by deploying measurement robots to continuously observe the slope of the hydropower station, three-dimensional observation data containing spatial location information and observation time information of each observation point on the slope surface is obtained. Based on the three-dimensional observation data, deformation transmission links between all observation points are established, clarifying the direction and correlation information of deformation transmission between observation points, mining cross-regional deformation transmission characteristics, and revealing the deformation transmission sequence and correlation status of observation points in different observation areas. This helps to grasp the spatial transmission law of slope deformation. By integrating cross-regional deformation transmission characteristics and the spatial location change trajectory of each observation point, a three-dimensional deformation evolution path is constructed. This path can intuitively present the continuous transmission trend of spatial location changes of observation points and the deformation diffusion path between regions, accurately predict the evolution trend of slope deformation, and generate three-dimensional deformation early warning information containing deformation evolution trend description and early warning triggering identifier based on the three-dimensional deformation evolution path. This information is then promptly sent to the hydropower station safety monitoring terminal, realizing real-time early warning of hydropower station slope deformation and effectively improving the safety monitoring level of hydropower station slopes. Attached Figure Description

[0006] Figure 1 This is a schematic diagram of the execution flow of the three-dimensional deformation early warning method for hydropower station slopes based on a measurement robot provided in an embodiment of the present invention.

[0007] Figure 2 This is a schematic diagram of exemplary hardware and software components of a hydropower station slope three-dimensional deformation early warning system based on a measurement robot, provided in an embodiment of the present invention. Detailed Implementation

[0008] Figure 1 This is a flowchart illustrating a method for early warning of three-dimensional deformation of hydropower station slopes based on a measurement robot, provided in one embodiment of the present invention. A detailed description follows.

[0009] Step S110: Deploy a measurement robot to continuously observe the slope of the hydropower station and obtain three-dimensional observation data of the slope. The three-dimensional observation data includes the spatial location information and observation time information of each observation point on the slope surface.

[0010] In this embodiment, firstly, based on the topographic features of the hydropower station slope, N surveying robots are deployed on the slope surface in a grid layout with a gradient difference of H. The observation coverage of each surveying robot is adjusted according to the slope gradient to ensure no blind spots. All surveying robots achieve time synchronization through a satellite timing system, with an observation sampling interval set to T and a continuous observation duration of no less than 72 hours. The spatial location information of the observation points collected by the surveying robots is stored in the form of coordinate data in a three-dimensional rectangular coordinate system (O-XYZ), where the X-axis is along the transverse direction of the slope, the Y-axis is along the longitudinal direction of the slope, and the Z-axis is along the vertical direction, with the coordinate data accuracy retained to 6 decimal places. The observation time information is stored in the form of UTC timestamps and is bound one-to-one with the corresponding spatial location information. To protect sensitive data such as the slope's geographical location, differential privacy technology is used to add noise perturbation to the coordinate data, with the perturbation amplitude controlled within the coordinate accuracy range. At the same time, the TLS 1.3 encryption protocol is used to achieve end-to-end data transmission between the surveying robots and the data processing center to prevent data leakage. The data processing center collects all uploaded data in the order of observation time, and finally generates a three-dimensional observation data set containing M observation points and K observation time points for each observation point. The set is stored in CSV format, and each row of data contains five types of fields: observation point ID, UTC timestamp, X coordinate, Y coordinate, and Z coordinate.

[0011] Step S120: Based on the spatial location information and observation time information of each observation point in the three-dimensional observation data, establish a deformation transmission link between all observation points on the hydropower station slope. The deformation transmission link includes the deformation transmission direction and transmission association information between the observation points.

[0012] Step S121: Extract the spatial location information of each observation point from the three-dimensional observation data, calculate the actual spatial distance between any two observation points based on the spatial location information, and generate a set of spatial distances between observation points. The set of spatial distances between observation points contains the actual spatial distance data between all pairs of observation points.

[0013] In this embodiment, firstly, all spatial coordinates of each observation point are extracted from the 3D observation data. The mean coordinate of each observation point over a continuous observation period is calculated using the sliding window method, which serves as the reference spatial coordinate P_i = (X_i_avg, Y_i_avg, Z_i_avg) for that observation point, where i is the observation point ID. For any two observation points i and j, the actual spatial distance D_ij is obtained using the 3D Euclidean distance calculation logic: First, the three coordinate differences (X_i_avg - X_j_avg), (Y_i_avg - Y_j_avg), and (Z_i_avg - Z_j_avg) are calculated respectively. Then, each difference is squared, and the three squared results are added together to obtain the sum of squares. Finally, the square root of the sum of squares is taken to obtain the value of D_ij, and the result is retained to three decimal places. Calculate the actual spatial distances between all M observation points in sequence using the method described above, generating a total of M×(M-1) / 2 sets of distance data. Store these data as a unique identifier for each observation point pair (i, j), creating a set of spatial distances between observation points. This set of spatial distances is stored in JSON format, with each data entry containing two fields: "point_pair" and "distance". The "point_pair" field stores the ID combination of the observation point pair, and the "distance" field stores the corresponding D_ij value.

[0014] Step S122: Combining the observation time information in the three-dimensional observation data, extract the spatial position change sequence of each observation point in chronological order within a continuous observation period. The spatial position change sequence includes the spatial position offset data of each observation point at different observation time points.

[0015] In this embodiment, the spatial coordinates of each observation point are first sorted in ascending order of UTC timestamps to form a spatial position time series for that observation point. Using the reference spatial coordinates P_i=(X_i_avg, Y_i_avg, Z_i_avg) for each observation point, the offsets of the observation point's coordinates from the reference coordinates at each observation time point are calculated sequentially: horizontal offset ΔX_ik=X_ik-X_i_avg, vertical offset ΔY_ik=Y_ik-Y_i_avg, and vertical offset ΔZ_ik=Z_ik-Z_i_avg, where k is the sequence number of the observation time point. The K sets of offset data for each observation point are arranged in timestamp order to generate a spatial position change sequence for that observation point. Each data entry in the sequence contains four fields: observation timestamp, ΔX_ik, ΔY_ik, and ΔZ_ik. All spatial position change sequences for all observation points are stored in a distributed database, with each sequence corresponding to a unique observation point ID index for easy and quick subsequent retrieval.

[0016] Step S123: Based on the spatial position change sequence of each observation point, extract the deformation correlation factor of the observation point. The deformation correlation factor includes the consistent state of the spatial position change direction and the coordinated state of the change rate of the observation point. It is generated by comparing the spatial position change direction and change rate of different observation points in the same observation period.

[0017] Step S1231: Divide the observation period into equal time intervals from the spatial position change sequence of each observation point. Each observation period contains a preset number of consecutive observation time points, and generate a segmented spatial position change sequence for each observation point.

[0018] In this embodiment, the entire observation period is first divided into L equal-time observation periods based on the total duration of continuous observation and a preset rule for dividing observation periods. Each observation period contains K_L consecutive observation time points, and the value of K_L must ensure that the observation data within each period reflects the deformation trend of that period. For the spatial position change sequence of each observation point, all offset data and corresponding observation timestamps within each of the divided observation periods are extracted sequentially to generate a segmented spatial position change sequence for that observation point. Each segmented sequence corresponds to an observation period and includes K_L sets of offset data within that period, as well as the start and end timestamps of the period. All segmented sequences are indexed and stored according to the observation point ID and period sequence number, facilitating subsequent batch processing of data by period.

[0019] Step S1232: For each segmented spatial position change sequence, calculate the spatial position change direction vector of the observation point within the observation period. The spatial position change direction vector is generated by connecting the spatial position coordinates of the start and end observation time points of the observation period.

[0020] Step S12321: Extract the spatial coordinates corresponding to the starting observation time point of each observation period from the segmented spatial position change sequence of each observation point, and record them as starting coordinate data. The starting coordinate data includes the horizontal coordinate, vertical coordinate and vertical coordinate in three-dimensional space.

[0021] In this embodiment, for each segmented spatial position change sequence of an observation point, the starting observation time point of that observation period is first located, and the corresponding spatial position coordinates S_i = (X_i_start, Y_i_start, Z_i_start) are extracted, where X_i_start is the horizontal coordinate of the starting time, Y_i_start is the vertical coordinate of the starting time, and Z_i_start is the vertical coordinate of the starting time. The starting coordinate data is stored in the form of triplets and associated with the observation point ID and the time period number to ensure that the corresponding observation time period and observation point can be accurately matched during subsequent calculations, avoiding data confusion.

[0022] Step S12322: Extract the spatial coordinates corresponding to the end observation time point of the observation period and record them as end coordinate data. The end coordinate data includes the horizontal coordinate, vertical coordinate and the vertical coordinate in three-dimensional space.

[0023] In this embodiment, for each observation point's segmented spatial position change sequence, the end observation time point of that observation period is located, and the corresponding spatial position coordinates E_i = (X_i_end, Y_i_end, Z_i_end) are extracted, where X_i_end is the horizontal coordinate of the end time, Y_i_end is the vertical coordinate of the end time, and Z_i_end is the vertical coordinate of the end time. The end coordinate data is also stored in the form of triplets, bound to the corresponding start coordinate data, observation point ID, and period sequence number, ensuring that each set of start and end coordinates corresponds to the same observation period of the same observation point.

[0024] Step S12323: Calculate the difference between the starting coordinate data and the ending coordinate data in the horizontal coordinate direction. The horizontal coordinate difference is generated by subtracting the horizontal coordinate value of the starting coordinate data from the horizontal coordinate value of the ending coordinate data.

[0025] In this embodiment, for the same observation period at each observation point, the lateral coordinate difference ΔX_dir = X_i_end - X_i_start is calculated, where X_i_end is the lateral coordinate at the end time and X_i_start is the lateral coordinate at the start time. The sign of the lateral coordinate difference represents the lateral offset direction of the observation point within that period; a positive value indicates offset along the positive X-axis, and a negative value indicates offset along the negative X-axis. The absolute value of the difference represents the magnitude of the lateral offset.

[0026] Step S12324: Calculate the difference between the starting coordinate data and the ending coordinate data in the vertical direction. The vertical coordinate difference is generated by subtracting the vertical coordinate value of the starting coordinate data from the vertical coordinate value of the ending coordinate data.

[0027] In this embodiment, for the same observation period at each observation point, the longitudinal coordinate difference ΔY_dir = Yi_i_end - Yi_i_start is calculated, where Yi_i_end is the longitudinal coordinate at the end time and Yi_i_start is the longitudinal coordinate at the start time. The sign of the longitudinal coordinate difference represents the longitudinal offset direction of the observation point within that period; a positive value indicates offset along the positive Y-axis, and a negative value indicates offset along the negative Y-axis. The absolute value of the difference represents the magnitude of the longitudinal offset.

[0028] Step S12325: Calculate the vertical coordinate difference between the starting coordinate data and the ending coordinate data. The vertical coordinate difference is generated by subtracting the vertical coordinate value of the starting coordinate data from the vertical coordinate value of the ending coordinate data.

[0029] In this embodiment, for the same observation period at each observation point, the vertical coordinate difference ΔZ_dir = Z_i_end - Z_i_start is calculated, where Z_i_end is the vertical coordinate at the end time and Z_i_start is the vertical coordinate at the start time. The sign of the vertical coordinate difference represents the vertical offset direction of the observation point within that period; a positive value indicates an offset along the positive Z-axis (upward), and a negative value indicates an offset along the negative Z-axis (downward). The absolute value of the difference represents the magnitude of the vertical offset.

[0030] Step S12326: Based on the horizontal coordinate difference, vertical coordinate difference and vertical coordinate difference, construct the original data of the spatial position change vector. The original data contains coordinate difference information in the three coordinate directions.

[0031] In this embodiment, the calculated horizontal coordinate difference ΔX_dir, vertical coordinate difference ΔY_dir, and vertical coordinate difference ΔZ_dir are combined in sequence to form a three-dimensional vector form of raw data V_dir_raw=(ΔX_dir, ΔY_dir, ΔZ_dir). This raw data directly reflects the offset of the observation point in the three coordinate directions during the observation period, with the value of each dimension corresponding to the magnitude and direction of the offset of the observation point in that direction.

[0032] Step S12327: Calculate the vector magnitude of the original data, which is generated by integrating the square root of the sum of the squares of the horizontal coordinate difference, the vertical coordinate difference, and the vertical coordinate difference, reflecting the total offset of the spatial position change.

[0033] In this embodiment, the squares of the horizontal coordinate difference ΔX_dir, the vertical coordinate difference ΔY_dir, and the vertical coordinate difference ΔZ_dir are first calculated to obtain (ΔX_dir)², (ΔY_dir)², and (ΔZ_dir)². These three squares are then added together to obtain the sum of squares S_dir = (ΔX_dir)² + (ΔY_dir)² + (ΔZ_dir)². Finally, the square root of the sum of squares S_dir is performed to obtain the vector magnitude L_dir = √S_dir (where √ represents the square root operation). The value of the vector magnitude L_dir reflects the total offset of the spatial position change of the observation point during the observation period; a larger magnitude indicates a larger total offset during that period.

[0034] Step S12328: Divide the difference in horizontal coordinates, the difference in vertical coordinates, and the difference in vertical coordinates in the original data by the vector magnitude to obtain the standardized horizontal component, vertical component, and vertical component, so that the vector magnitude is unified to a preset value.

[0035] In this embodiment, the horizontal coordinate difference ΔX_dir in the original data V_dir_raw is divided by the vector magnitude L_dir to obtain the standardized horizontal component X_dir_norm = ΔX_dir / L_dir; the vertical coordinate difference ΔY_dir is divided by the vector magnitude L_dir to obtain the standardized vertical component Y_dir_norm = ΔY_dir / L_dir; and the vertical coordinate difference ΔZ_dir is divided by the vector magnitude L_dir to obtain the standardized vertical component Z_dir_norm = ΔZ_dir / L_dir. After standardization, the sum of the squares of the three components is 1, i.e., (X_dir_norm)² + (Y_dir_norm)² + (Z_dir_norm)² = 1, making the vector magnitude uniform to 1, eliminating the difference in total offset amplitude between different observation points and different time periods, and retaining only the offset direction information. The standardized components are bound and stored with the corresponding original vector, vector magnitude, observation point ID, and time period number to ensure data traceability.

[0036] Step S12329: Based on the standardized horizontal, vertical and longitudinal components, construct the spatial position change direction vector of the observation point during the observation period. The direction vector contains three standardized component data.

[0037] In this embodiment, the standardized horizontal component X_dir_norm, vertical component Y_dir_norm, and vertical component Z_dir_norm are combined sequentially to form a spatial position change direction vector V_dir_norm = (X_dir_norm, Y_dir_norm, Z_dir_norm). This direction vector is a unit vector, reflecting only the direction of spatial position change of the observation point within the observation period, and does not include offset magnitude information. The direction vector is bound and stored with the corresponding observation point ID, time period number, original vector, and vector magnitude, and each observation point corresponds to a unique direction vector for each observation period.

[0038] Step S12330: Perform vector verification on the constructed spatial position change direction vector. By confirming that the sum of squares of the three standardized components meets the preset numerical requirements, the final spatial position change direction vector is generated.

[0039] In this embodiment, for the constructed direction vector V_dir_norm=(X_dir_norm, Y_dir_norm, Z_dir_norm), the sum of squares of the three normalized components S_norm=(X_dir_norm)²+(Y_dir_norm)²+(Z_dir_norm)² is calculated. The calculation result is compared with a preset value of 1, with an allowable error range of ±1×10^-6. If the sum of squares is within the error range, the direction vector is deemed valid, and the vector is stored as the final spatial position change direction vector. If the sum of squares exceeds the error range, a calculation error is determined in the construction process, and all steps from coordinate difference calculation to vector construction are re-executed until a direction vector that meets the requirements is obtained. Verified direction vectors are uniformly stored in a vector database, indexed by observation point ID and time period number for easy and rapid subsequent retrieval.

[0040] Step S12331: Extract the spatial position change direction vectors of all observation points within the same observation period, and construct a direction vector set, which contains the spatial position change direction data of all observation points within the observation period.

[0041] In this embodiment, for each observation period, the final spatial position change direction vectors of all observation points within that period are extracted from the vector database. These vectors are then sorted by observation point ID to construct a direction vector set, Set_V_norm. Each data entry in the set contains the observation point ID, the period number, and the corresponding direction vector V_dir_norm=(X_dir_norm, Y_dir_norm, Z_dir_norm). The set is stored in key-value pairs, where the key is the observation point ID and the value is the corresponding direction vector. Each observation period corresponds to an independent set of direction vectors, ensuring that direction data from different periods are not mixed up.

[0042] Step S12332: Calculate the angle between any two direction vectors in the set of direction vectors to generate direction angle data. Use the direction angle data to determine whether the spatial position changes of the two observation points are in the same direction during the observation period.

[0043] In this embodiment, for the set of direction vectors Set_V_norm during the same observation period, the direction vectors V_i_norm=(X_i_norm, Y_i_norm, Z_i_norm) and V_j_norm=(X_j_norm, Y_j_norm, Z_j_norm) of any two observation points i and j are selected. The angle θ_ij between the two vectors is calculated: first, the dot product of the two vectors Dot_ij=X_i_norm×X_j_norm+Y_i_norm×Y_j_norm+Z_i_norm×Z_j_norm is calculated, and then the inverse cosine operation is performed on Dot_ij to obtain θ_ij=arccos(Dot_ij). The value of θ_ij ranges from 0 to π (radians). The smaller the direction angle θ_ij, the more consistent the direction of change of the spatial position of the two observation points; the larger the θ_ij, the greater the difference in direction.

[0044] Step S12333: Based on the direction angle data, generate a spatial position change direction consistency state parameter. When the direction angle data is within the preset direction angle range, generate the spatial position change direction consistency state parameter according to the preset direction consistency parameter mapping rule.

[0045] In this embodiment, the preset directional angle interval is [0, α], where α is a preset angle threshold, ranging from 0 to π / 6 radians. For any two observation points, the directional angle θ_ij is determined to be consistent if θ_ij ≤ α, generating a consistent direction state parameter C_dir_ij = 1; if α < θ_ij ≤ π / 2, the directions are basically consistent, generating a consistent direction state parameter C_dir_ij = 0.5; if θ_ij > π / 2, the directions are inconsistent, generating a consistent direction state parameter C_dir_ij = 0. The consistent direction state parameter C_dir_ij ranges from 0 to 1, with a larger value indicating higher directional consistency. After the parameter is generated, it is bound and stored with the corresponding observation point pair (i, j), time period number, and directional angle θ_ij to form a consistent direction parameter dataset.

[0046] Step S12334: For each segmented spatial position change sequence, calculate the rate of change of the spatial position of the observation point within the observation period. The rate of change of the spatial position is generated by the ratio of the total spatial position offset within the observation period to the duration of the observation period.

[0047] In this embodiment, for each observation point's segmented spatial position change sequence, the duration T_segment is first calculated based on the start and end timestamps of the observation period, where T_segment = timestamp_end - timestamp_start, in seconds. Then, combined with the vector magnitude L_dir (i.e., the total spatial offset) within that period, the spatial position change rate V_rate = L_dir / T_segment is calculated, in meters per second (m / s). The value of the change rate V_rate reflects the rate of deformation of the observation point within that period; a larger value indicates a faster deformation rate.

[0048] Step S12335: Extract the rate of change of spatial position of all observation points within the same observation period, and construct a rate set, which contains the rate of change of spatial position of all observation points within the observation period.

[0049] In this embodiment, for each observation period, the spatial location change rate V_rate of all observation points within that period is extracted from the rate database and sorted by observation point ID to construct a rate set Set_V_rate. Each data entry in the set contains the observation point ID, the period number, and the corresponding change rate V_rate. The set is stored in key-value pair format, where the key is the observation point ID and the value is the corresponding change rate. Each observation period corresponds to an independent rate set, ensuring that rate data from different periods are not mixed up.

[0050] Step S12336: Calculate the ratio of any two rate data in the rate set to generate rate ratio data, and determine the coordinated state of the spatial position change rate of the two observation points during the observation period through the rate ratio data.

[0051] In this embodiment, for the rate set Set_V_rate within the same observation period, the change rates V_rate_i and V_rate_j of any two observation points i and j are selected, and the rate ratio R_rate_ij = min(V_rate_i, V_rate_j) / max(V_rate_i, V_rate_j) is calculated. The value of R_rate_ij ranges from 0 to 1. The closer the rate ratio R_rate_ij is to 1, the more coordinated the spatial position change rates of the two observation points are; the closer R_rate_ij is to 0, the greater the rate difference.

[0052] Step S12337: Based on the rate ratio data, generate the spatial position change rate coordination state parameter. When the difference between the rate ratio data and the preset coordination ratio is within the preset rate ratio difference range, generate the spatial position change rate coordination state parameter according to the preset rate coordination parameter mapping rule.

[0053] In this embodiment, the preset coordination ratio is β, ranging from 0.8 to 1, and the preset rate ratio difference interval is [-β, β]. For any two observation points, if |R_rate_ij-β|≤0.1, the spatial position change rates of the two observation points are determined to be coordinated, and a rate coordination state parameter C_rate_ij=1 is generated; if 0.1<|R_rate_ij-β|≤0.3, the rates are determined to be basically coordinated, and a rate coordination state parameter C_rate_ij=0.5 is generated; if |R_rate_ij-β|>0.3, the rates are determined to be uncoordinated, and a rate coordination state parameter C_rate_ij=0 is generated. The rate coordination state parameter C_rate_ij ranges from 0 to 1, with a larger value indicating higher rate coordination. After the parameter is generated, it is bound and stored with the corresponding observation point pair (i, j), time period number, and rate ratio R_rate_ij to form a rate coordination parameter dataset.

[0054] Step S12338: Take the consistent spatial position change direction state parameter and the coordinated spatial position change rate state parameter of any two observation points within the same observation period as independent dimensions of the deformation correlation factor of the observation point pair, and generate the deformation correlation factor of the observation point. The deformation correlation factor includes two independent parameter groups: the consistent spatial position change direction state and the coordinated change rate state of the observation point.

[0055] In this embodiment, for any two observation points i and j within the same observation period, the generated direction consistency state parameter C_dir_ij and rate coordination state parameter C_rate_ij are sequentially combined to form the deformation correlation factor F_ij=(C_dir_ij, C_rate_ij) for the observation point pair (i, j). The two dimensions of the deformation correlation factor reflect the state of the observation point pair in terms of direction consistency and rate coordination, respectively, providing two core quantitative parameters for subsequent calculation of deformation transmission intensity. The deformation correlation factors of all observation point pairs are grouped and stored according to the observation period. Each period corresponds to a set of correlation factors, and each data in the set of correlation factors contains the observation point pair (i, j), the period number, and the corresponding deformation correlation factor F_ij.

[0056] Step S124: Compare the actual spatial distance data in the observation point spatial distance set with the preset deformation correlation distance threshold, filter out observation point pairs whose actual spatial distance data is within the preset deformation correlation distance threshold range, and generate a potential deformation correlation observation point pair set.

[0057] In this embodiment, a preset deformation correlation distance threshold D_thresh is first set based on the rock and soil characteristics and historical deformation data of the hydropower station slope. The value of this preset deformation correlation distance threshold is determined based on parameters such as the thickness of the loose layer and the degree of rock fissure development of the slope, ensuring that only observation point pairs that are spatially sufficiently close and have a physical basis for deformation transmission are included in the screening range. The actual spatial distance D_ij of each observation point pair is extracted one by one from the set of spatial distances of observation points, and D_ij is compared with D_thresh: if D_ij≤D_thresh, it is determined that the observation point pair has a potential deformation transmission correlation and is added to the screening list; if D_ij>D_thresh, it is determined that the spatial distance of the observation point pair is too large and there is no possibility of direct deformation transmission, and it is excluded. After completing the comparison of all observation point pairs, the observation point pairs in the screening list are deduplicated and sorted according to the observation point ID combination to generate a set of potential deformation-correlated observation point pairs. The set of potential deformation-related observation point pairs is stored in JSON format. Each data point contains the combination of observation point pair IDs, the actual spatial distance D_ij, and the corresponding observation time period information, ensuring that cross-time period analysis can be performed in conjunction with deformation-related factors.

[0058] Step S125: For each observation point pair in the potential deformation-related observation point pair set, calculate the deformation transfer intensity between the observation point pairs in combination with the corresponding deformation-related factor. The deformation transfer intensity is generated by weighting the spatial position change direction consistent state parameter and the change rate coordinated state parameter with preset weight coefficients, and then mapping them to a unified dimensionless evaluation scale.

[0059] In this embodiment, firstly, based on the deformation transmission mechanism of the hydropower station slope, a weighting coefficient W_dir for the directional consistency state parameter and a weighting coefficient W_rate for the rate coordination state parameter are set, satisfying W_dir + W_rate = 1. The value of W_dir can be adjusted according to the dominant deformation direction of the slope; if the slope is dominated by lateral deformation, then the value of W_dir is greater than W_rate. For each observation point pair (i, j) in the potential deformation-related observation point pair set, the directional consistency state parameter C_dir_ij and the rate coordination state parameter C_rate_ij for the corresponding observation period are extracted. These are then multiplied by their respective weighting coefficients to obtain the weighted directional parameter W_dir × C_dir_ij and the weighted rate parameter W_rate × C_rate_ij. The two weighted parameters are then added to obtain the initial deformation transmission intensity S_ij_raw = W_dir × C_dir_ij + W_rate × C_rate_ij, where the value of S_ij_raw ranges from 0 to 1. To further enhance the discriminative power of deformation transfer intensity, S_ij_raw is mapped to a dimensionless evaluation scale of 0 to 10 using a normalized mapping function. The final deformation transfer intensity S_ij = 10 × S_ij_raw, with the result rounded to one decimal place. The calculated S_ij is then bound and stored with the corresponding observation point pair (i, j), time period number, C_dir_ij, C_rate_ij, W_dir, and W_rate.

[0060] Step S126: Based on the preset deformation initiation end determination rule, the observation point whose deformation transmission intensity value satisfies the deformation initiation end determination rule is taken as the deformation initiation end, and the other observation point is taken as the deformation receiving end, thus generating a one-way transmission relationship between the observation point pairs.

[0061] In this embodiment, the preset deformation initiator determination rule is as follows: within the same observation period, if the deformation transmission intensity S_ij of observation point i relative to observation point j is greater than the deformation transmission intensity S_ji of observation point j relative to observation point i, and the difference between S_ij and S_ji is greater than the preset intensity difference threshold ΔS_thresh, then observation point i is determined to be the deformation initiator and observation point j is the deformation receiver; if the difference between S_ij and S_ji is less than or equal to ΔS_thresh, then the deformation transmission of the observation point pair is determined to be a bidirectional interactive state, and no explicit initiator and receiver are set for the time being. First, for each unordered observation point pair (i, j) in the potential deformation-related observation point pair set, the corresponding deformation transmission intensity S_ij and S_ji are extracted, and the intensity difference ΔS_ij = S_ij - S_ji is calculated. The ΔS_ij and ΔS_thresh are compared: if ΔS_ij > ΔS_thresh, a one-way transfer relationship from i to j is generated and recorded as (i→j); if ΔS_ij < -ΔS_thresh, a one-way transfer relationship from j to i is generated and recorded as (j→i); if |ΔS_ij| ≤ ΔS_thresh, the observation point pair is marked as a two-way interactive state, which will be further determined by combining historical deformation data. All generated one-way transfer relationships are grouped and stored according to the observation period. Each relationship includes the initiator ID, receiver ID, and deformation transfer intensity S_ij or S_ji.

[0062] Step S127: Based on the unidirectional transmission relationship of all potential deformation-related observation point pairs, construct the initial deformation transmission link of the hydropower station slope. The initial deformation transmission link includes the transmission direction and corresponding deformation transmission intensity data of all observation point pairs.

[0063] In this embodiment, an initial deformation transfer link is first constructed using a graph structure model. Each observation point is treated as a node in the graph, with node attributes including observation point ID and reference spatial coordinates. The generated unidirectional transfer relationships are treated as directed edges in the graph, with edge attributes including initiator ID, receiver ID, deformation transfer intensity S_ij, and observation period number. For each observation period, all nodes and directed edges within that period are imported into the graph structure model to generate the initial deformation transfer sub-link for that period. Then, the sub-links of all observation periods are associated in chronological order to form an initial deformation transfer link covering the entire observation period. The initial deformation transfer link is stored in the form of an adjacency list, with each node corresponding to a list containing information on all directed edges originating from that node. The information for each edge includes receiver ID, deformation transfer intensity S_ij, and the corresponding observation period. Simultaneously, for convenient visualization, the link data is converted to GraphML format, including a set of nodes, a set of edges, and corresponding attribute data, ensuring that the deformation transfer relationship can be intuitively presented using graph visualization tools.

[0064] Step S128: Optimize the initial deformation transfer link. Remove the connections of the observation point pairs whose deformation transfer intensity values are outside the preset transfer intensity threshold range, and supplement the connections of the observation point pairs that are omitted due to spatial distance screening but whose deformation correlation factors meet the requirements of the preset correlation factors.

[0065] In this embodiment, first set the preset transfer intensity threshold range [S_low, S_high], where S_low is the lowest effective transfer intensity and S_high is the highest effective transfer intensity. The threshold values are determined based on the statistical data of the deformation transfer intensity in historical deformation events. For each directed edge in the initial deformation transfer link, extract the corresponding deformation transfer intensity S_ij, and compare S_ij with the threshold range: If S_ij < S_low, it is determined that the deformation transfer intensity corresponding to this edge is too low and has no actual deformation conduction significance, and it is removed from the link; if S_ij > S_high, it is determined that the deformation transfer intensity corresponding to this edge is too high and there may be data anomalies, and it needs to be rechecked in combination with the original observation data. If the data is confirmed to be correct after rechecking, this edge is retained, otherwise it is removed. After completing the removal operation, start the supplement operation: Traverse all the observation point pairs that have not been included in the set of potential deformation correlation observation point pairs, and extract their deformation correlation factors F_ij = (C_dir_ij, C_rate_ij). If C_dir_ij ≥ 0.8 and C_rate_ij ≥ 0.8, that is, the directions are highly consistent and the rates are highly coordinated, it is determined that although the spatial distance of this observation point pair exceeds the threshold, there may be an indirect deformation transfer correlation, and it is added to the supplement list. For each observation point pair in the supplement list, recalculate the deformation transfer intensity S_ij, generate a one-way transfer relationship, and then add it to the deformation transfer link. Finally, an optimized deformation transfer link is obtained. This deformation transfer link not only retains the effective deformation transfer relationship but also supplements the potential indirect associations, and is more in line with the actual conduction law of slope deformation.

[0066] Step S129: Classify and integrate the optimized connections of the observation point pairs according to the observation areas to generate sub-region deformation transfer sub-links. Each sub-region deformation transfer sub-link contains the transfer relationship and transfer intensity data of all the observation points in the corresponding observation area.

[0067] In this embodiment, based on the geographical zoning and safety management requirements of the hydropower station slope, the entire slope is first divided into M independent observation areas. The boundary of each observation area is determined by the slope's topographic features, soil and rock types, or the location of safety protection facilities, ensuring a roughly balanced number of observation points within each area. All directed edge information is extracted from the optimized deformation transfer link. Based on the regions to which the initiating and receiving observation points belong, each directed edge is assigned to its corresponding observation area: if the initiating and receiving ends belong to the same region, the edge is included in the sub-regional deformation transfer sub-link of that region; if the initiating and receiving ends belong to different regions, it is not included in the sub-regional sub-link for now, but reserved for subsequent cross-regional connections. For each observation area, all nodes and directed edges within that area are imported into a graph structure model to generate the sub-regional deformation transfer sub-link for that region. The sub-links are also stored in an adjacency list format. Node attributes include the observation point ID and the region ID, while edge attributes include the initiating end ID, the receiving end ID, the deformation transfer intensity S_ij, and the observation period. At the same time, statistical information is generated for each sub-link in the region, including the number of observation points in the region, the number of transmission relationships, and the average deformation transmission intensity.

[0068] Step S1210: Connect all sub-regional deformation transfer sub-links across regions. Based on the deformation transfer relationship of observation points at the boundary of adjacent observation areas, construct a complete deformation transfer link covering all observation points on the entire hydropower station slope. The complete deformation transfer link includes the deformation transfer direction and transfer association information between observation points.

[0069] In this embodiment, firstly, for each pair of adjacent observation regions A and B, the set of boundary observation points for region A and the set of boundary observation points for region B are extracted. The criterion for determining boundary observation points is that the distance between the reference spatial coordinates of the observation point and the region boundary is less than a preset boundary distance threshold D_bound. From the optimized deformation transfer link, all directed edges whose initiating end belongs to region A and whose receiving end belongs to region B, or vice versa, are extracted. These edges represent the cross-regional deformation transfer relationships. For each cross-regional directed edge, if its initiating end is a boundary observation point of region A and its receiving end is a boundary observation point of region B, then this edge is directly used as the cross-regional connection edge between region A and region B. If the initiating end or the receiving end is not a boundary observation point, then, in conjunction with the transfer relationship in the sub-regional deformation transfer sub-link, the transfer source or transfer endpoint of this edge is traced to ensure that both ends of the cross-regional connection edge are boundary observation points. All cross-regional connection edges are associated with the corresponding sub-regional deformation transfer sub-links to form a complete deformation transfer link covering the entire slope. The complete deformation propagation link is stored in a multi-layer graph structure. The first layer is a regional topology graph, where nodes represent observation areas and edges represent cross-regional propagation relationships. The second layer is a regional topology graph, where each region corresponds to a subgraph containing observation points and propagation relationships within the region. Additionally, propagation association information is added to each edge, including deformation propagation intensity, observation period, and the associated sub-regional link ID.

[0070] Step S130: Based on the deformation transmission link, mine the cross-regional deformation transmission characteristics of the hydropower station slope. The cross-regional deformation transmission characteristics include the deformation transmission sequence and transmission association status data of observation points in different observation areas.

[0071] Step S131: Analyze the regional deformation transmission sub-links in the deformation transmission link, and extract the core observation points in each regional deformation transmission sub-link. The core observation point refers to the observation point in the regional deformation transmission sub-link whose deformation transmission intensity data meets the core observation point judgment criteria and whose number of connected observation points meets the core observation point connection requirements.

[0072] In this embodiment, the preset core observation point determination criteria are as follows: in the regional deformation transmission sub-link, the average deformation transmission intensity S_avg of the observation point ≥ S_core_thresh, and the number of connections N_out as the initiating end of the observation point ≥ N_core_thresh, while the number of connections N_in as the receiving end ≥ N_core_in_thresh. First, for each regional deformation transmission sub-link, the average deformation transmission intensity S_avg of each observation point is calculated: the deformation transmission intensity of all directed edges originating from the observation point is extracted, and their average value is calculated to obtain S_avg. Then, the number of connections N_out as the initiating end of the observation point is counted, that is, the number of directed edges originating from the observation point; the number of connections N_in as the receiving end is counted, that is, the number of directed edges pointing to the observation point. The observation point is compared with S_avg and S_core_thresh, N_out and N_core_thresh, and N_in and N_core_in_thresh respectively. If S_avg ≥ S_core_thresh, N_out ≥ N_core_thresh, and N_in ≥ N_core_in_thresh, then the observation point is determined to be a core observation point in the region. If only two or one of these conditions are met, it is marked as a potential core observation point, which will be further confirmed by combining historical deformation data. The set of core observation points for each region is bound and stored with the corresponding sub-regional deformation transmission sub-link, and the attributes of the core observation points are labeled with a "core identifier".

[0073] Step S132: Based on the core observation point of each sub-region deformation transmission sub-link, trace the deformation transmission source of the core observation point, and determine the initial deformation initiation observation point of the core observation point by traversing the transmission relationship in the sub-region deformation transmission sub-link in reverse.

[0074] In this embodiment, for each core observation point c in the deformation transmission sub-link of a sub-region, starting from this observation point, the directed edges in the sub-link of deformation transmission are traversed in reverse, that is, all directed edges pointing to the core observation point c are found, resulting in the set of direct predecessor observation points P1. For each observation point p1 in set P1, its predecessor observation points are traversed in reverse to obtain the set of secondary predecessor observation points P2, and so on, until an observation point without a predecessor transmission relationship is reached, or the number of traversed levels reaches the preset maximum number of levels L_max. During the traversal, the transmission path length (i.e., the number of edges) from each predecessor observation point to the core observation point c and the average deformation transmission intensity on the path are recorded. The rule for determining the initial deformation initiation observation point is: among all traced predecessor observation points, the observation point with the longest path length and the largest average deformation transmission intensity on the path is selected as the initial deformation initiation observation point of the core observation point c; if multiple observation points meet the conditions, the observation point that first shows a deformation signal (i.e., the observation point whose observation timestamp shows the earliest obvious shift) is selected. Each core observation point is associated with and stored with its corresponding initial deformation initiation observation point, while the complete transmission path and path attribute data obtained through tracing are recorded.

[0075] Step S133: Extract the transmission path between the core observation point and the initial deformation initiation observation point in each sub-regional deformation transmission sub-link, record all intermediate observation points and corresponding deformation transmission intensity data in the transmission path, and generate the main deformation transmission path in the region.

[0076] In this embodiment, for each core observation point c and its corresponding initial deformation initiation observation point s, the directed edges in the sub-regional deformation transmission sub-link are traversed forward from s until the core observation point c is reached, thus obtaining a transmission path from s to c. If multiple transmission paths exist from s to c, the path with the highest average deformation transmission intensity is selected as the main deformation transmission path within the region; if the average intensity of the paths is the same, the path with the shortest length is selected. All observation points in the main transmission path are extracted and arranged in the transmission order as s→p1→p2→…→pn→c, where p1 to pn are intermediate observation points. The deformation transmission intensity data of each directed edge in the path is recorded, generating a path intensity sequence [S_s_p1, S_p1_p2, ..., S_pn_c], and the observation time period corresponding to each edge is also recorded. The main deformation transmission path within the region is bound and stored with the corresponding sub-regional deformation transmission sub-links, core observation points, and initial deformation initiation observation points. The path data is stored in the form of an ordered list, which includes the observation point ID sequence and the corresponding deformation transmission intensity sequence.

[0077] Step S134: Analyze the main deformation transmission path within different observation areas, and identify the boundary observation points in the main deformation transmission path within adjacent observation areas. The boundary observation point refers to the observation point located at the junction of two adjacent observation areas.

[0078] In this embodiment, the boundary information of all adjacent observation areas is first obtained, with each pair of adjacent areas (A, B) corresponding to a shared boundary line. For the main deformation transmission path within each observation area, the reference spatial coordinates of all observation points in the path are extracted, and the vertical distance d from each observation point to the shared boundary line of the adjacent areas is calculated. If d ≤ D_bound_point (a preset boundary observation point distance threshold), the observation point is determined to be a boundary observation point and marked as the boundary core transmission node of area A or B. For the adjacent area pair (A, B), the set of boundary observation points B_A in the main deformation transmission path within area A and the set of boundary observation points B_B in the main deformation transmission path within area B are extracted respectively. The observation points in B_A and B_B are paired, and the spatial distance between the paired observation points is calculated. If the distance is less than the preset cross-regional association distance threshold D_cross_thresh, the paired observation points are marked as a potential cross-regional transmission node pair.

[0079] Step S135: Calculate the deformation transmission correlation degree between boundary observation points of adjacent observation areas. The deformation transmission correlation degree is generated by comparing the deformation transmission direction, transmission intensity data and spatial position relationship of the boundary observation points, and integrating the multi-dimensional correlation parameters of the boundary observation points.

[0080] Step S1351: Extract the sub-links of deformation transfer in adjacent observation areas, select the boundary observation points located at the boundary of the observation area from the sub-links of deformation transfer in the observation areas, and record the observation area affiliation and spatial location coordinates of each boundary observation point.

[0081] In this embodiment, for adjacent observation region pairs (A, B), the sub-links of deformation transmission in regions A and B are read respectively. Nodes marked as boundary observation points are selected from these sub-links, resulting in the boundary observation point set B_A for region A and the boundary observation point set B_B for region B. The observation region affiliation (A or B) and the reference spatial coordinates P=(X, Y, Z) of each boundary observation point are recorded. Simultaneously, the deformation transmission direction vector V_dir_norm (the direction vector from the initial deformation initiation observation point to this observation point) and the average deformation transmission intensity S_avg in the main deformation transmission path within the region are extracted. The above data is stored according to the boundary observation point ID, forming a boundary observation point attribute dataset.

[0082] Step S1352: Obtain deformation transfer direction data for each boundary observation point in the deformation transfer link. The deformation transfer direction data includes the transfer direction vector of the boundary observation point as the deformation initiator or receiver.

[0083] In this embodiment, for each boundary observation point b, the propagation direction vectors of all directed edges originating from b are extracted from the regional deformation propagation sub-links. The average value of these vectors is calculated to obtain the average propagation direction vector V_out_avg of b as the originating point. Similarly, the propagation direction vectors of all directed edges originating from b are extracted, and the average value of these vectors is calculated to obtain the average propagation direction vector V_in_avg of b as the receiving point. If b is only the originating point or only the receiving point, only the corresponding average propagation direction vector is retained; if b is both the originating point and the receiving point, both direction vectors are retained.

[0084] Step S1353: Compare the deformation transmission direction vectors of the boundary observation points of adjacent observation areas, and calculate the direction vector coincidence. The direction vector coincidence is generated by integrating the corresponding parameters of the two direction vectors, reflecting the matching state of the transmission direction of the boundary observation points.

[0085] In this embodiment, for a potential cross-regional propagation node pair (b_A, b_B) of adjacent region pair (A, B), where b_A∈B_A and b_B∈B_B, the average propagation direction vector V_in_A (i.e., the average propagation direction pointing to b_A) of b_A as the receiving end and the average propagation direction vector V_out_B (i.e., the average propagation direction starting from b_B) of b_B as the initiating end are extracted. The dot product of the two direction vectors is calculated as Dot = V_in_A.x × V_out_B.x + V_in_A.y × V_out_B.y + V_in_A.z × V_out_B.z. Since the direction vectors are unit vectors, the range of the dot product is [-1, 1]. The direction vector overlap C_dir_cross = (Dot+1) / 2, with a value range of [0, 1]. The closer C_dir_cross is to 1, the more closely the two direction vectors match, that is, the more consistent the input direction of b_A is with the output direction of b_B. The closer C_dir_cross is to 0, the lower the direction overlap.

[0086] Step S1354: Extract the deformation transfer intensity data corresponding to each boundary observation point. The deformation transfer intensity data is the transfer intensity value of the boundary observation point in the deformation transfer link, which reflects the deformation transfer capability of the observation point.

[0087] In this embodiment, for each boundary observation point b, the deformation transfer intensity of all directed edges related to b is extracted from the deformation transfer sub-link of the sub-region, and the average value is calculated to obtain the average deformation transfer intensity S_avg_b. The value range of S_avg_b is [0, 10], and the larger the value, the stronger the deformation transfer capability of the boundary observation point. At the same time, the deformation transfer intensity S_path_b of b on the main deformation transmission path in the region is extracted, that is, the intensity value of the edge related to b in the path. This intensity value reflects the transmission capability of b on the main transmission path.

[0088] Step S1355: Calculate the deformation transfer intensity matching degree of the boundary observation points of adjacent observation areas. This is generated by comparing the difference between the deformation transfer intensity values ​​of the two boundary observation points with the preset intensity difference standard. The matching degree meets the requirements when the difference is within the preset intensity difference standard range.

[0089] In this embodiment, for a potential cross-regional transmission node pair (b_A, b_B), the corresponding average deformation transmission intensities S_avg_A and S_avg_B are extracted, and the intensity difference ΔS_cross = |S_avg_A - S_avg_B| is calculated. The preset standard for the intensity difference is ΔS_cross_thresh. If ΔS_cross ≤ ΔS_cross_thresh, the deformation transmission intensities of the two boundary observation points are considered to match, and the intensity matching degree C_rate_cross = 1 - ΔS_cross / ΔS_cross_thresh; if ΔS_cross > ΔS_cross_thresh, the intensity is considered to be mismatched, and the intensity matching degree C_rate_cross = 0. The value range of the intensity matching degree C_rate_cross is [0, 1], with a larger value indicating a higher intensity matching degree. Simultaneously, a secondary matching calculation can be performed by combining the deformation transfer intensities S_path_A and S_path_B on the path to obtain the path intensity matching degree C_path_cross. The final deformation transfer intensity matching degree is the weighted average of C_rate_cross and C_path_cross.

[0090] Step S1356: Measure the actual spatial distance between the boundary observation points of adjacent observation areas. The actual spatial distance is calculated by the spatial position coordinate data of the boundary observation points and reflects the spatial proximity of the boundary observation points.

[0091] In this embodiment, for a potential cross-regional transmission node pair (b_A, b_B), the reference spatial coordinates P_A=(X_A, Y_A, Z_A) of b_A and the reference spatial coordinates P_B=(X_B, Y_B, Z_B) of b_B are extracted. The actual spatial distance D_cross is obtained using three-dimensional Euclidean distance calculation logic: first, the differences between the three coordinates (X_A-X_B), (Y_A-Y_B), and (Z_A-Z_B) are calculated; then, each difference is squared; the three squared results are added together to obtain a sum of squares; finally, the square root of the sum of squares is taken to obtain the value of D_cross, and the result is retained to three decimal places. The value of D_cross reflects the spatial proximity of the two boundary observation points; the smaller the value, the closer the spatial distance, and the more likely there is a deformation transmission correlation.

[0092] Step S1357: Compare the actual spatial distance with the preset boundary association distance threshold to generate spatial location association parameters. When the actual spatial distance is within the preset boundary association distance threshold range, generate spatial location association parameters according to the preset spatial association parameter mapping rules.

[0093] In this embodiment, the preset boundary association distance threshold is D_cross_thresh. If the actual spatial distance D_cross ≤ D_cross_thresh, then the spatial location association parameter C_space_cross = 1 - D_cross / D_cross_thresh is generated; if D_cross > D_cross_thresh, then the spatial location association parameter C_space_cross = 0 is generated. The value range of the spatial location association parameter C_space_cross is [0, 1]. The larger the value, the higher the spatial proximity of the two boundary observation points, and the more sufficient the physical basis of deformation transmission. The generated C_space_cross is bound and stored with the corresponding node pair (b_A, b_B) and D_cross.

[0094] Step S1358: Based on the direction vector coincidence, deformation transmission intensity matching degree and spatial position correlation parameters, set the weight coefficients of each parameter. The weight coefficients are determined by the deformation influence priority analysis of the boundary observation points.

[0095] For example, step S13581: collect historical deformation transmission case data of hydropower station slope, the historical deformation transmission case data includes deformation transmission records of observation points at the boundary of adjacent observation areas in multiple historical deformation events.

[0096] In this embodiment, all deformation event records occurring within the past 5 years are extracted from the historical database of the hydropower station slope safety monitoring system. Each deformation event record includes information such as the event occurrence time, the observation area involved, the spatial location data of the boundary observation points, the deformation transmission direction data, the deformation transmission intensity data, and the deformation transmission result (e.g., whether it was successfully transmitted to adjacent areas). For each historical deformation event, transmission records involving boundary observation points of adjacent observation areas are selected and organized into a structured historical deformation transmission case dataset. Each case includes fields such as boundary observation point pair ID, direction vector coincidence, deformation transmission intensity matching degree, spatial location association parameters, transmission success identifier (1 indicates successful transmission, 0 indicates unsuccessful transmission), transmission stability parameters (e.g., the fluctuation coefficient of transmission intensity), and transmission timeliness parameters (e.g., the time required for transmission).

[0097] Step S13582: Extract the correlation between the direction vector coincidence data of the boundary observation point and the deformation transmission success rate in the historical deformation transmission case data for each historical deformation event. The deformation transmission success rate refers to the proportion of cases in which the boundary observation point successfully transmits deformation.

[0098] In this embodiment, historical deformation propagation case data are grouped according to the value range of the direction vector coincidence degree C_dir_cross, with grouping intervals of [0, 0.2), [0.2, 0.4), [0.4, 0.6), [0.6, 0.8), and [0.8, 1.0]. For each group, the number of cases with a propagation success marker of 1 and the total number of cases within that group are counted, and the deformation propagation success rate R_dir for each group is calculated as: number of successful cases / total number of cases. A correlation curve between C_dir_cross and R_dir is plotted, and the trend of the curve is analyzed: if R_dir increases significantly with the increase of C_dir_cross, it indicates that the direction vector coincidence degree has a large impact on the deformation propagation success rate; if the curve changes gently, it indicates that the impact is small.

[0099] Step S13583: Analyze the impact of directional vector coincidence data on deformation transmission success rate. By statistically analyzing the changes in deformation transmission success rate corresponding to different directional vector coincidence intervals, determine the basic value of the influence weight of directional vector coincidence.

[0100] In this embodiment, based on the deformation transmission success rate R_dir obtained from group statistics, the difference ΔR_dir between R_dir of each group and the minimum R_dir is calculated, and the maximum value of ΔR_dir is used as the influence intensity benchmark. The influence weight base value W_dir_base of the direction vector coincidence degree is = ΔR_dir_max / ΔR_total, where ΔR_total is the sum of the influence intensity benchmarks corresponding to all parameters. For example, if ΔR_dir_max corresponding to the direction vector coincidence degree is 0.7, ΔR_rate_max corresponding to the deformation transmission intensity matching degree is 0.5, and ΔR_space_max corresponding to the spatial location association parameter is 0.3, then ΔR_total = 0.7 + 0.5 + 0.3 = 1.5, and W_dir_base = 0.7 / 1.5 ≈ 0.47.

[0101] Step S13584: Extract the correlation between the deformation transmission intensity matching degree data and the deformation transmission stability of the boundary observation points in each historical deformation event from the historical deformation transmission case data. Deformation transmission stability refers to the stable state of intensity fluctuation during deformation transmission.

[0102] In this embodiment, the deformation transfer intensity matching degree C_rate_cross and the deformation transfer stability parameter S_stab are extracted from historical deformation transfer case data for each case. S_stab is the coefficient of variation of the transfer intensity (coefficient of variation = standard deviation / mean), and a smaller S_stab indicates higher transfer stability. The case data are grouped according to the value range of C_rate_cross, with the grouping intervals coinciding with the grouping intervals of the direction vector coincidence. For each group, the average value of S_stab within that group is calculated to obtain the average stability S_stab_avg. The correlation curve between C_rate_cross and S_stab_avg is plotted, and the trend of the curve is analyzed: if S_stab_avg decreases significantly with the increase of C_rate_cross, it indicates that the deformation transfer intensity matching degree has a significant impact on deformation transfer stability; if the curve changes gently, it indicates a smaller impact.

[0103] Step S13585: Analyze the influence of deformation transmission strength matching degree data on deformation transmission stability. By statistically analyzing the changes in deformation transmission stability corresponding to different deformation transmission strength matching degree intervals, determine the basic value of the influence weight of deformation transmission strength matching degree.

[0104] In this embodiment, based on the average stability S_stab_avg obtained from group statistics, the difference ΔS_stab between S_stab_avg and the maximum S_stab_avg for each group is calculated, and the maximum value of ΔS_stab is used as the influence intensity benchmark ΔS_rate_max. The influence weight base value W_rate_base = ΔS_rate_max / ΔS_total, where ΔS_total is the sum of the influence intensity benchmarks corresponding to all parameters. For example, if ΔS_rate_max is 0.6, ΔS_dir_max is 0.8, and ΔS_space_max is 0.4, then ΔS_total = 0.8 + 0.6 + 0.4 = 1.8, and W_rate_base = 0.6 / 1.8 ≈ 0.33.

[0105] Step S13586: Extract the spatial location correlation parameters of the boundary observation points in each historical deformation event from the historical deformation transmission case data and the correlation relationship between deformation transmission timeliness. The deformation transmission timeliness refers to the time it takes for deformation to be transmitted from one boundary observation point to another.

[0106] In this embodiment, the spatial location correlation parameter C_space_cross and the deformation transmission timeliness parameter T_trans are extracted from historical deformation transmission case data for each case. T_trans represents the time difference between deformation transmission from one boundary observation point to another; a smaller T_trans indicates higher transmission timeliness. The case data are grouped according to the value range of C_space_cross, with the grouping interval consistent with the previous two parameters. For each group, the average value of T_trans within that group is calculated to obtain the average transmission time T_trans_avg. The correlation curve between C_space_cross and T_trans_avg is plotted, and the trend of the curve is analyzed: if T_trans_avg decreases significantly with the increase of C_space_cross, it indicates that the spatial location correlation parameter has a significant impact on the deformation transmission timeliness; if the curve changes gently, it indicates a smaller impact.

[0107] Step S13587: Analyze the influence of spatial location-related parameters on the timeliness of deformation transmission. By statistically analyzing the changes in the timeliness of deformation transmission corresponding to different spatial location-related parameter intervals, determine the basic value of the influence weight of the spatial location-related parameters.

[0108] In this embodiment, based on the average transmission time T_trans_avg obtained from group statistics, the difference ΔT_trans between T_trans_avg and the maximum T_trans_avg for each group is calculated, and the maximum value of ΔT_trans is used as the influence intensity benchmark ΔT_space_max. The base value of the influence weight of the spatial location association parameter is W_space_base = ΔT_space_max / ΔT_total, where ΔT_total is the sum of the influence intensity benchmarks corresponding to all parameters. For example, if ΔT_space_max is 0.5, ΔT_dir_max is 0.7, and ΔT_rate_max is 0.4, then ΔT_total = 0.7 + 0.4 + 0.5 = 1.6, and W_space_base = 0.5 / 1.6 ≈ 0.31.

[0109] Step S13588: Based on the safety protection level of the hydropower station slope, adjust the basic values ​​of the influence weights of the direction vector coincidence, deformation transmission intensity matching degree and spatial location correlation parameters. The safety protection level is divided according to the preset level standard, and the adjustment range of the parameter weights that affect the deformation transmission success rate in accordance with the preset influence standard is determined according to the preset adjustment rules.

[0110] In this embodiment, the safety protection level of the hydropower station slope is divided into three levels: Level 1 (highest protection level, involving the core dam area), Level 2 (medium protection level, involving the dam abutment area), and Level 3 (general protection level, involving the downstream slope area). The preset adjustment rules are as follows: Under Level 1 protection, if a parameter has the greatest impact on the deformation transmission success rate, its base weight value is increased by 20%; under Level 2 protection, the base weight value is increased by 10%; under Level 3 protection, the base weight value is not adjusted. For example, if the direction vector coincidence degree has the greatest impact on the deformation transmission success rate, and the current slope area is at Level 1 protection, then W_dir_adjusted = W_dir_base × 1.2; if it is at Level 2 protection, then W_dir_adjusted = W_dir_base × 1.1; if it is at Level 3 protection, then W_dir_adjusted = W_dir_base. For the deformation transmission intensity matching degree and spatial location correlation parameters, adjustments are made according to the importance of their corresponding influence indicators (stability, timeliness) based on similar rules. The adjusted weight coefficients are normalized to ensure that W_dir_final+W_rate_final+W_space_final=1.

[0111] Step S13589: Normalize the basic values ​​of the influence weights of the adjusted direction vector coincidence, deformation transmission strength matching degree and spatial location correlation parameters to generate initial weight coefficients; combine the real-time observation environmental data of the hydropower station slope to fine-tune the initial weight coefficients. The real-time observation environmental data includes slope surface humidity data and soil and rock mass state data. Finally, determine the weight coefficients of the direction vector coincidence, deformation transmission strength matching degree and spatial location correlation parameters.

[0112] In this embodiment, the adjusted base weight values ​​W_dir_adjusted, W_rate_adjusted, and W_space_adjusted are first normalized, and the initial weight coefficients are calculated as follows: W_dir_initial = W_dir_adjusted / (W_dir_adjusted + W_rate_adjusted + W_space_adjusted), W_rate_initial = W_rate_adjusted / (W_dir_adjusted + W_rate_adjusted + W_space_adjusted), W_space_initial = W_space_adjusted / (W_dir_adjusted + W_rate_adjusted + W_space_adjusted). Then, fine-tuning is performed based on the real-time monitoring data of the current slope: if the slope surface humidity exceeds the preset humidity threshold, it indicates that the soil and rock mass has a high water content, and the resistance to deformation transmission may increase. In this case, the weight of the spatial location correlation parameter is increased by 5%, the weight of the direction vector coincidence is decreased by 3%, and the weight of the deformation transmission strength matching degree is decreased by 2%. If the real-time monitoring shows crack propagation signals in the soil and rock mass, the weight of the deformation transmission strength matching degree is increased by 5%, the weight of the direction vector coincidence is decreased by 2%, and the weight of the spatial location correlation parameter is decreased by 3%. After fine-tuning, normalization is performed again to ensure that the sum of the three weight coefficients is 1, resulting in the final weight coefficients W_dir_final, W_rate_final, and W_space_final.

[0113] Step S1359: Based on the weight coefficients of each parameter, the normalized dimensionless evaluation values ​​are weighted and fused to generate a preliminary deformation transmission correlation degree between observation points at the boundary of adjacent observation areas; the preliminary deformation transmission correlation degree is normalized to generate the final deformation transmission correlation degree between observation points at the boundary of adjacent observation areas.

[0114] In this embodiment, for a potential cross-regional transmission node pair (b_A, b_B), the direction vector coincidence degree C_dir_cross, deformation transmission intensity matching degree C_rate_cross, and spatial location association parameter C_space_cross are extracted. These are then multiplied by their corresponding final weighting coefficients W_dir_final, W_rate_final, and W_space_final, respectively, to obtain the weighted direction association degree W_dir_final×C_dir_cross, the weighted intensity association degree W_rate_final×C_rate_cross, and the weighted spatial association degree W_space_final×C_space_cross. The three weighted association degrees are then summed to obtain the preliminary deformation transmission association degree R_cross_raw = W_dir_final×C_dir_cross + W_rate_final×C_rate_cross + W_space_final×C_space_cross. Since the value range of each parameter is [0, 1], and the sum of the weighting coefficients is 1, the value range of R_cross_raw is [0, 1]. To improve the discriminative power of the correlation, R_cross_raw is mapped to the range [0, 10] to obtain the final deformation transmission correlation R_cross = R_cross_raw × 10, and the result is retained to one decimal place.

[0115] Step S136: Based on the preset deformation transmission order determination rule, the observation area whose deformation transmission correlation value satisfies the deformation transmission order determination rule is taken as the deformation first transmission area, and the other observation area is taken as the deformation later transmission area, and the cross-region deformation transmission direction is generated.

[0116] In this embodiment, the preset deformation transmission order determination rule is as follows: for adjacent observation region pairs (A, B), the deformation transmission correlation degree R_cross of all boundary observation point pairs (b_A, b_B) is counted, and the average correlation degree R_AB_avg from region A to region B and the average correlation degree R_BA_avg from region B to region A are calculated. If R_AB_avg - R_BA_avg > ΔR_thresh (preset correlation difference threshold), then region A is determined to be the region where deformation is conducted first, and region B is determined to be the region where deformation is conducted later. A cross-region deformation conduction direction from A to B is generated and recorded as (A→B). If R_BA_avg - R_AB_avg > ΔR_thresh, then region B is determined to be the region where deformation is conducted first, and region A is determined to be the region where deformation is conducted later. A cross-region deformation conduction direction from B to A is generated and recorded as (B→A). If |R_AB_avg - R_BA_avg| ≤ ΔR_thresh, then the deformation conduction of the adjacent region pair is determined to be in a bidirectional synchronous state, and no explicit conduction direction is set for the time being.

[0117] Step S137: For each cross-regional deformation transmission direction, extract the corresponding boundary observation point transmission relationship, integrate the main deformation transmission paths within adjacent observation areas, and generate a cross-regional deformation transmission main path. The cross-regional deformation transmission main path includes all observation points and transmission order of cross-regional transmission.

[0118] In this embodiment, for the generated cross-regional deformation propagation direction (A→B), the endpoint boundary observation point b_A_end (i.e., the boundary observation point closest to region B) is extracted from the main deformation propagation path within region A, and the starting boundary observation point b_B_start (i.e., the boundary observation point closest to region A) is extracted from the main deformation propagation path within region B. If the deformation propagation correlation R_cross corresponding to b_A_end and b_B_start is ≥ R_cross_thresh (preset effective correlation threshold), then the main deformation propagation path within region A, the cross-regional connection edge from b_A_end to b_B_start, and the main deformation propagation path within region B are integrated to form a complete cross-regional deformation propagation main path from the initial deformation initiation observation point s_A in region A to the core observation point c_B in region B, with the path order being s_A→…→b_A_end→b_B_start→…→c_B. If multiple boundary observation point pairs satisfy the condition, the observation point pair with the largest R_cross is selected as the cross-regional connection point. The main path of cross-regional deformation propagation is bound and stored with the corresponding observation area pair, propagation direction, and connection edge attributes. The path data is stored in the form of an ordered list, which includes the observation point ID sequence and the corresponding deformation propagation intensity sequence.

[0119] Step S138: Based on the cross-regional deformation transmission main path, calculate the transmission correlation status data between different observation areas. The transmission correlation status data is generated by integrating the deformation transmission intensity data of all observation point pairs in the cross-regional deformation transmission main path, reflecting the close relationship between the regions in deformation transmission.

[0120] In this embodiment, for the main path of cross-regional deformation transmission, deformation transmission intensity data of all observation point pairs in the path are extracted, and the average deformation transmission intensity on the path is calculated as R_cross_avg = ΣS_ij / n, where n is the number of observation point pairs on the path, and S_ij is the deformation transmission intensity of each observation point pair. Simultaneously, the coefficient of variation of the deformation transmission intensity on the path is calculated as CV_cross = σ_cross / R_cross_avg, where σ_cross is the standard deviation of the deformation transmission intensity on the path. The average deformation transmission intensity R_cross_avg and the coefficient of variation CV_cross are combined according to the observation region pairs to generate transmission correlation state data S_cross = (R_cross_avg, CV_cross). A larger R_cross_avg indicates a higher overall intensity of deformation transmission between regions, while a smaller CV_cross indicates higher stability of deformation transmission between regions.

[0121] Step S139: Extract the transmission sequence information of all cross-regional deformation transmission main paths, organize the deformation transmission start time of different observation areas according to the observation time, and generate deformation transmission sequence data of observation points in different observation areas.

[0122] In this embodiment, for each cross-regional deformation propagation main path, the initial deformation signal occurrence time of each observation point in the path is extracted (i.e., the observation timestamp when the spatial position offset of the observation point first exceeds a preset offset threshold). The start times of each observation point are organized according to the propagation order on the path to obtain the deformation propagation order sequence of observation points within the region: s_A_start_time→p1_start_time→…→b_A_end_start_time→b_B_start_start_time→…→c_B_start_time. For each observation region, the start times of all observation points within that region are statistically analyzed and sorted chronologically to obtain the deformation propagation order data of the observation points within the region, including the observation point ID sequence and the corresponding start timestamp sequence. Simultaneously, the deformation propagation start time difference between different observation regions is calculated, i.e., the difference between the earliest start time of region A and the earliest start time of region B, reflecting the time delay of deformation propagation between regions.

[0123] Step S1310: Integrate the deformation transmission sequence data and the inter-regional transmission correlation status data of the observation points in different observation areas to generate cross-regional deformation transmission characteristics of the hydropower station slope. The cross-regional deformation transmission characteristics include the deformation transmission sequence and transmission correlation status data of the observation points in different observation areas.

[0124] In this embodiment, the deformation transmission sequence data of each observation area is integrated with the corresponding inter-regional transmission correlation status data to form a structured cross-regional deformation transmission feature dataset. The feature data for each area includes: area ID, deformation transmission sequence of observation points within the area (observation point ID + start timestamp), transmission correlation status data with adjacent areas (average transmission intensity, coefficient of variation), cross-regional transmission direction, and transmission time delay. The feature data of all observation areas are correlated according to regional topology to form a cross-regional deformation transmission feature network covering the entire slope. The feature dataset is stored in JSON format and simultaneously converted to GraphML format for easy visualization.

[0125] Step S140: Integrate the cross-regional deformation transmission characteristics and the spatial position change trajectory of each observation point in the three-dimensional observation data to construct a three-dimensional deformation evolution path of the hydropower station slope. The three-dimensional deformation evolution path includes the continuous transmission trend of spatial position change of observation points and the deformation diffusion path between regions.

[0126] Step S141: Extract all spatial location information and corresponding observation time information of each observation point from the three-dimensional observation data, concatenate the spatial location information of each observation point in the order of observation time, and generate the spatial location change trajectory of each observation point. The spatial location change trajectory contains the continuous spatial location change data of the observation point throughout the entire observation period.

[0127] In this embodiment, firstly, all spatial coordinates (X_ik, Y_ik, Z_ik) and corresponding observation timestamps t_k for each observation point are extracted from the 3D observation data, where i is the observation point ID and k is the observation time point sequence number. The spatial coordinates of each observation point are sorted according to the order of the observation timestamps, forming an ordered coordinate sequence [(X_i1, Y_i1, Z_i1, t1), (X_i2, Y_i2, Z_i2, t2), ..., (X_iK, Y_iK, Z_iK, tK)], where K is the total number of observation time points for that observation point. This sequence is converted into a continuous spatial position change trajectory. Cubic spline interpolation is used to smooth the coordinate sequence, filling in position gaps within the observation time interval, resulting in continuous trajectory data with a time resolution of T / 10, ensuring the continuity and smoothness of the trajectory. The spatial position change trajectory of each observation point is stored in a two-dimensional array, with the first dimension being the timestamp and the second dimension being the corresponding 3D coordinate data. The start time, end time, and total duration of the trajectory are also recorded.

[0128] Step S142: Analyze the deformation transmission sequence of observation points in different observation areas in the cross-regional deformation transmission characteristics, and calibrate the spatial position change trajectory of each observation point according to the deformation transmission sequence on the time axis so that the spatial position change trajectory of observation points on the same transmission path remains synchronized in the time dimension.

[0129] In this embodiment, the deformation transmission sequence of observation points within each observation region is first extracted from the cross-regional deformation transmission characteristics, and the deformation start time stamp t_start_i (i.e., the time stamp of the first significant deformation offset) of each observation point is obtained. Using the start time stamp t_start_s of the observation point initiating the initial deformation within the region as the reference time, the time offset Δt_i = t_start_i - t_start_s of each observation point is calculated. For the spatial position change trajectory of each observation point, Δt_i is subtracted from all timestamps t_k to obtain the calibrated timestamp t_k' = t_k - Δt_i, ensuring that the deformation start time of observation points on the same transmission path is aligned to the reference time t_start_s on the calibrated time axis. For observation points on cross-regional transmission paths, the reference time of the first transmission region in the cross-regional transmission direction is used as a unified reference to calculate the time offset of observation points in the subsequent transmission regions (including the inter-regional transmission delay time), and the trajectory timestamps are calibrated accordingly.

[0130] Step S143: Based on the calibrated spatial position change trajectory of the observation point, extract the continuous trend of spatial position change of each observation point. The continuous trend of spatial position change includes the continuous state of the direction of spatial position change and the stable state of the rate of change of the observation point during the continuous observation period.

[0131] Step S1431: Divide the spatial location change trajectory of the calibrated observation points into multiple consecutive analysis periods in chronological order. Each analysis period contains multiple consecutive observation time points, so that the time span of each analysis period remains the same.

[0132] In this embodiment, the time span of each analysis period is set as ΔT. The calibrated spatial position change trajectory is divided into M consecutive analysis periods, starting from the initial timestamp and spaced at intervals of ΔT. If the last period is less than ΔT, it is merged with the previous period. For the trajectory of each observation point, the time range [t_m_start, t_m_end] corresponding to each analysis period is calculated, where m is the analysis period number, t_m_start = t_start + (m-1) × ΔT, t_m_end = t_start + m × ΔT, and t_start is the start time of the calibrated trajectory. All spatial position coordinates and corresponding calibration timestamps within each analysis period are extracted to form sub-trajectory data for each period, ensuring that the time span of each period is consistent.

[0133] Step S1432: For each analysis period, extract the direction of spatial position change of the observation point between all adjacent observation time points within the analysis period, and generate the direction sequence of the analysis period. The direction sequence contains multiple consecutive spatial position change direction data.

[0134] In this embodiment, for the sub-trajectory data of each analysis period, the spatial coordinates P_k=(X_k, Y_k, Z_k) and P_{k+1}=(X_{k+1}, Y_{k+1}, Z_{k+1}) of two adjacent observation time points are extracted sequentially, and the change direction vector V_k=(X_{k+1}-X_k, Y_{k+1}-Y_k, Z_{k+1}-Z_k) from P_k to P_{k+1} is calculated. Each direction vector V_k is standardized (divided by its magnitude) to obtain the unit direction vector V_k_norm, ensuring that the direction data contains only direction information and eliminating the influence of amplitude. All unit direction vectors in the analysis period are arranged in chronological order to form a direction sequence Seq_V=[V_1_norm, V_2_norm, ..., V_{n-1}_norm], where n is the number of observation time points in the period. The directional sequence is bound to the corresponding analysis period and observation point ID for storage.

[0135] Step S1433: Calculate the directional deviation of adjacent directional data in the directional sequence of each analysis period. The directional deviation is generated by calculating the angle between the changing directions of two adjacent spatial positions, reflecting the fluctuation state of directional changes.

[0136] In this embodiment, for each analysis period's direction sequence Seq_V, two adjacent unit direction vectors V_k_norm and V_{k+1}_norm are selected sequentially, and their included angle θ_k = arccos(V_k_norm·V_{k+1}_norm) is calculated, where "·" represents the vector dot product operation, and the value of θ_k ranges from 0 to π (radians). θ_k represents the direction deviation between adjacent direction data. The smaller θ_k is, the more consistent the deformation direction is between adjacent time points, and the smaller the fluctuation; the larger θ_k is, the more drastic the direction change is, and the greater the fluctuation. All direction deviations θ_k within the analysis period are arranged in chronological order to form a direction deviation sequence Seq_θ = [θ_1, θ_2, ..., θ_{n-2}_norm], which is then bound and stored with the corresponding direction sequence, analysis period, and observation point ID.

[0137] Step S1434: Based on the direction deviation data for each analysis period, calculate the direction continuous state parameters. When the statistical results of the direction deviation data are within the preset direction deviation range, generate the direction continuous state parameters according to the preset direction continuous parameter mapping rules.

[0138] In this embodiment, the statistical characteristics of the directional deviation sequence Seq_θ within each analysis period are first calculated, including the average deviation θ_avg = Σθ_k / (n-2) and the standard deviation σ_θ = √[Σ(θ_k-θ_avg)² / (n-2)]. The preset directional deviation interval is [θ_low, θ_high], where θ_low is the minimum allowable average deviation and θ_high is the maximum allowable average deviation; at the same time, the preset standard deviation threshold σ_θ_thresh is also used. If θ_avg∈[θ_low, θ_high] and σ_θ≤σ_θ_thresh, then the direction change is considered continuous and stable within this time period, and the direction continuity state parameter C_dir_cont=1 is generated; if θ_avg∈[θ_low, θ_high] but σ_θ>σ_θ_thresh, then the direction is considered generally consistent but with large local fluctuations, and C_dir_cont=0.7 is generated; if θ_avg∉[θ_low, θ_high] but σ_θ≤σ_θ_thresh, then the direction is considered generally offset but with small fluctuations, and C_dir_cont=0.5 is generated; if θ_avg∉[θ_low, θ_high] and σ_θ>σ_θ_thresh, then the direction change is considered disordered, and C_dir_cont=0 is generated. The value range of the direction continuity state parameter is from 0 to 1. The larger the value, the better the direction continuity. It is stored in conjunction with the corresponding analysis time period and deviation statistical characteristics.

[0139] Step S1435: For each analysis period, calculate the rate of change of spatial position of the observation point among all adjacent observation time points within the analysis period, and generate a rate sequence for the analysis period, wherein the rate sequence contains multiple consecutive spatial position change rate data.

[0140] In this embodiment, for the sub-trajectory data of each analysis period, the spatial coordinates P_k and P_{k+1} of two adjacent observation time points, as well as the corresponding calibration timestamps t_k' and t_{k+1}', are extracted sequentially. The time difference Δt_k = t_{k+1}' - t_k' between the two time points is calculated, and the Euclidean distance of the spatial position change D_k = √[(X_{k+1} - X_k)² + (Y_{k+1} - Y_k)² + (Z_{k+1} - Z_k)²] is calculated. Then, the rate of change of spatial position v_k = D_k / Δt_k, in meters per second (m / s). All rates of change v_k within the analysis period are arranged in chronological order to form a rate sequence Seq_v = [v_1, v_2, ..., v_{n-1}], which is then bound and stored with the corresponding analysis period and observation point ID.

[0141] Step S1436: Calculate the rate difference between adjacent rate data in the rate sequence of each analysis period. The rate difference is generated by the numerical difference between the rates of change of two adjacent spatial locations, reflecting the fluctuation state of rate change.

[0142] In this embodiment, for the rate sequence Seq_v of each analysis period, two adjacent rates of change, v_k and v_{k+1}, are selected sequentially, and the rate difference Δv_k = |v_{k+1} - v_k| is calculated. The value of Δv_k ranges from 0 to the maximum rate difference. The smaller Δv_k is, the more stable the deformation rate is at adjacent time points, and the smaller the fluctuation; the larger Δv_k is, the more drastic the rate change is, and the greater the fluctuation. All rate differences Δv_k within the analysis period are arranged in chronological order to form a rate difference sequence Seq_Δv = [Δv_1, Δv_2, ..., Δv_{n-2}], which is then bound and stored with the corresponding rate sequence, analysis period, and observation point ID.

[0143] Step S1437: Based on the rate difference data for each analysis period, calculate the rate steady state parameter. When the statistical result of the rate difference data is within the preset rate difference range, generate the rate steady state parameter according to the preset rate steady parameter mapping rule.

[0144] In this embodiment, the statistical characteristics of the rate difference sequence Seq_Δv within each analysis period are first calculated, including the average rate difference Δv_avg = ΣΔv_k / (n-2) and the standard deviation of the difference σ_Δv = √[Σ(Δv_k-Δv_avg)² / (n-2)]. The preset rate difference interval is [Δv_low, Δv_high], where Δv_low is the minimum allowable average rate difference and Δv_high is the maximum allowable average rate difference; at the same time, the preset standard deviation threshold σ_Δv_thresh is also calculated. If Δv_avg∈[Δv_low, Δv_high] and σ_Δv≤σ_Δv_thresh, then the rate change is considered stable during this period, and a rate stability state parameter C_rate_stab=1 is generated; if Δv_avg∈[Δv_low, Δv_high] but σ_Δv>σ_Δv_thresh, then the rate is considered stable overall but with large local fluctuations, and C_rate_stab=0.7 is generated; if Δv_avg∉[Δv_low, Δv_high] but σ_Δv≤σ_Δv_thresh, then the rate is considered to be deviated overall but with small fluctuations, and C_rate_stab=0.5 is generated; if Δv_avg∉[Δv_low, Δv_high] and σ_Δv>σ_Δv_thresh, then the rate change is considered disordered, and C_rate_stab=0 is generated. The value range of the rate stability parameter is 0 to 1. The larger the value, the better the rate stability. It is stored in conjunction with the corresponding analysis period and difference statistical characteristics.

[0145] Step S1438: Arrange the direction continuous state parameters of all analysis periods in chronological order to generate a direction continuous state parameter sequence, which contains the direction continuous change data of the observation point throughout the entire observation period.

[0146] In this embodiment, for each observation point, the corresponding directional continuous state parameter C_dir_cont is arranged according to the time sequence of the analysis period, forming a directional continuous state parameter sequence Seq_C_dir=[C_dir_cont_1, C_dir_cont_2, ..., C_dir_cont_M], where M is the number of analysis periods. Each parameter corresponds to the directional continuous state of one analysis period, and the sequence as a whole reflects the directional continuity change trend of the observation point throughout the entire observation period.

[0147] Step S1439: Arrange the rate steady-state parameters of all analysis periods in chronological order to generate a rate steady-state parameter sequence, which contains the rate steady-state change data of the observation point throughout the entire observation period.

[0148] In this embodiment, for each observation point, the corresponding rate steady-state parameter C_rate_stab is arranged according to the time sequence of the analysis period, forming a rate steady-state parameter sequence Seq_C_rate=[C_rate_stab_1, C_rate_stab_2, ..., C_rate_stab_M], where M is the number of analysis periods. Each parameter corresponds to the rate steady-state of one analysis period, and the sequence as a whole reflects the rate stability change trend of the observation point throughout the entire observation period.

[0149] Step S14310: Integrate the direction continuous state parameter sequence and the rate stable state parameter sequence to generate the spatial position change continuous situation of each observation point. The spatial position change continuous situation includes the spatial position change direction continuous state and change rate stable state data of the observation point during the continuous observation period.

[0150] In this embodiment, the directional continuous state parameter sequence Seq_C_dir and the velocity stable state parameter sequence Seq_C_rate for each observation point are integrated to form continuous situational data of spatial location change S_cont=(Seq_C_dir, Seq_C_rate). Simultaneously, time range data for each analysis period is added, enabling the situational data to correspond to specific observation periods. The continuous situational data is stored in a two-dimensional array, with the first dimension being the analysis period number and the second dimension being a structure containing the directional continuous parameters, velocity stable parameters, and time range.

[0151] Step S144: Combining the transmission correlation state data in the cross-regional deformation transmission characteristics, the continuous trend of spatial position changes of observation points on the same cross-regional deformation transmission main path is weighted and integrated, and the weights corresponding to the transmission correlation state data are used to strengthen the influence of the changing trend of key transmission nodes.

[0152] In this embodiment, the transmission correlation strength data of all observation points on the same cross-regional deformation transmission main path is extracted from the cross-regional deformation transmission characteristics. Based on this data, the weight coefficient W_i of each observation point is calculated: the transmission correlation strength S_i of each observation point is divided by the sum of the transmission correlation strengths of all observation points on the path, resulting in W_i = S_i / ΣS_j, where j is the observation point number on the path, ensuring that ΣW_i = 1. For each analysis period, the directional continuous state parameter C_dir_cont_i of all observation points on the path is multiplied by the corresponding weight coefficient W_i, and the weighted sum is obtained to obtain the path-level directional continuous parameter C_dir_path = Σ(W_i × C_dir_cont_i); similarly, the weighted sum is obtained to obtain the path-level rate stability parameter C_rate_path = Σ(W_i × C_rate_stab_i). The path-level parameters of each analysis period are arranged in chronological order to form a continuous trend sequence of spatial position changes at the path level, which strengthens the influence of key nodes with high transmission correlation strength in the path trend and better reflects the actual dominant logic of deformation transmission.

[0153] Step S145: Extract the inter-regional deformation transmission nodes in the main path of cross-regional deformation transmission. The inter-regional deformation transmission nodes refer to the deformation transmission observation points at the boundaries of adjacent observation areas. Integrate the continuous trend of spatial position changes of the inter-regional deformation transmission nodes to generate the inter-regional deformation transition trend.

[0154] In this embodiment, deformation transmission nodes located at the boundaries of adjacent observation areas (i.e., the end boundary node of the first transmission area and the starting boundary node of the subsequent transmission area) are extracted from the main path of cross-regional deformation transmission, and continuous situation data of the spatial position changes of these two nodes are obtained. For each analysis period, the average value of the directional continuous state parameters of the two nodes is calculated as the transition direction parameter C_dir_trans=(C_dir_cont_A+C_dir_cont_B) / 2, and the average value of the rate steady state parameter is calculated as the transition rate parameter C_rate_trans=(C_rate_stab_A+C_rate_stab_B) / 2, where A is the boundary node of the first transmission area and B is the boundary node of the subsequent transmission area. The transition direction parameter and transition rate parameter of each analysis period are arranged in chronological order to form the inter-regional deformation transition situation sequence S_trans=(Seq_C_dir_trans, Seq_C_rate_trans). At the same time, the time axis of the transition situation is adjusted by combining the cross-regional transmission delay time so that the transition situation can accurately reflect the process change of deformation from one region to another.

[0155] Step S146: Based on the deformation transition situation between regions, construct an inter-regional deformation diffusion model. The inter-regional deformation diffusion model includes the path direction and diffusion rate data of deformation diffusion from one observation region to adjacent observation regions, and is generated by integrating the changing situation of boundary observation points.

[0156] In this embodiment, for the inter-regional deformation transition state S_trans, the transition direction parameter C_dir_trans and the transition rate parameter C_rate_trans for each analysis period are extracted. Combined with the spatial coordinate data of the boundary nodes, an inter-regional deformation diffusion model M_diff=(Dir_diff, Rate_diff) is constructed. The path direction Dir_diff is obtained by time-weighted averaging of the transition direction parameter, with the weights representing the proportion of each analysis period's duration to the total duration. Dir_diff is a three-dimensional unit vector reflecting the dominant direction of deformation diffusion. The diffusion rate Rate_diff is obtained by time-weighted averaging of the transition rate parameter, combined with the inter-regional propagation delay time, and converted into the deformation diffusion distance per unit time, reflecting the average rate of deformation diffusion from one region to another. Simultaneously, a diffusion rate fluctuation range parameter is added, setting upper and lower limits for rate fluctuation based on the standard deviation of the transition rate parameter, enabling the model to reflect the uncertainty of the diffusion process.

[0157] Step S147: For each observation area, a three-dimensional deformation evolution sub-path is constructed within the area by combining the main deformation transmission path within the area and the corresponding continuous trend of spatial position change. The three-dimensional deformation evolution sub-path within the area includes the continuous trend of spatial position change of all observation points within the observation area and the transmission path within the area.

[0158] In this embodiment, for each observation area, the sequence of observation points and the corresponding continuous spatial position change data of the main deformation transmission path within the area are extracted. Starting from the initial deformation initiation observation point and ending at the core observation point, the continuous situation data of each observation point is bound to its position in the path according to the transmission order, forming a three-dimensional deformation evolution sub-path P_sub=(Seq_P, Seq_S_cont) within the area, where Seq_P is the observation point ID sequence and Seq_S_cont is the continuous spatial position change data corresponding to each observation point. Simultaneously, attributes such as transmission intensity data, average transmission rate data, and total deformation offset data are added to the sub-path to enrich its feature information. For observation points on non-main transmission paths within the area, their continuous situation data are linked to the corresponding position on the main transmission path through the association relationship with the nearest observation point on the main transmission path, ensuring that the sub-path can cover the deformation situation of all observation points within the area. The three-dimensional deformation evolution sub-path within the area is stored in a graph structure, with nodes representing observation points and edges representing transmission relationships. Edge attributes include transmission intensity and transmission direction.

[0159] Step S148: Connect the three-dimensional deformation evolution sub-paths within each observation area across regions using the inter-regional deformation diffusion model, so that the evolution sub-paths of adjacent observation areas are connected according to the deformation diffusion path, and generate the initial three-dimensional deformation evolution path.

[0160] In this embodiment, following the cross-regional deformation propagation direction, the endpoints (core observation points or boundary nodes) of the three-dimensional deformation evolution sub-paths within the region of the first propagation area are connected to the starting points (initial deformation initiation observation points or boundary nodes) of the three-dimensional deformation evolution sub-paths within the region of the subsequent propagation area through corresponding inter-regional deformation diffusion models. During the connection process, the starting direction of the sub-paths in the subsequent propagation area is adjusted according to the diffusion path direction of the model to ensure that the deformation diffusion direction is consistent with the model prediction; the time axis of the sub-paths in the subsequent propagation area is adjusted according to the diffusion rate of the model, and an inter-regional propagation delay is added to ensure that the deformation status of adjacent sub-paths is continuous in time and space. For adjacent regions with bidirectional synchronous propagation, the sub-paths of the two regions are connected according to their spatial positions to ensure the synergy of the deformation status. After completing the connection of all adjacent regions, an initial three-dimensional deformation evolution path covering the entire hydropower station slope is formed. The path is stored in a multi-layer graph structure, with the first layer being the regional level connection and the second layer being the regional observation point level connection.

[0161] Step S149: Optimize the initial three-dimensional deformation evolution path by adjusting the deformation diffusion rate data and path direction data between regions based on the deformation evolution records in the historical three-dimensional observation data, so that the evolution path matches the actual deformation evolution process.

[0162] In this embodiment, historical deformation evolution records with similar deformation characteristics to the current deformation over the past three years are extracted from the historical database of the hydropower station slope safety monitoring system. These records include data on the transmission path, diffusion rate, and directional changes of historical deformation. The initial three-dimensional deformation evolution path is compared with the historical records, and the deviation angle of the path direction and the deviation ratio of the diffusion rate are calculated. If the deviation of the current path direction from the average direction of the historical records exceeds a preset angle threshold, the path direction of the inter-regional deformation diffusion model is fine-tuned according to the directional distribution of the historical records, with an adjustment range of 30% of the deviation angle. If the deviation of the current diffusion rate from the average rate of the historical records exceeds a preset ratio threshold, the diffusion rate is fine-tuned according to the rate distribution of the historical records, with an adjustment range of 20% of the deviation ratio. Simultaneously, combined with the real-time soil and rock mass status data of the current slope (such as crack development and water content changes), the fine-tuned model parameters are adjusted a second time. If real-time monitoring shows the expansion of soil and rock mass cracks, the diffusion rate is appropriately increased. If real-time monitoring shows a significant increase in water content, the diffusion direction is appropriately adjusted to tilt towards the weak layer of the soil and rock mass.

[0163] Step S1410: Integrate the optimized three-dimensional deformation evolution sub-paths within the region and cross-regional connection relationships to generate a complete three-dimensional deformation evolution path covering the entire hydropower station slope. The complete three-dimensional deformation evolution path includes the continuous transmission trend of spatial position changes of observation points and the deformation diffusion path between regions.

[0164] In this embodiment, all optimized 3D deformation evolution sub-paths within the region are reconnected through an adjusted inter-regional deformation diffusion model, ensuring complete continuity of sub-paths in adjacent regions in terms of spatial direction, time axis, and deformation status. The integrated complete 3D deformation evolution path comprises three core layers: the first layer is a regional-level deformation diffusion network, displaying the diffusion paths, directions, and rates between observation areas; the second layer is a regional transmission path network, displaying the transmission order and correlation strength between observation points within the region; and the third layer is observation point-level continuous status data, displaying the directional continuity and rate stability of each observation point throughout the entire observation period. Simultaneously, a time-dimensional animation attribute is added to the path, enabling dynamic visualization of the deformation evolution process. The complete 3D deformation evolution path is stored in both standard Geographic Information System (GIS) format and graph structure format, supporting analysis in both spatial geographic dimensions and transmission relationship dimensions.

[0165] Step S150: Generate three-dimensional deformation early warning information for the hydropower station slope based on the three-dimensional deformation evolution path. The three-dimensional deformation early warning information includes a description of the deformation evolution trend and an early warning trigger identifier. Send the three-dimensional deformation early warning information to the hydropower station safety monitoring terminal.

[0166] Step S151: Analyze the continuous transmission trend of the spatial position change of the observation points in the three-dimensional deformation evolution path, and extract the overall deformation transmission direction, average transmission rate data and cumulative deformation offset data of each observation area according to the observation area classification.

[0167] In this embodiment, for each observation region in the complete three-dimensional deformation evolution path, the spatial position change trajectory data of all observation points within the region are extracted. The overall deformation transmission direction of the region, Dir_region, is calculated: the change direction vector of each observation point is weighted and summed according to the transmission correlation strength, and then standardized to obtain the unit direction vector, which serves as the overall transmission direction of the region. The average transmission rate of the region, Rate_region, is calculated: the change rate of each observation point is weighted and summed according to the transmission correlation strength to obtain the average transmission rate of the region, in meters per second. The cumulative deformation offset data, Disp_region, is calculated: the total offset of each observation point over the entire observation period is weighted and summed according to the transmission correlation strength to obtain the cumulative deformation offset of the region, in meters. Simultaneously, the fluctuation coefficient of the deformation state within the region is statistically analyzed, and the standard deviation of the direction continuity parameter and the standard deviation of the rate stability parameter are calculated as auxiliary indicators of the deformation stability of the region.

[0168] Step S152: Based on the overall deformation transmission direction, average transmission rate data and cumulative deformation offset data of each observation area, generate a deformation evolution trend description of the observation area. The deformation evolution trend description includes the extension trend of the deformation transmission direction and the rate change trend.

[0169] In this embodiment, for each observation area, the overall deformation transmission direction (Dir_region) is described using topographic data of the slope, for example, "The overall deformation of the region extends along the positive X-axis (lateral direction of the slope), pointing towards the left side of the dam body." For the average transmission rate (Rate_region), historical rate data is used to describe the rate change trend, for example, "The average transmission rate of the region has increased by 20% compared to the same period in history, showing an accelerating transmission trend." For the cumulative deformation offset data (Disp_region), a safety threshold is used to describe the offset trend, for example, "The cumulative deformation offset of the region has reached 75% of the safety threshold, approaching the warning threshold." Simultaneously, an auxiliary description of deformation stability is added, for example, "The regional deformation direction fluctuation coefficient is 0.1, indicating overall stability; the rate fluctuation coefficient is 0.2, showing signs of local acceleration." These descriptions are integrated into a structured deformation evolution trend description text, which is then polished using a natural language generation model to ensure accuracy, clarity, and compliance with engineering expression habits.

[0170] Step S153: Obtain the safety deformation control standard for the hydropower station slope. The safety deformation control standard includes the upper limit of cumulative deformation offset and the upper limit of average conduction rate corresponding to different observation areas, which are determined by the hydropower station slope safety design code.

[0171] In this embodiment, safety deformation control standards for different observation areas are extracted from the hydropower station slope safety design documents and industry standards: For the primary protection area around the core dam body, a cumulative deformation offset upper limit Disp_thresh_1 and an average conduction rate upper limit Rate_thresh_1 are set; for the secondary protection area of ​​the dam abutment, a cumulative deformation offset upper limit Disp_thresh_2 and an average conduction rate upper limit Rate_thresh_2 are set, where Disp_thresh_2 > Disp_thresh_1 and Rate_thresh_2 > Rate_thresh_1; for the tertiary protection area of ​​the downstream slope, a cumulative deformation offset upper limit Disp_thresh_3 and an average conduction rate upper limit Rate_thresh_3 are set, with the thresholds further relaxed. Simultaneously, a trend warning threshold is added; for example, a trend warning is triggered when the cumulative deformation offset reaches 80% of the upper limit or the average conduction rate reaches 90% of the upper limit.

[0172] Step S154: Compare the cumulative deformation offset data of each observation area with the corresponding upper limit of cumulative deformation offset, and at the same time compare the average conduction velocity data of each observation area with the corresponding upper limit of average conduction velocity.

[0173] In this embodiment, for each observation region, the cumulative deformation offset data Disp_region is compared with the corresponding upper limit of cumulative deformation offset Disp_thresh, and the offset ratio Disp_ratio = Disp_region / Disp_thresh is calculated; the average conduction velocity data Rate_region is compared with the corresponding upper limit of average conduction velocity Rate_thresh, and the rate ratio Rate_ratio = Rate_region / Rate_thresh is calculated. The Disp_ratio and Rate_ratio for each region are recorded, along with a preliminary determination of whether the threshold is exceeded: if Disp_region > Disp_thresh, it is marked as "displacement exceeds the limit"; if Rate_region > Rate_thresh, it is marked as "rate exceeds the limit"; if neither exceeds the limit, it is marked as "normal". Simultaneously, the triggering conditions for trend warnings are calculated: if Disp_ratio ≥ 0.8, it is marked as "displacement trend warning"; if Rate_ratio ≥ 0.9, it is marked as "rate trend warning".

[0174] Step S155: For observation areas where the cumulative deformation offset data is outside the corresponding upper limit of the cumulative deformation offset, mark the displacement over-limit warning trigger flag; for observation areas where the average conduction velocity data is outside the corresponding upper limit of the average conduction velocity, mark the velocity over-limit warning trigger flag; the warning trigger flag includes the observation area flag and specific over-limit type information.

[0175] In this embodiment, for each observation region, if Disp_region > Disp_thresh, a displacement exceeding the limit warning trigger flag is generated. The flag includes the observation region ID, region name, current cumulative deformation offset, corresponding upper limit value, and exceeding ratio (Disp_ratio-1) × 100%. If Rate_region > Rate_thresh, a rate exceeding the limit warning trigger flag is generated. The flag includes the observation region ID, region name, current average conduction rate, corresponding upper limit value, and exceeding ratio (Rate_ratio-1) × 100%. If both exceed the limit, two warning trigger flags are generated simultaneously. For regions triggering trend warnings, a trend warning trigger flag is generated. The flag includes the observation region ID, region name, current offset ratio or rate ratio, and corresponding trend warning threshold. The warning trigger flags are stored in structured JSON format, and corresponding visual icons and color codes are generated (e.g., red indicates exceeding the limit warning, yellow indicates a trend warning).

[0176] Step S156: Collect information on all observation areas marked with warning trigger indicators, including observation area indicators, exceedance type information and corresponding deformation evolution trend descriptions, and generate a preliminary three-dimensional deformation warning information set.

[0177] In this embodiment, all observation area information marked with warning trigger identifiers is collected. The information for each area includes: observation area ID, area name, warning trigger identifier (type, exceedance ratio, threshold data), corresponding deformation evolution trend description text, and the area's geographical boundary data. This information is integrated into a preliminary three-dimensional deformation warning information set, and initially sorted by warning level: exceedance warnings (red) take precedence over trend warnings (yellow), and warnings with higher exceedance ratios are ranked higher than warnings of the same level. The preliminary warning information set is stored in list form, with each element representing complete warning information for one area.

[0178] Step S157: Based on the deformation evolution trend description of the observation area of ​​each marked early warning trigger identifier, extract deformation diffusion risk parameters, which include deformation diffusion range prediction data and diffusion rate prediction data.

[0179] In this embodiment, for each observation area marked with an early warning trigger, the current transmission direction and rate data are extracted from the deformation evolution trend description. Combined with an inter-regional deformation diffusion model, the deformation diffusion range within the future ΔT_pred time period is predicted: the current transmission direction is extended, and the geographical boundary of the predicted diffusion is delineated based on the topographical constraints of the slope; the current average transmission rate is multiplied by ΔT_pred to obtain the predicted diffusion distance, and the fluctuation range of the diffusion distance is adjusted (±15%) based on the regional soil and rock mass state data. Simultaneously, based on the historical acceleration pattern of deformation, the trend of diffusion rate change is predicted: if the current rate shows an upward trend, the predicted future rate will increase by 10%-20% compared to the current rate; if the current rate is stable, the predicted future rate will remain unchanged; if the current rate shows a downward trend, the predicted future rate will decrease by 5%-10% compared to the current rate.

[0180] Step S158: Sort the preliminary three-dimensional deformation early warning information set according to the deformation diffusion risk parameters. Early warning information that conforms to the preset sorting rules is arranged at the front, and the display priority of early warning information in different observation areas is determined.

[0181] In this embodiment, the sorting rules for early warning information are set as follows: First, they are sorted by warning level, with over-limit warnings taking precedence over trend warnings; warnings of the same level are sorted by the predicted value of deformation diffusion range, with larger diffusion ranges listed first; if the diffusion ranges are similar, they are sorted by the predicted increase in diffusion rate, with larger increase rates listed first; if the above parameters are similar, they are sorted by the cumulative deformation offset ratio or rate ratio, with higher exceedance ratios listed first. For each early warning information in the preliminary three-dimensional deformation early warning information set, the corresponding deformation diffusion risk parameters and warning level are extracted, and weighted according to the above rules to calculate the sorting score for each early warning information: level score (over-limit warnings get 10 points, trend warnings get 5 points) + diffusion range score (linearly mapped to 0-3 points based on diffusion range size) + rate trend score (linearly mapped to 0-2 points based on increase rate) + exceedance ratio score (linearly mapped to 0-3 points based on exceedance ratio), with a total score range of 0-18 points. The early warning information is sorted from highest to lowest score to determine the final display priority, ensuring that the warning information with the highest risk is noticed by monitoring personnel as soon as possible.

[0182] Step S159: Organize the preliminary three-dimensional deformation early warning information set according to the display priority, supplement the geographical boundary data of the observation area corresponding to each early warning information and the observation record data of the associated measurement robot, and generate complete three-dimensional deformation early warning information.

[0183] In this embodiment, the initial set of 3D deformation early warning information is reorganized according to the final determined display priority. For each early warning information, the geographic boundary data (latitude and longitude coordinates) of the observation area is supplemented to ensure accurate location of the early warning information on the GIS map; the observation record data of the associated measurement robot is supplemented, including the original observation coordinates, data acquisition time, measurement robot ID, and status (normal / abnormal) within the corresponding time period; the early warning information generation time and data update time are supplemented to ensure the timeliness of the information. At the same time, a processing suggestion module is added to each early warning information, generating targeted processing suggestions based on industry standards and historical handling experience, such as "It is recommended to increase the observation frequency of this area to once every 15 minutes and increase the number of manual patrols." The above information is integrated into a complete 3D deformation early warning information, stored in both HTML and JSON formats, supporting both visual display and system interface calls.

[0184] Step S1510: Establish a dedicated data transmission link between the complete three-dimensional deformation early warning information and the hydropower station safety monitoring terminal. Using a preset encrypted transmission method, send the complete three-dimensional deformation early warning information to the hydropower station safety monitoring terminal, so that the monitoring terminal displays the early warning information according to the display priority.

[0185] In this embodiment, a dedicated transmission link is established between the data processing center and the hydropower station safety monitoring terminal using Virtual Private Network (VPN) technology to ensure the independence and security of data transmission. Before transmission, the complete 3D deformation early warning information is encrypted using the AES-256 encryption algorithm, with the encryption key employing a dynamic negotiation mechanism and updated hourly. During transmission, a breakpoint resumption mechanism is used to ensure the integrity of large file transfers; a checksum mechanism is also included, where the receiving end verifies the received information to ensure data integrity. Upon receiving the encrypted information, the monitoring terminal decrypts it and displays the early warning information on the GIS interface according to display priority: the geographical boundaries of each early warning area are highlighted with a bright color (red / yellow). Clicking on a highlighted area brings up a detailed early warning information panel, including trend descriptions, early warning indicators, risk parameters, original observation data, and processing suggestions. Simultaneously, audible and visual early warning prompts are triggered; the highest priority early warning information triggers a red audible and visual alarm, while trend early warnings trigger a yellow audible and visual alarm, ensuring timely response by monitoring personnel. Log information during the transmission and display process is recorded in real-time to the security audit system.

[0186] Based on the same inventive concept, please refer to Figure 2 The diagram shows a schematic block diagram of a three-dimensional deformation early warning system 100 for hydropower station slopes based on a measurement robot, provided in an embodiment of this application. The three-dimensional deformation early warning system 100 for hydropower station slopes based on a measurement robot may include a communication unit 110, a machine-readable storage medium 120, and a processor 130.

[0187] In this embodiment, the machine-readable storage medium 120 can also be integrated into the processor 130 and can communicate and interact with external systems through the communication unit 110.

[0188] The processor 130 is the control center of the three-dimensional deformation early warning system 100 for hydropower station slopes based on the measurement robot. It is connected to the entire system using various interfaces and lines. The machine-readable storage medium 120 is used to store machine-executable instructions for executing the scheme of this application. The processor 130 is used to execute the machine-executable instructions stored in the machine-readable storage medium 120 to implement the three-dimensional deformation early warning method for hydropower station slopes based on the measurement robot provided in the aforementioned method embodiment.

[0189] It should be noted that, in order to simplify the description of the present invention and thus help to understand one or more embodiments of the invention, multiple features may sometimes be grouped into one embodiment, drawing or description thereof in the foregoing description of the embodiments of the present invention.

Claims

1. A method for early warning of three-dimensional deformation of hydropower station slopes based on a measurement robot, characterized in that, The method includes: A measurement robot is deployed to continuously observe the slope of the hydropower station and obtain three-dimensional observation data of the slope. The three-dimensional observation data includes the spatial location information and observation time information of each observation point on the slope surface. Based on the spatial location information and observation time information of each observation point in the three-dimensional observation data, a deformation transmission link is established between all observation points of the hydropower station slope. The deformation transmission link includes the deformation transmission direction and transmission association information between the observation points. Based on the deformation transmission link, the cross-regional deformation transmission characteristics of the hydropower station slope are extracted. The cross-regional deformation transmission characteristics include the deformation transmission sequence and transmission association status data of observation points in different observation areas. By integrating the cross-regional deformation transmission characteristics and the spatial position change trajectory of each observation point in the three-dimensional observation data, a three-dimensional deformation evolution path of the hydropower station slope is constructed. The three-dimensional deformation evolution path includes the continuous transmission trend of spatial position change of the observation points and the deformation diffusion path between regions. Based on the three-dimensional deformation evolution path, three-dimensional deformation early warning information of the hydropower station slope is generated. The three-dimensional deformation early warning information includes a description of the deformation evolution trend and an early warning trigger identifier. The three-dimensional deformation early warning information is sent to the hydropower station safety monitoring terminal.

2. The method for early warning of three-dimensional deformation of hydropower station slopes based on a measurement robot according to claim 1, characterized in that, The process of establishing a deformation transfer link between all observation points on the hydropower station slope based on the spatial location and observation time information of each observation point in the three-dimensional observation data includes: The spatial location information of each observation point is extracted from the three-dimensional observation data. The actual spatial distance between any two observation points is calculated based on the spatial location information to generate a set of spatial distances between observation points. The set of spatial distances between observation points contains the actual spatial distance data between all pairs of observation points. Combining the observation time information in the three-dimensional observation data, the spatial position change sequence of each observation point in continuous observation period is extracted in chronological order. The spatial position change sequence includes the spatial position offset data of each observation point at different observation time points. Based on the spatial position change sequence of each observation point, the deformation correlation factor of the observation point is extracted. The deformation correlation factor includes the consistent state of the spatial position change direction and the coordinated state of the change rate of the observation point. It is generated by comparing the spatial position change direction and change rate of different observation points in the same observation period. The actual spatial distance data in the observation point spatial distance set is compared with the preset deformation correlation distance threshold, and observation point pairs whose actual spatial distance data are within the range of the preset deformation correlation distance threshold are selected to generate a potential deformation correlation observation point pair set. For each pair of observation points in the set of potential deformation-related observation points, the deformation transfer intensity between the observation point pairs is calculated by combining the corresponding deformation-related factor. The deformation transfer intensity is generated by weighting the state parameters of consistent spatial position change direction and the state parameters of coordinated change rate with preset weight coefficients, and then mapping them to a unified dimensionless evaluation scale. Based on the preset deformation initiation end determination rule, the observation point whose deformation transmission intensity value satisfies the deformation initiation end determination rule is taken as the deformation initiation end, and the other observation point is taken as the deformation receiving end, thus generating a one-way transmission relationship between the observation point pairs. Based on the unidirectional transmission relationship of all potential deformation-related observation point pairs, an initial deformation transmission link for the hydropower station slope is constructed. The initial deformation transmission link includes the transmission direction and corresponding deformation transmission intensity data of all observation point pairs. The initial deformation transmission link is optimized by removing observation point pairs whose deformation transmission intensity values ​​are outside the preset transmission intensity threshold range and supplementing observation point pairs that were missed due to spatial distance screening but whose deformation correlation factors meet the preset correlation factor requirements. The optimized observation point pairs are classified and integrated according to the observation area to generate regional deformation transfer sub-links. Each regional deformation transfer sub-link contains the transfer relationship and transfer intensity data of all observation points in the corresponding observation area. All regional deformation transfer sub-links are connected across regions. Based on the deformation transfer relationship of observation points at the boundary of adjacent observation areas, a complete deformation transfer link covering all observation points on the entire hydropower station slope is constructed. The complete deformation transfer link includes the deformation transfer direction and transfer association information between observation points.

3. The method for early warning of three-dimensional deformation of hydropower station slopes based on a measurement robot according to claim 1, characterized in that, The method for excavating the cross-regional deformation transmission characteristics of hydropower station slopes based on the deformation transmission link includes: The deformation transmission sub-links in the deformation transmission link are analyzed, and the core observation points in each sub-link are extracted. The core observation points refer to the observation points in the sub-links whose deformation transmission intensity data meet the core observation point judgment criteria and whose number of connected observation points meets the core observation point connection requirements. Based on the core observation point of each sub-region deformation transmission sub-link, trace the deformation transmission source of the core observation point, and determine the initial deformation initiation observation point of the core observation point by traversing the transmission relationship in the sub-region deformation transmission sub-link in reverse. Extract the transmission path between the core observation point and the initial deformation initiation observation point in each sub-regional deformation transmission sub-link, record all intermediate observation points and corresponding deformation transmission intensity data in the transmission path, and generate the main deformation transmission path in the region. Analyze the main deformation transmission path within different observation areas, and identify the boundary observation points in the main deformation transmission path within adjacent observation areas. The boundary observation point refers to the observation point located at the junction of two adjacent observation areas. The deformation transmission correlation degree between boundary observation points of adjacent observation areas is calculated. The deformation transmission correlation degree is generated by comparing the deformation transmission direction, transmission intensity data and spatial position relationship of the boundary observation points, and integrating the multi-dimensional correlation parameters of the boundary observation points. Based on the preset deformation transmission order determination rule, the observation area whose deformation transmission correlation value satisfies the deformation transmission order determination rule is taken as the deformation first transmission area, and the other observation area is taken as the deformation later transmission area, thus generating a cross-region deformation transmission direction. For each cross-regional deformation transmission direction, the corresponding boundary observation point transmission relationship is extracted, and the main deformation transmission paths within adjacent observation areas are integrated to generate a cross-regional deformation transmission main path. The cross-regional deformation transmission main path includes all observation points and transmission order of cross-regional transmission. Based on the cross-regional deformation transmission main path, the transmission correlation status data between different observation areas is calculated. The transmission correlation status data is generated by integrating the deformation transmission intensity data of all observation point pairs in the cross-regional deformation transmission main path, reflecting the close relationship between the regions in deformation transmission. Extract the transmission sequence information of all cross-regional deformation transmission main paths, organize the deformation transmission start time of different observation areas according to the observation time, and generate deformation transmission sequence data of observation points in different observation areas; By integrating the deformation transmission sequence data of observation points in different observation areas and the transmission correlation status data between areas, a cross-regional deformation transmission characteristic of the hydropower station slope is generated. The cross-regional deformation transmission characteristic includes the deformation transmission sequence and transmission correlation status data of observation points in different observation areas.

4. The method for early warning of three-dimensional deformation of hydropower station slopes based on a measurement robot according to claim 1, characterized in that, The method of integrating the cross-regional deformation transmission characteristics and the spatial position change trajectory of each observation point in the three-dimensional observation data to construct the three-dimensional deformation evolution path of the hydropower station slope includes: Extract all spatial location information and corresponding observation time information of each observation point from the three-dimensional observation data, concatenate the spatial location information of each observation point in the order of observation time, and generate the spatial location change trajectory of each observation point. The spatial location change trajectory contains the continuous spatial location change data of the observation point throughout the entire observation period. The deformation transmission sequence of observation points in different observation areas in the cross-regional deformation transmission characteristics is analyzed, and the spatial position change trajectory of each observation point is calibrated on the time axis according to the deformation transmission sequence, so that the spatial position change trajectory of observation points on the same transmission path remains synchronized in the time dimension. Based on the calibrated spatial position change trajectory of the observation point, the continuous trend of spatial position change of each observation point is extracted. The continuous trend of spatial position change includes the continuous state of the spatial position change direction and the stable state of the change rate of the observation point during the continuous observation period. Combining the transmission-related state data in the cross-regional deformation transmission characteristics, the continuous trend of spatial position changes of observation points on the same cross-regional deformation transmission main path is weighted and integrated, and the weights corresponding to the transmission-related state data are used to strengthen the influence of the changing trend of key transmission nodes. Extract inter-regional deformation transmission nodes from the main path of cross-regional deformation transmission. The inter-regional deformation transmission nodes refer to deformation transmission observation points at the boundaries of adjacent observation areas. Integrate the continuous trend of spatial position changes of the inter-regional deformation transmission nodes to generate the inter-regional deformation transition trend. Based on the deformation transition situation between regions, an inter-regional deformation diffusion model is constructed. The inter-regional deformation diffusion model includes the path direction and diffusion rate data of deformation diffusion from one observation region to adjacent observation regions, which is generated by integrating the changing situation of boundary observation points. For each observation area, a three-dimensional deformation evolution sub-path is constructed by combining the main deformation transmission path within the area and the corresponding continuous trend of spatial position change. The three-dimensional deformation evolution sub-path within the area includes the continuous trend of spatial position change of all observation points within the observation area and the transmission path within the area. The three-dimensional deformation evolution sub-paths within each observation area are connected across regions through an inter-regional deformation diffusion model, so that the evolution sub-paths of adjacent observation areas are connected according to the deformation diffusion path, generating an initial three-dimensional deformation evolution path. The initial three-dimensional deformation evolution path is optimized by adjusting the deformation diffusion rate data and path direction data between regions based on the deformation evolution records in historical three-dimensional observation data, so that the evolution path fits the actual deformation evolution process. By integrating and optimizing the sub-paths of three-dimensional deformation evolution within the region and the cross-regional connection relationships, a complete three-dimensional deformation evolution path covering the entire slope of the hydropower station is generated. The complete three-dimensional deformation evolution path includes the continuous transmission trend of spatial position changes of observation points and the deformation diffusion path between regions.

5. The method for early warning of three-dimensional deformation of hydropower station slopes based on a measurement robot according to claim 2, characterized in that, The extraction of deformation correlation factors for each observation point based on its spatial location change sequence includes: From the spatial position change sequence of each observation point, the observation period is divided into equal time intervals. Each observation period contains a preset number of continuous observation time points, generating a segmented spatial position change sequence for each observation point. For each segmented spatial position change sequence, calculate the spatial position change direction vector of the observation point within the observation period. The spatial position change direction vector is generated by connecting the spatial position coordinates of the start and end observation time points of the observation period. Extract the spatial position change direction vectors of all observation points within the same observation period, and construct a direction vector set, which contains the spatial position change direction data of all observation points within the observation period. Calculate the angle between any two direction vectors in the set of direction vectors to generate direction angle data, and use the direction angle data to determine whether the spatial position changes of the two observation points are in the same direction during the observation period. Based on the directional angle data, a spatial position change direction consistency state parameter is generated. When the directional angle data is within the preset directional angle range, the spatial position change direction consistency state parameter is generated according to the preset directional consistency parameter mapping rule. For each segmented spatial position change sequence, the rate of spatial position change of the observation point within the observation period is calculated. The rate of spatial position change is generated by the ratio of the total spatial position offset within the observation period to the duration of the observation period. Extract the rate of change of spatial location of all observation points within the same observation period and construct a rate set, which contains the rate of change of spatial location of all observation points within the observation period. The ratio of any two rate data in the rate set is calculated to generate rate ratio data. The rate ratio data is used to determine the coordinated state of the spatial position change rate of the two observation points during the observation period. Based on the rate ratio data, a coordinated state parameter for the rate of change of spatial position is generated. When the difference between the rate ratio data and the preset coordinated ratio is within the preset rate ratio difference range, the coordinated state parameter for the rate of change of spatial position is generated according to the preset rate coordinated parameter mapping rule. The spatial position change direction consistency state parameter and spatial position change rate coordination state parameter of any two observation points within the same observation period are used as independent dimensions of the deformation correlation factor of the observation point pair to generate the deformation correlation factor of the observation point. The deformation correlation factor includes two independent parameter groups: the spatial position change direction consistency state and the change rate coordination state of the observation point.

6. The method for early warning of three-dimensional deformation of hydropower station slopes based on a measurement robot according to claim 3, characterized in that, The deformation transmission correlation degree between boundary observation points of adjacent observation areas is calculated by comparing the deformation transmission direction, transmission intensity data, and spatial positional relationship of the boundary observation points, integrating multi-dimensional correlation parameters of the boundary observation points, including: Extract the sub-links of deformation transfer in adjacent observation areas, select the boundary observation points located at the boundary of the observation area from the sub-links of deformation transfer in the observation areas, and record the observation area affiliation and spatial location coordinates of each boundary observation point; Obtain deformation transfer direction data for each boundary observation point in the deformation transfer link. The deformation transfer direction data includes the transfer direction vector of the boundary observation point as the deformation initiator or receiver. By comparing the deformation transmission direction vectors of the boundary observation points of adjacent observation areas, the coincidence degree of the direction vectors is calculated. The coincidence degree of the direction vectors is generated by integrating the corresponding parameters of the two direction vectors, reflecting the matching state of the transmission direction of the boundary observation points. Extract the deformation transfer intensity data corresponding to each boundary observation point. The deformation transfer intensity data is the transfer intensity value of the boundary observation point in the deformation transfer link, which reflects the deformation transfer capability of the observation point. The deformation transfer intensity matching degree of adjacent observation area boundary observation points is calculated. It is generated by comparing the difference between the deformation transfer intensity values ​​of two boundary observation points with the preset intensity difference standard. When the difference is within the preset intensity difference standard range, the matching degree meets the requirements. The actual spatial distance between boundary observation points of adjacent observation areas is measured. The actual spatial distance is calculated using the spatial coordinate data of the boundary observation points and reflects the spatial proximity of the boundary observation points. The actual spatial distance is compared with a preset boundary association distance threshold to generate spatial location association parameters. When the actual spatial distance is within the preset boundary association distance threshold range, spatial location association parameters are generated according to the preset spatial association parameter mapping rules. Based on the aforementioned direction vector coincidence, deformation transmission intensity matching degree, and spatial location correlation parameters, weight coefficients for each parameter are set, and these weight coefficients are determined through deformation influence priority analysis of boundary observation points. The direction vector coincidence, deformation transmission intensity matching degree, and spatial location correlation parameters are normalized to the same preset evaluation interval to form comparable dimensionless evaluation values. Based on the weight coefficients of each parameter, the normalized dimensionless evaluation values ​​are weighted and fused to generate the preliminary deformation transmission correlation degree between observation points at the boundary of adjacent observation areas. The preliminary deformation transmission correlation is normalized to generate the final deformation transmission correlation between observation points at the boundary of adjacent observation areas.

7. The method for early warning of three-dimensional deformation of hydropower station slopes based on a measurement robot according to claim 4, characterized in that, The extraction of the continuous spatial position change trend of each observation point based on the calibrated observation point spatial position change trajectory includes: The spatial location change trajectory of the calibrated observation points is divided into multiple consecutive analysis periods in chronological order. Each analysis period contains multiple consecutive observation time points, so that the time span of each analysis period remains the same. For each analysis period, the spatial position change direction of the observation point between all adjacent observation time points within that analysis period is extracted to generate a direction sequence for that analysis period. The direction sequence contains multiple consecutive spatial position change direction data. The directional deviation of adjacent directional data in the directional sequence of each analysis period is calculated. The directional deviation is generated by the angle between the changing directions of two adjacent spatial positions, reflecting the fluctuation state of directional changes. Based on the direction deviation data for each analysis period, the direction continuous state parameter is calculated. When the statistical result of the direction deviation data is within the preset direction deviation interval, the direction continuous state parameter is generated according to the preset direction continuous parameter mapping rule. For each analysis period, the rate of change of spatial location of the observation point between all adjacent observation time points within that analysis period is calculated, and a rate sequence for that analysis period is generated, which contains multiple consecutive spatial location change rate data. The rate difference between adjacent rate data in the rate sequence of each analysis period is calculated. The rate difference is generated by the numerical difference between the rates of change of two adjacent spatial locations, reflecting the fluctuation state of rate change. Based on the rate difference data for each analysis period, the rate steady state parameter is calculated. When the statistical result of the rate difference data is within the preset rate difference range, the rate steady state parameter is generated according to the preset rate steady state parameter mapping rule. Arrange the continuous directional state parameters of all analysis periods in chronological order to generate a sequence of continuous directional state parameters. The sequence of continuous directional state parameters contains the continuous directional change data of the observation point throughout the entire observation period. Arrange the rate steady-state parameters of all analysis periods in chronological order to generate a rate steady-state parameter sequence, which contains the rate steady-state change data of the observation points throughout the entire observation period; By integrating the directional continuous state parameter sequence and the rate stable state parameter sequence, a continuous situation of spatial position change for each observation point is generated. The continuous situation of spatial position change includes the directional continuous state and rate stable state data of spatial position change of the observation point during the continuous observation period.

8. The method for early warning of three-dimensional deformation of hydropower station slopes based on a measurement robot according to claim 1, characterized in that, The process of generating three-dimensional deformation early warning information for the hydropower station slope based on the three-dimensional deformation evolution path, wherein the three-dimensional deformation early warning information includes a description of the deformation evolution trend and an early warning trigger identifier, and sending the three-dimensional deformation early warning information to the hydropower station safety monitoring terminal includes: The continuous transmission trend of spatial position changes of observation points in the three-dimensional deformation evolution path is analyzed, and the overall deformation transmission direction, average transmission rate data and cumulative deformation offset data of each observation area are extracted according to the observation area classification. Based on the overall deformation transmission direction, average transmission rate data, and cumulative deformation offset data of each observation area, a deformation evolution trend description of the observation area is generated. The deformation evolution trend description includes the extension trend of the deformation transmission direction and the rate change trend. The safety deformation control standard for hydropower station slopes is obtained. The safety deformation control standard includes the upper limit of cumulative deformation offset and the upper limit of average conduction rate corresponding to different observation areas, which are determined by the hydropower station slope safety design code. The cumulative deformation offset data of each observation area is compared with the corresponding upper limit of cumulative deformation offset, and the average conduction velocity data of each observation area is compared with the corresponding upper limit of average conduction velocity. For observation areas where the cumulative deformation offset data is outside the corresponding upper limit of the cumulative deformation offset range, a displacement over-limit warning trigger flag is marked; for observation areas where the average conduction velocity data is outside the corresponding upper limit of the average conduction velocity range, a velocity over-limit warning trigger flag is marked; the warning trigger flag includes the observation area flag and specific over-limit type information; Collect all observation area information marked with warning trigger indicators, including observation area indicators, exceedance type information and corresponding deformation evolution trend descriptions, and generate a preliminary three-dimensional deformation warning information set; Based on the deformation evolution trend description of the observation area for each marked early warning trigger identifier, deformation diffusion risk parameters are extracted, which include deformation diffusion range prediction data and diffusion rate prediction data. The preliminary three-dimensional deformation early warning information set is sorted according to the deformation diffusion risk parameters. Early warning information whose deformation diffusion risk parameters conform to the preset sorting rules is arranged at the front, and the display priority of early warning information in different observation areas is determined. Organize the preliminary three-dimensional deformation early warning information set according to the display priority, supplement the geographical boundary data of the observation area corresponding to each early warning information and the observation record data of the associated measurement robot, and generate complete three-dimensional deformation early warning information; Establish a dedicated data transmission link between complete three-dimensional deformation early warning information and the hydropower station safety monitoring terminal. Use a preset encrypted transmission method to send complete three-dimensional deformation early warning information to the hydropower station safety monitoring terminal, so that the monitoring terminal can display the early warning information according to the display priority.

9. The method for early warning of three-dimensional deformation of hydropower station slopes based on a measuring robot according to claim 5, characterized in that, For each segmented spatial position change sequence, the calculation of the spatial position change direction vector of the observation point within the observation period includes: From the segmented spatial position change sequence of each observation point, extract the spatial position coordinates corresponding to the starting observation time point of the observation period and record them as starting coordinate data. The starting coordinate data includes the horizontal coordinate, vertical coordinate and vertical coordinate in three-dimensional space. Extract the spatial coordinates corresponding to the end observation time point of the observation period and record them as end coordinate data. The end coordinate data includes the horizontal coordinate, vertical coordinate and vertical coordinate in three-dimensional space. Calculate the difference between the starting coordinate data and the ending coordinate data in the horizontal direction. The horizontal coordinate difference is generated by subtracting the horizontal coordinate value of the starting coordinate data from the horizontal coordinate value of the ending coordinate data. Calculate the difference between the starting coordinate data and the ending coordinate data in the vertical direction. The vertical coordinate difference is generated by subtracting the vertical coordinate value of the starting coordinate data from the vertical coordinate value of the ending coordinate data. Calculate the vertical coordinate difference between the starting and ending coordinate data. The vertical coordinate difference is generated by subtracting the vertical coordinate value of the starting coordinate data from the vertical coordinate value of the ending coordinate data. Based on the horizontal coordinate difference, vertical coordinate difference, and vertical coordinate difference, the original data for constructing the spatial position change vector is generated. The original data includes coordinate difference information in the three coordinate directions. The vector magnitude of the original data is calculated and generated by integrating the square root of the sum of the squares of the horizontal coordinate difference, the vertical coordinate difference, and the vertical coordinate difference, reflecting the total offset of the spatial position change; Divide the differences in the horizontal, vertical, and longitudinal coordinates in the original data by the vector magnitude to obtain the standardized horizontal, vertical, and longitudinal components, so that the vector magnitude is uniformly set to a preset value. Based on the standardized horizontal, vertical, and longitudinal components, a direction vector of spatial position change of the observation point during the observation period is constructed, and the direction vector contains three standardized component data. The constructed spatial position change direction vector is verified by confirming that the sum of the squares of the three standardized components meets the preset numerical requirements, and the final spatial position change direction vector is generated.

10. A three-dimensional deformation early warning system for hydropower station slopes based on a measurement robot, characterized in that, include: processor; A machine-readable storage medium for storing machine-executable instructions of the processor; The processor is configured to execute the three-dimensional deformation early warning method for hydropower station slopes based on a measurement robot as described in any one of claims 1 to 9 by executing the machine-executable instructions.

Images