Water conservancy monitoring and early warning method and system based on GIS platform

Abstract

The invention provides a water conservancy monitoring and early warning method and system based on a GIS platform, and the method comprises the following steps: laying distributed optical fiber sensors along the structural axis of a target water conservancy project, and collecting strain and sound wave data; a one-dimensional linear reference system is created in a GIS based on a structure axis, optical fiber measurement points are mapped into a plurality of spatial analysis units corresponding to an engineering physical structure, and accurate hooking of data and positions is realized. The internal erosion problem is identified by performing signal processing and acoustic fingerprint matching on the acoustic wave data of each unit. And dynamically calculating the real-time effective bearing capacity by combining the strain data and the structure attribute information of the unit. A risk evolution threshold model is constructed, and when and only when an internal erosion problem is monitored in the same spatial analysis unit, and a composite risk state passes through an evolution threshold, point location early warning information pointing to the specific spatial unit is generated and output. The method has the effect of high early warning accuracy.

Description

Technical Field

[0001] The present invention belongs to the technical field of water conservancy risk early warning, and specifically relates to a water conservancy monitoring and early warning method and system based on a GIS platform. Background Art

[0002] Large-scale water conservancy projects, such as dams and levees, are vital infrastructure that safeguards the nation's economy and people's livelihoods, and their long-term safe operation is crucial. Over their long service lives, these structures are subject to multiple factors, including water level fluctuations, material aging, and geological changes. These factors can lead to safety hazards such as internal erosion, structural deformation, and even instability. Therefore, real-time, effective health monitoring of water conservancy projects is crucial for preventing catastrophic accidents. Existing monitoring technologies primarily rely on point sensors, such as piezometers, displacement gauges, and strain gauges, deployed at specific locations, combined with regular manual inspections. However, these traditional methods have significant limitations. Point sensors have a limited monitoring range and low spatial resolution, making it difficult to fully cover the entire structure and easily overlooking localized defects occurring between two monitoring points. Manual inspections, on the other hand, are subject to high subjectivity, low efficiency, and inability to detect early signs of structural degradation.

[0003] To overcome the shortcomings of point-based monitoring, distributed fiber optic sensing technology has begun to be introduced into the field of water conservancy project monitoring. It can provide continuous measurement data along the fiber optic path. However, even if a large amount of continuous data is obtained, how to effectively analyze it and convert it into accurate risk warnings remains the core challenge currently faced. Existing data analysis methods usually set a fixed threshold for a single physical quantity such as strain, temperature, or vibration, and trigger an alarm once the measured value exceeds the threshold. Although this method is simple, it ignores the inherent correlation between different types of diseases and the dynamic evolution of risks, resulting in low accuracy of early warnings. Summary of the Invention

[0004] The present invention provides a water conservancy monitoring and early warning method and system based on a GIS platform to solve the problem of low early warning accuracy.

[0005] In a first aspect, the present invention provides a water conservancy monitoring and early warning method based on a GIS platform, the method comprising the following steps:

[0006] Lay distributed fiber optic sensors along the structural axis of the target water conservancy project, and integrate the geospatial information and structural attribute information of the target water conservancy project into the GIS platform;

[0007] Based on the structural axis, a one-dimensional linear reference system is created in the GIS platform to map the measurement points of the distributed fiber optic sensors. Taking the mapping point corresponding to the measurement point on the one-dimensional linear reference system as the center, multiple spatial analysis units corresponding to the physical structure of the target water conservancy project are projected perpendicular to the structural axis.

[0008] Collect sensor data from distributed fiber optic sensors in real time and assign the sensor data to corresponding spatial analysis units based on mapping relationships. The sensor data includes strain data and acoustic wave data.

[0009] The acoustic wave data of each spatial analysis unit is processed and pattern matched with a preset acoustic fingerprint library to identify internal erosion problems caused by internal structural deterioration of the target water conservancy project;

[0010] The effective bearing capacity of each spatial analysis unit is dynamically calculated by combining the strain data of each spatial analysis unit and the structural attribute information of the spatial analysis unit in the GIS platform;

[0011] A risk evolution threshold model is constructed. When and only when internal erosion problems are monitored simultaneously in the same spatial analysis unit, and the risk state point formed by the decrease rate of effective bearing capacity and the energy growth rate of internal erosion problems crosses the risk evolution threshold model, point warning information pointing to the spatial analysis unit is generated and output.

[0012] Optionally, integrating the geospatial information and structural attribute information of the target water conservancy project in the GIS platform includes the following steps:

[0013] Obtain and digitize the design center line of the target water conservancy project as the structural axis;

[0014] Import the digital elevation model, geological exploration layer, and historical hazard point layer of the target water conservancy project area as geospatial information;

[0015] Extract soil type and permeability coefficient from geological exploration layers as structural attribute information;

[0016] Perform spatial registration of structural axes, geographic spatial information, and structural attribute information in a unified geographic coordinate system to build a unified GIS data base.

[0017] Optionally, creating a one-dimensional linear reference system in the GIS platform based on the structural axis to map the measurement points of the distributed optical fiber sensor, and projecting perpendicularly to the structural axis with the mapping point corresponding to the measurement point on the one-dimensional linear reference system as the center to generate multiple spatial analysis units corresponding to the physical structure of the target water conservancy project includes the following steps:

[0018] Map the measurement points of the distributed fiber optic sensor to a one-dimensional linear reference system at preset intervals, and establish a corresponding relationship between the measurement points and the project mileage;

[0019] The mapping points corresponding to each measurement point on the one-dimensional linear reference system are used as the reference;

[0020] Query the cross-sectional design parameters of the corresponding project mileage in the GIS data base;

[0021] Dynamically calculate projection width and projection shape based on cross-sectional design parameters;

[0022] The projection width and projection shape are combined to generate polygons perpendicular to the structural axis that cover the entire cross section of the target water conservancy project, and each polygon is defined as a spatial analysis unit;

[0023] Assign a unique identifier to each spatial analysis unit and associate the geographic spatial information and structural attribute information corresponding to the spatial analysis unit.

[0024] Optionally, the signal processing of the acoustic wave data of each spatial analysis unit and the identification of internal erosion problems caused by internal structural deterioration of the target water conservancy project by pattern matching with a preset acoustic fingerprint library include the following steps:

[0025] The acoustic wave data assigned to each spatial analysis unit is intercepted according to a time window to obtain an acoustic wave data segment;

[0026] Applying a bandpass filter to remove environmental background noise in the acoustic wave data segment that is not related to internal structural degradation;

[0027] Performing wavelet packet decomposition on the filtered sound wave data segment to obtain energy distribution characteristics of the sound wave data segment in multiple preset frequency bands;

[0028] Calculate vector similarity between energy distribution features and patterns in the acoustic fingerprint library;

[0029] When the similarity calculation result exceeds the preset matching threshold, it is determined that an internal erosion problem has occurred and the energy intensity of the internal erosion problem is recorded.

[0030] Optionally, the acoustic fingerprint library is pre-built by following the steps below:

[0031] Construct a flume model in the laboratory to simulate the physical process of soil erosion within the target water conservancy project under different working conditions;

[0032] Use distributed fiber optic sensors to collect acoustic wave signals generated by physical processes;

[0033] Process the collected acoustic wave signals, extract the energy distribution characteristics of the acoustic wave signals in the key frequency bands, and form the initial acoustic fingerprint;

[0034] Collect background acoustic wave data at the actual site of the target water conservancy project during the non-flood season and under no-risk conditions;

[0035] The background sound wave data is used to perform noise adaptation and pattern optimization on the initial acoustic fingerprint to generate the final acoustic fingerprint library.

[0036] Optionally, the dynamically calculating the effective bearing capacity of each spatial analysis unit by combining the strain data of each spatial analysis unit and the structural attribute information of the spatial analysis unit in the GIS platform comprises the following steps:

[0037] The theoretical bearing capacity of each spatial analysis unit is calculated based on the structural attribute information of each spatial analysis unit;

[0038] Quantitatively evaluate the vulnerability index of each spatial analysis unit based on the historical records of dangerous conditions and geological conditions in the structural attribute information;

[0039] Construct a structural degradation function that takes the energy intensity and duration of the internal erosion problem as input;

[0040] Substitute the output values ​​of strain data, theoretical bearing capacity, fragility index and structural degradation function into the preset geotechnical mechanics correction model;

[0041] The real-time effective bearing capacity of the spatial analysis unit is calculated.

[0042] Optionally, the quantitative evaluation of the vulnerability index of each spatial analysis unit based on the historical dangerous situation records and geological conditions in the structural attribute information includes the following steps:

[0043] Select at least three key factors affecting structural stability from the structural attribute information;

[0044] The analytic hierarchy process is used to determine the weight coefficient of each key influencing factor;

[0045] Normalize the original parameter values ​​of each key influencing factor in each spatial analysis unit;

[0046] Perform weighted summation on the normalized parameter value and the corresponding weight coefficient;

[0047] The final weighted summation result is used as the vulnerability index of the spatial analysis unit.

[0048] Optionally, the step of constructing a risk evolution threshold model includes the following steps:

[0049] Define a two-dimensional risk coordinate space, where the two coordinate axes are the rate of decrease of effective bearing capacity and the rate of energy growth of internal erosion problem respectively.

[0050] Mapping historical dangerous situation data and laboratory simulated failure data into multiple failure sample points in a two-dimensional risk coordinate space;

[0051] The support vector machine algorithm is used to train the failure sample points and generate a classification hyperplane that divides the safe area and the dangerous area as the risk evolution threshold model;

[0052] Continuously calculate the real-time position of the risk status point of each spatial analysis unit in the two-dimensional risk coordinate space;

[0053] When the risk state point crosses the classification hyperplane from the safe area to the dangerous area, the warning triggering condition is met.

[0054] Optionally, the step of generating and outputting point warning information directed to the spatial analysis unit when and only when internal erosion problems are simultaneously monitored within the same spatial analysis unit, and a risk state point formed by a decrease rate of the effective bearing capacity and an energy growth rate of the internal erosion problem crosses a risk evolution threshold model, comprises the following steps:

[0055] Inside the classification hyperplane of the two-dimensional risk coordinate space, at least two warning level surfaces are defined according to the distance from the hyperplane to construct a graded warning area;

[0056] Calculate the distance between each risk status point and the nearest warning level surface in real time;

[0057] Divide the spatial analysis unit into different warning states according to distance;

[0058] Present different warning states in different visualization methods on the GIS platform;

[0059] If and only if the risk status point of the spatial analysis unit crosses the outermost classification hyperplane, the point warning information of the spatial analysis unit is generated.

[0060] In the second aspect, the present invention also provides a water conservancy monitoring and early warning system based on a GIS platform, comprising a memory, a processor, and a computer program stored on the memory and runnable on the processor. When the processor executes the computer program, it implements the water conservancy monitoring and early warning method based on the GIS platform as described in the first aspect.

[0061] The beneficial effects of the present invention are:

[0062] The present invention utilizes distributed optical fibers laid along the axis of the structure to achieve full-line, continuous, high-density real-time monitoring of engineering structure strain and acoustic wave signals, overcoming the defects of traditional point sensors with limited monitoring range and inability to capture early local diseases. Secondly, by creating a one-dimensional linear reference system in the GIS platform and generating spatial analysis units, the present invention accurately spatially maps abstract sensor data with the physical structure of the water conservancy project, making the monitoring results visual and intuitive, and being able to directly locate early warning information to specific structural units, providing precise spatial guidance for subsequent inspection and maintenance. More importantly, the present invention ingeniously constructs a risk evolution threshold model based on dual condition verification. It does not rely solely on a single parameter threshold, but rather comprehensively judges the dynamic evolution relationship between the occurrence of internal erosion problems and the rate of effective bearing capacity reduction, thereby achieving dynamic and correlation analysis of risk status. This early warning mechanism can effectively filter out false alarms caused by instantaneous disturbances caused by a single factor or stable minor defects, greatly improving the accuracy and reliability of the early warning, ensuring that an alarm is only issued when the structural health status shows a substantial and continuous deterioration trend, thus achieving a leap from static monitoring to dynamic early warning, and providing strong technical support for ensuring the safe operation of water conservancy projects and preventing catastrophic accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 This is a flow chart of a water conservancy monitoring and early warning method based on a GIS platform in one of the implementation methods of this application.

[0064] Figure 2 This is a flow chart of constructing a risk evolution threshold model in one embodiment of the present application.

[0065] Figure 3 This is a schematic diagram of the process of outputting warning information in one embodiment of the present application. DETAILED DESCRIPTION

[0066] The following will be combined with the accompanying drawings in the embodiments of the present application to clearly describe the technical solutions in the embodiments of the present application. Obviously, the embodiments described are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field are within the scope of protection of this application.

[0067] The terms "first," "second," and the like in the specification and claims of this application are used to distinguish similar objects, and are not used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments of this application can be implemented in an order other than that illustrated or described herein, and that the objects distinguished by "first," "second," and the like are generally of the same type, and do not limit the number of objects; for example, the first object can be one or more. In addition, the term "and / or" in the specification and claims refers to at least one of the connected objects, and the character " / " generally indicates that the objects connected are in an "or" relationship.

[0068] Figure 1 FIG. 1 is a flow chart of a water conservancy monitoring and early warning method based on a GIS platform in one embodiment. It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. In addition, Figure 1 At least part of the steps in the above process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but may be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but may be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps. Figure 1 As shown, the water conservancy monitoring and early warning method based on the GIS platform disclosed in the present invention specifically includes the following steps:

[0069] S101. Lay distributed fiber optic sensors along the structural axis of the target water conservancy project, and integrate the geospatial information and structural attribute information of the target water conservancy project into a GIS platform.

[0070] To achieve comprehensive awareness of a water conservancy project, distributed optical fibers equipped with integrated strain and acoustic wave sensors are laid along its core stress paths, such as the crest and abutments of a dam. These fibers, like sensitive nerves, can detect minute structural deformations and internal sounds. Simultaneously, a variety of spatial information, including the project's design drawings, a digital elevation model of the area, geological survey data, and historical records of hazardous conditions, are integrated within a digital mapping platform, the Geographic Information System (GIS). Precisely aligned within a unified geographic coordinate system, this creates a comprehensive digital foundation encompassing topography, geological conditions, and structural parameters. The ultimate result of this step is a digital twin of the water conservancy project, precisely mapping and linking the physical structure with its complex environment in virtual space. This provides a unified data foundation for subsequent spatial analysis and early warning.

[0071] S102. Create a one-dimensional linear reference system in the GIS platform based on the structural axis to map the measurement points of the distributed fiber optic sensor, and project perpendicularly to the structural axis with the mapping point corresponding to the measurement point on the one-dimensional linear reference system as the center to generate multiple spatial analysis units corresponding to the physical structure of the target water conservancy project.

[0072] To accurately link one-dimensional fiber optic sensor data with the three-dimensional engineering entity, a one-dimensional linear reference system must be created within the GIS platform, using the project's design centerline as the structural axis. This reference system acts as a mileage stake number for the linear structure, measured from its starting point. Each measurement point on the distributed optical fiber is assigned a unique mileage value based on its physical location, establishing a mapping between the sensor data and the specific location of the project. Next, with each mileage point as the center, the GIS digital base is queried for the corresponding cross-sectional design parameters of the project at that location. Based on these parameters, such as crest width and slope ratio, a two-dimensional polygon is generated perpendicular to the structural axis, completely covering the physical structure of the project at that location. This polygon is defined as a spatial analysis unit, which not only has a clear geographic scope but also inherits all geospatial information and structural properties of that location. Ultimately, the entire water conservancy project is divided into a series of continuous and seamless spatial analysis units, enabling spatial and grid-based management of monitoring data and giving each data point clear physical and geographic significance.

[0073] S103. Collect sensor data of the distributed optical fiber sensor in real time, and distribute the sensor data to the corresponding spatial analysis unit according to the mapping relationship, wherein the sensor data includes strain data and acoustic wave data.

[0074] The distributed fiber optic sensors, using their demodulators, continuously collect strain and acoustic signals at every point along the route. Strain data reflects minute stretches or compressions caused by soil pressure, water pressure, or structural deformation, while acoustic data captures subtle sounds generated by processes such as water infiltration and frictional movement of soil particles. The collected raw data stream carries precise location information (distance along the fiber). Using the mapping established in the previous step, the system automatically and accurately assigns the strain and acoustic data collected by each fiber segment to its corresponding spatial analysis unit. This process transforms data collection from linear to partitioned. The result is a dynamically updated data warehouse in which each spatial analysis unit is linked in real time to its strain value reflecting the current structural state and acoustic signals reflecting the internal acoustic environment. This provides immediate and accurate data input for subsequent independent health diagnosis of each unit.

[0075] S104. Perform signal processing on the acoustic wave data of each spatial analysis unit, and identify internal erosion problems caused by internal structural deterioration of the target water conservancy project by performing pattern matching with a preset acoustic fingerprint library.

[0076] Among them, in order to identify the core hidden danger of internal erosion, it is necessary to deeply process the acoustic wave data of each spatial analysis unit. Since the acoustic signal generated by internal soil erosion is extremely weak and easily drowned out by environmental noise, the continuous acoustic wave data is first divided into short-time windows, and a bandpass filter is applied to filter out interference from irrelevant frequency bands such as wind, rain or traffic vibrations. Then, the wavelet packet decomposition technology is used on the filtered pure signal. This is a powerful signal analysis tool that can extract its energy distribution characteristics in different preset frequency bands to form a multi-dimensional energy feature vector. This vector is like a fingerprint of sound. By matching the vector with the acoustic fingerprint library that has been simulated in advance in the laboratory for various soil erosion conditions and collected and optimized, the vector similarity between the two is calculated. A commonly used calculation method is cosine similarity. ,in is the energy vector monitored in real time, is the fingerprint vector in the library. When the similarity S exceeds the preset matching threshold, it is determined that an internal corrosion problem has occurred within the spatial analysis unit, and its energy intensity is recorded. This method can accurately identify early signs of structural degradation from complex background noise.

[0077] S105. Dynamically calculate the effective bearing capacity of each spatial analysis unit by combining the strain data of each spatial analysis unit and the structural attribute information of the spatial analysis unit in the GIS platform.

[0078] In order to dynamically evaluate the actual health status of each engineering unit, its effective bearing capacity needs to be calculated. This process does not rely solely on theoretical calculations, but integrates multiple information. First, based on the structural attribute information of the unit such as soil type, density, moisture content, etc. stored in GIS, the theoretical bearing capacity under ideal conditions is calculated using geotechnical formulas. At the same time, the inherent vulnerability index of the unit is quantitatively assessed by combining historical records of dangerous situations and geological conditions. More importantly, the strain data monitored in real time and the degree of structural deterioration defined by the energy intensity and duration of the identified internal corrosion problem As dynamic correction factors. These parameters are substituted into a pre-set geomechanical correction model, such as ,in is the effective carrying capacity, is the fragility influence coefficient, and \(f\) is a function that decreases with strain and erosion degradation. This step results in a dynamically changing quantitative indicator that truly reflects the current structural integrity and residual strength, completing the transition from static design to dynamic health assessment.

[0079] S106. Construct a risk evolution threshold model. Generate and output point warning information pointing to the spatial analysis unit when and only when internal erosion problems are monitored simultaneously in the same spatial analysis unit, and the risk state point formed by the rate of decrease of effective bearing capacity and the rate of increase of energy of internal erosion problems crosses the risk evolution threshold model.

[0080] The final warning decision is a rigorous multi-condition triggering process designed to maximize warning accuracy and minimize false alarms. This method constructs a two-dimensional risk evolution threshold model. The two axes of the model's coordinate system are the growth rate of internal erosion energy and the decline rate of effective carrying capacity. These two rates together define a risk state point, characterizing the speed of the hazard's development. Using machine learning algorithms such as support vector machines (SVMs), trained on a large amount of historical hazard data and laboratory simulated failure data, a classification hyperplane is generated in the two-dimensional risk space that optimally demarcates safe and dangerous zones. This plane serves as the risk evolution threshold. During real-time monitoring, a spatial analysis unit will only generate and output a point warning message for that specific spatial analysis unit if two conditions are simultaneously met: first, it is clearly identified as having an internal erosion problem; second, its corresponding risk state point, due to accelerated erosion and a rapid decline in carrying capacity, moves, ultimately crossing the threshold hyperplane. This method ensures that the early warning is based on the dual confirmation of the risk status and its evolution trend, so that it can accurately capture the critical point from quantitative change to qualitative change, and provide extremely valuable decision-making support for disaster prevention and rescue.

[0081] In one embodiment, integrating the geospatial information and structural attribute information of the target water conservancy project in the GIS platform includes the following steps:

[0082] Obtain and digitize the design center line of the target water conservancy project as the structural axis;

[0083] Import the digital elevation model, geological exploration layer, and historical hazard point layer of the target water conservancy project area as geospatial information;

[0084] Extract soil type and permeability coefficient from geological exploration layers as structural attribute information;

[0085] Perform spatial registration of structural axes, geographic spatial information, and structural attribute information in a unified geographic coordinate system to build a unified GIS data base.

[0086] In this implementation, a digital core skeleton is first established for the water conservancy project. This requires obtaining the project centerline determined during the design phase and converting it into digital vector data that can be recognized by the GIS platform. This process is usually done by scanning the original design drawings with high precision, then performing georeferencing and vectorization tracing in GIS software, or directly converting and importing from engineering design software such as CAD. This line is defined as the structural axis, which is a series of points with precise geographic coordinates in three-dimensional space. The effect of this step is to transform the abstract design blueprint representing the direction and elevation of the project into a digital route with precise spatial location information. Next, a series of key geospatial information layers need to be imported into the GIS platform. This includes a digital elevation model (DEM) that can accurately describe the surface undulations and stores the altitude of each point in the area in the form of a raster matrix; a geological exploration layer that reveals the composition and distribution of underground rock and soil, usually polygonal vector data; and a historical hazard point layer that marks the specific locations of historical hazards such as pipe bursts and landslides.

[0087] These data provide the topographic, geological, and empirical background that is crucial to the stability of the project. After importing, the slope and aspect of any point can be calculated based on the DEM, providing a basis for analyzing surface runoff and slope stability. The effect of this step is to build a multi-dimensional virtual environment around the target project in GIS, placing the engineering entity in its real topographic and geological background, and providing rich environmental context information for comprehensive risk assessment. In order to parameterize the physical properties of engineering materials, it is necessary to extract key structural attribute information from the imported geological exploration layer. The geological layer contains valuable data obtained through field drilling, sampling, and indoor testing, which are associated with the geological zoning graphics in the form of attribute tables.

[0088] Through spatial query and attribute extraction operations, we can obtain parameters such as soil type, density, internal friction angle, etc. in each geological zone, especially the permeability coefficient which has a great impact on engineering safety. It is the core indicator that describes the difficulty of water penetration in soil and directly affects the occurrence and development of internal erosion. Its importance is reflected in Darcy's law. The law states that the penetration rate of water The effect of this step is to convert the qualitative geological description into quantitative parameters that can be used for mechanical and hydraulic calculations, providing the necessary physical property basis for subsequent bearing capacity calculations and seepage analysis.

[0089] Finally, to ensure that all data from different sources and formats can work together, they must be spatially aligned within a unified geographic coordinate system, ultimately creating a comprehensive GIS data base. Because raw data such as design drawings, remote sensing images, and geological maps may be based on different coordinate systems and map projections, direct overlaying can lead to positional confusion. Spatial registration involves selecting well-defined control points on the ground and applying mathematical transformation models (such as affine transformations) to calibrate the positions of all layers to a unified geodetic or projected coordinate system. Registration accuracy is evaluated by calculating the root mean square error (RMS) to ensure that the error is within the acceptable range. The ultimate result of this step is a unified GIS data base in which all data layers are precisely aligned and spatial relationships are completely correct. This data base, like a rich and clearly indexed digital atlas, eliminates barriers and inconsistencies between data, providing a reliable and consistent data foundation for all subsequent complex spatial overlay analysis, querying, and visualization.

[0090] In one embodiment, a one-dimensional linear reference system is created in a GIS platform based on the structural axis to map the measurement points of the distributed optical fiber sensor, and multiple spatial analysis units corresponding to the physical structure of the target water conservancy project are generated by projecting perpendicular to the structural axis with the mapping point corresponding to the measurement point on the one-dimensional linear reference system as the center, including the following steps:

[0091] Map the measurement points of the distributed fiber optic sensor to a one-dimensional linear reference system at preset intervals, and establish a corresponding relationship between the measurement points and the project mileage;

[0092] The mapping points corresponding to each measurement point on the one-dimensional linear reference system are used as the reference;

[0093] Query the cross-sectional design parameters of the corresponding project mileage in the GIS data base;

[0094] Dynamically calculate projection width and projection shape based on cross-sectional design parameters;

[0095] The projection width and projection shape are combined to generate polygons perpendicular to the structural axis that cover the entire cross section of the target water conservancy project, and each polygon is defined as a spatial analysis unit;

[0096] Assign a unique identifier to each spatial analysis unit and associate the geographic spatial information and structural attribute information corresponding to the spatial analysis unit.

[0097] In this embodiment, in order to spatially align the one-dimensional sensor signal with the three-dimensional engineering entity, it is necessary to map the measurement points resolved by the sensor demodulator on the distributed optical fiber according to their physical distance along the optical fiber through a mapping function. Transform to a one-dimensional linear reference system, where is the distance of the measurement point along the fiber, It is the corresponding mileage on the axis of the engineering structure. This linear reference system is like assigning a set of continuous mileage stakes, starting from zero, to the entire linear structure of the water conservancy project. This step, by establishing a one-to-one correspondence between measurement points and engineering mileage, transforms the abstract physical sensor locations into geographic coordinates with clear engineering significance. Each mileage point mapped from the sensor measurement point on the linear reference system serves as a reference point, the starting point for constructing two-dimensional spatial analysis units. This step is implemented as a logical positioning process. Its core principle is to establish an infinitesimal geometric point on the continuous structural axis as the central origin for all subsequent vertical operations. In a GIS environment, this reference point is a digital entity with a unique mileage value and three-dimensional coordinates. This step itself does not involve complex calculations, but it plays a crucial role. It provides a stable geometric anchor for the expansion of one-dimensional linear information into two-dimensional cross-sectional information, ensuring that the subsequently generated spatial analysis units are precisely located at their intended engineering locations.

[0098] Based on the established mileage benchmarks, spatial queries are used to retrieve cross-sectional design parameters that exactly correspond to the mileage value within a pre-built GIS data base. The morphology of a hydraulic project varies from section to section; its crest width, internal and external slope ratio, and other parameters can vary with location. This step utilizes the mileage value as a unique index key to accurately match and search within a database containing design drawings, extracting all relevant design dimension data for that location. This results in a set of engineering parameters tailored to that specific location, such as crest width, height, and upstream and downstream slope ratios. This provides all the necessary raw data for accurately reproducing the physical configuration of the project at that location. Based on the retrieved cross-sectional design parameters, dynamic geometric calculations are performed to determine the projected width and specific shape of the spatial analysis unit at that location. For a typical trapezoidal dam, the total projected width is determined by the crest width and the horizontal projections of the side slopes.

[0099] For example, the total base width of a cross section The formula can be Calculated, where is the width of the embankment given in the design drawings, H is the embankment height, and and Representing the slope ratio coefficients of the upstream slope and the downstream slope respectively. Combined with the projection width and shape information calculated in the previous step, with the mileage reference point as the center and perpendicular to the direction of the structural axis, a polygon that can completely cover the physical cross-section of the target water conservancy project is automatically generated in the GIS environment. This generation process is a geometric construction operation. The system connects the calculated vertex coordinates into a closed figure, and this polygon is defined as a spatial analysis unit. In this way, the originally linear engineering structure is discretized into a series of closely connected and seamlessly spliced ​​two-dimensional analysis grids. The effect of this step is to successfully expand the one-dimensional monitoring line into a two-dimensional management unit that can represent the engineering entity, realizing the leap from line to surface monitoring range.

[0100] To effectively manage and trace information for each spatial analysis unit, each generated polygon needs to be assigned a globally unique identifier (ID) and linked to all geospatial and structural attribute information within the unit's spatial range. This process is achieved by establishing relational connections in the GIS database, binding each polygon ID to the soil type, permeability, historical hazard records, elevation data, and other data within its coverage area. The ultimate effect is that each spatial analysis unit is transformed from a simple geometric shape into an information-rich intelligent object. It not only knows its location and shape, but also understands the geology beneath it, its own material properties, and its historical health status, forming a highly integrated basic data unit that facilitates complex analysis and precise early warning.

[0101] In one embodiment, signal processing is performed on the acoustic wave data of each spatial analysis unit, and pattern matching is performed with a preset acoustic fingerprint library to identify internal erosion problems caused by internal structural deterioration of the target water conservancy project, including the following steps:

[0102] The acoustic wave data assigned to each spatial analysis unit is intercepted according to a time window to obtain an acoustic wave data segment;

[0103] Applying a bandpass filter to remove environmental background noise in the acoustic wave data segment that is not related to internal structural degradation;

[0104] Performing wavelet packet decomposition on the filtered sound wave data segment to obtain energy distribution characteristics of the sound wave data segment in multiple preset frequency bands;

[0105] Calculate vector similarity between energy distribution features and patterns in the acoustic fingerprint library;

[0106] When the similarity calculation result exceeds the preset matching threshold, it is determined that an internal erosion problem has occurred and the energy intensity of the internal erosion problem is recorded.

[0107] In this embodiment, the acoustic wave data is first divided from a continuous time stream into independent, processable data segments. This process is achieved by applying a time window of fixed length, just like cutting short segments from a long tape. By setting a time window length T and an overlap rate between windows, it can be ensured that any short and critical acoustic events will not be cut out and missed because they fall exactly on the boundary of two windows. Next, in order to highlight the core signals related to structural degradation, a bandpass filter needs to be applied to each acoustic wave data segment to filter out the interference of environmental background noise. The acoustic wave signals generated by internal erosion are usually concentrated in a specific frequency range, while noises such as wind, rain or traffic vibrations generated by human activities in the natural environment are distributed in other frequency bands. The bandpass filter is like a precise sieve that only allows preset signals related to the characteristic frequency of internal erosion (for example, at the frequency arrive between) through, and will be below and higher than All irrelevant frequency components of the signal are greatly weakened or completely removed. The effect of this step is to greatly improve the signal-to-noise ratio of the signal, making the weak erosion signal stand out from the noisy background.

[0108] Wavelet packet decomposition is required for the filtered sound wave data segment to extract its intrinsic energy distribution characteristics that can characterize its essence. Compared with the traditional Fourier transform, wavelet packet analysis can simultaneously analyze the signal in two dimensions, time and frequency, and is particularly good at capturing the details of transient signals. This process recursively decomposes the signal into a series of narrower preset frequency bands and calculates the energy of the signal in each frequency band. The energy of the jth frequency band is The wavelet packet decomposition coefficients within the frequency band can be The sum of the squares of The energy values ​​of all frequency bands are combined to form a multi-dimensional energy feature vector.

[0109] Matching the extracted energy eigenvectors against a pre-built acoustic fingerprint library is a key step in identifying the problem. This library stores a large number of known standard energy eigenvectors generated by internal erosion problems of varying types and degrees. By calculating the vector similarity between the real-time monitored eigenvectors and each standard pattern in the fingerprint library, the degree of pattern similarity between the two can be quantified. The commonly used cosine similarity calculation method has a value range of -1 to 1, with the closer the value is to 1, the more similar the acoustic patterns represented by the two vectors are.

[0110] Finally, a clear judgment criterion is set based on the similarity calculation results. When the calculated similarity exceeds a pre-set matching threshold that has been verified by a large number of experiments, a judgment can be made that an internal erosion event has occurred within the spatial analysis unit. The setting of this threshold is crucial, as it balances the sensitivity and reliability of the early warning and avoids false alarms caused by noise misjudgment. Once it is determined to be an erosion event, the energy intensity of the event is recorded, usually quantified by the total energy of the acoustic wave data segment. This step ultimately transforms the continuous analysis process into a discrete, meaningful event judgment result, which not only confirms the existence of hidden dangers, but also preliminarily quantifies their severity, providing trigger signals and key input parameters for initiating higher-level risk assessment models.

[0111] In one embodiment, the acoustic fingerprint library is pre-built through the following steps:

[0112] Construct a flume model in the laboratory to simulate the physical process of soil erosion within the target water conservancy project under different working conditions;

[0113] Use distributed fiber optic sensors to collect acoustic wave signals generated by physical processes;

[0114] Process the collected acoustic wave signals, extract the energy distribution characteristics of the acoustic wave signals in the key frequency bands, and form the initial acoustic fingerprint;

[0115] Collect background acoustic wave data at the actual site of the target water conservancy project during the non-flood season and under no-risk conditions;

[0116] The background sound wave data is used to perform noise adaptation and pattern optimization on the initial acoustic fingerprint to generate the final acoustic fingerprint library.

[0117] In this embodiment, a physical water tank model is constructed in a controllable laboratory to reproduce the actual process of soil erosion inside a water conservancy project. This model will fill in the corresponding soil samples according to the actual soil parameters of the target project, and construct simulated internal defects such as tiny cracks or holes. By precisely controlling the water pumps and valves to apply different hydraulic gradients, that is, creating different water pressure differences at both ends of the soil sample, it is possible to induce a complete physical process from slow infiltration to the formation of concentrated seepage channels, and ultimately causing soil particles to be carried away by water. By systematically changing the soil type, density, and hydraulic gradient (in is the water head difference, is the seepage path length), comprehensively simulating various potential internal erosion scenarios. While simulating physical processes in the laboratory, distributed fiber optic sensors identical to those used for field monitoring are used to collect the acoustic signals generated by these processes. The sensing fibers are pre-embedded within the soil of the flume model, enabling them to closely and sensitively capture the acoustic vibrations generated by friction and collision between water and soil particles, as well as micro-fractures in the soil structure.

[0118] These vibrations will cause micro-strain in the optical fiber, and the demodulation equipment can convert these micro-strains into continuous digital acoustic signals and record them. The core of this step is to ensure the consistency of the sensing technology used in the characteristic signal acquisition stage and the actual monitoring stage in the future, thereby eliminating the systematic errors that may be introduced by sensor differences. The final effect is to obtain a batch of valuable original acoustic wave data sets that correspond one-to-one to specific physical erosion processes. These data are the direct raw materials for constructing acoustic fingerprints. The collected original acoustic wave signals need to be deeply processed to extract their core features, thereby forming the initial acoustic fingerprint. This process first applies signal processing technology to analyze the frequency distribution of acoustic wave energy under different erosion conditions and find several key frequency bands where the energy is most concentrated. Subsequently, for each acoustic wave signal, wavelet packet analysis and other methods are used to accurately calculate its energy distribution in these key frequency bands. These energy values ​​constitute a multidimensional vector, namely the initial acoustic fingerprint. ,in represents the energy of the signal in the nth critical frequency band. This vector numerically describes the acoustic nature of a specific erosion event in a compact and comprehensive way.

[0119] To ensure that the initial fingerprint generated in the laboratory is applicable to the noisy real world, background acoustic data must be collected at the actual site of the target water conservancy project. During the non-flood season, when the structure is assessed as safe, a distributed fiber optic sensing system is deployed for long-term continuous monitoring. This approach aims to comprehensively capture all environmental background noise at the project under normal, non-risk conditions. This background noise may originate from wind, rain, surrounding traffic, normal reservoir operation, and even biological activity. The collected data represents the unique and persistent noise profile of the site. This step results in a large library of background noise samples, providing a realistic basis for subsequent fingerprint optimization and crucial for distinguishing real risk signals from everyday environmental disturbances. Finally, using the collected field background acoustic data, the initial acoustic fingerprint generated in the laboratory is subjected to noise adaptation and pattern optimization to generate a final, practical acoustic fingerprint library. The core principle of this process is to modify the initial fingerprint by analyzing the spectral characteristics of the background noise, reducing the weight of frequency components that are easily drowned out by field noise in the pattern matching process.

[0120] In practice, this process can be regarded as an optimization function ,in is the initial fingerprint, is the background noise characteristic, and This is the optimized final fingerprint. This optimization process enhances the robustness and uniqueness of the fingerprint in strong noise environments. Ultimately, all optimized final fingerprints corresponding to different erosion conditions are combined to construct a final acoustic fingerprint library with a high signal-to-noise ratio tailored to the specific water conservancy project, providing an accurate and reliable pattern matching benchmark for on-site monitoring.

[0121] In one embodiment, dynamically calculating the effective bearing capacity of each spatial analysis unit by combining the strain data of each spatial analysis unit and the structural attribute information of the spatial analysis unit in the GIS platform includes the following steps:

[0122] The theoretical bearing capacity of each spatial analysis unit is calculated based on the structural attribute information of each spatial analysis unit;

[0123] Quantitatively evaluate the vulnerability index of each spatial analysis unit based on the historical records of dangerous conditions and geological conditions in the structural attribute information;

[0124] Construct a structural degradation function that takes the energy intensity and duration of the internal erosion problem as input;

[0125] Substitute the output values ​​of strain data, theoretical bearing capacity, fragility index and structural degradation function into the preset geotechnical mechanics correction model;

[0126] The real-time effective bearing capacity of the spatial analysis unit is calculated.

[0127] In this implementation, the theoretical bearing capacity of each spatial analysis unit is calculated and evaluated based on the inherent structural attribute information. This bearing capacity represents the ultimate resistance of the unit in an ideal, non-destructive state. This calculation relies entirely on classical geotechnical mechanics theory, taking into account the soil type, density, cohesion c, and internal friction angle extracted from the GIS data base. By applying recognized mechanical models such as the Mohr-Coulomb strength criterion, the shear strength of the soil can be determined. , this strength is directly related to its load-bearing capacity. For example, its basic form can be expressed as ,in is the normal stress acting on the shear plane. Next, to quantify the inherent risk of each spatial analysis unit, we need to combine its historical record of hazards with its geological conditions to generate a vulnerability index. This index does not measure current damage, but rather assesses a unit's propensity to failure due to its historical background and inherent environmental factors.

[0128] By extracting factors such as the frequency and severity of historical hazards and the proximity of adverse geological structures (such as weak interlayers and fault zones) from structural attribute information and assigning corresponding weights to these factors, a weighted sum model can be constructed. This model combines risk factors of different natures into a single numerical indicator. ,in where represents the normalized quantitative values ​​of historical, geological, and soil factors, respectively, and w represents the corresponding weights. This step creates an innate risk profile for each section of the project, identifying weak links that require significant attention even under normal circumstances.

[0129] In order to convert the internal erosion problem monitored in real time into a specific impact on the structural strength, a structural degradation function needs to be constructed. The core principle of this function is that the loss of structural strength is positively correlated with the severity and duration of internal erosion. The function is based on the internal erosion energy intensity determined in the acoustic analysis. and the duration of the erosion event As input variables, an effective degradation function can simulate the cumulative effect of damage, for example, the structural degradation degree It can be expressed as ,in Is a degradation coefficient related to the material's corrosion resistance. This exponential model reflects the physical process of damage growth in the early stage and gradually slowing down in the later stage. The effect of this step is to transform the abstract acoustic event into a dynamic variable that can quantify the degree of structural damage. , which varies between 0 and 1, intuitively representing the degree from intact to complete failure.

[0130] All key information is then integrated into a pre-set geomechanical correction model, which takes the theoretical bearing capacity as a basis and then reduces it with a series of coefficients reflecting negative impacts. It combines the vulnerability index representing the inherent deficiencies , structural degradation representing acquired damage, and real-time strain data representing the current true response of the structure . This correction model organically combines these independent parameters to form a unified evaluation framework. For example, the model can be in the form of a product, where each reduction factor is multiplied and applied to the theoretical bearing capacity. Finally, by executing the geotechnical mechanics correction model, the real-time effective bearing capacity of each spatial analysis unit is calculated. This step is the final output of the entire evaluation process. It substitutes the results of all the previous steps - theoretical bearing capacity, vulnerability index, structural deterioration and real-time strain - as instant parameters into the correction model formula for calculation. For example, the final effective bearing capacity Available through To determine, is a function that decreases with increasing strain, indicating that increasing strain further reduces bearing capacity. Because the energy and strain of internal erosion are updated in real time, this calculation is dynamic and continuous. The net effect is the generation of a real-time effective bearing capacity curve for each engineering unit. This curve accurately reflects the dynamic evolution of structural health and provides the core quantitative basis for early warning decisions.

[0131] In one embodiment, quantitatively evaluating the vulnerability index of each spatial analysis unit based on historical risk records and geological conditions in the structural attribute information includes the following steps:

[0132] Select at least three key factors affecting structural stability from the structural attribute information;

[0133] The analytic hierarchy process is used to determine the weight coefficient of each key influencing factor;

[0134] Normalize the original parameter values ​​of each key influencing factor in each spatial analysis unit;

[0135] Perform weighted summation on the normalized parameter value and the corresponding weight coefficient;

[0136] The final weighted summation result is used as the vulnerability index of the spatial analysis unit.

[0137] In this implementation, several key influencing factors with the most significant impact on structural stability are selected from a vast database of structural attribute information through expert experience and engineering mechanism analysis. Typically, factors from at least three different dimensions are selected to ensure a comprehensive assessment. For example, the soil permeability coefficient reflects internal seepage risk, the density of dangerous situations reflects historical experience, and the distance to unfavorable geological bodies (such as faults) characterizes inherent geological defects. The core of this step is to narrow the complex stability problem down to a few core and representative variables. To scientifically determine the relative importance of each key influencing factor, the Analytic Hierarchy Process (AHP) is used to calculate their weight coefficients. This method decomposes the decision problem into a hierarchy of objectives and criteria (i.e., individual influencing factors). Domain experts are invited to compare each factor pairwise, judging their relative importance on a scale of 1 to 9, thereby constructing a judgment matrix. By calculating the maximum eigenvalue of this judgment matrix and its corresponding eigenvector and normalizing the eigenvector, the weight coefficient for each factor is obtained. The entire process also requires consistency testing to ensure the logical consistency of the expert judgment.

[0138] Since the original parameter values ​​of the selected key influencing factors have different units and numerical ranges, for example, the unit of permeability coefficient is meter / second, while the hazard density is times / square kilometer, mathematical operations cannot be performed directly between them. Therefore, the original parameter values ​​of each factor in each spatial analysis unit must be normalized and mapped to a unified, dimensionless interval, usually [0, 1]. For positive factors, the larger the value, the higher the risk. After completing the weight determination and data normalization of all factors, the two results need to be weighted and summed to form a comprehensive evaluation score. For each independent spatial analysis unit, multiply the normalized values ​​of each key influencing factor it contains by the globally unified weight coefficient of the factor, and then add up all the product results. This process can be done by the formula To express, is the overall score of the unit. is the weight of the i-th factor, is the normalized value of the i-th factor of the unit.

[0139] Finally, the weighted summation result calculated in the previous step is formally defined as the final vulnerability index of the spatial analysis unit. This index is a value between 0 and 1, which highly condenses all the inherent weak link information of the unit. An index value close to 1 means that the soil conditions, geological environment or historical performance of the unit have a higher potential risk; conversely, an index value close to 0 means that its innate conditions are relatively superior and the structure is more stable. The ultimate effect of this step is to give each section of the entire water conservancy project a clear, intuitive and horizontally comparable quantitative vulnerability label, providing important static base information for subsequent dynamic risk assessment and differentiated management.

[0140] Constructing a risk evolution threshold model includes the following steps:

[0141] Define a two-dimensional risk coordinate space, where the two coordinate axes are the rate of decrease of effective bearing capacity and the rate of energy growth of internal erosion problem respectively.

[0142] Mapping historical dangerous situation data and laboratory simulated failure data into multiple failure sample points in a two-dimensional risk coordinate space;

[0143] The support vector machine algorithm is used to train the failure sample points and generate a classification hyperplane that divides the safe area and the dangerous area as the risk evolution threshold model;

[0144] Continuously calculate the real-time position of the risk status point of each spatial analysis unit in the two-dimensional risk coordinate space;

[0145] When the risk state point crosses the classification hyperplane from the safe area to the dangerous area, the warning triggering condition is met.

[0146] In this implementation, a two-dimensional risk coordinate space must first be defined. This space does not describe static quantities but rather captures rates of change. The horizontal axis is defined as the rate of growth of the internal erosion energy, quantifying the rate of intensification of internal damage; the vertical axis is defined as the rate of decline of effective bearing capacity, quantifying the rate of structural strength loss. Using these two rate metrics, the health status of any unit in the project can be converted from a static numerical value to a dynamic vector with direction and velocity. The fundamental principle of this definition is that the ultimate instability of a project is not determined by a single instantaneous state, but rather by the cumulative evolution of the deterioration rate and strength decay rate. To give the risk coordinate space practical physical meaning, known failure case data must be mapped into it to serve as sample points for demarcating hazardous areas. This data comes from two sources: first, real-world project failures or major hazard cases recorded in a historical hazard database. By analyzing monitoring data immediately before the failure, the energy growth rate and bearing capacity decline rate at that time can be calculated; second, laboratory structural failure experiments simulated using water tank models. These experiments provide comprehensive data on structures reaching their ultimate limit state under controlled conditions.

[0147] These rate data, obtained from real-world and controlled experiments, represent the ultimate failure conditions. , , as failure sample points, are mapped one by one into the two-dimensional risk coordinate space. The effect of this step is to use real empirical data to mark out an area representing danger in the originally abstract coordinate space, providing a data basis for the subsequent division of safety and danger boundaries. In order to draw a clear and reliable boundary between many failure sample points and safe sample points under normal operating conditions, it is necessary to use support vector machine (SVM), an advanced machine learning algorithm, for training. The goal of the support vector machine algorithm is to find an optimal classification boundary between two types of sample points that can maximize the blank area (i.e., the interval) on both sides. This boundary appears as a straight line or curve in two-dimensional space and is called a classification hyperplane. Its mathematical expression is ,in is any point in the coordinate space, and b are the model parameters determined by the algorithm after learning sample data. This hyperplane is the final constructed risk evolution threshold model.

[0148] After the monitoring and early warning method is put into operation, it is necessary to continuously calculate the real-time position of the risk status point of each spatial analysis unit in the two-dimensional risk coordinate space. For each independent analysis unit, the current energy growth rate can be calculated in real time by taking the time derivative of the continuously monitored internal erosion energy and effective bearing capacity data. and the rate of decrease in bearing capacity These two instantaneous rate values ​​together form a coordinate point , that is, the risk state point of the unit at the current moment. With the dynamic changes of monitoring data, this risk state point will continue to move in the two-dimensional risk coordinate space, and its movement trajectory intuitively reflects the evolution path of the unit's risk state. Finally, when the risk state point of a spatial analysis unit crosses from the safe area defined by the classification hyperplane into the dangerous area during its movement, the triggering condition of the early warning is met. This crossing process can be accurately determined mathematically, that is, when the risk state point makes the discriminant function The value of changes from negative (representing safety) to positive (representing danger). The core advantage of this judgment mechanism lies in its ability to detect the turning point from quantitative change to qualitative change earlier, rather than the absolute magnitude of a static value. This step ultimately establishes a logically rigorous and responsive early warning trigger mechanism that effectively identifies dangerous evolution processes with genuine destabilizing tendencies, thereby generating highly reliable point-by-point warning information.

[0149] In one embodiment, when and only when internal erosion problems are simultaneously monitored within the same spatial analysis unit, and the risk state point formed by the rate of decrease in effective bearing capacity and the rate of increase in energy of the internal erosion problem crosses the risk evolution threshold model, generating and outputting point warning information directed to the spatial analysis unit includes the following steps:

[0150] Inside the classification hyperplane of the two-dimensional risk coordinate space, at least two warning level surfaces are defined according to the distance from the hyperplane to construct a graded warning area;

[0151] Calculate the distance between each risk status point and the nearest warning level surface in real time;

[0152] Divide the spatial analysis unit into different warning states according to distance;

[0153] Present different warning states in different visualization methods on the GIS platform;

[0154] If and only if the risk status point of the spatial analysis unit crosses the outermost classification hyperplane, the point warning information of the spatial analysis unit is generated.

[0155] In this embodiment, the graded warning area is constructed by defining an equidistant surface inside the classification hyperplane as the final risk boundary. As a benchmark, we can generate multiple parallel warning lines by moving different distances inside the safe area. For example, we can define the first-level warning surface as and the second-level warning area ,in and is a preset distance constant, and In this way, multiple nested buffer zones such as attention and warning are divided within the original safety zone. The effect of this step is to transform the risk assessment from a black-and-white switch to a risk dashboard with multiple scales, which can prompt the approach of risks earlier. In order to determine the specific warning level of each spatial analysis unit, it is necessary to calculate the positional relationship between its risk status point and these warning level surfaces in real time. For any risk status point in the two-dimensional risk coordinate space , which reaches any warning surface The directed distance can be expressed by the formula Calculated, where is the intercept corresponding to different warning surfaces. The absolute value of the distance indicates the degree of proximity, while the sign indicates which side of the warning surface it is located on. By continuously calculating the distance between a point and all warning surfaces, its dynamic position within the graded warning area can be accurately tracked.

[0156] Based on the relative position of risk points and warning level surfaces, each spatial analysis unit can be automatically classified into multiple different warning states. This classification process is based on interval judgment, assigning corresponding status labels by examining which two warning surfaces a risk point falls between. For example, if a point has crossed the first-level warning surface but has not yet reached the second-level warning surface, it is labeled as a first-level warning state; if it has crossed the second-level warning surface, it is labeled as a second-level warning state. This step effectively converts continuous distance calculation results into discrete warning levels with clear management implications, such as normal, concern, severe, and critical, enabling managers to clearly understand the risk level of each engineering unit. To convey warning information intuitively and efficiently, different warning states need to be presented in different visualization methods within the GIS platform. Since each spatial analysis unit is a polygon with a geographic location in GIS, the warning state can be directly linked to the visual attributes of the graphic. For example, units in a normal state can be rendered green, units in a concern state can be rendered blue, units in a severe state can be rendered yellow, and units in a critical state can be rendered orange. This situation map based on the geographic information system can display the risk levels of all units along the entire line in real time and dynamically on the map.

[0157] To ensure the seriousness and accuracy of the highest-level alert, a point-level warning for a spatial analysis unit is generated and output only when the risk status point of that unit crosses the outermost final classification hyperplane trained from historical failure data. This crossing event indicates that the risk evolution trend of the unit has entered the confirmed dangerous category, marking the critical transition from quantitative change to qualitative change. At this point, the system triggers the highest-level alert mechanism, generating a formal warning message containing the unique identifier and precise geographic coordinates of the spatial analysis unit. This step effectively establishes a clear action trigger, strictly distinguishing the early multi-level warnings from the final highest-level alert requiring immediate emergency measures, ensuring the hierarchical and precise decision-making response.

[0158] The present invention also discloses a water conservancy monitoring and early warning system based on a GIS platform, comprising a memory, a processor, and a computer program stored on the memory and runnable on the processor. When the processor executes the computer program, it implements the water conservancy monitoring and early warning method based on the GIS platform as described in any one of the above embodiments.

[0159] Among them, the processor can adopt a central processing unit (CPU). Of course, according to actual usage, other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. can also be adopted. The general-purpose processor can adopt a microprocessor or any conventional processor, etc., and this application does not impose any restrictions on this.

[0160] Among them, the memory can be an internal storage unit of a computer device, such as a hard disk or memory of a computer device, or an external storage device of a computer device, such as a plug-in hard disk, smart memory card (SMC), secure digital card (SD) or flash memory card (FC) equipped on the computer device. In addition, the memory can also be a combination of an internal storage unit and an external storage device of a computer device. The memory is used to store computer programs and other programs and data required by the computer device. The memory can also be used to temporarily store data that has been output or is to be output. This application does not impose any restrictions on this.

[0161] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of protection of the present application is limited to these examples. In line with the present application, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of different aspects of one or more embodiments of the present application as described above, which are not provided in detail for the sake of simplicity.

[0162] The one or more embodiments of this application are intended to encompass all such substitutions, modifications, and variations that fall within the broad scope of this application. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of one or more embodiments of this application should be included in the scope of protection of this application.

Claims

1. A water conservancy monitoring and early warning method based on a GIS platform, characterized in that: The steps include: Lay distributed fiber optic sensors along the structural axis of the target water conservancy project, and integrate the geospatial information and structural attribute information of the target water conservancy project into the GIS platform; Based on the structural axis, a one-dimensional linear reference system is created in the GIS platform to map the measurement points of the distributed fiber optic sensors. Taking the mapping point corresponding to the measurement point on the one-dimensional linear reference system as the center, multiple spatial analysis units corresponding to the physical structure of the target water conservancy project are projected perpendicular to the structural axis. Collect sensor data from distributed fiber optic sensors in real time and assign the sensor data to corresponding spatial analysis units based on mapping relationships. The sensor data includes strain data and acoustic wave data. The acoustic wave data of each spatial analysis unit is processed and pattern matched with a preset acoustic fingerprint library to identify internal erosion problems caused by internal structural deterioration of the target water conservancy project; The effective bearing capacity of each spatial analysis unit is dynamically calculated by combining the strain data of each spatial analysis unit and the structural attribute information of the spatial analysis unit in the GIS platform; A risk evolution threshold model is constructed. When and only when internal erosion problems are monitored simultaneously in the same spatial analysis unit, and the risk state point formed by the decrease rate of effective bearing capacity and the energy growth rate of internal erosion problems crosses the risk evolution threshold model, point warning information pointing to the spatial analysis unit is generated and output.

2. The water conservancy monitoring and early warning method based on the GIS platform according to claim 1 is characterized in that: The integration of the geospatial information and structural attribute information of the target water conservancy project in the GIS platform includes the following steps: Obtain and digitize the design center line of the target water conservancy project as the structural axis; Import the digital elevation model, geological exploration layer, and historical hazard point layer of the target water conservancy project area as geospatial information; Extract soil type and permeability coefficient from geological exploration layers as structural attribute information; Perform spatial registration of structural axes, geographic spatial information, and structural attribute information in a unified geographic coordinate system to build a unified GIS data base.

3. The water conservancy monitoring and early warning method based on the GIS platform according to claim 2 is characterized in that: The method of creating a one-dimensional linear reference system in the GIS platform based on the structural axis to map the measurement points of the distributed optical fiber sensor, and projecting perpendicularly to the structural axis with the mapping point corresponding to the measurement point on the one-dimensional linear reference system as the center to generate multiple spatial analysis units corresponding to the physical structure of the target water conservancy project includes the following steps: Map the measurement points of the distributed fiber optic sensor to a one-dimensional linear reference system at preset intervals, and establish a corresponding relationship between the measurement points and the project mileage; The mapping points corresponding to each measurement point on the one-dimensional linear reference system are used as the reference; Query the cross-sectional design parameters of the corresponding project mileage in the GIS data base; Dynamically calculate projection width and projection shape based on cross-sectional design parameters; The projection width and projection shape are combined to generate polygons perpendicular to the structural axis that cover the entire cross section of the target water conservancy project, and each polygon is defined as a spatial analysis unit; Assign a unique identifier to each spatial analysis unit and associate the geographic spatial information and structural attribute information corresponding to the spatial analysis unit.

4. The water conservancy monitoring and early warning method based on the GIS platform according to claim 3 is characterized in that: The signal processing of the acoustic wave data of each spatial analysis unit and the identification of internal erosion problems caused by internal structural deterioration of the target water conservancy project by pattern matching with a preset acoustic fingerprint library include the following steps: The acoustic wave data assigned to each spatial analysis unit is intercepted according to a time window to obtain an acoustic wave data segment; Applying a bandpass filter to remove environmental background noise in the acoustic wave data segment that is not related to internal structural degradation; Performing wavelet packet decomposition on the filtered sound wave data segment to obtain energy distribution characteristics of the sound wave data segment in multiple preset frequency bands; Calculate vector similarity between energy distribution features and patterns in the acoustic fingerprint library; When the similarity calculation result exceeds the preset matching threshold, it is determined that an internal erosion problem has occurred and the energy intensity of the internal erosion problem is recorded.

5. The water conservancy monitoring and early warning method based on the GIS platform according to claim 4 is characterized in that: The acoustic fingerprint library is pre-built through the following steps: Construct a flume model in the laboratory to simulate the physical process of soil erosion within the target water conservancy project under different working conditions; Use distributed fiber optic sensors to collect acoustic wave signals generated by physical processes; Process the collected acoustic wave signals, extract the energy distribution characteristics of the acoustic wave signals in the key frequency bands, and form the initial acoustic fingerprint; Collect background acoustic wave data at the actual site of the target water conservancy project during the non-flood season and under no-risk conditions; The background sound wave data is used to perform noise adaptation and pattern optimization on the initial acoustic fingerprint to generate the final acoustic fingerprint library.

6. The water conservancy monitoring and early warning method based on the GIS platform according to claim 1 is characterized in that: The dynamic calculation of the effective bearing capacity of each spatial analysis unit by combining the strain data of each spatial analysis unit and the structural attribute information of the spatial analysis unit in the GIS platform includes the following steps: The theoretical bearing capacity of each spatial analysis unit is calculated based on the structural attribute information of each spatial analysis unit; Quantitatively evaluate the vulnerability index of each spatial analysis unit based on the historical records of dangerous conditions and geological conditions in the structural attribute information; Construct a structural degradation function that takes the energy intensity and duration of the internal erosion problem as input; Substitute the output values ​​of strain data, theoretical bearing capacity, fragility index and structural degradation function into the preset geotechnical mechanics correction model; The real-time effective bearing capacity of the spatial analysis unit is calculated.

7. The water conservancy monitoring and early warning method based on the GIS platform according to claim 6 is characterized in that: The quantitative evaluation of the vulnerability index of each spatial analysis unit based on the historical dangerous situation records and geological conditions in the structural attribute information includes the following steps: Select at least three key factors affecting structural stability from the structural attribute information; The analytic hierarchy process is used to determine the weight coefficient of each key influencing factor; Normalize the original parameter values ​​of each key influencing factor in each spatial analysis unit; Perform weighted summation on the normalized parameter value and the corresponding weight coefficient; The final weighted summation result is used as the vulnerability index of the spatial analysis unit.

8. The water conservancy monitoring and early warning method based on the GIS platform according to claim 1 is characterized in that: The construction of the risk evolution threshold model includes the following steps: Define a two-dimensional risk coordinate space, where the two coordinate axes are the rate of decrease of effective bearing capacity and the rate of energy growth of internal erosion problem respectively. Mapping historical dangerous situation data and laboratory simulated failure data into multiple failure sample points in a two-dimensional risk coordinate space; The support vector machine algorithm is used to train the failure sample points and generate a classification hyperplane that divides the safe area and the dangerous area as the risk evolution threshold model; Continuously calculate the real-time position of the risk status point of each spatial analysis unit in the two-dimensional risk coordinate space; When the risk state point crosses the classification hyperplane from the safe area to the dangerous area, the warning triggering condition is met.

9. The water conservancy monitoring and early warning method based on the GIS platform according to claim 8 is characterized in that: The step of generating and outputting point warning information directed to the spatial analysis unit when and only when internal erosion problems are simultaneously monitored in the same spatial analysis unit and the risk state point formed by the decrease rate of the effective bearing capacity and the energy growth rate of the internal erosion problem crosses the risk evolution threshold model comprises the following steps: Inside the classification hyperplane of the two-dimensional risk coordinate space, at least two warning level surfaces are defined according to the distance from the hyperplane to construct a graded warning area; Calculate the distance between each risk status point and the nearest warning level surface in real time; Divide the spatial analysis unit into different warning states according to distance; Present different warning states in different visualization methods on the GIS platform; If and only if the risk status point of the spatial analysis unit crosses the outermost classification hyperplane, the point warning information of the spatial analysis unit is generated.

10. A water conservancy monitoring and early warning system based on a GIS platform, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the water conservancy monitoring and early warning method based on the GIS platform as described in any one of claims 1 to 9 is implemented.

Images