Power transmission and transformation project self-adaptive water and soil conservation intelligent decision-making system and method based on meteorological-landform dynamic coupling

Abstract

The invention discloses a power transmission and transformation project self-adaptive water and soil conservation intelligent decision-making system and method based on meteorological-landform dynamic coupling. The system comprises a multi-source data fusion module, a dynamic risk assessment and prediction module, a self-adaptive measure intelligent decision module and an early warning push module. The core innovation lies in that the self-adaptive measure intelligent decision-making module is internally provided with a parameterization measure knowledge base, and can dynamically generate a personalized prevention and control scheme containing an accurate engineering size by combining real-time landform parameters and executing predefined computational logic according to the dynamically predicted risk type and grade. According to the method, short-term weather forecast is deeply fused into engineering decision, normal form transformation from static experience design to dynamic parameterization intelligent design is achieved, the problem that water and soil conservation measures are disjointed with real-time climate-landform conditions is solved, and the perspectiveness, accuracy and automation level of power transmission and transformation project water and soil loss prevention and control are remarkably improved.

Description

Technical Field

[0001] This invention relates to the fields of ecological engineering, disaster prevention and mitigation, and power engineering. Specifically, it relates to an intelligent system and method for soil and water conservation in power transmission and transformation projects that combines climate change adaptation theory with complex geomorphological features. In particular, it relates to a decision support system and method that can integrate short-term weather forecasts, high-precision geographic information, and real-time engineering disturbance data to achieve dynamic prediction of soil erosion risk, parameterized intelligent generation of prevention and control measures, and adaptive execution. Background Technology

[0002] As typical linear engineering projects, power transmission and transformation projects generate numerous artificial disturbance surfaces during tower foundation excavation and construction road building, significantly altering regional micro-topography and hydrological processes. Under extreme weather events, these projects are highly susceptible to becoming sources of soil erosion. Soil erosion not only causes the loss of soil resources and damage to the ecological environment but also directly threatens the stability of tower foundations, leading to major safety accidents such as tower collapse and power line breakage.

[0003] Currently, soil and water conservation technologies for power transmission and transformation projects mainly follow standards such as the "Technical Standard for Soil and Water Conservation of Production and Construction Projects" (GB50433-2018), forming a prevention and control system primarily based on engineering and vegetation measures. However, with the intensification of global climate change, the frequent occurrence of extreme rainfall and abnormal freeze-thaw events, coupled with the expansion of power transmission and transformation projects into areas with more complex terrain, the traditional technical system faces severe challenges in terms of scientific rigor, accuracy, and timeliness. Its limitations are mainly reflected in the following three aspects:

[0004] First, the design paradigm is static and rigid, severely disconnected from dynamic climate risks. Existing technologies rely on historical climate statistics (such as multi-year average rainfall) for "one-off" designs, using fixed return periods to determine engineering parameters. This static design paradigm is completely incapable of responding to short-term extreme weather events with high uncertainty. For example, a sudden torrential rainstorm far exceeding historical statistical values ​​can cause a drainage system designed according to specifications to fail instantly. Furthermore, current design methods view single rainfall events in isolation, ignoring the complex erosion pattern of "high soil moisture content in the early stages + subsequent torrential rain," which leads to a dramatic increase in the runoff coefficient, resulting in systematic biases in risk prediction. While related patent literature attempts to incorporate historical data analysis, it does not address the issue of responding to specific future weather events.

[0005] Second, the decision-making process for mitigation measures is often based on experience and is too rudimentary, lacking sufficient coupling with the micro-characteristics of the site. Existing measures often employ a "classification-matching" model based on macro-characteristics such as slope and landform type. Key dimensional parameters of the measures (such as drainage ditch cross-section and lattice beam spacing) are usually given within empirical ranges in specifications or rely on the subjective judgment of designers. For a long time, this field has lacked a technical method and system capable of quantitatively and automatically calculating optimal engineering parameters based on the micro-characteristics of a specific site, such as catchment area, soil infiltration rate, and slope geometry. This leads to a tendency for mitigation measures to be applied in a "one-size-fits-all" manner, resulting in either insufficient protection leading to safety hazards or over-design causing resource waste. The optimization models proposed in related research papers are mostly based on expert scoring, essentially remaining semi-qualitative decisions and failing to achieve continuous and precise generation of engineering parameters.

[0006] Third, the management process is fragmented, lacks intelligence, and suffers from delayed emergency response. In the existing management process, risk monitoring, analysis and early warning, solution development, and on-site execution are relatively independent, relying on manual coordination and experience-based decision-making. Although some so-called "smart water conservation" systems have achieved visualization of monitoring data and simple over-threshold alarms, their core flaw lies in "having perception but no decision-making, having early warning but no solution." After a system alarm, risks still need to be assessed manually, countermeasures formulated based on experience, and construction organized. The entire process is time-consuming and labor-intensive, and the response is severely delayed in emergency situations such as extreme weather, failing to meet the timeliness requirements for ensuring the safety of critical infrastructure.

[0007] In summary, the urgent technical problem to be solved in this field is: how to break through the existing static, experience-based, and fragmented technical framework, and create an integrated system and method that can dynamically integrate real-time and future meteorological information, automatically generate quantitatively implementable soil and water conservation engineering schemes with optimal parameters based on the refined topographic and soil characteristics of specific engineering sites, and achieve intelligent early warning and rapid response, so as to scientifically address the increasingly severe soil erosion risks of power transmission and transformation projects under the background of climate change. Summary of the Invention

[0008] Purpose of the invention

[0009] The purpose of this invention is to overcome the aforementioned deficiencies of the prior art and provide an adaptive intelligent decision-making system and method for soil and water conservation in power transmission and transformation projects based on dynamic coupling of meteorology and geomorphology. This system aims to achieve the following objectives:

[0010] 1) Dynamic risk prediction: Using short-term weather forecasts (6-72 hours in the future) as the core input, we can achieve advanced, accurate and dynamic assessment of soil erosion risk under specific future weather processes.

[0011] 2) Parameterization and precision of measures generation: Abandoning experience-based selection, a knowledge base of soil and water conservation measures that can be calculated parametrically is constructed to achieve "algorithmic" design that dynamically outputs engineering parameters at the construction drawing level based on micro-site conditions.

[0012] 3) Closed-loop and intelligent management processes: This involves streamlining the entire process from data perception, risk prediction, intelligent decision-making, and solution delivery to automatic execution and feedback, forming an intelligent management closed loop and significantly improving the automation level and timeliness of emergency response. To achieve the above objectives, this invention adopts the following technical solution:

[0013] Technical solution

[0014] To achieve the above-mentioned objectives, the present invention adopts the following technical solution:

[0015] In a first aspect, the present invention provides an adaptive intelligent decision-making system for soil and water conservation in power transmission and transformation projects based on dynamic coupling of meteorology and geomorphology.

[0016] The system includes a multi-source data fusion module that is logically connected and interconnected, a dynamic risk assessment and prediction module, an adaptive measures intelligent decision-making module, a visualization interaction and early warning push module, and an optional intelligent execution feedback module.

[0017] 1. Multi-Source Data Fusion Module: Serving as the system's data foundation, this module is configured to acquire and process multi-dimensional spatiotemporal data from external data sources in real-time or periodically. The data includes at least: real-time monitoring data and high-resolution spatiotemporal gridded forecast data for the next 6-72 hours from meteorological departments; vegetation index and soil moisture inversion products from remote sensing platforms; high-precision digital elevation models (DEMs) and soil type and soil erodibility factor (K-value) data from geographic information systems; and building information model (BIM) or geographic information system (GIS) data from engineering management systems, used to characterize construction stages, disturbance range, and intensity. This module performs coordinate system matching, spatiotemporal scale matching, missing value imputation, and quality checks on the aforementioned multi-source heterogeneous data, ultimately fusing and constructing an "engineering digital twin" capable of synchronously mapping the actual physical state of the engineering project and the future short-term meteorological environment.

[0018] The specific implementation of the multi-source data fusion module in constructing the 'engineering digital twin' includes the following steps:

[0019] Step 1: Coordinate System 1

[0020] All spatial data are unified to the CGCS2000 National Geodetic Coordinate System, and the elevation datum adopts the 1985 National Elevation Datum.

[0021] For WGS84 coordinate data, a seven-parameter transformation method is used, and the transformation parameters are determined based on the control points of the engineering area;

[0022] For data in local coordinate systems, first convert to WGS84, then convert to CGCS2000;

[0023] The conversion residuals are controlled within ±0.5m in the horizontal plane and ±0.3m in the vertical plane.

[0024] Step 2: Spatiotemporal scale matching

[0025] Establish a unified spatial grid and time reference:

[0026] Spatial grid: A regular 30m×30m grid is used, covering a 500m range on both sides of the project line;

[0027] Time base: Coordinated Universal Time (UTC) is the base, and all time data is converted to UTC time;

[0028] Resampling methods:

[0029] Meteorological data: interpolated from the meteorological department's 0.01° grid to a 30m grid using bilinear interpolation;

[0030] Remote sensing data: resampled from the original resolution to a 30m grid using nearest neighbor or bilinear interpolation;

[0031] Engineering data: rasterized from vector data to a 30m grid.

[0032] Step 3: Time Synchronization and Alignment

[0033] For real-time data: Establish a data buffer queue to solve the transmission delay problem;

[0034] For periodic data: the most recent data is filled in backwards, with the hour as the baseline;

[0035] Establish a time index: use 10-minute time slices, and data within the same slice is considered synchronized.

[0036] Step 4: Handling Missing Values

[0037] A layered interpolation strategy is adopted:

[0038] 1) First, try using linear interpolation of data from adjacent time points using the same sensor;

[0039] 2) If the data is still missing, use relevant data from other sensors at the same grid point to perform multivariate regression interpolation;

[0040] 3) If still missing, perform kriging interpolation using data from similar grid points in the spatial neighborhood;

[0041] 4) Data points with an interpolation ratio exceeding 20% ​​are marked as low-quality data.

[0042] Step 5: Quality Inspection and Control

[0043] Establish a three-level quality control system:

[0044] Level 1 inspection (format inspection): Checks whether the data format, units, and range conform to the specifications;

[0045] Secondary inspection (logic inspection): Check the logical rationality of the data, such as: rainfall cannot be negative, soil moisture is between 0-100%, and the slope does not change abruptly at the same location (slope difference between adjacent grids <15°).

[0046] Level 3 verification (consistency test): cross-validation of the same parameter from different data sources, such as: the difference between remote sensing soil moisture and ground sensor data ≤15%, and the correlation coefficient between meteorological station rainfall and radar-retrieved rainfall ≥0.7.

[0047] Step 6: Construction of Digital Twin Data Structure

[0048] Construct a multi-layered spatiotemporal data cube.

[0049] 2. Dynamic Risk Assessment and Prediction Module: As the core of the system's intelligent perception, its input is connected to the output of the multi-source data fusion module. This module incorporates a machine learning prediction model trained and validated using extensive historical disaster case data. The model is configured to: use real-time and forecast meteorological elements, static landform and soil properties, and dynamic engineering disturbance information provided by the "engineering digital twin" as input features; through model inference, it continuously calculates and outputs the probability value of soil erosion risk for each spatial evaluation unit (e.g., a 30m×30m grid) along the project route, divided according to rules, within multiple preset forecast periods (e.g., 6h, 12h, 24h, 72h); then, based on preset probability threshold intervals, it quantifies the risk into multiple levels (e.g., low, medium, high, extremely high); simultaneously, based on the analysis of the importance or combination rules of the input features, it intelligently identifies the dominant risk type for each high-risk unit. These types include at least "hydraulic erosion" dominated by rainfall runoff, "freeze-thaw erosion" dominated by freeze-thaw cycles, and "composite erosion" resulting from a coupling of both.

[0050] 3. Adaptive Measures Intelligent Decision-Making Module: This is the core innovation of the invention, namely the system's intelligent decision-making engine, whose input is connected to the dynamic risk assessment and prediction module. The core of this module is to construct and maintain a parameterized, computable knowledge base of soil and water conservation measures. This knowledge base is fundamentally different from traditional "measures lists" or "standard atlases":

[0051] 1) The database stores “measure calculation units”, each unit corresponding to a type of soil and water conservation technology (such as “ecological interception ditch unit”, “honeycomb infiltration pond unit”, and “freeze-thaw resistant cover unit”).

[0052] 2) Each "measure calculation unit," in addition to being associated with its applicable topographical type and risk type conditions, is crucially associated with a set of predefined "parametric calculation logic." This logic clarifies the quantitative relationship between the key design parameters of the measure and environmental variables. Specifically, it includes:

[0053] (1) For measures based on physical models (such as drainage facilities):

[0054] Use the appropriate engineering calculation formulas, for example:

[0055] - Ecological interception trench: The relationship between B, H and Qp, n, i, m is established based on Manning's formula.

[0056] - Infiltration tank volume: based on the water balance formula V = k×C×I×A, where k is the safety factor or adjustment factor.

[0057] The methods for determining each parameter are as follows:

[0058] Design peak flow Qp: Calculated using the SCS-CN method, Qp = 0.278 × C × I × A, where C is the runoff coefficient, I is the rainfall intensity, and A is the catchment area.

[0059] Roughness coefficient n: Determined by referring to a table based on the lining type.

[0060] Slope coefficient m: determined based on soil type and stability requirements.

[0061] (2) For measures based on experience or rules (such as covering materials):

[0062] Decision rule trees or empirical formulas can be used, for example:

[0063] - Covering material usage Q = α × S × (1 + β × T).

[0064] Where α is the basic dosage coefficient, β is the temperature correction coefficient, which is determined through regression analysis of experimental data; S is the area to be covered (m²), and T is the actual ambient temperature.

[0065] (3) For measures that need optimization:

[0066] Establish an optimization model with the objective function of minimizing cost or maximizing benefit, and constraints based on engineering requirements.

[0067] Numerical methods (such as iterative search and linear programming) are used to solve for the optimal parameters.

[0068] This module is configured to: upon receiving high-risk unit information, first intelligently match one or more applicable "measure calculation units" from the knowledge base based on the unit's dominant risk type and geomorphological and soil properties, and determine their combination. Then, it executes the core "parameter dynamic calculation" step: extracting the current specific environmental variable values ​​of the high-risk unit and substituting them as input parameters into the "parametric calculation logic" of each matched "measure calculation unit." By running the built-in calculation formula or rule engine, it calculates in real-time the optimal engineering parameter values ​​for the measure in the current specific scenario. The implementation of the parameterized calculation logic includes:

[0069] 1. For hydraulic engineering measures: establish the objective function and constraints based on hydraulic formulas, and solve them using numerical optimization methods;

[0070] 2. For structural engineering measures: calculations are performed based on structural mechanics formulas and material mechanics parameters, in accordance with relevant design specifications;

[0071] 3. For biological measures: Parameters are determined based on plant growth models and ecological principles, combined with local climate and soil conditions.

[0072] Taking the calculation unit of ecological interception ditch measures as an example, its parameterized calculation logic is based on the hydraulic Manning formula, specifically including:

[0073] (1) Input parameters: design flood peak flow Qp, channel roughness n, longitudinal slope i, and slope coefficient m;

[0074] (2) Calculation objective: To minimize the cross-sectional area A = B·H + m·H² under the condition that the flow capacity Q ≥ Qp is satisfied;

[0075] (3) Calculation method: An iterative algorithm is used to search within the range of B∈[0.3,1.0]m and H∈[0.3,0.8]m with a step size of 0.05m to find the (B,H) combination that satisfies Q≥Qp and minimizes A;

[0076] (4) Output parameters: optimal bottom width B*, depth H*, cross-sectional area A*.

[0077] Finally, these calculated parameter values ​​are filled into a standardized measure design template, automatically assembling a personalized soil and water conservation measure configuration plan that includes precise engineering quantities, material specifications, construction key points, and schematic diagrams. Optionally, the module can also call hydrological erosion models (such as RUSLE and SWAT) to simulate, compare, and evaluate the benefits of different measure combinations.

[0078] 4. Visual Interaction and Early Warning Push Module: Serving as the system's human-computer interaction and command and dispatch center, this module receives and integrates the dynamic risk level spatial distribution map (in the form of heat maps, etc.) and the generated personalized measure configuration schemes, forming an integrated "risk-measure" linked view. Simultaneously, this module, through message middleware or application programming interfaces (APIs), pushes risk warning information of different levels and their corresponding detailed, directly actionable measures to user terminals (including web management platforms, mobile apps, etc.) of different roles such as project management, design, construction, and operation and maintenance in a structured and automated manner.

[0079] 5. Optional Intelligent Execution Feedback Module: Serving as the system's on-site execution and optimization arm, this module consists of Internet of Things (IoT) hardware devices pre-deployed at key engineering locations, including but not limited to: electrically adjustable weirs / valvees installed at drainage ditch outlets or inlet reservoirs; automatically deployable / retractable protective fabric systems for temporary slope covering; and a network of soil temperature, humidity, seepage pressure, and displacement monitoring sensors. This module receives automated control commands from the adaptive measures intelligent decision-making module (such as "open a certain weir to α% opening degree at time T"), driving on-site equipment to execute predetermined control actions. Simultaneously, it continuously collects equipment status and environmental feedback data and transmits it back to the system, providing data support for measure effectiveness evaluation and model iterative optimization.

[0080] Furthermore, the machine learning model in the dynamic risk assessment and prediction module preferably employs ensemble learning algorithms such as random forest and gradient boosting decision tree (GBDT), as these algorithms can effectively handle high-dimensional nonlinear relationships and output feature importance. The feature vector input to the model needs to be carefully constructed and should at least cover: the forecast rainfall intensity (e.g., maximum 1-hour rainfall intensity) and cumulative rainfall for key future periods, the previous soil moisture index based on remote sensing or model assimilation, the slope and aspect extracted from the DEM, the soil erodibility K value, the vegetation cover (NDVI) and its rate of change, and quantitative indicators representing the intensity of engineering disturbances such as excavation, backfilling, and dumping.

[0081] Furthermore, the construction of the parameterized soil and water conservation measures knowledge base combines domain knowledge with data-driven approaches. For example:

[0082] For the "ecological interception ditch unit" targeting "hydraulic erosion type" risks, its parametric calculation logic is established based on the hydraulic Manning formula. The function inputs are the predicted design peak flow rate Qp (calculated by the SCS-CN method or inference formula based on the predicted rainfall intensity and catchment area), the ditch roughness n (determined according to the lining type), and the longitudinal slope i. By iteratively solving the Manning formula, the ditch's flow capacity (design drainage capacity) Q = (A·R^(2 / 3)·i^(1 / 2)) / n (where A is the flow area, for a trapezoidal cross-section: A = B·H + m·H², and R is the hydraulic radius, both of which are functions of the bottom width B and the depth H). Under the constraint of flow capacity, with the goal of minimizing the engineering workload, the optimal values ​​of B and H are dynamically output.

[0083] For "freeze-thaw erosion type" risks, the parameterized calculation logic of the "freeze-thaw resistant cover unit" can be based on heat conduction models or empirical rules. The function inputs are the predicted negative temperature index, number of freeze-thaw cycles, slope aspect, and soil texture; through table lookup or calculation, the output parameters include recommended cover material type (such as hardener, fiber blanket), dosage per unit area, and laying thickness.

[0084] Furthermore, in the intelligent execution feedback module, the electric regulating weir gate can be driven by a stepper motor or a servo motor and has an opening feedback function; the automatic coverage system can integrate a meteorological sensor to realize triggering based on local real-time weather.

[0085] Secondly, this invention provides an adaptive intelligent decision-making method for soil and water conservation in power transmission and transformation projects based on dynamic coupling of meteorology and geomorphology.

[0086] This method is executed by the aforementioned system and includes the following steps:

[0087] S100. Digital Twin Construction and Update: Continuously aggregate and integrate multi-source spatiotemporal data to establish and dynamically update a high-fidelity "digital twin" along the power transmission and transformation project, serving as a unified data base for all analysis and decision-making.

[0088] S200. Dynamic Risk Simulation and Prediction: Using a pre-trained AI prediction model, the forecast meteorological field for different future periods is applied to the "digital twin" to simulate and output a spatiotemporal dynamic evolution picture of soil erosion risk, and identify high-risk areas and their dominant erosion types.

[0089] S300. Parametric Intelligent Solution Decision-Making: This step is the core of this method and includes:

[0090] S310. Measure Unit Matching: For each identified high-risk unit, based on its risk label (type and level) and site attribute label, retrieve and combine applicable basic "measure calculation units" from the parametric measure knowledge base.

[0091] S320. Parameter Dynamic Calculation: The specific environmental parameter values ​​of the current high-risk unit are used as variables and substituted into the predefined mathematical function or rule set of the "measure calculation unit" matched in step S310. The calculation process is executed to directly solve for the optimal engineering parameter value of the measure in the current scenario.

[0092] S330. Scheme Generation and Optimization: The parameter values ​​calculated in S320 are automatically filled into the corresponding standardized engineering drawings or construction instruction templates to form a customized prevention and control scheme that can be implemented immediately. Optionally, multi-scheme hydraulic and hydrological simulation and cost-benefit analysis are performed to recommend the comprehensive optimal solution.

[0093] S400. Collaborative Early Warning and Precise Push: The visualized risk distribution and structured, actionable prevention and control solutions are packaged and pushed to front-line construction, supervision, and management decision-makers in a targeted and real-time manner through a collaborative work platform, realizing "early warning as a contingency plan".

[0094] S500. Optional Intelligent Execution and Closed-Loop Learning: Through an IoT control gateway, the automatically executable parts of the solution are converted into control commands to drive field equipment to perform preventive and control actions (such as pre-flood discharge and coverage deployment). Simultaneously, execution process and post-event monitoring data are collected and fed back to the system for evaluating effectiveness, calibrating models, and optimizing knowledge base rules, forming a closed loop of continuous improvement.

[0095] Thirdly, the present invention provides a computer-readable storage medium.

[0096] The storage medium stores a computer program that, when executed by a processor, can perform all or part of the steps of the method as described in the second aspect.

[0097] Beneficial effects

[0098] Compared with the closest existing technology, the substantial features and significant progress brought about by this invention are reflected in:

[0099] 1. This invention pioneered a dynamic adaptive design paradigm for soil and water conservation driven by short-term weather forecasts, resolving the fundamental contradiction between static design and dynamic risk. Existing technologies are static designs oriented towards historical averages, while this invention, for the first time, uses deterministic and probabilistic weather forecasts for the next 6-72 hours as core dynamic input variables, driving the entire process from risk identification to the generation of measure parameters. This endows the soil and water conservation system with "foresight," enabling proactive deployment for specific weather events "about to occur," achieving a paradigm shift from passive disaster relief to proactive disaster prevention, and significantly enhancing the ability to cope with the uncertainties of climate change.

[0100] 2. This invention proposes an intelligent decision-making method based on "parametric calculation of measures," achieving a leap from "experience-based selection" to "model generation" and solving the problem of extensive measure implementation. Unlike existing qualitative / semi-qualitative decision-making methods based on classification matching or expert scoring, this invention creatively proposes the concept of a "parametric measure knowledge base." Within this framework, each soil and water conservation technology is abstracted as a "mathematical function" or "algorithm module." The decision-making process is no longer about choosing "which measure," but rather calculating "how this measure should be implemented under current conditions." For example, an "ecological interception ditch" is defined as the function Ditch = F(Qp, n,i, cost), and the system directly outputs the optimal cross-sectional dimensions by solving this function. This method, for the first time in the field, achieves infinitely refined and quantitative customization of engineering measures, fundamentally overcoming the drawbacks of "one-size-fits-all" design and greatly improving resource utilization efficiency while ensuring safety.

[0101] 3. An end-to-end intelligent decision-making and execution closed loop has been constructed, bridging the gap between risk perception and on-site action, and solving the problems of fragmented management and delayed response. This invention is not a simple improvement to monitoring systems or design tools, but rather the construction of a full-chain intelligent system integrating "perception-prediction-decision-push-execution-feedback." It completely breaks down the barriers between data and action, enabling the system not only to detect risks but also to automatically generate executable "battle plans" and directly command "frontline equipment" to respond automatically. This highly integrated and automated closed-loop management capability reduces emergency response time from "hours" to "minutes," significantly improving the timeliness and reliability of critical infrastructure security protection, representing a substantial leap forward in the intelligent development of this field.

[0102] 4. The technical solution is complete, highly feasible, and has broad prospects for promotion and application. This invention provides detailed technical specifications, including data fusion methods, AI model training frameworks, examples of constructing a parameterized measures library, and the interaction logic of each module in the system. Those skilled in the art, based on the teachings of this invention and combined with mature GIS, cloud computing, and IoT technologies, can fully reproduce and implement this invention. The modular design of the system and the scalability of the knowledge base make it easily adaptable to power transmission and transformation projects in different climate zones and landform types, and can be extended to similar problems in other linear engineering projects such as highways, railways, and pipelines, demonstrating significant practical value. Attached Figure Description

[0103] Figure 1 This is a schematic diagram of the overall system architecture provided in an embodiment of the present invention.

[0104] Figure 2 This is a flowchart of the dynamic risk assessment and prediction module in an embodiment of the present invention.

[0105] Figure 3 This is a flowchart of the adaptive soil and water conservation intelligent decision-making method provided in the embodiments of the present invention. Detailed Implementation

[0106] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments. The following embodiments are for illustrative purposes only and are not intended to limit the invention.

[0107] Example 1: Specific application of the system in responding to sudden rainstorms in the Liaodong mountainous area

[0108] See Figure 1-3 In this embodiment, the system is deployed on a cloud service platform. The example is a 500kV transmission line project under construction in the eastern mountainous region of Liaoning Province.

[0109] Step A: Multi-source data fusion and digital twin construction.

[0110] The multi-source data fusion module runs automatically at predetermined intervals (e.g., every hour). It obtains 0.01° resolution grid precipitation forecasts for the next 72 hours from the provincial meteorological bureau via an API interface; acquires the latest surface soil moisture inversion data from the European Space Agency (ESA) Sentinel satellite data platform; and synchronizes a BIM-based construction progress GIS layer from the EPC project management platform, which precisely marks the excavation status and soil piling location of each tower foundation. The module performs coordinate transformation (to the CGCS2000 coordinate system), spatial interpolation (unified to a 30m grid), and time alignment on all data, generating a digital twin of the project containing multi-dimensional information such as "heavy precipitation forecast for the next 24-48 hours," "current high soil moisture distribution," and "tower foundation T32 is in the deep foundation pit excavation stage."

[0111] Step B: Dynamic risk prediction and identification.

[0112] The dynamic risk assessment and prediction module loads a trained Gradient Boosting Decision Tree (GBDT) model. The model extracts feature vectors from the digital twin of the grid containing base T32, such as: [Maximum rainfall intensity in the next 24 hours: 65 mm / h, previous soil moisture index: 0.87, slope: 30°, soil K-value: 0.03, disturbance intensity: 3 (heavy excavation)...]. After inference, the model outputs: the probability of severe soil erosion occurring in this grid is 92%. Based on a preset threshold (>85% is extremely high risk), the risk level of this point is determined to be "extremely high". Simultaneously, the model analyzes the importance of features, determining that rainfall intensity and previous moisture are the dominant factors, thus identifying the dominant risk type as "torrential rain erosion".

[0113] Step C: Parameterized intelligent measures decision-making.

[0114] The adaptive measures intelligent decision-making module is triggered. First, the measures unit is matched: based on the conditions of "rainstorm erosion type" + "steep mountain slope", the "mountain rainstorm emergency prevention and control combination" is called from the knowledge base. This combination includes "ecological interception ditch unit" and "honeycomb infiltration pond unit" by default.

[0115] Subsequently, the core parameter dynamic calculation is performed:

[0116] For the "ecological interception ditch unit," the module calls its associated calculation function `calculate_ditch_params(Qp, slope, soil_type)`. This function first calculates the design peak flow rate `Qp = 0.52 m³ / s` based on the grid's catchment area (250 m² from DEM analysis) and the predicted rainfall process line using an inference formula. Then, based on Manning's formula `Q = (A·R^(2 / 3)·i^(1 / 2)) / n`, the ecological ditch is set as a trapezoidal cross-section (slope coefficient m = 1.5), roughness coefficient n = 0.035 (considering vegetation), and longitudinal slope i = 0.03. Internally, the function uses an iterative algorithm to search for a combination of `B` (bottom width) and `H` (depth) that minimizes the cross-sectional area (A) under the constraint that `Q_capacity ≥ Qp`. The calculated recommended parameters are: `B* = 0.45m`, `H* = 0.40m`.

[0117] For the "honeycomb infiltration storage unit," the module calls the function `calculate_pool_volume(W, infiltration_rate)`. Based on the predicted initial rainfall (2 hours prior) and catchment area, the flood volume W = 15.6 m³ is calculated. Considering the local soil infiltration rate, a target retention rate η = 0.7 is set, along with a safety factor. Dynamic calculation yields the required total infiltration storage volume V = 11.5 m³, and it is recommended to use five hexagonal sub-pools (each with a volume of approximately 2.3 m³) arranged in a quincunx pattern.

[0118] The module automatically fills the calculated parameters into the standard "Design and Construction Drawing of Ecological Interception Ditch and Infiltration Pond" template, and supplements the material list (such as dry masonry volume and geotextile area) and construction precautions, generating a complete "Tall Base T32 Rainstorm Emergency Protection Special Plan".

[0119] Step D: Precise delivery of early warnings and solutions.

[0120] The visualization, interaction, and early warning push module marks the location of Tower Base T32 as a prominent flashing red icon on the WebGIS platform. Simultaneously, the generated "Special Plan" PDF file and a brief early warning description are pushed to the project's safety director, site construction foreman, and supervising engineer via the WeChat Work robot API and the project management app's message center. The construction foreman can directly view the key parameters and diagrams in the plan on the mobile app.

[0121] Step E: Intelligent Execution.

[0122] Assume that a honeycomb-type infiltration storage tank has been pre-constructed upstream of the tower foundation and an electric inlet valve has been installed. Three hours before the start of the rainstorm forecast, the intelligent execution feedback module automatically sends a command to the inlet valve: "Open the valve to 60% opening at 14:00 today and continue until the warning is lifted." This aims to proactively lower the water level in the tank before rainfall, freeing up storage capacity to maximize the absorption of the upcoming flood peak. After the valve executes the command, a confirmation signal is fed back to the system.

[0123] Example 2: Application of the system in preventing spring freeze-thaw erosion in the hilly area of ​​western Liaoning

[0124] In early spring, the system predicts that a significant freeze-thaw cycle will occur along a 220kV power line in western Liaoning, coupled with a small amount of rainfall, and outputs a risk type of "freeze-thaw-hydraulic composite erosion".

[0125] The adaptive measures intelligent decision-making module matches the "composite erosion control combination." For the freeze-thaw phase, it invokes the "liquid soil stabilizer spraying unit." Based on the predicted diurnal temperature variation curve (determining the number of freeze-thaw cycles), the slope aspect of the area (southwest, sunny slope), and soil texture (sandy loam), the module consults the built-in expert rule table: when freeze-thaw cycles > 5 times / day, sunny slope, and sandy soil, a permeable soil stabilizer with a concentration of 3.5% is recommended. The module dynamically calculates the spraying rate per unit area should be 1.8 L / m². Simultaneously, for potential accompanying rainfall, simple drainage measures are matched.

[0126] The system will push the "Spring Freeze-Thaw Protection Operation Guidelines" containing precise proportions and dosages to the maintenance teams, guiding them to carry out precise and economical maintenance operations, and avoiding material waste and poor results.

[0127] The above embodiments fully demonstrate how the present invention transforms dynamic weather forecasts and detailed geomorphological data into specific and quantifiable engineering instructions, thereby achieving a unity of scientific rigor, precision, and intelligence in soil and water conservation work.

[0128] Although the present invention has been described in detail through preferred embodiments, those skilled in the art will understand that various modifications, substitutions, and alterations can be made to the present invention without departing from its spirit and scope. The scope of protection of the present invention is defined by the appended claims.

Claims

1. An adaptive intelligent decision-making system for soil and water conservation in power transmission and transformation projects based on dynamic coupling of meteorology and geomorphology, characterized in that, include: The multi-source data fusion module is used to acquire and fuse meteorological forecast data, remote sensing data, geographic information data and engineering disturbance data along the power transmission and transformation project to construct a spatiotemporally consistent digital twin of the project. The dynamic risk assessment and prediction module is connected to the multi-source data fusion module and has a built-in machine learning model. It is used to predict and output the soil and water loss risk level and dominant risk type of the spatial grid along the project in the future period based on the digital twin of the project. The adaptive measures intelligent decision-making module is connected to the dynamic risk assessment and prediction module. It has a built-in parameterized measures knowledge base, which stores multiple measures calculation units. Each measures calculation unit is associated with parameterized calculation logic. This module is used to call the matching measures calculation unit according to the risk level and the dominant risk type, and substitute the current environmental parameters into its parameterized calculation logic to dynamically calculate the key design parameters of the prevention and control measures and generate a personalized soil and water conservation plan. The visualization and interactive early warning push module is used to display risk information and the personalized soil and water conservation plan, and push early warnings and plan instructions to the terminal.

2. The system according to claim 1, characterized in that, The machine learning model in the dynamic risk assessment and prediction module is an ensemble learning model, and its input features include: the forecast rainfall intensity and duration for the next 6-72 hours, the previous soil moisture, slope, soil erodibility factor K value, vegetation coverage, and quantitative indicators of engineering disturbance intensity.

3. The system according to claim 1, characterized in that, In the parameterized measures knowledge base, the parameterized calculation logic of the measure calculation unit for the risk dominated by rainstorm erosion includes a hydraulic calculation function based on Manning's formula, which is used to dynamically solve the cross-sectional dimensions of drainage or interception facilities based on the predicted peak flow. For the measure calculation unit for the risk dominated by freeze-thaw damage, the parameterized calculation logic includes decision rules or functions based on temperature, slope aspect, and soil texture, which are used to dynamically determine the parameters of the cover material or the parameters of micro-topography modification.

4. The system according to claim 3, characterized in that, The hydraulic calculation function based on Manning's formula is configured to take the predicted design peak flow rate Qp, channel roughness n, and design longitudinal slope i as inputs, and iteratively solve the equations relating Manning's formula and cross-sectional geometry to output the channel bottom width B and depth H that meet the flow capacity requirements and are economically efficient.

5. The system according to claim 1, characterized in that, It also includes an intelligent execution feedback module, which includes an IoT actuator deployed at the engineering site to receive and execute automated control commands from the adaptive measures intelligent decision-making module. The IoT actuator includes an electric regulating weir gate, an automatic slope covering device, and environmental monitoring sensors.

6. An adaptive intelligent decision-making method for soil and water conservation in power transmission and transformation projects based on dynamic coupling of meteorology and geomorphology, characterized in that, include: S1. Used to acquire and integrate meteorological forecast data, remote sensing data, geographic information data and engineering disturbance data along the power transmission and transformation project, and to construct and update the digital twin of the project; S2. Based on the aforementioned engineering digital twin, use machine learning models to predict the spatiotemporal distribution and type of soil erosion risk; S3. For identified high-risk areas, based on their risk type and site attributes, match the measure calculation unit from the parametric measure knowledge base, and execute the parametric calculation logic of the unit to dynamically generate a prevention and control plan containing specific values ​​of key design parameters with the current environmental parameters as input. S4. Visualize and push the risk distribution information and the generated prevention and control plan.

7. The method according to claim 6, characterized in that, The "execution of the parameterized calculation logic of the unit" mentioned in step S3 means: substituting the real-time environmental parameter values ​​of the high-risk area into a predefined mathematical function or rule engine for calculation, and directly outputting the specific values ​​of the key engineering dimensions, material usage, or process parameters of the prevention and control measures.

8. The method according to claim 6 or 7, characterized in that, It also includes step S5: based on the generated prevention and control plan, sending control commands to pre-deployed IoT actuators to automatically execute some prevention and control measures and collect feedback data.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by a processor, it implements the steps of the method as described in any one of claims 6 to 8.

Images