Power distribution facility rush-arrangement scheduling method and system based on risk early warning

Abstract

The invention provides a power distribution facility rush-arrangement scheduling method and system based on risk early warning, and the method comprises the steps: obtaining a corrected rainfall forecast field based on the observation data of a rainfall station in a region to be subjected to rush-arrangement scheduling, the grid forecast data and a topographic factor; constructing a local flood forecasting model and inputting a corrected rainfall forecasting field on the basis of topographic and geomorphic characteristics and drainage system layout in a local area of the power distribution facility to obtain water depth data, water flow velocity and submerging duration of the position of the power distribution facility, and calculating an equipment risk score; and determining a candidate power distribution facility based on the equipment risk score, distributing a corresponding robbing scheduling resource through a preset optimization algorithm, generating a robbing scheduling sequence, distributing the robbing scheduling sequence to the power distribution facility, and executing a corresponding robbing scheduling action on the to-be-robbing scheduling region. According to the method, the equipment and regional risks are quantified after the extreme environment of the power distribution facility region is predicted, and the power distribution facility emergency scheduling is optimized in combination with risk assessment, so that the reliability of the power distribution facility emergency scheduling is improved.

Description

Technical Field

[0001] This invention relates to the field of emergency dispatching technology, and in particular to a method and system for emergency dispatching of power distribution facilities based on risk early warning. Background Technology

[0002] Power distribution facilities are key infrastructure of the power system. They are highly susceptible to flooding in harsh environments, which can affect the stability of power supply in the power supply area. Therefore, how to reliably dispatch power distribution facilities in extreme environments has become a technical problem that needs to be studied.

[0003] Currently, existing technologies are mostly focused on weather forecasting and real-time monitoring of power distribution facility areas. Then, based on the weather forecast data and real-time monitoring data, emergency dispatching is carried out on the power distribution facility areas according to the pre-set emergency dispatching plan. However, the existing technologies, which adopt static plan execution, lack the ability to make dynamic adjustments and lack the ability to quantitatively assess the risk index in the face of extreme environments, resulting in poor reliability of emergency dispatching of power distribution facilities. Summary of the Invention

[0004] To address the aforementioned issues, this invention proposes a method and system for emergency dispatching of power distribution facilities based on risk warning. This method quantifies equipment and regional risks by predicting extreme environmental conditions in the power distribution facility area and optimizes emergency dispatching based on risk assessment, thereby improving the reliability of emergency dispatching of power distribution facilities.

[0005] To achieve the above objectives, embodiments of the present invention provide a method for emergency drainage scheduling of power distribution facilities based on risk warning, comprising: obtaining a modified rainfall forecast field based on pre-acquired rain gauge observation data, gridded forecast data, and topographic factors within the area to be urgently drained; constructing a local flood forecast model based on pre-acquired topographic features and drainage system layout within a local area of ​​the power distribution facility, and inputting the modified rainfall forecast field into the local flood forecast model to obtain water depth data, water flow velocity, and inundation duration at the location of the power distribution facility; calculating equipment risk scores based on water depth data, water flow velocity, and inundation duration; determining candidate power distribution facilities based on the equipment risk scores, and allocating corresponding emergency drainage scheduling resources to the candidate power distribution facilities through a preset optimization algorithm to generate an emergency drainage scheduling sequence; allocating the emergency drainage scheduling sequence to the power distribution facilities, and executing corresponding emergency drainage scheduling actions within the area to be urgently drained.

[0006] This invention proposes a risk-based emergency drainage scheduling method for power distribution facilities. It combines pre-acquired rain gauge observation data, gridded forecast data, and topographic factors to generate a corrected rainfall forecast field, providing reliable input for subsequent flood forecasting. Secondly, a flood forecasting model is constructed based on local topography and drainage systems. By dynamically predicting water depth, flow velocity, and inundation time at the location of power distribution facilities, the flood risk faced by the equipment is quantified. Based on this, high-risk candidate facilities are identified by calculating equipment risk scores, and an optimization algorithm is used to rationally allocate emergency drainage resources, generating an executable emergency drainage scheduling sequence. Finally, the emergency drainage scheduling sequence is distributed to the power distribution facilities for automatic or assisted emergency drainage scheduling, solving the problem of insufficient reliability in emergency drainage scheduling of power distribution facilities due to the lack of risk quantification assessment in existing methods, and improving the reliability of emergency drainage scheduling of power distribution facilities under extreme weather conditions.

[0007] Furthermore, based on the pre-acquired rain gauge observation data, gridded forecast data, and topographic factors, a revised rainfall forecast field is obtained, including: dividing the power distribution facility area into several forecast grids; acquiring rain gauge observation data of each rain gauge station and gridded forecast data of each forecast grid within the area to be urgently dispatched based on preset time intervals; correcting outliers in the rain gauge observation data and gridded forecast data to obtain standard rain gauge observation data and standard gridded forecast data; performing time registration of rain gauge stations and forecast grids based on preset time references and time scales, and calculating the initial station forecast deviation field; constructing a topographic deviation relationship using the pre-acquired topographic factors and the initial station forecast deviation field; spatially expanding the initial station forecast deviation field based on the topographic deviation relationship using a preset interpolation algorithm to obtain the target station forecast deviation field; and superimposing and correcting the standard gridded forecast data and the target station forecast deviation field based on the standard rain gauge observation data to obtain the revised rainfall forecast field.

[0008] In the above scheme, the power distribution facility area is divided into several forecast grids, and the measured data from rain gauge stations and the gridded forecast data are integrated to effectively correct the observed and forecast data. Then, outlier correction and time registration processes reduce data noise and temporal inconsistencies, improving data usability. Next, topographic factors are introduced to construct topographic deviation relationships, and the forecast deviation fields of stations are spatially interpolated and extended according to the topographic deviation relationships to obtain the forecast deviation fields of target stations. Finally, the standard gridded forecast data and the forecast deviation fields of target stations are superimposed and corrected according to the observation data of standard rain gauge stations, so that the correction results can more realistically reflect the spatial distribution characteristics of rainfall under complex terrain, resulting in a corrected rainfall forecast field. This corrected rainfall forecast field provides high-quality driving data for subsequent local flood simulation, improves the accuracy and spatial resolution of rainfall forecasts, enhances applicability in complex geographical environments, and thus improves the reliability of emergency dispatching of power distribution facilities under extreme weather conditions.

[0009] Furthermore, based on the terrain deviation relationship, the initial site forecast deviation field is spatially expanded using a preset interpolation algorithm to obtain the target site forecast deviation field. This includes: calculating the terrain relief index based on the terrain deviation relationship and determining the search range based on the terrain relief index; obtaining a list of adjacent sites and the three-dimensional distance between each adjacent site using a preset three-dimensional distance algorithm and the search range; fitting the semi-variogram function to several fitting models using a preset semi-variogram function and a preset fitting algorithm based on the list of adjacent sites and the three-dimensional distance between each adjacent site, and selecting the fitting model that meets the preset fitting requirements as the variogram function model; obtaining the anisotropy type of each site using a preset anisotropy solution algorithm based on the list of adjacent sites; constructing the interpolation matrix of each forecast grid using a preset interpolation algorithm, the variogram function model, and the anisotropy type of each site based on the initial site forecast deviation field; and solving the interpolation matrix of each forecast grid using a preset decomposition algorithm to obtain the target site forecast deviation field.

[0010] In the above scheme, the topographic relief index is calculated using the terrain deviation relationship to determine the dynamic search range. Then, a list of adjacent stations and the three-dimensional distances between each adjacent station are constructed using a preset three-dimensional distance algorithm, fully considering the influence of elevation on rainfall deviation. Next, a variogram model is constructed using a preset semi-variogram function and a fitting model to accurately characterize the spatial correlation structure of rainfall deviation. Furthermore, a preset anisotropy solving algorithm is introduced to solve for the anisotropy type of each station, identifying and processing the variation characteristics of deviation in different directions. Finally, the interpolation matrix of each forecast grid is solved by combining a preset interpolation algorithm, the variogram model, and the anisotropy type of each station to obtain the forecast deviation field for the target station. This results in a target station forecast deviation field with greater spatial resolution and physical meaning, improving the accuracy of the corrected rainfall forecast field, thereby enhancing the accuracy of subsequent flood risk prediction and contributing to improved reliability of power distribution facility emergency dispatch under extreme weather conditions.

[0011] Furthermore, based on the pre-acquired topographic features and drainage system layout within the local area of ​​the power distribution facility, a local flood forecasting model is constructed, and the corrected rainfall forecast field is input into the local flood forecasting model to obtain water depth data, water flow velocity, and inundation duration at the location of the power distribution facility. This includes: acquiring topographic features within the local area of ​​the power distribution facility based on a preset area range; acquiring the drainage system layout within the local area of ​​the power distribution facility based on preset drainage design drawings; acquiring real-time water accumulation monitoring data within the local area of ​​the power distribution facility, and selecting initial scenarios that meet preset similarity requirements based on a preset historical water accumulation scenario database and a preset similarity algorithm; constructing a local flood forecasting model based on the initial scenarios and a preset finite difference algorithm; inputting the corrected rainfall forecast field into the local flood forecasting model, and solving the water accumulation process in the local area of ​​the power distribution facility through the local flood forecasting model to obtain water depth data, water flow velocity, and inundation duration at the location of the power distribution facility.

[0012] The aforementioned scheme integrates topographic features, drainage system layout, and real-time waterlogging monitoring data, and combines this with a historical waterlogging scenario database for scene matching. This provides a solid physical foundation and scenario adaptability for the subsequent construction of local flood forecasting models. Then, a finite difference algorithm is used to construct these local flood forecasting models, effectively simulating surface runoff, waterlogging evolution, and drainage processes. Inputting the corrected rainfall forecast field into the local flood forecasting model allows for the direct output of key risk parameters such as water depth, flow velocity, and inundation duration at facility locations. This provides customized risk inputs for each power distribution facility, making equipment risk score calculations more reliable and contributing to improved reliability of emergency drainage and dispatching of power distribution facilities under extreme weather conditions.

[0013] Furthermore, based on water depth data, water flow velocity, and inundation duration, an equipment risk score is calculated, including: matching corresponding first weight values ​​to water depth data, water flow velocity, and inundation duration based on a preset first weight coefficient, and calculating a hazard index; obtaining population density and economic impact data within the power supply range of the power distribution facility, and matching corresponding second weight values ​​to population density and economic impact data based on a preset second weight coefficient, and calculating an impact degree index; obtaining historical flooding data within a local area of ​​the power distribution facility, and calculating a resistance index; obtaining the importance mapping value of each power distribution facility, and calculating an equipment importance index; and calculating an equipment risk score based on the hazard index, impact degree index, resistance index, and equipment importance index.

[0014] The aforementioned scheme considers not only the inherent hazard of the flood itself, but also the impact on the affected areas of power distribution facilities, the disaster resistance of the facilities themselves, and the equipment importance of the power distribution facilities. By assigning weights to each dimension of the indicators, the risk to equipment and the region is quantified. The resulting comprehensive score effectively distinguishes the risk levels of different facilities, avoiding the limitations of ranking based solely on a single hydrological parameter. This ensures that emergency drainage scheduling prioritizes the most critical facilities with the highest risk and greatest impact, thus improving the reliability of emergency drainage scheduling for power distribution facilities under extreme weather conditions.

[0015] Furthermore, candidate power distribution facilities are determined based on equipment risk scores, and corresponding emergency drainage scheduling resources are allocated to these facilities using a preset optimization algorithm to generate an emergency drainage scheduling sequence. This process includes: acquiring the real-time status of power distribution facilities, the status of fixed drainage equipment, and the status of mobile emergency drainage resources; calculating the urgency score of power distribution facilities based on their real-time status and obtaining several candidate power distribution facilities based on these scores; allocating corresponding emergency drainage scheduling resources to these candidate facilities based on the status of fixed drainage equipment and mobile emergency drainage resources to construct an initial emergency drainage scheduling sequence; and performing feasibility optimization on the initial emergency drainage scheduling sequence based on the preset optimization algorithm and equipment risk scores to obtain an emergency drainage scheduling sequence that meets the preset emergency drainage scheduling requirements.

[0016] In the above scheme, the real-time status of facilities is acquired, and the urgency score of each facility is dynamically calculated to determine candidate facilities. Based on this, resource constraints such as the availability of fixed drainage equipment and mobile emergency drainage resources are considered to allocate appropriate emergency drainage scheduling resources to the candidate facilities, constructing an initial scheduling sequence. Finally, based on the equipment risk score, a preset optimization algorithm is used to perform feasibility optimization on the initial emergency drainage scheduling sequence to obtain an emergency drainage scheduling sequence that meets the preset emergency drainage scheduling requirements. This ensures that scheduling decisions are both risk-oriented and closely integrated with real-time situation and resource conditions, avoiding a disconnect between decision-making and execution. It also transforms the scheduling scheme from a static contingency plan into a dynamic response, improving the ability to formulate feasible and efficient emergency drainage plans in extreme environments and resource-limited scenarios, and enhancing the reliability of emergency drainage scheduling for power distribution facilities under extreme weather conditions.

[0017] Furthermore, based on a preset optimization algorithm and equipment risk scoring, the initial emergency scheduling sequence is optimized for feasibility to obtain an emergency scheduling sequence that meets the preset emergency scheduling requirements. This includes: performing a feasibility check on the initial emergency scheduling sequence to obtain the feasibility check result; if the feasibility check result does not meet the preset feasibility check requirements, then constraining the initial emergency scheduling sequence to obtain a feasible emergency scheduling sequence; and based on the preset emergency scheduling objective, performing a lightweight optimization on the feasible emergency scheduling sequence using a preset optimization algorithm and equipment risk scoring to obtain an emergency scheduling sequence that meets the preset emergency scheduling requirements.

[0018] The above scheme incorporates a feasibility check step, which automatically identifies potential resource conflicts, time conflicts, or logical contradictions in the initial sequence. For initial emergency dispatch sequences that do not meet feasibility requirements, automatic constraint repair is performed to generate feasible emergency dispatch sequences. Based on this, lightweight optimization is performed using a preset optimization algorithm and equipment risk scoring, based on preset emergency dispatch objectives, to find a better solution in the feasible solution space, resulting in an emergency dispatch sequence that meets the preset emergency dispatch requirements. This ensures that the final output emergency dispatch sequence is not only executable but also of high quality and efficiency, improving the reliability of emergency dispatch of power distribution facilities under extreme weather conditions.

[0019] This invention also provides a risk-based power distribution facility emergency drainage scheduling system, comprising: a rainfall forecast field correction module, used to obtain a corrected rainfall forecast field based on pre-acquired rain gauge observation data, gridded forecast data, and topographic factors within the area to be urgently drained; a flood data acquisition module, used to construct a local flood forecast model based on pre-acquired topographic features and drainage system layout within a local area of ​​the power distribution facility, and input the corrected rainfall forecast field into the local flood forecast model to obtain water depth data, water flow velocity, and inundation duration at the location of the power distribution facility; a risk score calculation module, used to calculate the equipment risk score based on the water depth data, water flow velocity, and inundation duration; an emergency drainage scheduling sequence generation module, used to determine candidate power distribution facilities based on the equipment risk score, and allocate corresponding emergency drainage scheduling resources to the candidate power distribution facilities through a preset optimization algorithm to generate an emergency drainage scheduling sequence; and an emergency drainage scheduling execution module, used to allocate the emergency drainage scheduling sequence to the power distribution facilities and execute corresponding emergency drainage scheduling actions within the area to be urgently drained.

[0020] This invention proposes a risk-based emergency drainage scheduling system for power distribution facilities. It combines pre-acquired rain gauge observation data, gridded forecast data, and topographic factors to generate a corrected rainfall forecast field, providing reliable input for subsequent flood forecasting. Secondly, a flood forecasting model is constructed based on local topography and drainage systems. By dynamically predicting water depth, flow velocity, and inundation time at the location of power distribution facilities, the system quantifies the flood risk faced by the equipment. Based on this, high-risk candidate facilities are identified by calculating equipment risk scores, and an optimization algorithm is used to rationally allocate emergency drainage resources, generating an executable emergency drainage scheduling sequence. Finally, the emergency drainage scheduling sequence is distributed to power distribution facilities for automatic or assisted emergency drainage scheduling, addressing the reliability issues of existing methods that lack risk quantification assessment, thus improving the reliability of emergency drainage scheduling for power distribution facilities under extreme weather conditions.

[0021] Furthermore, the rainfall forecast field correction module is used to obtain a corrected rainfall forecast field based on the pre-acquired rain gauge observation data, gridded forecast data, and topographic factors within the area to be expedited for emergency drainage and dispatch. This includes: an initial data acquisition unit, used to divide the power distribution facility area into several forecast grids, and acquire rain gauge observation data from each rain gauge station and gridded forecast data from each forecast grid within the area to be expedited for emergency drainage and dispatch based on preset time intervals; an outlier repair unit, used to correct outliers in the rain gauge observation data and gridded forecast data to obtain standard rain gauge observation data and standard gridded forecast data; and an initial deviation field construction unit. The system is divided into four parts: a time registration unit for rainfall stations and forecast grids based on a preset time reference and time scale, and a topographic deviation relationship acquisition unit for constructing topographic deviation relationships using pre-acquired topographic factors and the initial station forecast deviation field; a target deviation field construction unit for spatially expanding the station forecast deviation field based on the topographic deviation relationship using a preset interpolation algorithm to obtain the target station forecast deviation field; and an overlay correction unit for overlaying and correcting the standard gridded forecast data and the target station forecast deviation field based on standard rainfall station observation data to obtain the corrected rainfall forecast field.

[0022] Furthermore, the flood data acquisition module is used to construct a local flood forecast model based on pre-acquired topographic features and drainage system layout within a local area of ​​the power distribution facility, and input the corrected rainfall forecast field into the local flood forecast model to obtain water depth data, water flow velocity, and inundation duration at the location of the power distribution facility. This includes: a topographic feature acquisition unit, used to acquire topographic features within a local area of ​​the power distribution facility based on a preset area range; a drainage system layout acquisition unit, used to acquire the drainage system layout within a local area of ​​the power distribution facility based on preset drainage design drawings; an initial scene matching unit, used to acquire real-time water accumulation monitoring data within a local area of ​​the power distribution facility, and to select initial scenes that meet preset similarity requirements based on a preset historical water accumulation scenario database and a preset similarity algorithm; a local flood forecast model construction unit, used to construct a local flood forecast model based on the initial scenes and a preset finite difference algorithm; and a flood data solving unit, used to input the corrected rainfall forecast field into the local flood forecast model, and to solve the water accumulation process in the local area of ​​the power distribution facility through the local flood forecast model to obtain water depth data, water flow velocity, and inundation duration at the location of the power distribution facility. Attached Figure Description

[0023] Figure 1 A flowchart illustrating the steps of a risk warning-based emergency dispatching method for power distribution facilities, provided in a certain embodiment of the present invention; Figure 2 This is a schematic diagram of the module structure of a power distribution facility emergency dispatching system based on risk warning, provided in one embodiment of the present invention. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] Example 1 See Figure 1 , Figure 1 This is a flowchart illustrating the steps of a risk-based emergency dispatching method for power distribution facilities, provided in one embodiment of the present invention. Figure 1 As shown in the figure, this embodiment of the invention proposes a method for emergency dispatching of power distribution facilities based on risk warning, including steps 101 to 105, each step of which is as follows: Step 101: Based on the pre-acquired rain gauge observation data, gridded forecast data and topographic factors in the area to be urgently dispatched, a corrected rainfall forecast field is obtained; Step 102: Based on the pre-acquired topographic features and drainage system layout in the local area of ​​the power distribution facility, construct a local flood forecast model and input the corrected rainfall forecast field into the local flood forecast model to obtain water depth data, water flow velocity and inundation duration at the location of the power distribution facility; Step 103: Calculate the equipment risk score based on water depth data, water flow velocity, and inundation duration; Step 104: Based on the equipment risk score, candidate power distribution facilities are determined, and corresponding emergency scheduling resources are allocated to the candidate power distribution facilities through a preset optimization algorithm to generate an emergency scheduling sequence. Step 105: Assign the emergency dispatch sequence to the power distribution facilities and execute the corresponding emergency dispatch actions for the areas to be dispatched.

[0026] One specific implementation method involves acquiring observational data from various rain gauge stations and gridded forecast data within the area to be expedited for flood control. The gridded forecast data is then corrected using topographic factors, typically employing topographic-based statistical interpolation fusion techniques to obtain a corrected rainfall forecast field. This addresses the uncertainty in rainfall forecast input by providing a topographic-oriented statistical interpolation fusion process to create a continuously updated corrected rainfall forecast field, which serves as the driving input for subsequent localized flood forecasting models for power distribution facilities. Next, the topographic features and drainage system layout within the local area of ​​the power distribution facility are acquired. A model is then constructed to model the water accumulation process in a small-scale area surrounding the power distribution facility, comprehensively considering local topographic features, station drainage capacity, soil permeability, and rainwater runoff processes. This localized flood forecasting model is then deployed in the cloud for GPU parallel computation, outputting water depth data, flow velocity, and inundation duration at the location of the power distribution facility. The water depth, flow velocity, and inundation duration at the location of the power distribution facilities are used as indicators of the flood's inherent hazard. The impact on the affected area, the facility's own disaster resistance, and the equipment's importance are also incorporated to calculate an equipment risk score. In this embodiment, the RI (Responsive Analysis) method is used. After obtaining the equipment risk score, candidate power distribution facilities are identified based on the availability of resources. A pre-defined optimization algorithm is then used to allocate corresponding emergency drainage resources to these candidate facilities, resulting in an emergency drainage sequence. In this embodiment, the pre-defined optimization algorithm can be implemented using a heuristic optimization algorithm. Finally, the emergency drainage sequence is distributed to each power distribution facility, and corresponding emergency drainage actions are executed according to the sequence to complete the emergency drainage of the affected area.

[0027] In this embodiment, after executing the emergency drainage scheduling action, an execution receipt can be generated, and the scheduling measures can be updated on a rolling basis based on the new round of forecast results. The issuance of the emergency drainage action sequence and the rolling update scheduling are explained using a three-layer distributed scheduling architecture of cloud-edge-terminal. The cloud scheduling center, as the decision-making core, is responsible for global optimization calculations and scheduling instruction generation, possessing powerful computing and data processing capabilities. Edge computing nodes are deployed locally at the power distribution facilities, possessing a certain degree of autonomous decision-making capability and maintaining basic emergency scheduling functions during communication interruptions. Terminal devices include drainage pump station controllers, mobile emergency drainage equipment, and mobile terminals for on-site personnel, responsible for the final execution of instructions and status feedback. This multi-layered fault-tolerance mechanism ensures the system can still operate normally under various abnormal conditions. Specifically, actions within the recent locked window are partially frozen, generating the instruction sequence and expected effects for the next planning period, which are then issued to the pump stations and mobile teams, and execution receipts are recorded. The instruction issuance uses JSON format, containing key elements such as task ID, power distribution facility ID, start time, duration, pump group list, and mobile pump truck information, ensuring the integrity and executability of the instructions. The execution receipt includes information such as instruction reception confirmation time, actual equipment start-up time, operating parameters, and abnormal situation reports. Upon receiving new observation data, forecast data, or status changes, the optimization process of the emergency scheduling sequence in step 104 is repeated according to the rolling clock, inheriting actions that were not completed but were feasible in the previous period; and exponential smoothing is used to update risk parameters. One interpretation is that the rolling update cycle is dynamically set according to the situation, updating once every 15 minutes under normal circumstances, once every 5 minutes in emergency situations, and continuously updating in extreme cases. The risk parameter update uses an exponential smoothing formula, expressed as: In the formula, This represents the weight value at the previous moment, i.e., the historical weight. This represents the weight value calculated based on actual observation data at the current moment, i.e., the observation weight. The smoothing coefficient λ=0.8 can be adaptively adjusted according to the actual situation.

[0028] It is worth mentioning that, in response to equipment / pump truck failures, road blockages, or extreme scenarios, backup equipment is activated or nearby pump trucks are reassigned. If necessary, a conservative strategy driven by the worst or second-worst scenario is adopted, and the replanning cycle is shortened. This embodiment of the invention also proposes an emergency response mechanism covering three aspects: fault handling, extreme scenario response, and conservative strategy adjustment. Fault handling includes automatically assigning the nearest backup pump truck when a pump truck fails, recalculating the optimal path when a road is blocked, and using backup communication methods such as satellite phones when communication is interrupted. Extreme scenario response includes activating all backup resources in the event of a once-in-a-century flood, deploying 24-hour advance warnings for typhoons and heavy rains, and implementing a power grid black start plan in the event of cascading equipment failures. Conservative strategy adjustment includes increasing the safety margin from 10% to 20%, shortening the replanning cycle from 15 minutes to 5 minutes, and increasing the reserve ratio of backup resources from 20% to 40%.

[0029] This invention proposes a risk-based emergency drainage scheduling method for power distribution facilities. It combines pre-acquired rain gauge observation data, gridded forecast data, and topographic factors to generate a corrected rainfall forecast field, providing reliable input for subsequent flood forecasting. Secondly, a flood forecasting model is constructed based on local topography and drainage systems. By dynamically predicting water depth, flow velocity, and inundation time at the location of power distribution facilities, the flood risk faced by the equipment is quantified. Based on this, high-risk candidate facilities are identified by calculating equipment risk scores, and an optimization algorithm is used to rationally allocate emergency drainage resources, generating an executable emergency drainage scheduling sequence. Finally, the emergency drainage scheduling sequence is distributed to the power distribution facilities for automatic or assisted emergency drainage scheduling, solving the problem of insufficient reliability in emergency drainage scheduling of power distribution facilities due to the lack of risk quantification assessment in existing methods, and improving the reliability of emergency drainage scheduling of power distribution facilities under extreme weather conditions.

[0030] A preferred scheme for obtaining a corrected rainfall forecast field based on pre-acquired rain gauge observation data, gridded forecast data, and topographic factors includes: dividing the power distribution facility area into several forecast grids; acquiring rain gauge observation data of each rain gauge station and gridded forecast data of each forecast grid within the area to be dispatched for emergency drainage based on a preset time interval; correcting outliers in the rain gauge observation data and gridded forecast data to obtain standard rain gauge observation data and standard gridded forecast data; performing time registration of the rain gauge stations and forecast grids based on a preset time reference and time period scale, and calculating the initial station forecast deviation field; constructing a topographic deviation relationship using pre-acquired topographic factors and the initial station forecast deviation field; spatially expanding the initial station forecast deviation field based on the topographic deviation relationship using a preset interpolation algorithm to obtain the target station forecast deviation field; and superimposing and correcting the standard gridded forecast data and the target station forecast deviation field based on the standard rain gauge observation data to obtain the corrected rainfall forecast field.

[0031] One preferred implementation involves dividing the power distribution facility area into several forecast grids. Within each grid, rain gauge observation data from each rain gauge station and gridded forecast data from each forecast grid are acquired according to a preset time interval. The size of each grid is generally set to 1 km × 1 km. The preset time interval can be hourly, 15-minute, or 5-minute intervals.

[0032] For rain gauge observation data, a classification interpolation method was used to process outlier data. First, negative data (less than 0 mm / h), data exceeding twice the maximum historical rainfall range, and missing data due to equipment malfunction were identified as outliers. Then, a classification processing strategy was applied to the selected outliers. Specifically, short-term missing values ​​(less than or equal to 30 minutes) were filled using cubic spline interpolation; medium-term missing values ​​(30 minutes to 2 hours) were filled using linear interpolation combined with historical data correction; and long-term missing values ​​(greater than 2 hours) were estimated using distance-weighted regression with neighboring stations, with the regression model considering altitude differences and terrain similarity. Finally, outlier correction employed median filtering to remove random noise and maintain the temporal continuity of the rainfall process. The processed data were labeled with quality levels to provide a reliable observational benchmark for subsequent bias correction.

[0033] For the gridded forecast data of each forecast grid, a hierarchical quality control and adaptive correction method is used to process outlier data. First, a specific quality control process is developed for the characteristics of numerical weather prediction products. The quality control process checks include numerical range checks, gradient rationality checks, and temporal continuity checks. For example, if the numerical range check requires rainfall to be within the range of [0, 300] mm / h, then values ​​exceeding the upper limit are replaced with the 98th quantile, and negative values ​​are directly set to 0. If the gradient rationality check requires that the rainfall intensity gradient between adjacent grids is less than 100 mm / h / km, then anomalous gradients are spatially smoothed using bidirectional Laplace filtering. For the temporal continuity check, the local mean absolute deviation (MAD) of the historical time period is first calculated. When the change between adjacent times exceeds 5 times the MAD, cubic spline interpolation combined with Kalman filtering is used for time series correction.

[0034] After outlier processing of rain gauge observation data and gridded forecast data from each forecast grid, the rain gauge stations and forecast grids are time-registered according to a preset time base and time scale, and the initial station forecast bias field is calculated. For example, using the start time and lead of the ensemble forecast as a benchmark, the rain gauge station sequence is aligned to a unified time forecast grid. Generally, observation data is extended from 5-minute data to minute-by-minute levels using cubic spline interpolation to ensure continuity. Nearest neighbor values ​​are used for boundary processing at the start and end times of the data. The time period alignment strategy uses sliding window matching, with the forecast product's time as the benchmark. Then, various bias calculation methods are employed to calculate the initial station forecast bias field. In this embodiment, the bias calculation uses three methods: absolute bias, relative bias, and standardized bias. Before calculating the initial site forecast bias field, the rainfall stations are first matched to the nearest forecast grid using spatial matching methods, typically the nearest neighbor method. For rainfall stations not located in the center of the forecast grid, bilinear interpolation is performed using the values ​​from the surrounding four forecast grids. The elevation difference between the rainfall station and the average elevation of the forecast grid is considered, and elevation correction for rainfall is applied using the orographic rainfall effect coefficient. This correction coefficient is determined based on regional climate characteristics and topographic conditions. It should be noted that the absolute bias calculation formula is as follows: The formula for calculating relative deviation is as follows: The formula for calculating standardized deviation is: In the formula; These are station observations. For grid forecast values, This represents the climatological standard deviation.

[0035] To improve applicability in complex geographical environments, topographic factors such as altitude, elevation undulation, and aspect are introduced as auxiliary variables to establish topographic deviation relationships. In this embodiment, the extraction of topographic factors includes: altitude H is extracted from 10m resolution DEM data; elevation undulation R is obtained by calculating the elevation standard deviation within a 3km × 3km window; slope S is calculated using the Horn algorithm; and aspect A is represented using 8-azimuth encoding. The extraction is then performed using the formula... The topographic wetness index (TWI) and the degree of topographic obstruction in the prevailing wind direction are used to obtain the wind direction exposure (WE), where α is the catchment area and S is the slope. Then, the topographic deviation relationship is modeled using multiple linear regression, expressed as follows: In the formula, is the regression coefficient, where i = 0, 1, 2, 3, 4, 5, 6, 7; ε is the residual term; it is worth mentioning that the model optimization strategy includes using stepwise regression to screen significant topographic factors for variable selection, performing piecewise linear modeling for elevation less than 500m, 500 - 1500m, and greater than 1500m to handle non - linear relationships, considering the interaction between elevation and slope direction, especially the differences between the windward slope and the leeward slope, and establishing regression models for each of the four seasons of spring, summer, autumn, and winter for seasonal adjustment.

[0036] Next, according to the above - mentioned topographic deviation relationship, the initial station forecast deviation field is spatially extended through a preset interpolation algorithm to obtain the target station forecast deviation field; in this embodiment, the preset interpolation algorithm can adopt the local Kriging interpolation algorithm. Specifically, the local Kriging interpolation algorithm realizes the correction of rainfall forecast deviation in complex terrain areas through intelligent neighborhood construction and terrain - adaptive weight calculation, robust variogram modeling and calculation optimization, and constructing and solving the local Kriging equations. After obtaining the target station forecast deviation field, based on the observed data of standard rain gauges, the standard gridded forecast data and the target station forecast deviation field are superimposed and corrected to obtain the corrected rainfall forecast field. Specifically, the rainfall intensity is obtained based on the observed data of standard rain gauges and the weighting coefficient is set. The light rain is corrected weakly, the heavy rain is corrected strongly, and the standard gridded forecast data and the target station forecast deviation field are superimposed to obtain the corrected rainfall forecast field. During the superimposed correction process, a 3 - hour sliding window is used for time smoothing to avoid sudden changes in the correction coefficient, and then a 5×5 Gaussian filter is used for the weight field to ensure spatial continuity. When correcting the regional boundary, distance - weighted fusion is also used to avoid boundary effects; after completing the superimposed correction, the corrected rainfall forecast field is also subjected to correction quality assessment and consistency check. Among them, the correction quality assessment includes uncertainty propagation calculation, and the consistency check includes checking the consistency of the spatial gradient and time change rate of the corrected rainfall field. The final output results include the corrected rainfall forecast field, the uncertainty field providing the corrected forecast confidence interval, and the quality marks for each forecast grid. The quality marks include multiple quality levels such as excellent, good, general, and poor.

[0037] In the embodiment of the present invention, a deviation correction strategy with intensity adaptability is designed. The intensity - graded correction sets a non - linear weight function according to the rainfall intensity P: when P ≤ 1 mm / h, the weight is set to 0.7 for weak correction of light rain; when 1 < P ≤ 5 mm / h, the weight is set to 0.8 for weak correction of small rain; when 5 < P ≤ 15 mm / h, the weight is set to 1.0 to keep the moderate rain unchanged; when 15 < P ≤ 30 mm / h, the weight is set to 1.2 for strong correction of heavy rain; when P > 30 mm / h, the weight is set to 1.3 for enhanced correction of heavy rainstorm.

[0038] In the above scheme, the power distribution facility area is divided into several forecast grids, and the measured data from rain gauge stations and the gridded forecast data are integrated to effectively correct the observed and forecast data. Then, outlier correction and time registration processes reduce data noise and temporal inconsistencies, improving data usability. Next, topographic factors are introduced to construct topographic deviation relationships, and the forecast deviation fields of stations are spatially interpolated and extended according to the topographic deviation relationships to obtain the forecast deviation fields of target stations. Finally, the standard gridded forecast data and the forecast deviation fields of target stations are superimposed and corrected according to the observation data of standard rain gauge stations, so that the correction results can more realistically reflect the spatial distribution characteristics of rainfall under complex terrain, resulting in a corrected rainfall forecast field. This corrected rainfall forecast field provides high-quality driving data for subsequent local flood simulation, improves the accuracy and spatial resolution of rainfall forecasts, enhances applicability in complex geographical environments, and thus improves the reliability of emergency dispatching of power distribution facilities under extreme weather conditions.

[0039] A preferred scheme involves spatially expanding the initial site forecast deviation field using a preset interpolation algorithm based on terrain deviation relationships to obtain the target site forecast deviation field. This includes: calculating a terrain relief index based on the terrain deviation relationships and determining the search range based on the terrain relief index; obtaining a list of adjacent sites and the 3D distances between each adjacent site using a preset 3D distance algorithm and the search range; fitting a semi-variogram function to several fitting models using a preset semi-variogram function and a preset fitting algorithm based on the list of adjacent sites and the 3D distances between each adjacent site, and selecting the fitting model that meets the preset fitting requirements as the variogram function model; obtaining the anisotropy type of each site using a preset anisotropy solution algorithm based on the list of adjacent sites; constructing the interpolation matrix of each forecast grid using a preset interpolation algorithm, the variogram function model, and the anisotropy type of each site based on the initial site forecast deviation field; and solving the interpolation matrix of each forecast grid using a preset decomposition algorithm to obtain the target site forecast deviation field.

[0040] One preferred implementation involves spatially expanding the initial station forecast deviation field using a preset interpolation algorithm based on the aforementioned terrain deviation relationship to obtain the target station forecast deviation field. In this embodiment, the preset interpolation algorithm can be a local kriging interpolation algorithm. Specifically, the local kriging interpolation algorithm corrects rainfall forecast deviations in complex terrain areas through intelligent neighborhood construction and terrain-adaptive weight calculation, robust variogram modeling and computational optimization, and the construction and solution of local kriging equations. Specifically, firstly, using 10m resolution DEM data, the elevation standard deviation is calculated using a 3km × 3km moving window to obtain the terrain relief index TRI, expressed as: In the formula, Let i be the elevation value of the i-th cell within the window. Let be the average elevation within the window, and n be the number of effective pixels. Then, a hierarchical strategy is used to set differentiated radii for plains (TRI < 50m, R = 5km), hills (TRI 50-200m, R = 10km), and mountains (TRI > 200m, R = 20km) to determine the search range. Specifically, when TRI < 50m and station density > 0.1 / km², the neighborhood radius R = 5km is set, representing plain terrain; when TRI is in the 50-200m range or station density is 0.05-0.1 / km², the neighborhood radius R = 10km is set, representing hilly terrain; when TRI > 200m or station density < 0.05 / km², the neighborhood radius R = 10km is set, representing hilly terrain. 20km represents mountainous terrain; generally, the number of observation stations is limited to 8-20 to ensure statistical stability. Then, a preset three-dimensional distance algorithm and search range are used to obtain a list of adjacent stations and the three-dimensional distance between each adjacent station. In this embodiment, the preset three-dimensional distance algorithm can use a three-dimensional anisotropic distance metric formula, expressed as: Introducing the vertical-horizontal correlation scale ratio Quantifying the impact of terrain can also be combined with a leave-one-out cross-validation parameter optimization strategy for optimal weight allocation under complex terrain conditions. In this embodiment, the optimal weight allocation is determined. The optimal range for this value is 5-20km. The optimal value range is 200-1000m. The value typically varies between 0.25 and 100, thus completing the intelligent neighborhood construction and terrain-adaptive weight calculation.

[0041] Then, the semivariogram is fitted to several fitting models using a preset semivariogram function and a preset fitting algorithm. The fitting model that meets the preset fitting requirements is selected as the variogram function model. Specifically, a competitive fitting mechanism for three types of theoretical models—spherical, exponential, and Gaussian—is established using a robust experimental semivariogram function designed for the non-normal distribution characteristics of rainfall data. The optimal model is automatically identified using a weighted least squares method combined with a distance-sample pair dual-weighting strategy. In this embodiment, the preset semivariogram function can be the experimental semivariogram function, expressed as: In the formula, The number of station pairs within a distance range h. Let be the value of the i-th observation point. Let h be the observation value at a distance h from the i-th point. Then, three types of theoretical models are established: spherical, exponential, and Gaussian. The spherical model is expressed as: The exponential model is represented as: The Gaussian model is represented as: In the formula, For the nugget effect; Here, 'a' represents the partial sill value, and 'a' represents the range, indicating the effective distance for spatial correlation. Beyond this distance, the spatial correlation between observations disappears. The range parameter 'a' is determined as follows: First, empirical semivariogram values ​​for different distance intervals are calculated based on observation station data. The distance corresponding to 95% of the sill value is used as the initial value of 'a'. In this embodiment, the typical range is 5-50 km. Then, a weighted least squares method combined with a distance-sample pair dual-weighting method is used to determine the range parameter 'a'. , The three parameters, 'a', 'a', and 'a', are jointly optimized and fitted to ensure that the range parameter 'a' satisfies the physical constraint 'a>0' while minimizing the weighted sum of squared residuals between the theoretical model and the empirical semivariogram. Then, an overweighted least squares method combined with a distance-sample pair size dual-weighting strategy is used to automatically identify the optimal model. During model selection, leave-one-out cross-validation is used to calculate the standardized root mean square error (RMSE). The fitted model that meets the preset fitting requirements is selected as the variogram model, where the preset fitting requirement is to minimize the standardized RMSE. The formula for calculating the standardized RMSE is expressed as: In the formula, To leave one predicted value, These are actual observations. denoted as Kriging standard deviation.

[0042] Then, based on the list of adjacent stations, the anisotropy type of each station is obtained through a preset anisotropy solution algorithm. Specifically, the preset anisotropy solution algorithm adopts an elliptical anisotropy ratio test. By calculating the ratio of the principal axis range to the secondary axis range, when the ratio is >1.5, it is considered that there is significant geometric anisotropy and the principal axis azimuth angle θ needs to be estimated to describe the main direction of anisotropy. In addition, through a four-dimensional variogram parameter pre-calculation cache system, an intelligent lookup table with 108 combinations is established according to terrain type × rainfall level × model type × anisotropy type to improve computational efficiency. Among them, terrain type includes plains, hills and mountains, rainfall level is divided into 0-1, 1-5, 5-10 and 10+ mm / h, model type includes spherical, exponential and Gaussian, and anisotropy type includes isotropic, geometric anisotropy and zonal anisotropy.

[0043] Finally, based on the initial site forecast bias field, interpolation matrices for each forecast grid are constructed using a pre-defined interpolation algorithm, a variogram model, and the anisotropy type of each site. The interpolation matrices for each forecast grid are then solved using a pre-defined decomposition algorithm to obtain the target site forecast bias field. Specifically, the pre-defined interpolation algorithm employs local kriging interpolation, constructing a standard kriging equation matrix for each forecast grid. The pre-defined decomposition algorithm uses the Cholesky decomposition method, solving the equation system. For unobserved and boundary areas more than 30 km from the nearest rainfall station, a neighborhood radius gradually increasing strategy is adopted, expanding from the initial radius R to 1.5R, 2R, and 3R until at least four sufficient rainfall observation stations are found. When the kriging variance exceeds the threshold of 2.0 (mm / h), the bias field is determined. 2 When marked as a high uncertainty region, uncertainty clipping is performed. The final output includes three sets of raster products: a high-precision deviation raster with the same resolution as the original forecast grid, an uncertainty raster expressed in the form of standard deviation, and a quality-labeled raster divided into four levels: excellent, good, average, and poor, based on Kriging variance and cross-validation error.

[0044] In the above scheme, the topographic relief index is calculated using the terrain deviation relationship to determine the dynamic search range. Then, a list of adjacent stations and the three-dimensional distances between each adjacent station are constructed using a preset three-dimensional distance algorithm, fully considering the influence of elevation on rainfall deviation. Next, a variogram model is constructed using a preset semi-variogram function and a fitting model to accurately characterize the spatial correlation structure of rainfall deviation. Furthermore, a preset anisotropy solving algorithm is introduced to solve for the anisotropy type of each station, identifying and processing the variation characteristics of deviation in different directions. Finally, the interpolation matrix of each forecast grid is solved by combining a preset interpolation algorithm, the variogram model, and the anisotropy type of each station to obtain the forecast deviation field for the target station. This results in a target station forecast deviation field with greater spatial resolution and physical meaning, improving the accuracy of the corrected rainfall forecast field, thereby enhancing the accuracy of subsequent flood risk prediction and contributing to improved reliability of power distribution facility emergency dispatch under extreme weather conditions.

[0045] A preferred approach involves constructing a local flood forecasting model based on pre-acquired topographic features and drainage system layout within a local area of ​​a power distribution facility. A modified rainfall forecast field is then input into the local flood forecasting model to obtain water depth data, water flow velocity, and inundation duration at the location of the power distribution facility. This includes: acquiring topographic features within a pre-defined area; acquiring the drainage system layout within the local area of ​​the power distribution facility based on pre-defined drainage design drawings; acquiring real-time water accumulation monitoring data within the local area of ​​the power distribution facility; selecting initial scenarios that meet pre-defined similarity requirements based on a pre-defined historical water accumulation scenario database and a pre-defined similarity algorithm; constructing a local flood forecasting model based on the initial scenarios and a pre-defined finite difference algorithm; inputting the modified rainfall forecast field into the local flood forecasting model; and solving the water accumulation process within the local area of ​​the power distribution facility using the local flood forecasting model to obtain water depth data, water flow velocity, and inundation duration at the location of the power distribution facility.

[0046] One preferred implementation method involves acquiring high-precision DEM (Digital Elevation Model), surface roughness, soil permeability coefficient, and groundwater level, among other topographic features, within a preset area, typically within 1 km of the power distribution facility. Specifically, the surface roughness coefficient is obtained through field surveys or UAV aerial surveys, combined with five layers of soil permeability data from geological data, and the initial groundwater level is obtained. Then, the drainage system layout, including drainage ditch layout, pipe diameter parameters, drainage outlet location, drainage capacity, and station area ground plan layout, is acquired. Specifically, data such as the bottom elevation, cross-sectional dimensions, pipe routing, drainage outlet coordinates, and drainage capacity curves of the complete design drawings of the power distribution facility drainage system are obtained, along with the distribution of hardened ground and surface cover in the station area. Then, real-time monitoring data of water accumulation in local areas of power distribution facilities is acquired. The scenario most similar to the current rainfall intensity and meteorological conditions is retrieved from the historical water accumulation scenario database. Based on similarity analysis, the optimal initial field is selected as the initial condition for this calculation. Generally, based on historical data and experience, a scenario database containing multi-dimensional features such as rainfall intensity, duration, previous soil moisture content and wind speed is established. Euclidean distance is used to calculate similarity, and the historical scenario with the highest similarity is selected as the initial scenario. The preset similarity requirement is represented as the highest similarity. Then, a local flood forecasting model is constructed using a pre-defined finite difference algorithm. In this embodiment, the pre-defined finite difference algorithm can be implemented using two-dimensional shallow water equations. The grid resolution is set to 0.2m × 0.2m in areas with dense power distribution facilities and equipment, and 0.5m × 0.5m in the outer areas. The time step is adaptively adjusted using CFL conditions. The solution domain is set to a range of 500m around the power distribution facilities. The boundary conditions are open or permeable boundaries. The inner boundary considers the water-blocking effect of buildings and equipment and adopts impermeable boundary conditions, while the outer boundary is set as an open boundary, permeable boundary, or periodic boundary according to the terrain features. By comprehensively considering rainfall input, surface evaporation, soil infiltration, and drainage system discharge, a complete water balance equation is established to obtain the local flood forecasting model. Then, the modified rainfall forecast field is input into the local flood forecasting model, and the local flood forecasting model is used to solve the local water accumulation process of the power distribution facilities, obtaining the water depth data, water flow velocity, and inundation duration at the location of the power distribution facilities.

[0047] After the solution is completed, the water depth data, water flow velocity and inundation duration of the power distribution facility location are obtained based on the spatial correspondence and model output results.

[0048] The aforementioned scheme integrates topographic features, drainage system layout, and real-time waterlogging monitoring data, and combines this with a historical waterlogging scenario database for scene matching. This provides a solid physical foundation and scenario adaptability for the subsequent construction of local flood forecasting models. Then, a finite difference algorithm is used to construct these local flood forecasting models, effectively simulating surface runoff, waterlogging evolution, and drainage processes. Inputting the corrected rainfall forecast field into the local flood forecasting model allows for the direct output of key risk parameters such as water depth, flow velocity, and inundation duration at facility locations. This provides customized risk inputs for each power distribution facility, making equipment risk score calculations more reliable and contributing to improved reliability of emergency drainage and dispatching of power distribution facilities under extreme weather conditions.

[0049] A preferred embodiment calculates equipment risk scores based on water depth data, water flow velocity, and inundation duration, including: calculating a hazard index by matching corresponding first weight values ​​to water depth data, water flow velocity, and inundation duration based on a preset first weight coefficient; obtaining population density and economic impact data within the power supply range of the power distribution facility, and matching corresponding second weight values ​​to the population density and economic impact data based on a preset second weight coefficient; obtaining historical flooding data within a local area of ​​the power distribution facility and calculating a resistance index; obtaining the importance mapping value of each power distribution facility and calculating an equipment importance index; and calculating an equipment risk score based on the hazard index, impact index, resistance index, and equipment importance index.

[0050] One preferred implementation method is for water depth data. Water flow velocity and flood duration Match the corresponding first weight value, the preset first weight coefficient is represented as , and Calculate the hazard index H w , is represented as: Population density within the power supply range of the power distribution facilities and economic impact data Match the corresponding second weight value, the preset second weight coefficient is represented as and The influence level index E is calculated using the following expression: Furthermore, historical flooding data for localized areas of power distribution facilities were analyzed. The resistance index is calculated and expressed as: Then, the RI calculation method calculates the equipment risk score. It also applies a monotonically non-decreasing constraint to ensure that the RI does not decrease when any component increases. The equipment risk score calculation expression is as follows: In the formula, For monotonically non-decreasing constraints, it is expressed as: ; I represents the importance of the device, with a value range of (0, 1]; (i=0, 1, 2, 3, 4) are weighting coefficients, determined through historical event backtracking and expert weighting, and can be calibrated on a rolling basis by region / site. The default is to use the average value. In this embodiment, RI is divided into four levels according to the threshold: less than or equal to 0.25 is low, (0.25–0.50] is medium, (0.50–0.75] is high and greater than 0.75 is very high, which are used to drive the subsequent scheduling priority.

[0051] The aforementioned scheme considers not only the inherent hazard of the flood itself, but also the impact on the affected areas of power distribution facilities, the disaster resistance of the facilities themselves, and the equipment importance of the power distribution facilities. By assigning weights to each dimension of the indicators, the risk to equipment and the region is quantified. The resulting comprehensive score effectively distinguishes the risk levels of different facilities, avoiding the limitations of ranking based solely on a single hydrological parameter. This ensures that emergency drainage scheduling prioritizes the most critical facilities with the highest risk and greatest impact, thus improving the reliability of emergency drainage scheduling for power distribution facilities under extreme weather conditions.

[0052] A preferred approach involves determining candidate power distribution facilities based on equipment risk scores, and allocating corresponding emergency drainage scheduling resources to these candidate facilities using a preset optimization algorithm to generate an emergency drainage scheduling sequence. This includes: acquiring the real-time status of the power distribution facilities, the status of fixed drainage equipment, and the status of mobile emergency drainage resources; calculating the urgency score of the power distribution facilities based on their real-time status and obtaining several candidate power distribution facilities based on the urgency score; allocating corresponding emergency drainage scheduling resources to the several candidate power distribution facilities based on the status of fixed drainage equipment and mobile emergency drainage resources to construct an initial emergency drainage scheduling sequence; and performing feasibility optimization on the initial emergency drainage scheduling sequence based on the preset optimization algorithm and equipment risk scores to obtain an emergency drainage scheduling sequence that meets the preset emergency drainage scheduling requirements.

[0053] One preferred implementation method involves acquiring the real-time status of power distribution facilities, including the rated capacity of the main transformer, current load, historical highest water level record, voltage level and protection level (IP65 / IP67) of switchgear, and installation height and waterproofing measures of control equipment; acquiring the status of fixed drainage equipment, including the rated flow rate, head, operating status, and cumulative operating time of drainage pumps, and the opening degree, flow coefficient, and maintenance records of drainage valves; and acquiring the status of mobile emergency drainage resources, including the location coordinates, rated flow rate, and estimated arrival time of mobile pump trucks, and the number of personnel, equipment configuration, and current task status of emergency response teams; and reading the current predicted water depth of each device based on the latest monitoring data and short-term forecasts. Safe water level and importance The factors include the available discharge capacity and start / stop status of fixed pumps / valvees, the location and available capacity of mobile pump trucks, road accessibility and station area accessibility, and the safety boundaries of power grid operation. Then, based on the current predicted water depth... Safe water level and importance The urgency score of the calculated equipment is expressed as: The urgency scores of each power distribution facility are sorted in descending order. A hysteresis mechanism with minimum duration and threshold hysteresis is introduced for equipment with jitter to obtain several candidate power distribution facilities, forming the candidate power distribution facility set for this round. The hysteresis mechanism with minimum duration and threshold hysteresis includes a minimum duration of 15 minutes and a threshold hysteresis of 0.05m to prevent frequent start-stop.

[0054] Then, targeting the candidate power distribution facility set, the remaining discharge capacity of the station's fixed pumps / valve is allocated from high to low according to the equipment urgency score, satisfying constraints such as discharge capacity limits, start-up and shutdown intervals, and minimum operating time. Allocations exceeding constraints are partially rolled back and redistributed. For example, if the fixed drainage system configuration within the rain gauge station includes two main drainage pumps with a flow rate of 500 m³ / h and a total head of 15 m, one standby drainage pump with a flow rate of 300 m³ / h (emergency start-up required), and a DN800 drainage pipeline with a design flow rate of 1200 m³ / h; then the constraints include a minimum start-up and shutdown interval of 5 minutes, a minimum operating time of 10 minutes, and a limit of no more than two pumps operating simultaneously; the allocation algorithm processes the equipment in the candidate set in descending order of urgency score. If the equipment's required flow rate is less than or equal to the total available flow rate, the fixed pump flow rate is allocated and the available flow rate is updated; otherwise, it is marked as requiring mobile support.

[0055] For mobile pumping resources, for equipment that remains high-risk after a fixed discharge rate, the nearest reach priority is given, taking into account travel time, road accessibility, station accessibility, and mutual exclusion service constraints. The nearest neighbor or shortest path heuristic is used to generate pump truck-task matching and arrival times. One interpretation is that mobile pump truck scheduling uses a matrix management approach. For example, pump truck PC01 is located at coordinates (x1, y1), with a flow rate of 800 m³ / h, an arrival time of 15 minutes, and an idle status; pump truck PC02 is located at coordinates (x2, y2), with a flow rate of 1200 m³ / h, an arrival time of 25 minutes, and an operating status; pump truck PC03 is located at coordinates (x3, y3), with a flow rate of 600 m³ / h, an arrival time of 8 minutes, and an idle status. In this embodiment, the nearest reachability priority includes four steps: reachability determination, arrival time correction, comprehensive priority calculation and constraint conflict handling. Among them, reachability determination refers to the reachability assessment of the path from each pump truck to each demand site. Generally, a Boolean reachability matrix R[i,j] is used to represent whether pump truck i can reach site j. Reachable is marked as 1, and unreachable is marked as 0. Then, arrival time correction refers to calculating the actual arrival time T for a reachable path R[i,j]=1 based on a multi-factor correction model. arrival The corrected formula is as follows: In the formula, Based on travel time; This is a real-time traffic correction factor; This is a correction factor for road waterlogging. This is a correction factor for the bridge's load-bearing capacity. For queuing and waiting time; The comprehensive priority calculation takes into account accessibility, arrival time, and pump truck capacity for dispatch priority. The calculation formula is as follows: In the formula, Assign a score to the dispatch priority (the higher the score, the higher the priority); To correct arrival time; The rated flow rate of the pump truck, Traffic is required for the site; Path distance; weighting coefficient =0.5、 =0.3、 =0.2, indicating that arrival time is the most critical factor, followed by capability matching, with distance serving as a supplement.

[0056] Finally, constraint conflict handling refers to the process of handling mutually exclusive service constraints, such as when a pump truck is simultaneously needed by multiple sites. In this case, the urgency priority principle is applied, which includes comparing the urgency scores of the equipment at each site. The pump trucks are prioritized for allocation to the sites with the highest urgency. For sites where allocation is unsuccessful, the priority of the remaining available pump trucks is recalculated and they are reallocated. This process ultimately constructs the initial emergency scheduling sequence.

[0057] A feasibility check is performed on the initial preemptive scheduling sequence, and the feasibility check adopts a constraint check; if the feasibility check requirements are not met, local repairs are performed according to priority to minimize the costs of switching, delaying, and replacement.

[0058] Finally, based on the preset optimization algorithm and equipment risk score, the initial rush scheduling sequence is optimized for feasibility to obtain a rush scheduling sequence that meets the preset rush scheduling requirements.

[0059] In the above scheme, the real-time status of facilities is acquired, and the urgency score of each facility is dynamically calculated to determine candidate facilities. Based on this, resource constraints such as the availability of fixed drainage equipment and mobile emergency drainage resources are considered to allocate appropriate emergency drainage scheduling resources to the candidate facilities, constructing an initial scheduling sequence. Finally, based on the equipment risk score, a preset optimization algorithm is used to perform feasibility optimization on the initial emergency drainage scheduling sequence to obtain an emergency drainage scheduling sequence that meets the preset emergency drainage scheduling requirements. This ensures that scheduling decisions are both risk-oriented and closely integrated with real-time situation and resource conditions, avoiding a disconnect between decision-making and execution. It also transforms the scheduling scheme from a static contingency plan into a dynamic response, improving the ability to formulate feasible and efficient emergency drainage plans in extreme environments and resource-limited scenarios, and enhancing the reliability of emergency drainage scheduling for power distribution facilities under extreme weather conditions.

[0060] A preferred approach involves optimizing the initial emergency scheduling sequence based on a preset optimization algorithm and equipment risk scoring to obtain an emergency scheduling sequence that meets preset emergency scheduling requirements. This includes: performing a feasibility check on the initial emergency scheduling sequence to obtain a feasibility check result; if the feasibility check result does not meet the preset feasibility check requirements, then constraining and repairing the initial emergency scheduling sequence to obtain a feasible emergency scheduling sequence; and based on the preset emergency scheduling objective, performing lightweight optimization on the feasible emergency scheduling sequence using a preset optimization algorithm and equipment risk scoring to obtain an emergency scheduling sequence that meets the preset emergency scheduling requirements.

[0061] One preferred implementation involves performing a feasibility check on the initial emergency dispatch sequence, employing a constraint check. If the feasibility check requirements are not met, local repairs are performed to minimize the costs of switching, delays, and replacements based on priority. In this embodiment, the constraint check includes power grid operation safety constraints and station area hydraulic constraints. The power grid operation safety constraints require that the load transfer capacity not exceed 80% of the line thermal stability limit, the voltage stability margin not less than 10%, and the short-circuit current not exceed 90% of the switching equipment breaking capacity. The station area hydraulic constraints require that the total drainage volume not exceed the design drainage capacity multiplied by a safety factor of 0.9, the pump station inlet conditions meet the requirement that the water level difference not be less than the minimum starting head of 0.5m, and the pipeline pressure not exceed 1.2 times the design pressure.

[0062] If the feasibility check requirements are not met, local repairs are carried out according to priority to minimize the costs of switching, delaying, and replacing. Specific constraint repair strategies include delaying conflicting tasks by 5-10 minutes, replacing them with backup equipment or paths, switching them by adjusting the operation sequence, and degrading them by reducing drainage flow or extending operation time, to obtain a feasible emergency drainage scheduling sequence.

[0063] Finally, based on the preset optimization algorithm and equipment risk score, the feasible rush scheduling sequence is lightly optimized to obtain a rush scheduling sequence that meets the preset rush scheduling requirements. Specifically, using "total risk reduction during the planning period - switching cost - travel cost" as the objective function, a time-limited light optimization is performed using 2-opt, local redistribution, or few-generation swarm intelligence algorithms to obtain a rush scheduling sequence that meets the preset rush scheduling requirements. In this embodiment, the objective function is defined as: In the formula, The initial equipment risk score for device i. To determine the equipment risk score for equipment i at time t after measures are taken. For equipment switchover costs, The cost of the pump truck is calculated using a cost weighting coefficient δ=0.1 and φ=0.05.

[0064] It is worth mentioning that during the lightweight optimization process, the algorithm can be adaptively selected based on the problem size and time constraints.

[0065] The above scheme incorporates a feasibility check step, which automatically identifies potential resource conflicts, time conflicts, or logical contradictions in the initial sequence. For initial emergency dispatch sequences that do not meet feasibility requirements, automatic constraint repair is performed to generate feasible emergency dispatch sequences. Based on this, lightweight optimization is performed using a preset optimization algorithm and equipment risk scoring, based on preset emergency dispatch objectives, to find a better solution in the feasible solution space, resulting in an emergency dispatch sequence that meets the preset emergency dispatch requirements. This ensures that the final output emergency dispatch sequence is not only executable but also of high quality and efficiency, improving the reliability of emergency dispatch of power distribution facilities under extreme weather conditions.

[0066] Example 2 See Figure 2 , Figure 2 This is a schematic diagram of the module structure of a power distribution facility emergency dispatching system based on risk early warning, provided as an embodiment of the present invention. Figure 2 As shown, this embodiment of the invention also provides a power distribution facility emergency drainage scheduling system based on risk warning, including: a rainfall forecast field correction module 201, used to obtain a corrected rainfall forecast field based on pre-acquired rain gauge observation data, gridded forecast data, and topographic factors in the area to be urgently drained; a flood data acquisition module 202, used to construct a local flood forecast model based on pre-acquired topographic features and drainage system layout in a local area of ​​the power distribution facility, and input the corrected rainfall forecast field into the local flood forecast model to obtain water depth data, water flow velocity, and inundation duration at the location of the power distribution facility; a risk score calculation module 203, used to calculate the equipment risk score based on the water depth data, water flow velocity, and inundation duration; an emergency drainage scheduling sequence generation module 204, used to determine candidate power distribution facilities based on the equipment risk score, and allocate corresponding emergency drainage scheduling resources to the candidate power distribution facilities through a preset optimization algorithm to generate an emergency drainage scheduling sequence; and an emergency drainage scheduling execution module 205, used to allocate the emergency drainage scheduling sequence to the power distribution facilities and execute corresponding emergency drainage scheduling actions in the area to be urgently drained.

[0067] This invention proposes a risk-based emergency drainage scheduling system for power distribution facilities. It combines pre-acquired rain gauge observation data, gridded forecast data, and topographic factors to generate a corrected rainfall forecast field, providing reliable input for subsequent flood forecasting. Secondly, a flood forecasting model is constructed based on local topography and drainage systems. By dynamically predicting water depth, flow velocity, and inundation time at the location of power distribution facilities, the system quantifies the flood risk faced by the equipment. Based on this, high-risk candidate facilities are identified by calculating equipment risk scores, and an optimization algorithm is used to rationally allocate emergency drainage resources, generating an executable emergency drainage scheduling sequence. Finally, the emergency drainage scheduling sequence is distributed to power distribution facilities for automatic or assisted emergency drainage scheduling, addressing the reliability issues of existing methods that lack risk quantification assessment, thus improving the reliability of emergency drainage scheduling for power distribution facilities under extreme weather conditions.

[0068] Furthermore, the rainfall forecast field correction module 201 is used to obtain a corrected rainfall forecast field based on the pre-acquired rain gauge observation data, gridded forecast data, and topographic factors within the area to be expedited for emergency drainage and dispatching. This includes: an initial data acquisition unit 301, used to divide the power distribution facility area into several forecast grids, and acquire rain gauge observation data from each rain gauge station and gridded forecast data from each forecast grid within the area to be expedited for emergency drainage and dispatching based on a preset time interval; an outlier repair unit 302, used to correct outliers in the rain gauge observation data and gridded forecast data to obtain standard rain gauge observation data and standard gridded forecast data; and an initial deviation field construction unit 3... 03 is used to time-register rainfall stations and forecast grids based on a preset time reference and time period scale, and to calculate the initial station forecast deviation field; the terrain deviation relationship acquisition unit 304 is used to construct the terrain deviation relationship through the pre-acquired terrain factors and the initial station forecast deviation field; the target deviation field construction unit 305 is used to spatially expand the station forecast deviation field based on the terrain deviation relationship through a preset interpolation algorithm to obtain the target station forecast deviation field; the overlay correction unit 306 is used to overlay and correct the standard gridded forecast data and the target station forecast deviation field based on the standard rainfall station observation data to obtain the corrected rainfall forecast field.

[0069] Furthermore, the flood data acquisition module 202 is used to construct a local flood forecast model based on pre-acquired topographic features and drainage system layout within a local area of ​​the power distribution facility, and input the corrected rainfall forecast field into the local flood forecast model to obtain water depth data, water flow velocity, and inundation duration at the location of the power distribution facility. This includes: a topographic feature acquisition unit 401, used to acquire topographic features within a local area of ​​the power distribution facility based on a preset area; and a drainage system layout acquisition unit 402, used to acquire the drainage system layout within the local area of ​​the power distribution facility based on preset drainage design drawings; and initial... The initial scene matching unit 403 is used to acquire real-time water accumulation monitoring data in the local area of ​​the power distribution facility, and select the initial scene that meets the preset similarity requirements based on the preset historical water accumulation scenario database and preset similarity algorithm; the local flood forecast model construction unit 404 is used to construct a local flood forecast model based on the initial scene and preset finite difference algorithm; the flood data solving unit 405 is used to input the modified rainfall forecast field into the local flood forecast model, and solve the water accumulation process in the local area of ​​the power distribution facility through the local flood forecast model to obtain the water depth data, water flow velocity and inundation duration of the power distribution facility location.

[0070] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

[0071] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. Furthermore, the described specific features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0072] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "a plurality of" means two or more, unless otherwise explicitly specified.

Claims

1. A method for emergency dispatching of power distribution facilities based on risk early warning, characterized in that, include: Based on the pre-acquired rain gauge observation data, gridded forecast data and topographic factors in the area to be urgently dispatched, a corrected rainfall forecast field is obtained; Based on the pre-acquired topographic features and drainage system layout within the local area of ​​the power distribution facility, a local flood forecast model is constructed, and the modified rainfall forecast field is input into the local flood forecast model to obtain water depth data, water flow velocity, and inundation duration at the location of the power distribution facility. Based on the water depth data, the water flow velocity, and the inundation duration, a risk score for the equipment is calculated. Candidate power distribution facilities are determined based on the equipment risk score, and corresponding emergency scheduling resources are allocated to the candidate power distribution facilities through a preset optimization algorithm to generate an emergency scheduling sequence. The emergency response and dispatch sequence is allocated to the power distribution facilities, and the corresponding emergency response and dispatch actions are executed in the areas to be urgently dispatched.

2. The method for emergency dispatching of power distribution facilities based on risk early warning as described in claim 1, characterized in that, The revised rainfall forecast field is obtained based on pre-acquired rain gauge observation data, gridded forecast data, and topographic factors, including: The power distribution facility area is divided into several forecast grids, and the rain gauge observation data of each rain gauge station in the area to be dispatched and the gridded forecast data of each forecast grid are obtained based on the preset time interval. Outlier corrections are performed on the rain gauge observation data and gridded forecast data to obtain standard rain gauge observation data and standard gridded forecast data; Based on a preset time reference and time period scale, the rainfall stations and forecast grids are time-registered, and the initial station forecast deviation field is calculated. A terrain deviation relationship is constructed using the pre-acquired terrain factors and the initial station prediction deviation field; Based on the terrain deviation relationship, the initial station prediction deviation field is spatially expanded using a preset interpolation algorithm to obtain the target station prediction deviation field. Based on the observation data from the standard rain gauge stations, the standard gridded forecast data and the forecast deviation field of the target station are superimposed and corrected to obtain the corrected rainfall forecast field.

3. The method for emergency dispatching of power distribution facilities based on risk early warning as described in claim 2, characterized in that, Based on the aforementioned terrain deviation relationship, the initial station prediction deviation field is spatially extended using a preset interpolation algorithm to obtain the target station prediction deviation field, including: The terrain relief index is calculated based on the terrain deviation relationship, and the search range is determined based on the terrain relief index. By using a preset three-dimensional distance algorithm and the search range, a list of adjacent stations and the three-dimensional distance between each adjacent station are obtained. Based on the list of adjacent stations and the three-dimensional distance between each adjacent station, the semivariogram is fitted to several fitting models using a preset semivariogram and a preset fitting algorithm, and the fitting model that meets the preset fitting requirements is selected as the variogram model. Based on the list of adjacent sites, the anisotropy type of each site is obtained by using a preset anisotropy solving algorithm; Based on the initial site forecast bias field, the interpolation matrix of each forecast grid is constructed by using a preset interpolation algorithm, the variation function model, and the anisotropy type of each site. The interpolation matrix of each forecast grid is solved based on the preset decomposition algorithm to obtain the forecast deviation field of the target station.

4. The method for emergency dispatching of power distribution facilities based on risk early warning as described in claim 1, characterized in that, Based on pre-acquired topographic features and drainage system layout within a local area of ​​the power distribution facility, a local flood forecasting model is constructed. The corrected rainfall forecast field is then input into the local flood forecasting model to obtain water depth data, water flow velocity, and inundation duration at the location of the power distribution facility, including: Based on the preset area range, obtain the topographic features of the local area of ​​the power distribution facility; Based on the pre-designed drainage drawings, obtain the layout of the drainage system in a local area of ​​the power distribution facility; Real-time acquisition of water accumulation monitoring data in local areas of power distribution facilities; initial scenarios that meet preset similarity requirements are selected based on a preset historical water accumulation scenario database and a preset similarity algorithm. Based on the initial scenario and the preset finite difference algorithm, a local flood forecasting model is constructed; The modified rainfall forecast field is input into the local flood forecast model, and the local flood forecast model is used to solve the local water accumulation process of the power distribution facility, so as to obtain the water depth data, water flow velocity and flooding duration of the power distribution facility location.

5. The method for emergency dispatching of power distribution facilities based on risk early warning as described in claim 1, characterized in that, Based on the water depth data, the water flow velocity, and the inundation duration, a risk score for the equipment is calculated, including: Based on a preset first weighting coefficient, the water depth data, the water flow velocity, and the inundation duration are matched with corresponding first weighting values ​​to calculate the hazard index; Acquire population density and economic impact data within the power supply range of the power distribution facilities, and match the population density and economic impact data with corresponding second weight values ​​by a preset second weight coefficient to calculate the degree of impact index; Obtain historical flooding data for local areas of power distribution facilities and calculate resistance indicators; Obtain the importance mapping values ​​of each power distribution facility and calculate the equipment importance index; Based on the hazard index, the impact index, the resistance index, and the equipment importance index, the equipment risk score is calculated.

6. The method for emergency dispatching of power distribution facilities based on risk early warning as described in claim 1, characterized in that, Candidate power distribution facilities are determined based on the equipment risk score, and corresponding emergency scheduling resources are allocated to the candidate power distribution facilities using a preset optimization algorithm to generate an emergency scheduling sequence, including: Acquire the real-time status of power distribution facilities, fixed drainage equipment, and mobile emergency drainage resources; Based on the real-time status of the power distribution facilities, calculate the urgency score of the power distribution facilities and obtain several candidate power distribution facilities according to the urgency score of the power distribution facilities; Based on the status of the fixed drainage equipment and the status of the mobile emergency drainage resources, corresponding emergency drainage scheduling resources are allocated to several candidate power distribution facilities to construct an initial emergency drainage scheduling sequence. Based on the preset optimization algorithm and the equipment risk score, the initial rush scheduling sequence is optimized for feasibility to obtain a rush scheduling sequence that meets the preset rush scheduling requirements.

7. The method for emergency dispatching of power distribution facilities based on risk early warning as described in claim 6, characterized in that, Based on the preset optimization algorithm and the equipment risk score, the initial rush scheduling sequence is optimized for feasibility to obtain a rush scheduling sequence that meets the preset rush scheduling requirements, including: A feasibility check is performed on the initial preemptive scheduling sequence to obtain the feasibility check result; If the feasibility check result does not meet the preset feasibility check requirements, then the initial preemptive scheduling sequence is constrained and repaired to obtain a feasible preemptive scheduling sequence. Based on the preset emergency scheduling target, the feasible emergency scheduling sequence is lightly optimized using a preset optimization algorithm and the equipment risk score to obtain an emergency scheduling sequence that meets the preset emergency scheduling requirements.

8. A power distribution facility emergency dispatch system based on risk early warning, characterized in that, The method for emergency dispatching of power distribution facilities based on risk warning, as described in any one of claims 1 to 7, includes: The rainfall forecast field correction module is used to obtain a corrected rainfall forecast field based on the pre-acquired rain gauge observation data, gridded forecast data and topographic factors in the area to be urgently dispatched; The flood data acquisition module is used to construct a local flood forecast model based on the pre-acquired topographic features and drainage system layout in a local area of ​​the power distribution facility, and input the modified rainfall forecast field into the local flood forecast model to obtain water depth data, water flow velocity and inundation duration at the location of the power distribution facility; The risk score calculation module is used to calculate the equipment risk score based on the water depth data, the water flow velocity, and the inundation duration. The emergency dispatch sequence generation module is used to determine candidate power distribution facilities based on the equipment risk score, and to allocate corresponding emergency dispatch resources to the candidate power distribution facilities through a preset optimization algorithm to generate an emergency dispatch sequence. The emergency dispatching execution module is used to allocate the emergency dispatching sequence to the power distribution facilities and execute the corresponding emergency dispatching actions in the areas to be dispatched.

9. A power distribution facility emergency dispatch system based on risk early warning as described in claim 8, characterized in that, The rainfall forecast field correction module is used to obtain a corrected rainfall forecast field based on pre-acquired rain gauge observation data, gridded forecast data, and topographic factors within the area to be expedited for emergency drainage and scheduling. This includes: The initial data acquisition unit is used to divide the power distribution facility area into several forecast grids and acquire the rain gauge observation data of each rain gauge station and the gridded forecast data of each forecast grid in the area to be urgently dispatched based on a preset time interval. An outlier repair unit is used to correct outliers in the rain gauge observation data and gridded forecast data to obtain standard rain gauge observation data and standard gridded forecast data. The initial deviation field construction unit is used to perform time registration between the rainfall stations and the forecast grid based on a preset time reference and time period scale, and to calculate the initial station forecast deviation field. The terrain deviation relationship acquisition unit is used to construct the terrain deviation relationship using the pre-acquired terrain factors and the initial station predicted deviation field; The target deviation field construction unit is used to spatially expand the station prediction deviation field based on the terrain deviation relationship and through a preset interpolation algorithm to obtain the target station prediction deviation field. The superposition correction unit is used to superimpose and correct the standard gridded forecast data and the forecast deviation field of the target station based on the standard rain gauge observation data to obtain the corrected rainfall forecast field.

10. A power distribution facility emergency dispatch system based on risk early warning as described in claim 8, characterized in that, The flood data acquisition module is used to construct a local flood forecasting model based on pre-acquired topographic features and drainage system layout within a local area of ​​the power distribution facility, and input the corrected rainfall forecast field into the local flood forecasting model to obtain water depth data, water flow velocity, and inundation duration at the location of the power distribution facility, including: The terrain and geomorphology feature acquisition unit is used to acquire the terrain and geomorphology features of a local area of ​​the power distribution facility based on a preset area range; The drainage system layout acquisition unit is used to acquire the drainage system layout of a local area of ​​the power distribution facility based on the preset drainage design drawings. The initial scene matching unit is used to acquire water accumulation monitoring data in a local area of ​​the power distribution facility in real time, and to select initial scenes that meet the preset similarity requirements based on a preset historical water accumulation scenario database and a preset similarity algorithm. The local flood forecasting model construction unit is used to construct a local flood forecasting model based on the initial scenario and the preset finite difference algorithm. The flood data solving unit is used to input the modified rainfall forecast field into the local flood forecast model, and solve the local water accumulation process of the power distribution facility through the local flood forecast model to obtain the water depth data, water flow velocity and inundation duration of the power distribution facility location.

Images