A power material scheduling method, system, device and medium

Abstract

The application discloses a kind of electric power material scheduling method, system, equipment and medium, belong to electric power system field, method is: according to fire change tendency and base station position, the base station of existence power interruption risk is identified, obtain multiple risk base stations, the density data is obtained by carrying out density analysis to all risk base stations, combined with density data and the historical outage data of each risk base station is spatio-temporal joint analysis, obtain interruption probability distribution;The demand information of each risk base station is matched with inventory information and is analyzed, and resource coverage is obtained, if resource coverage is lower than preset threshold, then based on fire change tendency, interruption probability distribution and inventory information, from multiple risk base stations, determine that there is supply and demand imbalance in the target base station in the prediction time interval;Resource scheduling is carried out to each target base station, and the scheduling scheme station is obtained, therefore, by implementing the present application, the accuracy of base station interruption risk prediction can be improved, and then resource utilization and power safety are improved.

Description

Technical Field

[0001] This invention relates to the field of power systems, and more particularly to a method, system, equipment, and medium for dispatching power resources. Background Technology

[0002] In responding to large-scale natural disasters such as forest fires, ensuring power grid communication is fundamental for rescue command and disaster information transmission. Communication base stations, as key nodes in the power grid, are crucial to the power supply. However, the rapid spread of fires can directly lead to power outages at numerous communication base stations within the fire zone, while the availability of portable battery modules and other power supplies for emergency power generation in storage facilities is often limited. Therefore, optimizing the allocation of power resources during the dynamic process of fire occurrence and development is beneficial for improving emergency response efficiency, avoiding resource waste, and enhancing power grid security.

[0003] Currently, existing power emergency material dispatching methods mostly focus on immediate response after a disaster occurs. For example, an emergency resource dispatching scheme based on real-time disaster information and base station geographic information typically plans the shortest path to transport emergency materials (such as battery modules) after a communication base station experiences a power outage, based on its geographical location and the distance to storage points. However, due to the rapid and unpredictable spread of forest fires, existing methods cannot accurately and proactively predict which base stations will face the risk of power outages before they actually lose power. This can lead to the risk of power outages for critical communication base stations during periods of resource scarcity, resulting in low resource utilization and low security in the dispatching scheme. Summary of the Invention

[0004] This invention provides a method, system, equipment, and medium for power material dispatching, which can solve the problem of low accuracy in existing interruption risk prediction due to the uncertainty of disaster spread.

[0005] This invention provides a method for dispatching power resources, comprising: Based on the obtained fire change trend and communication base station location data of the target forest area, base stations with power outage risk are identified, and multiple risk base stations are obtained. Spatial density analysis is performed on all the risk base stations to obtain density data. Combined with the density data and the historical power outage data of each risk base station, a spatiotemporal joint analysis is performed to obtain the outage probability distribution. The demand information of each risk base station is matched and analyzed with the inventory information of each power material warehouse to obtain the resource coverage rate. If the resource coverage rate is lower than the preset threshold, the target base station with supply and demand imbalance in the predicted time interval is determined from the multiple risk base stations based on the fire change trend, the interruption probability distribution and the inventory information. Resource scheduling is performed on each target base station to obtain a scheduling plan, and power supplies are pre-scheduled to the target base stations within the predicted time interval according to the scheduling plan.

[0006] This invention integrates fire trend data with base station location data to identify base stations at risk of power outages, enabling pre-screening of fire-base station coupled risks and identifying sites that are truly likely to experience power outages in advance, thus avoiding blind adjustments across the entire network. By performing spatial density analysis on risky base stations and combining it with historical power outage data for spatiotemporal joint analysis, it achieves a density-time dual-corrected outage probability distribution, resolving misjudgments caused by "same distance, different historical data" and reducing the false identification rate. By matching base station demand with material inventory and calculating resource coverage, it achieves quantitative supply and demand hedging, clearly showing whether "enough is available" and "not enough time." When coverage is below a preset threshold, it identifies target base stations with supply and demand imbalances within a predicted time interval based on fire trends, outage probabilities, and inventory information, achieving precise hedging within the time window, transforming "potential future shortages" into "immediate adjustments," compressing the pre-emptive scheduling window from reactive to proactive. By generating scheduling plans and pre-scheduling power materials within the predicted time interval, it achieves proactive emergency hedging, completing battery deployment before the fire arrives and shortening the average time for base station power outage recovery. In summary, this embodiment realizes a four-dimensional integrated hedging scheduling of "fire situation-risk-supply and demand-time", which can improve resource coverage, reduce mis-schedule rate, shorten recovery time, improve resource utilization, and enhance scheduling security.

[0007] Furthermore, the process of identifying base stations at risk of power outages based on the acquired fire trend data and communication base station location data in the target forest area specifically involves: Acquire multispectral remote sensing image data of the target forest area, as well as meteorological data that is spatiotemporally synchronized with the multispectral remote sensing image data. The multispectral remote sensing image data includes thermal infrared band data, visible light band data, and near-infrared band data, and the meteorological data includes wind speed, wind direction, relative humidity, and ambient temperature. The temperature at each location in the target forest area is obtained by inversion based on thermal infrared band data, and the normalized differential vegetation index at each location in the target forest area is calculated based on the visible light band data and the near-infrared band data. The initial fire point is identified by comparing the temperature threshold with the temperature at each location, and by comparing the vegetation index threshold with the normalized difference vegetation index at each location. The initial fire state map is then generated by combining the vegetation index threshold with the normalized difference vegetation index at each location. The temperature threshold is dynamically determined based on the ambient temperature. Image segmentation and morphological processing are performed on the initial state map of the fire to obtain the spatial morphological information of the target fire area. Based on the spatial morphological information, the temperature at each location and the meteorological data, the fire change trend is predicted. The locations of all communication base stations are spatially compared with the fire change trend. If the distance between the communication base station and the fire boundary is less than a preset safety threshold within the predicted time interval, the communication base station is identified as the risk base station.

[0008] This approach achieves high accuracy in fire point identification and avoids single-source false alarms by acquiring multispectral remote sensing data combined with spatiotemporally synchronized meteorological data and integrating space-based and ground-based data. Initial fire points are identified using a dual-condition approach of thermal infrared temperature retrieval and NDVI thresholding, enabling joint vegetation-temperature triggering and filtering out false fire points such as bare ground hotspots and sunlight reflections. A dynamic temperature threshold (fluctuating with ambient temperature) enables seasonally adaptive fire point discrimination. Image segmentation and morphological processing are used to obtain the spatial morphology of the fire, achieving refined fire field contours with a boundary error of less than one pixel, providing high-precision input for subsequent spread prediction. By comparing the spatial relationship between the location of communication base stations and the fire's changing trend, base stations within a safe threshold are identified as high-risk base stations, achieving dual risk locking based on spatial distance and time window.

[0009] Furthermore, the prediction of the fire change trend based on the spatial morphology information, the temperature at each location, and the meteorological data specifically includes: Based on the temperature at each location, combined with the temperature threshold and area threshold, the fire intensity level is determined, and the fuel type is determined based on the multispectral remote sensing data. Based on the fire intensity level and the fuel type, the combustion rate of the target fire area is obtained by querying a preset combustion rate database. Based on the wind speed, relative humidity, and combustion rate, the first boundary expansion velocity of the target fire area is calculated, and the first boundary expansion velocity is adjusted based on the wind direction and terrain slope data to obtain the second boundary expansion velocity in different directions. Based on the current boundary information in the spatial morphology information, the fire change trend is obtained by iteratively calculating the boundary position sequence within the prediction time interval according to the second boundary expansion speed.

[0010] This approach quantifies fire intensity by using a combination of temperature and area thresholds, preventing underestimation of fire rates due to large fires. It automatically identifies fuel types by inverting fuel type based on smoke particle diameter, incorporating differences in burning rates between coniferous, broadleaf, and shrub forests to reduce rate errors. The first boundary expansion velocity is obtained by querying a database using wind speed, humidity, and burning rate, enabling meteorological-fuel coupled rate calculations that replace empirical coefficients and improve rate accuracy. The second boundary expansion velocity is obtained by applying directional correction to the rate using wind direction and terrain slope, achieving joint correction based on slope and wind direction. Finally, iterative calculation of the boundary position sequence yields the fire change trend within the predicted time interval, automatically generating a trend map and providing reliable input for base station power outage countdowns.

[0011] Furthermore, the spatiotemporal joint analysis combining the density data and historical power outage data of each risky base station yields the outage probability distribution, specifically as follows: Based on the density data, each risky base station is divided into different density intervals, and based on the historical power outage data, the average power outage frequency of each risky base station in each density interval is statistically calculated, and the spatial influence coefficient of each density interval is calculated. The spatial influence coefficient is used to characterize the influence coefficient of base station density on the power outage probability. Time series analysis is performed on the historical power outage data to identify the periodic characteristics of power outage events, and the time correction factor for each risky base station at the current moment is determined based on the periodic characteristics. Based on the spatial influence coefficient corresponding to the density interval of each risk base station, the initial interruption probability is spatially corrected to obtain a first corrected probability, and the initial interruption probability is temporally corrected based on the time correction factor to obtain a second corrected probability. The first corrected probability and the second corrected probability are weighted and fused to obtain the target interruption probability of each risk base station, and the interruption probability distribution is obtained based on each target interruption probability.

[0012] This approach uses density interval division and statistical average power outage frequency to obtain a spatial influence coefficient, achieving a quantitative mapping between density and power outage frequency and resolving the difference in outage probability caused by "same distance, different density". By using time series analysis to identify power outage cycle characteristics, it enables automatic extraction of seasonal / monthly power outage patterns, improving the timeliness of probability. By using spatial correction coefficient and time correction factor weighted fusion to obtain the target outage probability, it achieves spatiotemporal two-dimensional probability fusion, and the probability difference of the same base station in different months can reach 2 times, avoiding the misjudgment of "one-size-fits-all throughout the year".

[0013] Furthermore, the process of matching and analyzing the demand information of each risk base station with the inventory information of each power material warehouse to obtain the resource coverage rate is as follows: The initial transportation distance was calculated based on the geographical coordinates of each power material warehouse and each risk base station. The terrain and road data are acquired, the initial transportation distance is corrected to obtain the actual transportation length, and the actual transportation time from each power material warehouse to each risk base station is calculated in combination with the transportation speed. The actual transportation time is compared with the remaining power duration of each risk base station, and the risk base stations whose actual transportation time is less than the remaining power duration are identified as coverable risk base stations. The resource coverage rate is calculated based on the number of coverable risk base stations and the total number of all risk base stations.

[0014] This approach corrects the initial transport distance using terrain and road data, restoring the true path length in mountainous areas and reducing distance errors. By comparing the actual transport time with the remaining power time of the base station, timely coverage determination is achieved. If the battery delivery time exceeds the backup power depletion time, the system is deemed uncoverable, avoiding ineffective transport that is delivered but has already lost power, thus improving resource utilization. The resource coverage rate is calculated by dividing the number of coverable base stations by the total number of risky base stations, enabling a clear understanding of supply and demand matching. If the coverage rate falls below a threshold, a hedging warning is immediately triggered.

[0015] Furthermore, the step of identifying target base stations with supply-demand imbalances within the predicted time interval from the plurality of risk base stations based on the fire change trend, the interruption probability distribution, and the inventory information specifically involves: Based on the total demand for power materials of all risk base stations and the total inventory in the inventory information, the resource difference is calculated. At the same time, the degree of spatial mismatch between the inventory distribution of each power material warehouse and the demand distribution of all risk base stations is calculated to obtain the inventory distribution deviation. If the resource difference is greater than the preset difference, or the inventory distribution deviation is greater than the preset inventory deviation, then it is determined that there is a supply and demand contradiction. In response to the supply and demand imbalance, based on the interruption probability distribution and the resource coverage, a prediction time interval is determined, and base stations whose interruption probability is greater than a preset interruption threshold and which are not covered risk base stations within the prediction time interval are identified as the target base stations.

[0016] This approach uses a dual indicator of resource difference and inventory distribution deviation to determine supply and demand imbalances, enabling the detection of both total quantity and distribution imbalances and improving resource utilization. By identifying base stations with an interruption probability greater than a threshold and not belonging to a covered base station category, a dual locking mechanism of high risk and unreachability is achieved, improving the accuracy of risk base station identification and reducing ineffective scheduling. Furthermore, by dynamically generating predicted time intervals, a time window coupling fire speed and resource speed is realized, transforming the time interval from static to dynamic and resulting in higher scheduling granularity.

[0017] Furthermore, the resource scheduling for each target base station to obtain the scheduling scheme is specifically as follows: Construct an optimization function with the objectives of minimizing total transport time and maximizing the number of coverable risky base stations; The scheduling scheme is obtained by solving the optimization function, using the demand of each target base station as the demand constraint and the inventory of each power material warehouse as the supply constraint.

[0018] By constructing a dual-objective optimization function that minimizes total transportation time and maximizes the number of coverable base stations, a Pareto front solution for timeliness and coverage is achieved. A single calculation provides a trade-off solution, eliminating the need for repeated manual trials and improving scheduling efficiency. Furthermore, by using demand constraints and inventory constraints, hard constraints ensure feasibility, allowing the solution to be directly deployed and executed, thus improving its executability.

[0019] Another embodiment of the present invention provides a power material dispatching system, including: an interruption identification module, a supply and demand judgment module, and a dispatching module; The interruption identification module is used to identify base stations with power outage risk based on the fire change trend and communication base station location data of the target forest area, obtain multiple risk base stations, perform spatial density analysis on all the risk base stations to obtain density data, and perform spatiotemporal joint analysis by combining the density data and the historical power outage data of each risk base station to obtain the interruption probability distribution. The supply and demand judgment module is used to match and analyze the demand information of each risk base station with the inventory information of each power material warehouse to obtain the resource coverage rate. If the resource coverage rate is lower than a preset threshold, the target base station with supply and demand imbalance in the predicted time interval is determined from the multiple risk base stations based on the fire change trend, the interruption probability distribution and the inventory information. The scheduling module is used to perform resource scheduling on each target base station to obtain a scheduling plan, so as to pre-schedule power supplies to the target base station within the predicted time interval according to the scheduling plan.

[0020] Another embodiment of the present invention provides a terminal device, including: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the steps of the power material dispatching method of the present invention.

[0021] Another embodiment of the present invention provides a computer-readable storage medium item, including: a stored computer program, which, when the computer program is running, controls the device where the computer-readable storage medium is located to perform the steps of the power material dispatching method of the present invention. Attached Figure Description

[0022] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0023] Figure 1 This is a flowchart illustrating a power material dispatching method provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of a power material dispatching system provided in an embodiment of the present invention. Detailed Implementation

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

[0025] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0026] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0027] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0028] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0029] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0030] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. For those skilled in the art, the specific meaning of the above terms in the embodiments of this application can be understood according to the specific circumstances.

[0031] See Figure 1 To address the problem of low accuracy in existing power outage risk prediction due to the uncertainty of disaster spread, an embodiment of the present invention provides a power material dispatching method, comprising: Step 101: Based on the obtained fire change trend and communication base station location data of the target forest area, identify base stations with power outage risk to obtain multiple risk base stations. Perform spatial density analysis on all the risk base stations to obtain density data. Combine the density data with the historical power outage data of each risk base station to perform spatiotemporal joint analysis to obtain the outage probability distribution.

[0032] In this embodiment, remote sensing monitoring data and environmental meteorological data of the target area are acquired, and spatiotemporal alignment and fusion processing are performed to identify and extract the current spatial range and dynamic spread trend of the fire. The distribution data of communication base stations are spatially overlaid with the fire situation. Based on the spatiotemporal relationship between the base stations and the fire front, a set of base stations with potential power outage risks is initially identified. For the set of risky base stations, their spatial distribution density characteristics and historical operational reliability data are introduced for joint analysis to construct an evaluation model that integrates geospatial factors and time series characteristics, and outputs the quantitative power outage probability distribution of each risky base station.

[0033] Step 102: Match and analyze the demand information of each risk base station with the inventory information of each power material warehouse to obtain the resource coverage rate. If the resource coverage rate is lower than the preset threshold, then based on the fire change trend, the interruption probability distribution and the inventory information, determine the target base station with supply and demand imbalance in the predicted time interval from the multiple risk base stations.

[0034] In this embodiment, the inventory information and spatial distribution of emergency power supplies are acquired and matched with the geographical location and power demand of risky base stations. The emergency resource coverage rate is quantified by calculating the proportion of risky base stations to which supplies can be delivered within a limited time. When the resource coverage rate is lower than the system's set guarantee capacity threshold, a resource supply and demand imbalance is determined, and an early warning is triggered. Under the early warning status, the direction of fire spread, the interruption probability of each base station, and the inventory and distribution of supplies are comprehensively considered. Base stations with a high risk of interruption within the predicted time window and whose existing resources are difficult to effectively cover are selected from all risky base stations and identified as target base stations that need to be prioritized for scheduling.

[0035] Step 103: Perform resource scheduling on each target base station to obtain a scheduling plan, and pre-schedule power supplies to the target base station within the predicted time interval according to the scheduling plan.

[0036] In this embodiment, a resource scheduling model is constructed with the core objective of maximizing the overall power supply guarantee efficiency of the target base stations. The core constraints of this model include: the minimum demand of each target base station, the inventory capacity of each material depot, and the spatiotemporal constraints of material transportation. Solving the model outputs an executable scheduling configuration scheme. This scheme clearly specifies which material depot to use, which target base station to allocate, what type and quantity of power materials to use, the execution order, and the expected timeline. The scheduling scheme is then transformed into structured scheduling instructions and early warning reports, which are automatically distributed to relevant emergency dispatch centers and implementing units via the network. Based on the final determined scheduling scheme, the status of relevant materials in the material inventory management system is updated in real time (e.g., from "available" to "allocated") to ensure that system data is consistent with actual scheduling actions, providing an accurate basis for subsequent decision-making.

[0037] As an example of an embodiment of the present invention, the identification of base stations at risk of power outage based on the acquired fire change trend and communication base station location data of the target forest area specifically involves: acquiring multispectral remote sensing image data of the target forest area, and meteorological data spatiotemporally synchronized with the multispectral remote sensing image data. The multispectral remote sensing image data includes thermal infrared band data, visible light band data, and near-infrared band data, and the meteorological data includes wind speed, wind direction, relative humidity, and ambient temperature. The temperature at each location in the target forest area is retrieved from the thermal infrared band data, and the normalized differential vegetation at each location in the target forest area is calculated based on the visible light band data and the near-infrared band data. The system employs several methods: First, it compares a temperature threshold with the temperature at each location, and a vegetation index threshold with the normalized difference vegetation index at each location to identify initial fire points. Then, it generates an initial fire state map based on meteorological data. The temperature threshold is dynamically determined according to the ambient temperature. Next, it performs image segmentation and morphological processing on the initial fire state map to obtain spatial morphological information of the target fire area. Based on this spatial morphological information, the temperature at each location, and the meteorological data, it predicts the fire's change trend. Finally, it compares the spatial relationship between the locations of all communication base stations and the fire's change trend. If the distance between a communication base station and the fire boundary is less than a preset safety threshold within the predicted time interval, then the communication base station is identified as a risk base station.

[0038] In this embodiment, multispectral remote sensing image data covering the forest area is acquired from a satellite receiving station, including raw images in the visible light, near-infrared, and thermal infrared bands. Simultaneously, wind speed, wind direction, relative humidity, and ambient temperature data at corresponding times are acquired from a meteorological monitoring station. Through timestamp matching and geographic coordinate alignment, a spatial correspondence between the remote sensing images and meteorological data is established, resulting in a spatiotemporally synchronized multi-source monitoring dataset. For this multi-source monitoring dataset, the temperature value of each pixel is calculated based on the linear relationship between the thermal infrared band pixel value and the surface temperature. Non-vegetation-covered areas are identified using the normalized difference vegetation index (NDVI) in the visible and near-infrared bands. Areas with temperatures higher than a preset threshold and NDVI lower than a preset threshold are marked as suspected fire points. The smoke diffusion range is calculated by combining wind speed and direction data from the meteorological station, resulting in an initial fire state map including temperature field distribution and smoke coverage. The initial state image of the fire is subjected to threshold segmentation processing. Pixels with temperature values ​​greater than the preset fire point temperature threshold are marked as combustion areas. Edge noise is eliminated by expansion and erosion operations. The outer contour of the fire area is extracted by an edge detection operator. The distribution of temperature values ​​and smoke concentration values ​​of each pixel within the contour is statistically analyzed to obtain the contour and features of the fire area.

[0039] For example, the satellite receiving station acquires remote sensing image data of the forest area through polar-orbiting meteorological satellites and geostationary meteorological satellites. The polar-orbiting satellites pass overhead twice daily, providing multispectral data with a spatial resolution of 250 meters to 1000 meters, while the geostationary satellites provide observation data every 15 minutes. The multispectral remote sensing images include 11 spectral bands, with the visible light band covering a wavelength range of 0.4 to 0.7 micrometers, the near-infrared band covering 0.7 to 1.3 micrometers, and the thermal infrared band covering 3.7 to 12 micrometers. Meteorological monitoring stations are deployed in the area surrounding the forest, spaced approximately 20 kilometers apart, to collect real-time data on wind speed, wind direction, relative humidity, and ambient temperature. Time synchronization is achieved using GPS timestamps and Coordinated Universal Time (UTC) to align the satellite transit time with the ground meteorological observation time, controlling the time error to within 5 seconds.

[0040] It should be noted that the geographic coordinate alignment process employs an affine transformation method to unify the coordinate systems of different data sources to the WGS84 coordinate system. First, the four corner coordinates and center point coordinates of the remote sensing image are extracted. Then, a spatial index is established based on the latitude and longitude coordinates of the weather stations. Using a nearest neighbor interpolation algorithm, the discrete weather station data is extended to the entire image coverage area, forming a continuous meteorological element field. For each pixel, the three nearest weather stations are found, and the meteorological element value for that point is calculated using an inverse distance weighting method, where the weighting coefficient is inversely proportional to the square of the distance.

[0041] For example, the conversion between pixel values ​​in the thermal infrared band and surface temperature employs a two-step process: radiometric calibration and atmospheric correction. Radiometric calibration converts pixel grayscale values ​​into radiance values, with conversion coefficients read from satellite metadata files. Atmospheric correction uses the radiative transfer equation, considering the effects of atmospheric transmittance and upward atmospheric radiation. Surface temperature is calculated through the inversion of the Planck function, where emissivity is determined based on surface type: 0.97 for forest areas and 0.92 for bare soil areas. The normalized differential vegetation index is calculated by dividing the difference in reflectance between the near-infrared and red bands by the sum of the two values, with a range of -1 to 1. The index value for healthy vegetation is typically greater than 0.3, while the index value for burned areas is less than 0.1.

[0042] For example, the calculation of the smoke diffusion range comprehensively considers wind speed, wind direction, and topographical factors. According to the Gaussian plume diffusion model, the smoke concentration follows a normal distribution downwind, and the diffusion coefficient is related to atmospheric stability. When the wind speed is less than 2 meters per second, the smoke mainly diffuses vertically, and the affected area is circular; when the wind speed is greater than 5 meters per second, the smoke spreads rapidly along the wind direction, and the affected area is elliptical. The major axis of the diffusion range is aligned with the wind direction, and the ratio of the major to minor axes is directly proportional to the wind speed. By superimposing the smoke diffusion ranges at multiple times, a spatiotemporal distribution map of smoke coverage is formed.

[0043] For example, the fire point temperature threshold in the threshold segmentation process is dynamically adjusted according to the ambient temperature. In summer, when the ambient temperature may reach 35 degrees Celsius, the fire point temperature threshold is set to the ambient temperature plus 50 degrees Celsius; in winter, when the ambient temperature may be below 0 degrees Celsius, the fire point temperature threshold is set to a fixed value of 45 degrees Celsius. The dilation operation uses a 3×3 structuring element to extend the boundary of the burning area outward by one pixel, filling the internal voids; the erosion operation uses the same structuring element to eliminate isolated noise points at the edges. Edge detection uses the Sobel operator to calculate the gradients in the horizontal and vertical directions respectively. Pixels with gradient magnitudes greater than a preset threshold are marked as edge points, and edge points are connected through 8-neighborhood connectivity analysis to form a complete contour. For example, in a forest fire monitoring scenario, satellite images showed that the temperature in the core area of ​​the fire reached 800 degrees Celsius, the temperature in the surrounding burned area was between 200 and 400 degrees Celsius, and the smoke coverage extended to 15 kilometers downwind. By statistically analyzing the number of pixels in different temperature ranges, it was found that high-temperature zones above 600 degrees Celsius accounted for 15% of the total fire area, medium-temperature zones between 300 and 600 degrees Celsius accounted for 45%, and low-temperature zones below 300 degrees Celsius accounted for 40%. The ratio of the standard deviation to the mean of smoke concentration was 0.8, indicating that the smoke distribution was uneven and that multiple concentration centers existed.

[0044] For example, the temperature gradient is calculated using the central difference method. For each pixel, the temperature difference between it and its eight surrounding pixels is calculated, and the direction with the largest difference is selected as the temperature gradient direction. When the angle between the temperature gradient direction and the wind direction is less than 30 degrees, it indicates that the fire is spreading with the wind and the combustion intensity is high; when the angle is greater than 90 degrees, it indicates that the fire is developing against the wind and the combustion intensity is relatively low. The fire hazard level and spread rate are assessed by statistically analyzing the area ratio of high combustion intensity regions. The comprehensive characteristics of the fire area contain information in multiple dimensions, each reflecting different attributes of the fire. Temperature distribution characteristics are obtained through histogram statistics, dividing the temperature range into 10 intervals and statistically analyzing the pixel frequency of each interval. Smoke concentration characteristics are obtained through spatial autocorrelation analysis, calculating the local Moran's index to assess the degree of smoke accumulation. Combustion intensity characteristics are estimated through energy release rate, combining temperature values ​​and combustion area to calculate the heat release per unit time. These characteristics together constitute a complete description of the fire state.

[0045] As an example of an embodiment of the present invention, the prediction of the fire change trend based on the spatial morphology information, the temperature at each location, and the meteorological data specifically involves: determining the fire intensity level based on the temperature at each location, combined with the temperature threshold and the area threshold; determining the fuel type based on the multispectral remote sensing data; querying the combustion rate of the target fire area from a preset combustion rate database based on the fire intensity level and the fuel type; calculating the first boundary expansion speed of the target fire area based on the wind speed, the relative humidity, and the combustion rate; adjusting the first boundary expansion speed based on the wind direction and terrain slope data to obtain the second boundary expansion speed in different directions; and obtaining the fire change trend by iteratively calculating the boundary position sequence within the prediction time interval based on the current boundary information in the spatial morphology information and the second boundary expansion speed.

[0046] In this embodiment, based on the temperature distribution characteristics of the target fire area (i.e., the pixel area ratio of different temperature ranges and their corresponding median temperature ranges), a preset fire intensity classification standard is compared. Areas where the representative temperature (e.g., the median temperature of the high-temperature range) exceeds a preset high-temperature threshold and the area ratio of that high-temperature range exceeds a preset area threshold are classified as high-intensity fires; areas where both the representative temperature and area ratio are in the middle range are classified as medium-intensity fires; and the rest are classified as low-intensity fires. Simultaneously, using smoke particle diameter distribution data obtained from multispectral remote sensing data inversion, and based on the correspondence between smoke particle diameter and vegetation type (when the particle diameter is less than a preset threshold, it is identified as coniferous forest combustion; when it is greater than a preset threshold, it is identified as broadleaf forest combustion), the fuel type is obtained. For the fire level identifier and fuel type, the baseline combustion rate for the corresponding fuel type is queried from a preset combustion rate database. Combined with real-time wind speed data, the basic speed of fire line advancement is calculated through the product of wind speed and baseline combustion rate. If the relative humidity is lower than a preset dryness threshold, the basic speed is multiplied by a humidity correction factor to obtain the first boundary expansion speed of the fire boundary. Based on the first boundary expansion rate and terrain slope data, the expansion rate in the uphill direction is calculated as the first boundary expansion rate multiplied by the slope enhancement coefficient, and the expansion rate in the downhill direction is calculated as the first boundary expansion rate multiplied by the slope reduction coefficient. Simultaneously, based on the angles between the wind direction and each azimuth angle, the dominant and secondary directions of fire spread are determined, resulting in second boundary expansion rates in eight directions. Based on these second boundary expansion rates, a maximum expansion distance is set along the dominant direction, and a minimum expansion distance is set perpendicular to the dominant direction. By connecting the predicted boundary points in each direction to form a closed curve, the new fire boundary position is iteratively calculated at preset time intervals, yielding a forest fire trend map.

[0047] For example, the fire intensity classification standard is determined based on historical fire data. Specifically, through statistical analysis of forest fire cases over the past decade, temperature thresholds are set into three levels: high-intensity fires have a temperature threshold above 600 degrees Celsius and an area threshold above 100 hectares; medium-intensity fires have a temperature threshold of 300 to 600 degrees Celsius and an area threshold of 10 to 100 hectares; and low-intensity fires have a temperature threshold below 300 degrees Celsius and an area threshold below 10 hectares. Fuel type identification is based on the physical characteristics of smoke particles. The diameter of smoke particles produced by coniferous forest combustion is mainly concentrated in the range of 0.1 to 0.3 micrometers, the diameter of smoke particles produced by broadleaf forest combustion is mainly concentrated in the range of 0.5 to 1.0 micrometers, and mixed forests fall between the two.

[0048] It should be noted that the combustion rate database includes combustion characteristic parameters of different vegetation types under various environmental conditions. The baseline combustion rate for coniferous forests is 0.8 to 1.2 meters per hour, for broadleaf forests it is 0.4 to 0.6 meters per hour, and for shrub forests it is 1.5 to 2.0 meters per hour. These values ​​were obtained through controlled combustion experiments in the field and statistical analysis of actual fire cases. The effect of wind speed on the advance speed of the fire line is non-linear. When the wind speed is less than 3 meters per second, the increment of the fire line advance speed is approximately 0.3 times the wind speed; when the wind speed is in the range of 3 to 10 meters per second, the speed increment is approximately 0.5 times the wind speed; when the wind speed exceeds 10 meters per second, the speed increment tends to saturate, approximately 0.2 times the wind speed.

[0049] For example, the humidity correction factor is determined considering both relative humidity and fuel moisture content. When the relative humidity is below 30%, the humidity correction factor is 1.5, indicating that fire spread is accelerated under dry conditions; when the relative humidity is between 30% and 60%, the correction factor changes linearly between 1.0 and 1.5; when the relative humidity is above 60%, the correction factor is less than 1.0, indicating that fire spread is suppressed. The fuel moisture content is estimated using previous precipitation and evaporation. For each additional day of consecutive dry weather, the fuel moisture content decreases by approximately 2%, corresponding to an increase in combustion rate of approximately 5%.

[0050] For example, terrain slope data is extracted using a digital elevation model with a spatial resolution of 30 meters. The calculation of the slope enhancement and reduction coefficients is based on the flame heat transfer mechanism. When a fire spreads uphill, the flame tilt angle decreases, and the preheating efficiency increases. For every 10-degree increase in slope, the spread rate in the uphill direction increases by approximately 20%, corresponding to a slope enhancement coefficient of 1.2. On the downhill direction, the flame moves further away from the unburned area, reducing preheating efficiency. For every 10-degree increase in slope, the spread rate in the downhill direction decreases by approximately 15%, corresponding to a slope reduction coefficient of 0.85. When the slope exceeds 35 degrees, the coefficient values ​​tend to stabilize and no longer change significantly with increasing slope.

[0051] For example, the calculation process for the differentiated expansion speed in eight directions is as follows: Taking due north as 0 degrees, a directional point is set every 45 degrees clockwise, corresponding to the eight directions: north, northeast, east, southeast, south, southwest, west, and northwest. Based on the angle between the wind direction and each direction, the wind force influence factor for each direction is calculated. When the angle is 0 degrees, the wind force influence factor is 1.0, indicating a downwind direction; when the angle is 90 degrees, the influence factor is 0.5, indicating a crosswind direction; and when the angle is 180 degrees, the influence factor is 0.2, indicating a headwind direction. The expansion speed in each direction is equal to the first boundary expansion speed multiplied by the corresponding wind force influence factor and slope correction coefficient. For example, in a forest fire monitoring case, the fire level was identified as high intensity, the fuel type as coniferous forest, and the baseline burning rate as 1.0 m / s. The real-time wind speed was 8 m / s, the wind direction was northwest, and the relative humidity was 25%. Based on the wind speed influence relationship, the base speed of the fire line advance is the base speed plus a wind speed increment of 4.0 m / s, resulting in 5.0 m / s. Because the humidity was below the dryness threshold, the humidity correction factor was 1.6, and the first boundary expansion rate was 8.0 meters per hour. Specifically, based on topographic data analysis, the southeast direction of the fire was uphill with a slope of 25 degrees and a slope enhancement factor of 1.5; the northwest direction was downhill with a slope of 20 degrees and a slope reduction factor of 0.7. The southeast direction, being both downwind and uphill, became the dominant spread direction, with an expansion rate of 12.0 meters per hour; the northwest direction, being upwind and downhill, had an expansion rate of only 1.1 meters per hour. The rates in the other six directions were between these two, forming an asymmetrical fire expansion pattern. Furthermore, the trend map was constructed using an iterative prediction method. The prediction time interval was set to 30 minutes. Starting from the current fire boundary, the corresponding expansion distance was extended along each of the eight directions, with the distance equal to the expansion rate in that direction multiplied by the time interval. Connecting the eight prediction points formed a new fire boundary, and cubic spline interpolation was used to smooth the boundary curve to avoid sharp turns. After each iteration, the environmental conditions in each direction are reassessed, including changes in terrain, vegetation type, and meteorological conditions, and the expansion speed is dynamically adjusted for the next moment.

[0052] Understandably, by iteratively predicting over 24 consecutive hours, a spatiotemporal evolution sequence of fire spread is formed. The fire boundary at each moment is overlaid and displayed, with varying shades of color representing different times, creating an intuitive visualization of the spread trend. Comparing the prediction results with the actual fire development, the average location error is controlled within 15%, providing a reliable basis for emergency dispatch decisions.

[0053] For example, base station numbers, latitude and longitude coordinates, altitude, and service range radius are read from a preset communication base station location database. Based on the predicted boundary coordinates at each moment in the trend chart, all base stations located within the maximum predicted range of the fire area plus a preset buffer distance are selected to obtain a list of base stations to be evaluated. For each base station in the list of base stations to be evaluated, the Euclidean distance formula is used to calculate the distance from the base station coordinate point to all points on the fire boundary line at each predicted moment. The minimum value is taken as the shortest distance at that moment. If the shortest distance at a certain moment is less than a preset safe distance threshold, the difference between that moment and the current moment is recorded as the risk warning time, thus obtaining the risk warning time value for each base station. Based on the risk warning time value, base stations with risk warning times less than a preset emergency threshold are marked as high-risk, those between the emergency threshold and the warning threshold are marked as medium-risk, and those greater than the warning threshold are marked as low-risk. They are arranged in order of risk level from high to low and risk warning time from short to long to obtain a risk base station sequence.

[0054] For example, the communication base station location database pre-stores detailed information about all base stations within the area. Each base station record includes a unique ID, latitude and longitude coordinates, altitude, service radius, construction time, and equipment model. The database uses a spatial index structure, supporting fast range query operations. The trend map stores the fire boundary coordinates at each time point in time series format, with each time point containing hundreds of boundary points, forming a closed polygon.

[0055] It should be noted that the preset buffer distance is determined based on historical fire case statistics and is typically set to 5 to 10 kilometers. The filtering process first calculates the maximum expansion range of the fire within the predicted time period, and then uses the geometric center of this range as the center and the distance from the farthest boundary point plus the buffer distance as the radius to query all base stations falling within this circular area. This method avoids querying each time point separately, thus improving filtering efficiency.

[0056] For example, the Euclidean distance calculation uses a segmented approach. For each line segment on the fire boundary, it is first determined whether the foot of the perpendicular from the base station to the line segment lies within the line segment. If so, the perpendicular distance is calculated; otherwise, the distances to the two endpoints of the line segment are calculated, and the smaller value is taken. By traversing all boundary line segments, the precise shortest distance from the base station to the fire boundary is obtained. When the fire area has an irregular shape, this method is more accurate than simple point-to-point distance.

[0057] For example, the threshold setting for risk warning time takes into account emergency response capabilities. An emergency threshold of 2 hours indicates that immediate action is required; a warning threshold of 6 hours indicates that advance preparation is needed. A base station in a mountainous area is 8 kilometers from the current fire boundary. Based on the predicted fire spread rate, the fire boundary will expand to 2 kilometers from the base station in 3 hours, below the safe distance threshold of 3 kilometers. Therefore, the risk warning time for this base station is 3 hours, and it is marked as a medium-risk level.

[0058] For example, the prioritization employs a dual standard. First, base stations are grouped by risk level, with high-risk stations ranked first, followed by medium-risk stations, and finally low-risk stations. Within the same risk level, they are ranked in ascending order of risk warning time, with shorter warning times having higher priority. This prioritization method ensures the rational allocation of resources in emergency situations, allowing base stations most in need of support to receive power supplies first.

[0059] As an example of an embodiment of the present invention, the step of performing spatiotemporal joint analysis of the density data and historical power outage data of each risk base station to obtain the outage probability distribution specifically involves: dividing each risk base station into different density intervals based on the density data, and calculating the average power outage frequency of each risk base station within each density interval based on the historical power outage data, thereby calculating the spatial influence coefficient of each density interval, wherein the spatial influence coefficient is used to characterize the influence coefficient of base station density on the power outage probability; performing time series analysis on the historical power outage data to identify the periodic characteristics of power outage events, and determining the time correction factor of each risk base station at the current moment based on the periodic characteristics; spatially correcting the initial outage probability according to the spatial influence coefficient corresponding to the density interval of each risk base station to obtain a first corrected probability, and temporally correcting the initial outage probability according to the time correction factor to obtain a second corrected probability; weighted fusing the first corrected probability and the second corrected probability to obtain the target outage probability of each risk base station, and obtaining the outage probability distribution based on each target outage probability.

[0060] In this embodiment, based on the geographical coordinates of each base station in the risk base station sequence, the kernel density estimation method is used to calculate the number of base stations within a preset radius around each base station. The local base station density value is obtained by dividing the number of base stations by the coverage area. Simultaneously, records of the power outage time, duration, cause of failure, and recovery method of each base station within a preset time period are extracted from the power operation and maintenance database to obtain the base station density distribution and historical power outage event sequence. For the historical power outage event sequence, aggregation statistics are performed according to monthly time windows to calculate the number of power outages and average power outage duration for each base station in different months. Periodic variation patterns are identified by calculating the difference in the number of power outages between adjacent months. At the same time, the base station density distribution is divided into three intervals: high density, medium density, and low density. The average power outage frequency of base stations in each density interval is statistically analyzed to obtain a time feature vector and a density frequency correspondence table. Based on the density-frequency correspondence table, the ratio of the average power outage frequency of each density interval to the overall average power outage frequency of all base stations is calculated as the spatial influence coefficient. The linear correlation is verified by calculating the correlation coefficient between the base station density value and the power outage frequency. If the absolute value of the correlation coefficient is greater than a preset significance threshold, the spatial influence coefficient is adopted; otherwise, the spatial influence coefficient is set to a preset default value. According to the spatial influence coefficient and the risk level of the risky base station, the risk level is converted into an initial outage probability, which is then multiplied by the spatial influence coefficient of the corresponding density interval to obtain the first corrected probability. At the same time, the time feature vector is introduced to correct the initial outage probability in the time dimension to obtain the second corrected probability. The specific method is as follows: extract the periodic change pattern in the time feature vector (for example, the identified seasonal pattern, such as summer being a high-incidence month for power outages), and determine the time window in which the current time is located (such as the current month). If the current time is in a high-incidence period, a time enhancement factor (such as 1.1~1.3) is applied to the initial outage probability; if it is in a low-incidence period, a time reduction factor (such as 0.7~0.9) is applied. The first corrected probability, the second corrected probability, and the normalized value of the base station's historical power outage frequency are weighted and fused together. The weights are determined based on the data's reliability and timeliness (for example, real-time risk data has the highest weight, followed by historical statistics). This results in a more accurate density and time-adjusted outage probability distribution that integrates spatial density, time trend, and historical performance.

[0061] For example, the kernel density estimation method uses a Gaussian kernel function to achieve a continuous expression of the spatial density of base stations. For each base station location, a two-dimensional Gaussian distribution is constructed with that point as the center, and the bandwidth parameter is adaptively adjusted according to the average distance between base stations. When the base station distribution is dense, the bandwidth value is small, approximately 500 to 1000 meters; when the base station distribution is sparse, the bandwidth increases to 2000 to 3000 meters. By superimposing the kernel function values ​​of all base stations, a continuous density field is formed, and the density value represents the expected number of base stations per unit area. The construction of historical power outage event sequences requires cleaning and standardization of the original records. Power outage records in the power operation and maintenance database contain various fault types, such as equipment failure, line failure, and power outages caused by natural disasters. For forest fire scenarios, power outage events related to high temperature, smoke, and flames are extracted. Each record includes the base station number, power outage start time, recovery time, fault cause code, and maintenance method. The power outage duration is obtained by calculating the difference between the start time and the recovery time, and recorded in hours. The identification of periodic variation patterns is achieved through difference operations. The monthly power outage frequency series is first-differenced to calculate the change in the number of power outages between adjacent months. If the difference series exhibits a regular alternation of positive and negative values, and the alternation period is stable within a specific number of months, a seasonal pattern is considered to exist. For example, the number of power outages increases significantly during the hot summer months and decreases during the winter months, forming an annual cyclical fluctuation. By calculating the autocorrelation function, when the correlation coefficient with a 12-month lag exceeds 0.7, the existence of an annual cycle is confirmed.

[0062] For example, the base station density intervals are divided using a quantile method. All base station local density values ​​are sorted from smallest to largest, with the 33rd and 67th quantiles used as dividing points. Base stations with density values ​​below the 33rd quantile are classified into the low-density interval, those between the 33rd and 67th quantiles into the medium-density interval, and those above the 67th quantile into the high-density interval. This division method ensures that each interval contains approximately the same number of base stations, facilitating subsequent statistical analysis. In a practical application in a forest area, the base station density is 0.5 to 2 per square kilometer in the low-density interval, 2 to 5 per square kilometer in the medium-density interval, and 5 to 10 per square kilometer in the high-density interval.

[0063] For example, the calculation of the correlation coefficient involves standardization of two variable sequences. First, the local density value and historical average power outage frequency of each base station are calculated to form paired data points. The mean and standard deviation of the density value sequence and the power outage frequency sequence are calculated respectively, and zero-mean unit variance standardization is performed. Then, the covariance of the two standardized sequences is calculated, which is the Pearson correlation coefficient. The correlation coefficient ranges from -1 to 1, and the closer the absolute value is to 1, the stronger the linear correlation. When the absolute value of the correlation coefficient is greater than 0.5, it is considered to have a significant correlation, and the influence of density on the probability of power outage cannot be ignored. The spatial influence coefficient is determined based on the results of statistical analysis. If the average power outage frequency of the high-density interval is 8 times per year, the medium-density interval is 5 times per year, and the low-density interval is 3 times per year, while the overall average power outage frequency of all base stations is 5.3 times per year, then the influence coefficient of the high-density interval is 8 divided by 5.3, which is approximately 1.5; the influence coefficient of the medium-density interval is 5 divided by 5.3, which is approximately 0.94; and the influence coefficient of the low-density interval is 3 divided by 5.3, which is approximately 0.57. These coefficients reflect the amplification or reduction effect of base station density on the risk of power outages. For example, the conversion from risk level to initial outage probability uses a piecewise mapping method. The initial outage probability corresponding to high risk level is set to 0.7 to 0.9, medium risk level to 0.4 to 0.6, and low risk level to 0.1 to 0.3. The specific values ​​are determined by linear interpolation based on the relative position of the risk warning time within each level range. For example, a medium-risk base station with a risk warning time of 4 hours falls in the middle of the 2-6 hour medium-risk time range, and its initial outage probability is determined to be 0.5.

[0064] Understandably, the weights in the weighted summation were determined considering both the reliability and timeliness of the data sources. Historical power outage frequency, based on long-term statistical data, has high reliability but poor timeliness, and its weight was set to 0.3. Adjusted probability, based on real-time fire situation and density analysis, has good timeliness but contains prediction errors, and its weight was set to 0.7. Through weighted fusion, the normalized value of a base station's historical power outage frequency is 0.2, the density-adjusted probability is 0.6, and the final outage probability is 0.2 multiplied by 0.3 plus 0.6 multiplied by 0.7 equals 0.48. This fusion method comprehensively considers historical experience and current risks, improving the accuracy of probability assessment.

[0065] For example, the entire probability distribution adjustment process realizes the transformation from static risk assessment to dynamic probability prediction. By introducing the spatial dimension factor of base station density and the temporal dimension information of historical power outage data, the assessment of outage probability becomes more comprehensive and accurate.

[0066] As an example of an embodiment of the present invention, the step of matching and analyzing the demand information of each risk base station with the inventory information of each power material warehouse to obtain the resource coverage rate specifically involves: calculating the initial transportation distance based on the geographical coordinates of each power material warehouse and each risk base station; acquiring terrain and road data, correcting the initial transportation distance to obtain the actual transportation length, and calculating the actual transportation time from each power material warehouse to each risk base station in combination with the transportation speed; comparing the actual transportation time with the remaining power duration of each risk base station, and identifying risk base stations whose actual transportation time is less than the remaining power duration as coverable risk base stations; and calculating the resource coverage rate based on the number of coverable risk base stations and the total number of all risk base stations.

[0067] In this embodiment, the available quantity and location of battery modules are obtained from the power supply warehouse. The resource coverage rate is determined by comparing the risk base station list with the associated power supply warehouse location and calculating the transportation time based on the risk base station list and the power supply warehouse location.

[0068] For example, the number of battery modules and geographical coordinates of each reserve point are obtained from the power supply warehouse, the latitude and longitude information of the risk base station is extracted, and the spherical distance from each reserve point to each base station is calculated using the Haversine formula, where the spherical distance d=2r×arcsin(√(sin²(Δφ / 2)+cos(φ1)×cos(φ2)×sin²(Δλ / 2))), φ is the latitude, φ1 is the latitude of location 1 (e.g., the latitude of the power supply warehouse), φ2 is the latitude of location 2 (e.g., the latitude of the risk base station), Δφ is the latitude difference, Δφ=φ2-φ1, λ is the longitude, and r is the Earth's radius, thus obtaining the initial distance matrix between the power supply warehouse and the risk base station. Based on the initial distance matrix, mountain terrain undulation data and road tortuosity parameters are obtained. The actual transportation path is obtained by multiplying the initial distance by a path correction coefficient, where the path correction coefficient is obtained by weighted summation of terrain undulation and road tortuosity. The actual transportation time is calculated by dividing the actual transportation path by the average speed. If the actual transportation time is less than the remaining power duration of the base station, the base station is marked as having effective coverage. The proportion of base stations with effective coverage to the total number of risky base stations is counted to obtain the resource coverage rate. For example, when obtaining battery module information from the power material warehouse, the real-time inventory quantity, rated capacity specifications, and GPS coordinates of the warehouse for each reserve point are read. Specifically, for the base station power guarantee needs in forest fire scenarios, the location coordinates and historical power outage data of all communication base stations within a preset distance interval of the target forest area are extracted simultaneously. Based on these basic data, a spatial analysis foundation for resource scheduling is constructed. It should be noted that the Haversine formula fully considers the influence of the Earth's curvature when calculating the spherical distance between the reserve point and the base station. The latitude difference Δφ in the formula is obtained by subtracting the latitudes of two points, and the longitude difference Δλ is calculated similarly. The Earth's radius r is taken as 6371 kilometers. In practical applications, when a reserve point is located at 30.5 degrees North latitude and 104.2 degrees East longitude, while the target base station is located at 30.8 degrees North latitude and 104.6 degrees East longitude, the spherical distance calculated by this formula is approximately 47 kilometers. This initial distance provides a benchmark value for subsequent path correction. Preferably, when acquiring mountain terrain relief data, the system calls the digital elevation model to extract the elevation change information along the path from the reserve point to the base station. The relief coefficient is obtained by calculating the ratio of the cumulative elevation difference between adjacent sampling points along the path to the horizontal distance. The road tortuosity parameter is obtained by analyzing the ratio of the actual road trajectory to the straight-line distance. In mountainous environments, this ratio is usually between 1.3 and 1.8. The path correction coefficient is obtained by multiplying the relief coefficient by 0.4 and adding the tortuosity parameter by 0.6, reflecting the comprehensive influence of terrain on the actual transportation distance. When determining whether a base station can be effectively covered, the system first obtains the backup battery capacity and current remaining power of each base station, and calculates the remaining power duration based on the average power consumption of the base station.For example, a base station equipped with a 48V / 100Ah backup battery pack, currently with 80% remaining power, and a base station power consumption of 500W, has approximately 7.7 hours of remaining power. If the actual transport time for an emergency vehicle from the nearest reserve point is 8.5 hours, then the base station is marked as having ineffective coverage. This judgment mechanism ensures the timeliness of resource scheduling. When the system detects that the resource coverage rate is below the preset 65% threshold, it immediately triggers an orange-level resource shortage alarm.

[0069] In one possible implementation, the available quantity, storage location coordinates, and battery capacity data of battery modules are extracted from the power material warehouse. At the same time, the geographical location, power demand, and priority information of the risky base stations are obtained. Spatial analysis is performed through a geographic information system to match the base station locations with the lithium battery inventory locations. The straight-line distance and transportation feasibility from each base station to the nearest inventory point are analyzed to obtain the name and coordinates of the target inventory point corresponding to each risky base station, the straight-line distance between the base station and the inventory point, the estimated transportation time, and the number of battery modules that can be allocated.

[0070] For example, the battery module list for each warehouse is read from the power supply inventory, including the unique code, rated capacity, and current charging status of each module. Simultaneously, the latitude and longitude coordinates, backup power requirements, and risk priority levels of risky base stations are extracted from the base station management database. These coordinates are then unified to the same geographic reference system to obtain an inventory resource table and a base station demand table. Based on these tables, the straight-line distance from each base station to each inventory point is calculated using the spherical distance formula. Inventory points with distances less than a preset transportation radius are selected as candidate supply sources. For these candidate supply sources, the transportation network database is queried to obtain actual road mileage and road grade information, and the feasibility of the transportation route is determined, resulting in a set of reachable inventory points and transportation route parameters for each base station. Based on these reachable inventory point sets and transportation route parameters, the estimated transportation time is calculated by dividing the road mileage by the standard driving speed for the corresponding road grade. If multiple reachable inventory points exist, the one with the shortest transportation time is selected as the target inventory point. Simultaneously, the required total battery capacity is calculated based on the base station power requirements. The minimum number of modules required to meet the capacity requirements is allocated from the available modules at the target inventory point, resulting in a matching table between base stations and inventory points. For the matching relationship table, the target inventory point name, geographical coordinates, straight-line distance between the two points, actual transportation mileage, estimated arrival time, and allocated battery module number and quantity are summarized for each base station. The execution priority of each allocation record is marked to obtain comprehensive configuration information including spatial matching results and transportation plan.

[0071] For example, the power supply warehouse adopts a distributed database architecture, with data from each regional warehouse synchronized to the central server in real time. Battery modules are identified according to a unified coding rule, with the code including production batch, capacity level, manufacturing date, and warehouse location information. Each module is equipped with an electronic tag that records the current voltage, internal resistance, and remaining capacity percentage, and the charging status is divided into four levels: fully charged, good, low, and faulty. The base station management database stores the static attributes and dynamic status of all communication base stations in the area. Static attributes include base station number, operator identifier, construction time, tower height, and equipment model, while dynamic status includes current load, backup power duration, and power supply status.

[0072] It should be noted that the geographic coordinate system transformation involves the conversion of multiple coordinate systems. Commonly used coordinate systems in China include WGS84, GCJ02, and BD09, while different data sources may use different coordinate systems. The transformation process first identifies the type of the original coordinate system, and then transforms it to a unified target coordinate system using a seven-parameter or four-parameter transformation method. The seven parameters include three translation parameters, three rotation parameters, and one scale parameter, suitable for large-scale, high-precision transformations. The coordinate error after transformation is controlled within the meter range, meeting the positioning accuracy requirements for emergency dispatch. The spherical distance calculation uses the Haversine formula, which takes into account the influence of the Earth's curvature. For the latitude and longitude coordinates of two points, the angle is first converted to radians, then the sine values ​​of the latitude and longitude differences are calculated, the central angle is obtained through the arcsine function, and finally multiplied by the Earth's average radius of 6371 kilometers to obtain the spherical distance. This calculation method has an error within 0.5%, providing sufficient accuracy for distance calculations within a range of several hundred kilometers. The transportation network database contains road node and road segment information. Road node records the location coordinates of intersections, toll stations, and service areas. Road segment information includes road grade, design speed, current road conditions, and traffic restrictions. Road grades are divided into five levels: expressway, national highway, provincial highway, county road, and township road, with corresponding standard driving speeds of 100, 80, 60, 40, and 30 kilometers per hour, respectively. Trafficability assessment considers multiple factors, including whether the road is closed due to construction, whether height restrictions affect transport vehicle passage, and whether seasonal traffic restrictions apply. When a road segment is impassable, an alternative route is found through the road network topology.

[0073] For example, the transportation time is calculated using a segmented cumulative method. The transportation route is divided into multiple segments according to the road class change points, and the travel time for each segment is calculated based on its length and the corresponding class's speed. Considering the uncertainties in actual transportation, a 20% time margin is added to the calculated result. For complex routes that cross multiple road classes, waiting times at class transition points, such as highway toll station travel time and urban traffic light waiting times, also need to be considered.

[0074] For example, battery module allocation follows the principles of minimum quantity and capacity matching. Base station power requirements are calculated based on equipment power and backup power duration, expressed in kilowatt-hours (kWh). The inventory of battery modules comes in various capacity specifications; priority is given to large-capacity batteries where a single module can meet the requirements, avoiding the connection complexity caused by using multiple small-capacity batteries. When multiple modules are required, modules of the same model and capacity are selected to ensure current balance when used in parallel. The allocation process also considers the battery's lifespan and cycle life, prioritizing modules in good performance condition. For instance, a base station in a mountainous area is 45 kilometers in a straight line from the nearest county warehouse, with an actual road distance of 68 kilometers, including 15 kilometers of provincial highway and 53 kilometers of county road. At standard driving speeds, the provincial highway section takes 15 minutes, and the county road section takes 80 minutes. Adding a 20% time margin, the estimated transportation time is 114 minutes. This base station requires 3 kilowatts of power and 8 hours of backup power, totaling 24 kWh. The warehouse has eight 10 kWh modules and three 20 kWh modules. Allocating two 20 kWh modules will meet the demand, with the remaining capacity used as safety redundancy. The output format of the comprehensive configuration information uses a structured data table, including fields such as base station number, target warehouse, geographical coordinates, straight-line distance, road mileage, estimated time, allocated module number, number of modules, and execution priority. The execution priority is determined based on a combination of the base station's risk level and warning time; high-risk base stations with short warning times receive the highest priority.

[0075] As an example of an embodiment of the present invention, the step of determining the target base station with supply-demand imbalance within the prediction time interval from the plurality of risk base stations based on the fire change trend, the interruption probability distribution, and the inventory information specifically involves: calculating the resource difference based on the total demand for power materials of all risk base stations and the total inventory in the inventory information; simultaneously, calculating the degree of spatial mismatch between the inventory distribution of each power material warehouse and the demand distribution of all risk base stations to obtain the inventory distribution deviation; if the resource difference is greater than a preset difference, or the inventory distribution deviation is greater than a preset inventory deviation, then it is determined that there is a supply-demand contradiction; in response to the supply-demand contradiction, a prediction time interval is determined based on the interruption probability distribution and the resource coverage rate, and the base stations with an interruption probability greater than a preset interruption threshold and not belonging to the coverable risk base stations within the prediction time interval are determined as the target base stations.

[0076] In this embodiment, if the resource coverage value is lower than a preset coverage threshold or if more than a preset number of base stations have actual transport times exceeding a preset time threshold, a resource shortage alarm signal is triggered. The total number of battery modules required by each risky base station is aggregated. The required number of battery modules for each base station is calculated based on a 72-hour operating standard. This is compared with the existing inventory at each reserve point to calculate the resource gap. Based on the resource gap, coverage value, and the distribution of base stations with ineffective coverage, an assessment result is generated, including the required number of battery modules to be replenished and suggested locations for additional reserve points. According to the assessment result, the emergency battery module demand for each risky base station is extracted. The total number of battery modules required for each base station to maintain a preset operating time is aggregated, and the difference is calculated with the existing inventory at each reserve point to obtain the resource difference. Simultaneously, the deviation between the weighted distance from each reserve point to the demanding base station and the actual inventory distribution is calculated. The inventory distribution deviation is obtained by accumulating the inventory distribution deviations at each reserve point. If the resource difference exceeds a preset difference or the inventory distribution deviation exceeds a preset inventory deviation, a supply-demand imbalance is determined. Historical power outage data for each base station is obtained, and the frequency and duration of power outages for each base station in a forest fire scenario are statistically analyzed. The power outage probability of each base station is calculated based on the relationship between the fire trend and the base station location. The number of base stations with a probability greater than a preset probability threshold is multiplied by the aforementioned resource coverage rate to determine the predicted time interval for resource allocation in advance. Within the predicted time interval, base stations with an outage probability greater than a preset threshold (e.g., 0.7) are selected. Next, from these high-risk base stations, sites that can cover the risky base stations are excluded, and the remaining base stations are finally identified as target base stations with supply-demand imbalance.

[0077] For example, the total demand for power supplies is determined based on the communication equipment power parameters of each at-risk base station, the current remaining capacity of the backup battery, and the historical average duration of power outages. Specifically, for a site equipped with 5G base station equipment, its full-load power is typically 3.5 kW. Considering that power outages caused by forest fires may last from 48 to 96 hours, the required number of battery modules is calculated based on a standard of maintaining base station operation for 72 hours. Each standard 48V / 200Ah battery module can provide approximately 9.6 kWh of power, and a single 5G base station requires approximately 26 standard modules to maintain operation for 72 hours.

[0078] It should be noted that the calculation of inventory distribution deviation reflects the degree of unevenness in the spatial distribution of resources. The weighted distance is obtained by multiplying the actual transportation distance from each reserve point to the demand base station by the demand urgency weight of that base station. The demand urgency weight is determined comprehensively based on factors such as the number of users covered by the base station, whether it is an emergency communication node, and the speed of fire spread in the surrounding area. For example, if a reserve point is 30 km, 45 km, and 60 km away from three demand base stations, with corresponding demand urgency weights of 0.8, 0.6, and 0.5, then the weighted average distance of this reserve point is 30×0.8 + 45×0.6 + 60×0.5 divided by the sum of the weights, 1.9, which is approximately 42 km. The inventory distribution deviation is obtained by summing the absolute values ​​of the difference between the actual inventory ratio and the ideal inventory ratio of each reserve point. The ideal inventory ratio is determined based on the number of base stations covered by the reserve point and the average demand. Preferably, when determining the supply and demand imbalance, the gap threshold set by the system is usually 15% of the total demand, and the deviation threshold is 0.3.

[0079] For example, by retrieving power outage records from each base station over the past three years, the system statistically analyzes the frequency of power outages, the duration of each outage, and the distance from the fire line at the time of the outage in a forest fire scenario. Based on this historical data, the system establishes a mapping relationship between the probability of power outages and the distance to the fire line. When the fire line is within 10 kilometers of the base station, the probability of power outages typically exceeds 0.8; when the distance is between 10 and 20 kilometers, the probability is between 0.4 and 0.8; and when the distance exceeds 20 kilometers, the probability drops below 0.4. In one possible implementation, the determination of the hedging forecast time range fully considers the dynamic characteristics of fire spread and the timeliness requirements of resource allocation. The system obtains 72-hour wind speed and direction forecast data from meteorological departments and, combined with factors such as terrain slope, vegetation type, and water content, uses a forest fire spread simulation algorithm to predict the location of the fire line at different time points. Based on the predicted changes in the fire line location, the probability of power outages for each base station in different future time windows is dynamically assessed. For example, if a base station is currently 25 kilometers from the fire line, and the fire line is predicted to advance to 15 kilometers from the base station in 12 hours and to 8 kilometers in 24 hours, then the probability of the base station losing power increases from the current 0.3 to 0.6 after 12 hours and 0.9 after 24 hours. The number of base stations with a power outage probability exceeding a preset threshold of 0.7 is multiplied by the resource coverage rate. If the product is greater than the existing emergency response capacity, the corresponding time point is determined as the starting point of the prediction time interval, thus obtaining the prediction time interval.

[0080] As an example of an embodiment of the present invention, the resource scheduling of each target base station to obtain a scheduling scheme specifically involves: constructing an optimization function with the objectives of minimizing the total transportation time and maximizing the number of risky base stations that can be covered; using the demand of each target base station as the demand constraint and the inventory of each power material warehouse as the supply constraint, solving the optimization function to obtain the scheduling scheme.

[0081] In this embodiment, based on the urgency of resource demand, a scheduling optimization function is constructed with the objectives of minimizing the total transportation time and maximizing the number of base stations covered. The number of battery modules allocated from each reserve point to each base station is set as the variable to be solved. Under the conditions of satisfying the minimum demand constraints of each base station and the existing inventory constraints of each reserve point, the specific allocation quantity from each reserve point to each base station is obtained through iterative solution, generating a scheduling scheme that includes the allocation quantity, allocation order and actual transportation time.

[0082] For example, the construction process of the scheduling optimization function comprehensively considers multiple constraints and optimization objectives. The objective of minimizing the total transportation time is achieved by calculating the sum of the transportation times of all dispatch paths. The transportation time of each path is equal to the path length divided by the average speed of emergency vehicles on mountain roads, which is 35 km / h. The objective of maximizing the number of base stations covered is to count the total number of base stations that can deliver power in a timely manner and meet the minimum power requirements under resource constraints. The constraints include: the number of battery modules received by each base station is not less than the minimum requirement for maintaining 24-hour operation; the total number of modules dispatched from each reserve point does not exceed 90% of its existing inventory, with 10% reserved as an emergency reserve; priority is given to important base stations covering a population of more than 5,000 people. Through an iterative solution process, the system gradually adjusts the dispatch quantity from each reserve point to each base station until a solution that satisfies all constraints and achieves a relatively optimal objective function value is found. Specifically, the scheduling configuration scheme includes a detailed resource dispatch matrix, clearly indicating the specific number of battery modules dispatched from each reserve point to each base station, the estimated departure time, the selected transportation route, and the estimated arrival time. The plan also includes a priority ranking of emergency resource allocation to ensure that, in situations of limited capacity, resources are prioritized for base stations with high power outage risk and high communication support needs. Understandably, when the assessment results show that resources are sufficient and reasonably distributed, the system automatically records a snapshot of the current inventory status, including detailed information such as battery module model, quantity, charging status, and last maintenance time for each reserve point. The resource sufficiency report, in addition to confirming that current resources meet demand, also includes resource redundancy analysis, indicating which areas have relatively abundant reserves and which areas, while currently sufficient, have limited safety margins, providing a reference for subsequent resource allocation optimization. Furthermore, throughout the entire process of performing resource hedging demand assessment, changes in the fire situation and updates to the base station operating status are monitored in real time. Once a significant change in key parameters is detected, such as a sudden change in the fire spread direction or the addition of a new power outage base station, a reassessment process is immediately triggered, dynamically adjusting the scheduling configuration plan to ensure the timeliness and effectiveness of the emergency response.

[0083] Specifically, capacity demand calculations consider both base station equipment power and backup power duration requirements. The power of communication base station equipment varies from 2 kW to 10 kW depending on coverage area and number of users. Backup power duration requirements are determined based on the duration of the fire and expected grid recovery, generally set at 8 to 12 hours. Capacity demand equals power multiplied by backup power duration, plus a 20% safety margin. For example, a base station with 3 kW equipment requires 10 hours of backup power; adding the safety margin, the total demand is 36 kWh. Available battery modules are available in 10 kWh and 20 kWh specifications, requiring either two 20 kWh modules or four 10 kWh modules. For instance, supply status is determined by comprehensively considering inventory adequacy, timely delivery, and technical compatibility. Inventory adequacy checks whether the remaining battery capacity meets the base station's needs; timely delivery verifies whether delivery was completed before the power outage; and technical compatibility confirms that the battery module's voltage and interface match the base station equipment. When all three conditions are met, the system is marked as available; if any condition is not met, it is marked as difficult to supply. In practice, about 70% of base stations can be fully supplied, 20% face some difficulties, and 10% cannot be supplied in a timely manner.

[0084] Understandably, batch planning requires balancing multiple constraints. Vehicle loading capacity limits the number of batteries that can be transported per batch, mountainous road conditions restrict vehicle type and speed, and the number of delivery personnel limits the number of tasks that can be performed simultaneously. Batch division is based on geographical area and time urgency; base stations in the same area are grouped into the same batch to minimize round-trip time. Each batch contains 3 to 5 base stations, and the delivery route is optimized using a shortest path algorithm to ensure all delivery tasks are completed within the time window.

[0085] For example, to provide the target base station with battery models and quantities that match the available capacity, the delivery priority is determined by sorting them according to the probability of interruption. A batch plan is formulated based on the delivery route and time constraints. The departure time, loading list, target base station and estimated completion time of each batch are recorded to obtain the scheduling scheme.

[0086] Specifically, the probability of power outages is categorized using an equally spaced division method. The range of probability values ​​from 0 to 1 is divided into three levels: base stations with a probability value greater than 0.7 are classified as high-probability, those between 0.3 and 0.7 as medium-probability, and those less than 0.3 as low-probability. This categorization method directly reflects the differences in the degree of risk faced by base stations. Each level corresponds to a different resource allocation strategy: high-probability base stations receive priority protection, medium-probability base stations receive secondary protection, and low-probability base stations are used as supplementary targets. The probability values ​​are derived from a comprehensive evaluation of fire spread model predictions, historical power outage statistics, and real-time monitoring data.

[0087] For example, the correspondence between base stations and inventory points follows the principles of risk priority and shortest distance. High-probability base stations are first associated with the nearest inventory point, even if that inventory point is closer to medium- or low-probability base stations; high-risk base stations are prioritized for supply. This correspondence is not a fixed one-to-one relationship, but a dynamic many-to-many mapping. One inventory point can serve multiple base stations, and a base station can receive supplies from multiple inventory points. During the correspondence establishment process, inventory capacity limitations must also be considered to prevent over-allocation of any inventory point, which could lead to resource depletion.

[0088] For example, the predicted power outage time is based on the fire spread rate and the location of the base station. The fire boundary is expanding northeastward at a speed of 2 kilometers per hour. Base station A is 6 kilometers from the current fire boundary, and the fire is expected to reach its location in 3 hours. Considering the high temperatures and smoke effects as the fire approaches, the actual power outage time will be 30 minutes to 1 hour earlier; therefore, the predicted power outage time for base station A is 2 to 2.5 hours later. The resource arrival time is calculated based on the delivery route and transportation speed. Starting from the nearest storage point, it takes 1.5 hours to reach base station A via a 45-kilometer mountain road. The time difference of 0.5 to 1 hour indicates that resources can be delivered before the power outage.

[0089] As an example of an embodiment of the present invention, resource allocation details, delivery sequence, and gap information are extracted according to the scheduling scheme. The data is organized in a format including base station number, risk level, allocated battery quantity, and actual transportation time to form a structured document, which is then transmitted over the network to the emergency dispatch center server to obtain an early warning report containing complete scheduling information. Based on the receipt and confirmation of the early warning report, the status flags of allocated battery modules in the inventory database are updated, changing the available status to the occupied status and recording the target base station number. The power supply guarantee time nodes and execution status of all base stations are summarized to form a final hedging forecast result to guide the power restoration work of communication base stations.

[0090] For example, the data organization of the early warning report adopts a standardized format. The report includes a base station identification section, a resource allocation section, and a timing schedule section. The base station identification section records the base station number, geographical coordinates, and risk level; the resource allocation section lists the allocated battery model, quantity, and source inventory point; the timing schedule section indicates the expected departure time, transportation route, and arrival time. The data is organized in XML or JSON format to facilitate parsing and processing by the emergency dispatch center's information system.

[0091] It should be noted that network transmission uses the TCP / IP protocol to ensure reliable data delivery. Report files are compressed before sending to reduce the amount of data transmitted. A timeout retransmission mechanism is set during transmission; if no acknowledgment signal is received within a preset time, the report is automatically retransmitted. After receiving the report, the emergency dispatch center server returns an acknowledgment code to indicate successful reception.

[0092] For example, inventory status updates employ a transaction processing mechanism. The battery module's status is changed from available to occupied, while simultaneously recording the allocation time, target base station, and estimated return time. The update operation is atomic; either all operations succeed or all are rolled back, avoiding data inconsistency. The final hedging forecast result includes the protection status of all base stations, resource allocation, and execution progress, forming a complete power restoration guidance plan.

[0093] like Figure 2 As shown, based on the above-described method embodiments, another embodiment of the present invention also provides a power material dispatching system 200, including: an interruption identification module 201, a supply and demand judgment module 202, and a dispatching module 203; The interruption identification module 201 is used to identify base stations with power interruption risk based on the fire change trend and communication base station location data of the target forest area, obtain multiple risk base stations, perform spatial density analysis on all the risk base stations to obtain density data, and perform spatiotemporal joint analysis by combining the density data and the historical power outage data of each risk base station to obtain the interruption probability distribution. The supply and demand judgment module 202 is used to match and analyze the demand information of each risk base station with the inventory information of each power material warehouse to obtain the resource coverage rate. If the resource coverage rate is lower than a preset threshold, then based on the fire change trend, the interruption probability distribution and the inventory information, the target base station with supply and demand imbalance in the predicted time interval is determined from the multiple risk base stations. The scheduling module 203 is used to perform resource scheduling on each target base station to obtain a scheduling plan, so as to pre-schedule power materials to the target base station within the predicted time interval according to the scheduling plan.

[0094] It is understood that the above system item embodiments correspond to the method item embodiments of the present invention, and can implement the power material dispatching method provided by any of the above method item embodiments of the present invention.

[0095] It should be noted that the system embodiments described above are merely illustrative, and some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the system embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0096] For ease of description and brevity, the system embodiments of the present invention include all the implementation methods described in the above-described power material dispatching method embodiments, and will not be repeated here.

[0097] Based on the above embodiments of the power material dispatching method, another embodiment of the present invention provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the power material dispatching method of any embodiment of the present invention.

[0098] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the terminal device.

[0099] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0100] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0101] Based on the above-described method embodiments, another embodiment of the present invention provides a computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to execute the power material dispatching method described in any of the above-described method embodiments of the present invention.

[0102] The modules / units integrated in the device / terminal equipment, if implemented as software functional units and sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0103] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for dispatching power resources, characterized in that, include: Based on the obtained fire change trend and communication base station location data of the target forest area, base stations with power outage risk are identified, and multiple risk base stations are obtained. Spatial density analysis is performed on all the risk base stations to obtain density data. Combined with the density data and the historical power outage data of each risk base station, a spatiotemporal joint analysis is performed to obtain the outage probability distribution. The demand information of each risk base station is matched and analyzed with the inventory information of each power material warehouse to obtain the resource coverage rate. If the resource coverage rate is lower than the preset threshold, the target base station with supply and demand imbalance in the predicted time interval is determined from the multiple risk base stations based on the fire change trend, the interruption probability distribution and the inventory information. Resource scheduling is performed on each target base station to obtain a scheduling plan, and power supplies are pre-scheduled to the target base stations within the predicted time interval according to the scheduling plan.

2. The power material dispatching method as described in claim 1, characterized in that, The process involves identifying base stations at risk of power outages based on the acquired fire trend data and communication base station location data for the target forest area. Acquire multispectral remote sensing image data of the target forest area, as well as meteorological data that is spatiotemporally synchronized with the multispectral remote sensing image data. The multispectral remote sensing image data includes thermal infrared band data, visible light band data, and near-infrared band data, and the meteorological data includes wind speed, wind direction, relative humidity, and ambient temperature. The temperature at each location in the target forest area is obtained by inversion based on thermal infrared band data, and the normalized differential vegetation index at each location in the target forest area is calculated based on the visible light band data and the near-infrared band data. The initial fire point is identified by comparing the temperature threshold with the temperature at each location, and by comparing the vegetation index threshold with the normalized difference vegetation index at each location. The initial fire state map is then generated by combining the vegetation index threshold with the normalized difference vegetation index at each location. The temperature threshold is dynamically determined based on the ambient temperature. Image segmentation and morphological processing are performed on the initial state map of the fire to obtain the spatial morphological information of the target fire area. Based on the spatial morphological information, the temperature at each location and the meteorological data, the fire change trend is predicted. The locations of all communication base stations are spatially compared with the fire change trend. If the distance between the communication base station and the fire boundary is less than a preset safety threshold within the predicted time interval, the communication base station is identified as the risk base station.

3. The power material dispatching method as described in claim 2, characterized in that, The fire change trend is predicted based on the spatial morphology information, the temperature at each location, and the meteorological data, specifically as follows: Based on the temperature at each location, combined with the temperature threshold and area threshold, the fire intensity level is determined, and the fuel type is determined based on the multispectral remote sensing data. Based on the fire intensity level and the fuel type, the combustion rate of the target fire area is obtained by querying a preset combustion rate database. Based on the wind speed, relative humidity, and combustion rate, the first boundary expansion velocity of the target fire area is calculated, and the first boundary expansion velocity is adjusted based on the wind direction and terrain slope data to obtain the second boundary expansion velocity in different directions. Based on the current boundary information in the spatial morphology information, the fire change trend is obtained by iteratively calculating the boundary position sequence within the prediction time interval according to the second boundary expansion speed.

4. The power material dispatching method as described in claim 1, characterized in that, The spatiotemporal joint analysis, combining the density data and historical power outage data of each risky base station, yields the outage probability distribution, specifically: Based on the density data, each risky base station is divided into different density intervals, and based on the historical power outage data, the average power outage frequency of each risky base station in each density interval is statistically calculated, and the spatial influence coefficient of each density interval is calculated. The spatial influence coefficient is used to characterize the influence coefficient of base station density on the power outage probability. Time series analysis is performed on the historical power outage data to identify the periodic characteristics of power outage events, and the time correction factor for each risky base station at the current moment is determined based on the periodic characteristics. Based on the spatial influence coefficient corresponding to the density interval of each risk base station, the initial interruption probability is spatially corrected to obtain a first corrected probability, and the initial interruption probability is temporally corrected based on the time correction factor to obtain a second corrected probability. The first corrected probability and the second corrected probability are weighted and fused to obtain the target interruption probability of each risk base station, and the interruption probability distribution is obtained based on each target interruption probability.

5. The power material dispatching method as described in claim 1, characterized in that, The process involves matching and analyzing the demand information of each risk base station with the inventory information of each power supply depot to obtain the resource coverage rate. Specifically: The initial transportation distance was calculated based on the geographical coordinates of each power material warehouse and each risk base station. The terrain and road data are acquired, the initial transportation distance is corrected to obtain the actual transportation length, and the actual transportation time from each power material warehouse to each risk base station is calculated in combination with the transportation speed. The actual transportation time is compared with the remaining power duration of each risk base station, and the risk base stations whose actual transportation time is less than the remaining power duration are identified as coverable risk base stations. The resource coverage rate is calculated based on the number of coverable risk base stations and the total number of all risk base stations.

6. The power material dispatching method as described in claim 5, characterized in that, Based on the fire change trend, the interruption probability distribution, and the inventory information, the target base station with a supply-demand imbalance within the predicted time interval is identified from the multiple risk base stations. Specifically: Based on the total demand for power materials of all risk base stations and the total inventory in the inventory information, the resource difference is calculated. At the same time, the degree of spatial mismatch between the inventory distribution of each power material warehouse and the demand distribution of all risk base stations is calculated to obtain the inventory distribution deviation. If the resource difference is greater than the preset difference, or the inventory distribution deviation is greater than the preset inventory deviation, then it is determined that there is a supply and demand contradiction. In response to the supply and demand imbalance, based on the interruption probability distribution and the resource coverage, a prediction time interval is determined, and base stations whose interruption probability is greater than a preset interruption threshold and which are not covered risk base stations within the prediction time interval are identified as the target base stations.

7. The power material dispatching method as described in claim 1, characterized in that, The resource scheduling for each target base station, resulting in a scheduling scheme, is as follows: Construct an optimization function with the objectives of minimizing total transport time and maximizing the number of coverable risky base stations; The scheduling scheme is obtained by solving the optimization function, using the demand of each target base station as the demand constraint and the inventory of each power material warehouse as the supply constraint.

8. A power material dispatching system, characterized in that, include: Interruption identification module, supply and demand judgment module, and scheduling module; The interruption identification module is used to identify base stations with power outage risk based on the fire change trend and communication base station location data of the target forest area, obtain multiple risk base stations, perform spatial density analysis on all the risk base stations to obtain density data, and perform spatiotemporal joint analysis by combining the density data and the historical power outage data of each risk base station to obtain the interruption probability distribution. The supply and demand judgment module is used to match and analyze the demand information of each risk base station with the inventory information of each power material warehouse to obtain the resource coverage rate. If the resource coverage rate is lower than a preset threshold, the target base station with supply and demand imbalance in the predicted time interval is determined from the multiple risk base stations based on the fire change trend, the interruption probability distribution and the inventory information. The scheduling module is used to perform resource scheduling on each target base station to obtain a scheduling plan, so as to pre-schedule power supplies to the target base station within the predicted time interval according to the scheduling plan.

9. A terminal device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements the power material dispatching method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, include: A stored computer program, wherein, when the computer program is executed, it controls the device containing the computer-readable storage medium to perform the power material dispatching method as described in any one of claims 1-7.

Images