An early warning method for the supply of power materials in temporary construction warehouses

Abstract

This application provides an early warning method for the supply of power materials in temporary construction warehouses, comprising: establishing a three-dimensional electromagnetic field distribution model based on the relative positional relationship between temporary construction warehouses and high-voltage transmission lines on highways, and obtaining the theoretical distribution of the electromagnetic environment of each warehouse; within the temporary warehouses, deploying electromagnetic sensors for materials whose electronic components are highly sensitive to electromagnetic interference, and deploying large dynamic range electromagnetic sensors for materials whose metallic materials are susceptible to strong electromagnetic fields; transmitting the monitoring data collected by various electromagnetic sensors in the warehouses back to the power grid dispatch center in real time through a low-power wide-area Internet of Things, and correlating it with the structural parameters of the warehouse materials to construct a database; mining historical data on power grid load and electromagnetic environment, extracting the correlation between electromagnetic interference characteristics and the operating conditions of transmission lines, and establishing a collaborative prediction model of the electromagnetic safety status of multiple warehouses.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to an early warning method for the supply of power materials in temporary construction warehouses. Background Technology

[0002] The management of power supply materials in temporary warehouses faces unprecedented challenges under the unique environment of highway construction. Given the linear nature of the construction route, these warehouses are often scattered across a wide geographical area and are adjacent to high-voltage transmission lines, placing them within a strong electromagnetic field. This electromagnetic environment not only threatens the storage of power materials but also increases management complexity due to the significant differences in the sensitivity of materials to electromagnetic interference. On the one hand, some power materials may experience performance degradation due to electromagnetic interference; on the other hand, some materials require specific electromagnetic environments to maintain their optimal condition. Therefore, in material management, the impact of the electromagnetic environment must be accurately considered to ensure the integrity of materials. Meanwhile, considering the wide distribution of warehouses and the dynamic changes in the construction environment, traditional manual inspections and static monitoring methods are clearly insufficient. How to construct a dynamic, multi-dimensional early warning system that can monitor the electromagnetic environment in real time and flexibly adjust early warning strategies according to material characteristics and construction progress, under limited resources, has become a research focus in the field of power material management. This system needs to be highly adaptable and intelligent, providing a solid guarantee for material safety in the ever-changing construction site. In conclusion, developing a dynamic, multi-dimensional early warning system for power materials suitable for highway construction environments is a pressing technical problem that urgently needs to be solved, and innovative solutions are urgently needed to fill the gaps in existing technologies. Summary of the Invention

[0003] This invention provides an early warning method for the supply of power materials in temporary construction warehouses, mainly including:

[0004] Based on the relative positional relationship between temporary warehouses for highway construction and high-voltage transmission lines, a three-dimensional electromagnetic field distribution model based on actual environmental influences is established to obtain the theoretical distribution of the electromagnetic environment of each warehouse.

[0005] In the temporary warehouse, electromagnetic sensors are deployed for materials whose electronic components are highly sensitive to electromagnetic interference, and large dynamic range electromagnetic sensors are deployed for materials whose metallic materials are sensitive to strong electromagnetic fields.

[0006] Through low-power wide-area IoT, monitoring data collected by various electromagnetic sensors in the warehouse are transmitted back to the power grid dispatch center in real time, and associated with the warehouse material structure parameters to build a database.

[0007] Electromagnetic interference thresholds for various materials are extracted from the database, and the monitoring data is fused with the theoretical distribution to determine the distribution of electromagnetic field strength in the warehouse and identify the types of materials and specific areas with electromagnetic interference risks.

[0008] If the electromagnetic field strength in a certain area is found to exceed the electromagnetic susceptibility threshold of the stored materials in that area, an early warning message will be generated, and protective facilities will be configured according to the actual electromagnetic field, and corresponding suggestions for adjusting protective measures will be given.

[0009] If the overall electromagnetic environment of the warehouse deteriorates, the early warning information will be sent to the power grid dispatch center to coordinate the dynamic adjustment of the operating parameters of the high-voltage transmission lines and reduce the electromagnetic interference level generated by the transmission lines to a safe range that various materials can withstand.

[0010] By mining historical data on power grid load and electromagnetic environment, the correlation between electromagnetic interference characteristics and transmission line operating conditions is extracted, and a collaborative prediction model of electromagnetic security status for multiple warehouses is established.

[0011] Based on the prediction results of the prediction model, electromagnetic protection measures are adjusted through warehouse hardware facilities, the storage area of ​​sensitive materials is dynamically optimized, and the early warning information is fed back to the power grid dispatch center to coordinate and optimize the operation mode of transmission lines.

[0012] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0013] This invention discloses an early warning method for the supply of electrical materials in temporary construction warehouses. By real-time monitoring and analysis of electromagnetic field strength, it accurately locates potential electromagnetic interference risk areas, effectively preventing sensitive materials from being affected by electromagnetic fields and ensuring the integrity and normal operation of materials. Combining IoT technology and intelligent algorithms, it can issue early warnings, prompting managers to respond quickly, avoiding the risk of material damage, and enhancing the foresight and proactivity of warehouse management. Simultaneously, it dynamically adjusts protection strategies based on real-time monitoring data, providing customized storage solutions according to the characteristics and electromagnetic sensitivity of different materials, ensuring that materials are in optimal preservation conditions in complex electromagnetic environments. Furthermore, collaborative operations with the power grid dispatch center not only improve the operational stability of the power grid but also optimize the operating mode of transmission lines based on real-time monitoring data, reducing electromagnetic interference and achieving harmonious coexistence between power facilities and warehouse material management. It also promotes the optimal allocation of resources, achieving a double harvest of economic benefits and safety management by reducing material losses and maintenance costs, providing solid logistical support for highway construction projects. Finally, the generated real-time data and analysis reports provide key information for decision-makers, accelerating the optimization process of material allocation and warehouse layout, and greatly improving the efficiency and reliability of construction logistics. In summary, this invention significantly enhances the electromagnetic environment safety of temporary warehouses near high-voltage power lines, building a solid defense for the complete protection of sensitive materials. Attached Figure Description

[0014] Figure 1 This is a flowchart of an early warning method for the supply of power materials in a temporary construction warehouse according to the present invention.

[0015] Figure 2 This is a schematic diagram of an early warning method for the supply of power materials in a temporary construction warehouse according to the present invention.

[0016] Figure 3 This is another schematic diagram of an early warning method for the supply of power materials in a temporary construction warehouse according to the present invention. Detailed Implementation

[0017] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] like Figure 1-3 This embodiment of an early warning method for the supply of power materials in temporary construction warehouses may specifically include:

[0019] S101. Based on the relative positional relationship between the temporary warehouses for highway construction and high-voltage transmission lines, a three-dimensional electromagnetic field distribution model based on the actual environmental impact is established to obtain the theoretical distribution of the electromagnetic environment of each warehouse.

[0020] Three-dimensional coordinate data of temporary warehouses for highway construction and high-voltage transmission lines were acquired. This data included the warehouse locations and dimensions, as well as the tower coordinates of the transmission lines. The spatial distribution of the electromagnetic field sources was determined based on this data, determined by the voltage level, current load, phase line arrangement, and tower structure of the transmission lines. A three-dimensional electromagnetic field distribution model was established using the finite element method, with the transmission line parameters as boundary conditions. If topographic relief exists, a digital terrain model was established; if atmospheric influences exist, an atmospheric propagation model was constructed. The spatial distribution of electromagnetic field intensity was analyzed, taking into account the shielding and reflection effects of topographic relief on electromagnetic wave propagation, as well as the influence of atmospheric conditions on electromagnetic wave attenuation. For each warehouse's specific location and building structure, the distribution of the electromagnetic field inside the warehouse was calculated, determined by the electromagnetic shielding coefficient and reflection coefficient of the warehouse wall materials.

[0021] Specifically, three-dimensional coordinate data of the temporary construction warehouse for the highway and the high-voltage transmission line were obtained, including the warehouse location and dimensions, as well as the tower coordinates of the transmission line. Based on the voltage level, current load, phase line arrangement, and tower structure of the transmission line, the spatial distribution of the electromagnetic field source was determined. A three-dimensional electromagnetic field distribution model was established using the finite element method, with the transmission line parameters as boundary condition inputs.

[0022] A digital terrain model is established based on topographic elevation data, and an atmospheric propagation model is constructed using meteorological parameters such as atmospheric refractive index and humidity. These environmental factors are then parameterized and input into the electromagnetic field distribution model. The spatial distribution of electromagnetic field intensity is calculated by solving Maxwell's equations. The effects of terrain undulations on electromagnetic wave propagation, including shielding and reflection, as well as the influence of atmospheric conditions on electromagnetic wave attenuation, are considered.

[0023] For each warehouse, considering its specific location and building structure, the distribution of electromagnetic fields within the warehouse is calculated. Based on the electromagnetic shielding and reflection coefficients of the warehouse wall materials, the propagation attenuation and multipath effects of electromagnetic waves within the warehouse are calculated. Secondary scattering of the electromagnetic field by equipment and items inside the warehouse is considered to obtain more accurate data on the distribution of the electromagnetic environment inside the warehouse.

[0024] Using 3D visualization technology, the calculated electromagnetic field strength data is transformed into a 3D electromagnetic environment distribution map. Different colors and transparency levels are used to represent changes in electromagnetic field strength, visually presenting the theoretical electromagnetic environment distribution for each warehouse. On-site electromagnetic field strength measurements are conducted at typical locations, and the results are compared with the model calculations to evaluate the model's accuracy and perform necessary calibrations. Model parameters are adjusted according to different construction stages and seasonal changes to generate electromagnetic environment distribution maps for various scenarios, providing a comprehensive electromagnetic environment assessment.

[0025] When acquiring the three-dimensional coordinate data of the temporary warehouse for highway construction and the high-voltage transmission line, a GPS positioning system was used to accurately locate the warehouse, whose dimensions were measured to be 20 meters long, 15 meters wide, and 6 meters high. The coordinates of the transmission line towers were obtained using a laser rangefinder, with an adjacent tower spacing of 300 meters. The transmission line voltage level is 500kV, the current load is 1000A, and it uses a triangular arrangement with double-circuit straight-line towers. These parameters were input into the finite element software COMSOL Multiphysics to establish a three-dimensional electromagnetic field distribution model.

[0026] A terrain model with a resolution of 30 meters was constructed using digital elevation model (DEM) data. Atmospheric refractive index of 1.000325 and relative humidity of 60% were obtained from a weather station. An atmospheric propagation model was built using this data, considering the propagation characteristics of electromagnetic waves in different media. Maxwell's equations were solved using the finite-difference time-domain method (FDTD) to calculate the electromagnetic field intensity distribution. The shielding effect of terrain undulations on electromagnetic waves was considered in the calculations, and ray tracing algorithms were used to simulate electromagnetic wave reflection and diffraction.

[0027] For the warehouse building structure, the wall material is reinforced concrete, the electromagnetic shielding coefficient is 20dB, and the reflection coefficient is 0.8. The transmission line matrix method (TLM) is used to calculate the multipath propagation of electromagnetic waves within the warehouse, considering multiple reflections from the walls, floor, and ceiling. Monte Carlo simulation is then used to simulate the secondary scattering of electromagnetic fields by equipment and items inside the warehouse, generating more accurate data on the electromagnetic environment distribution within the warehouse.

[0028] 3D visualization was implemented using the OpenGL graphics library, transforming the calculated electromagnetic field strength data into a 3D electromagnetic environment distribution map. A red, yellow, and green color gradient was used to represent the change in electromagnetic field strength from high to low, with an opacity range of 30%–100%. Field electromagnetic field strength measurements were conducted at three typical locations—50 meters, 100 meters, and 200 meters from the transmission line. A spectrum analyzer was used to compare the model's calculation results with the actual values, evaluating the model's accuracy. Based on the comparison results, the model parameters were optimized and calibrated using the least squares method. Considering load variations during construction (daytime) and non-construction (nighttime) periods, as well as temperature and humidity variations throughout the four seasons (spring, summer, autumn, and winter), the model parameters were adjusted to generate electromagnetic environment distribution maps for eight different scenarios, providing a comprehensive electromagnetic environment assessment.

[0029] S102. In the temporary warehouse, electromagnetic sensors are deployed for materials whose electronic components are highly sensitive to electromagnetic interference; and large dynamic range electromagnetic sensors are deployed for materials whose metallic materials are susceptible to strong electromagnetic fields.

[0030] For the warehouse interior, a laser scanner is used to perform a 3D spatial scan to acquire point cloud data. Based on the point cloud data, a spatial model of the warehouse is constructed using 3D reconstruction software. Electromagnetic susceptibility is assessed for the electronic component and metal material areas within the warehouse spatial model, determining the coordinates of deployment points for high-sensitivity electromagnetic sensors and large dynamic range electromagnetic sensors. Based on these coordinates, Hall effect sensors and capacitive sensors are deployed in the electronic component areas, and Rogowski coil sensors and electrostatic field meters are deployed in the metal material areas. Sensor data from the Hall effect sensors, capacitive sensors, Rogowski coil sensors, and electrostatic field meters are collected. Kalman filter state equations and observation equations are established based on the sensor data. Optimal estimates are obtained through iterative calculations based on the state equations and observation equations. Based on the optimal estimates and the deployment point coordinates, a 3D model of the electromagnetic field distribution within the warehouse is constructed. If electromagnetic interference exceeding a preset threshold is detected in the 3D model of the electromagnetic field distribution, an alarm is triggered, and the location information of the interference source is provided using a triangulation algorithm.

[0031] Specifically, a laser scanner is used to perform a 3D spatial scan of the warehouse interior, generating point cloud data. A 3D reconstruction software is then used to construct a warehouse spatial model. Electronic components and metallic materials are categorized, their electromagnetic susceptibility is assessed, and the deployment coordinates of high-sensitivity and large dynamic range electromagnetic sensors are determined. In the electronic component area, Hall effect sensors with a sensitivity of 1 nT are deployed to measure weak magnetic fields, and capacitive sensors with a measurement range of 0.1 V / m to 100 V / m are used to detect low-intensity electric fields. The sampling frequency is set to 1 kHz, with an accuracy requirement of 0.1% of full scale. Spectral analysis is performed on the collected high-precision electromagnetic field data to identify electromagnetic interference characteristics in different frequency bands. In the metallic material area, Rogowski coil sensors with a measurement range of 0.1 mT to 1 T are installed to measure strong magnetic fields, and electrostatic field meters with a range of 1 kV / m to 100 kV / m are used to measure high-intensity electric fields. The sampling frequency is set to 10 kHz, with an accuracy requirement of 1% of full scale. Time-domain analysis was performed on the acquired wide-range electromagnetic field data to identify the instantaneous variation characteristics of the electromagnetic field intensity. All sensor data were collected, and the Kalman filter state equation x(k) = Ax(k-1) + w(k) and observation equation z(k) = Hx(k) + v(k) were established. Here, A is the state transition matrix, describing how the system state changes from time step k-1 to k; x(k) is the state vector at time step k; x(k-1) represents the state vector at the previous time step -1; z(k) is the observation vector at time step k; H is the observation matrix, describing the relationship between the system state and the observations; w(k) is the process noise, a random error introduced during the state transition; and v(k) is the observation noise, a random error introduced during the measurement. Noise covariance matrices Q and R were set, and the optimal estimate was obtained through iterative calculation. Based on the sensor location information and the filtered measurement data, a three-dimensional model of the electromagnetic field distribution within the warehouse was constructed. A heatmap was used to represent the electromagnetic field intensity distribution. When electromagnetic interference exceeding a preset threshold was detected, an alarm was triggered, and the location information of the interference source was provided through a triangulation algorithm. A FAROFocus S350 laser scanner was used to perform a 3D spatial scan of the interior of a 60m×40m×10m warehouse, with a scanning accuracy set to 2mm@10m. This means that the scanning accuracy at a distance of 10 meters is 2 millimeters, generating point cloud data containing approximately 500 million points. Autodesk ReCap software was used for point cloud registration and noise reduction to construct a warehouse spatial model accurate to the centimeter level. Based on the material list, electronic components were divided into three categories: high sensitivity (e.g., integrated circuits), medium sensitivity (e.g., resistors and capacitors), and low sensitivity (e.g., connectors). Metallic materials were divided into two categories: strong magnetic (e.g., iron products) and weak magnetic (e.g., aluminum products). The electromagnetic susceptibility of various materials was evaluated using the electromagnetic field simulation software ANSYS Maxwell, and the coordinates of the deployment points of 50 high-sensitivity electromagnetic sensors and 30 large dynamic range electromagnetic sensors were determined.In the electronic component area, a Honeywell HMC1001 magnetic sensor (sensitivity 1nT) and a Teledyne LeCroy AP015 electric field probe with a range of 0.1V / m-100V / m were deployed. The sampling frequency was set to 1kHz, and a 16-bit ADC ensured 0.1% full-scale accuracy. Fast Fourier Transform (FFT) was used to perform spectral analysis on the acquired electromagnetic field data to identify electromagnetic interference at characteristic frequencies such as 50Hz and 400Hz. For the metallic material area, a GMWA Associates 42B Rogowski coil with a range of 0.1mT-1T and a Monroe 247 electrostatic field meter with a range of 1kV / m-100kV / m were installed. The sampling frequency was 10kHz, and a 12-bit ADC ensured 1% full-scale accuracy. Wavelet transform was used to perform time-frequency analysis on the wide-range electromagnetic field data to identify the instantaneous changes in electromagnetic field intensity. All sensor data were acquired using a National Instruments DAQ-9188 data acquisition unit, and a Kalman filter model was established. The state transition matrix A and observation matrix H were set based on the electromagnetic field propagation characteristics, while the noise covariance matrices Q and R were determined based on sensor specifications and measured data. The Kalman filter algorithm was implemented using MATLAB, and the optimal estimate was obtained through iterative calculation. Combining sensor location information and filtered measurement data, a 0.5m × 0.5m × 0.5m grid resolution electromagnetic field distribution model within the warehouse was constructed using a 3D interpolation algorithm. An electromagnetic field intensity heatmap was generated using ParaView software, with colors ranging from blue to red indicating increasing field strength. When electromagnetic interference exceeding a preset threshold, such as electric field strength > 10V / m or magnetic field strength > 0.1mT, an audible and visual alarm was triggered, and a triangulation algorithm based on the least squares method was used to provide the interference source location information, with a positioning accuracy better than 1m.

[0032] S103. Through low-power wide-area IoT, the monitoring data collected by various electromagnetic sensors in the warehouse is transmitted back to the power grid dispatch center in real time, and associated with the warehouse material structure parameters to build a database.

[0033] Raw electromagnetic sensor data collected by LoRa node devices is acquired, and the difference between adjacent sampling points is calculated using the DPCM algorithm based on the raw data. For the difference information, an encrypted data packet is generated using the AES-128 encryption algorithm and transmitted to the power grid dispatch center via the LoRaWAN network. The encrypted data packet is decrypted using a preset key to obtain the raw monitoring data. A database is constructed, including acquiring material structure parameter information and converting the raw monitoring data and the material structure parameter information into JSON format data. If the JSON format data is structured material parameter information, it is stored in a MySQL database; if the JSON format data is semi-structured monitoring data, it is stored in a MongoDB database.

[0034] Specifically, LoRaWAN gateways and LoRa node devices are deployed within the warehouse, employing a star network topology to connect various electromagnetic sensors to the LoRa nodes. Low-power wide-area data transmission is achieved via the LoRaWAN protocol, with the node devices operating in the 470-510MHz frequency band and a transmit power of 14dBm to ensure coverage of the entire warehouse area. The raw data collected by the sensors is preprocessed and compressed using the DPCM (Differential Pulse Code Modulation) algorithm to reduce data volume. The DPCM algorithm calculates the difference between adjacent sampling points and transmits only the difference information, achieving data compression. The compressed data is encrypted using the AES-128 encryption algorithm to ensure transmission security. Encrypted data packets are transmitted to the power grid dispatch center in real time via the LoRaWAN network, with a transmission cycle of 5 minutes. The power grid dispatch center receives and parses data packets from each LoRa node, decrypts them using a preset key, and restores the original monitoring data. Data quality checks are performed to remove outliers, such as data exceeding the sensor's range. Material structure parameter information, including material type and electromagnetic shielding performance, is obtained from the warehouse management system via a RESTful API interface. Monitoring data and material parameters were uniformly converted into JSON format to prepare for database storage. A heterogeneous database was constructed using MySQL and MongoDB. Relational tables were created in MySQL to store structured material parameter information, while collections were created in MongoDB to store semi-structured monitoring data. Indexes were built to optimize query performance, such as creating B-tree indexes for material IDs in MySQL and hash indexes for timestamps in MongoDB. Foreign key relationships were used to establish data connections between MySQL and MongoDB. Database view technology was used to create a unified view containing monitoring data and corresponding material parameters, supporting fast relational queries. One Semtech SX1301 multi-channel LoRaWAN gateway and 50 LoRa node devices equipped with SX1276 chips were deployed in a 100m×50m×10m warehouse, using a star network topology. Twenty Hall effect magnetic sensors and 30 capacitive electric field sensors were connected to the LoRa nodes, operating at 470MHz, with a transmit power of 14dBm and a sensitivity of -137dBm, achieving full warehouse coverage. Each node collects data every 5 minutes, totaling 200 bytes. The DPCM algorithm is used to compress the data, achieving a compression ratio of 3:1. Data is represented by the difference between the current value Xn and the previous value Xn-1, ΔXn = Xn - Xn-1; only ΔXn is transmitted. The compressed data is encrypted using the AES-128 algorithm, with a 128-bit key length, a 128-bit block length, and CBC encryption mode. Encrypted data packets are transmitted via LoRaWAN Class A mode, with the uplink SF (spreading factor) set to 7-12, adjustable adaptively. The power grid dispatch center receives the data packets, decrypts them using the preset AES-128 key, and restores the original data.Perform quality checks on the data, removing outliers exceeding ±5% of the sensor's range. Obtain material structure parameters, including material ID, type, and electromagnetic shielding performance, from the warehouse management system via a RESTful API (port 8080, OAuth 2.0 authentication). Convert the monitoring data and material parameters into JSON format; the monitoring data structure is as follows.

[0035] The material parameter structure is {"t imestamp":1617235200,"sensor_id":"MAG001","value":0.05}, with the structure {"material_id":"M001","type":"metal","shielding_factor":30}. A material parameter table is created in a MySQL 8.0 database with the primary key "material_id" and a B+ tree index on it. A monitoring dataset is created in a MongoDB 4.4 database, with a hash index on the "t imestamp" field. The MySQL and MongoDB data are linked via the foreign key "material_id". A view, "view_monitoring_data", is created to connect the monitoring data and material parameters, supporting compound queries such as "SELECT * FROM view_monitoring_data WHERE t imestamp BETWEEN ? AND ? AND shielding_factor > ?", with a query time of less than 100ms.

[0036] S104. Extract the electromagnetic interference thresholds of various materials from the database, integrate the monitoring data with the theoretical distribution, determine the distribution of electromagnetic field strength in the warehouse, and identify the types of materials and specific areas with electromagnetic interference risks.

[0037] Electromagnetic interference (EMI) threshold data for various materials are obtained from a database. An EMI threshold table is established based on material type and sensitivity, including electric field strength thresholds and magnetic field strength thresholds. Real-time monitoring data within the warehouse is acquired. If the real-time monitoring data consists of discrete monitoring points, ordinary kriging is used to spatially interpolate the discrete monitoring point data to obtain continuous EMI distribution data. The continuous EMI distribution data is then fused with the theoretical distribution using a Markov chain Monte Carlo method to obtain a fused EMI distribution. The fused EMI distribution is compared with the EMI threshold table, and a C4.5 decision tree algorithm is used to identify regions exceeding the thresholds. The minimum leaf node sample size and maximum depth of the C4.5 decision tree algorithm are preset values. Historical data is updated using a sliding window method, and the EMI thresholds in the EMI threshold table are adjusted based on the updated historical data. The time window of the sliding window method is a preset window value.

[0038] Specifically, electromagnetic interference (EMI) thresholds for various materials are extracted from a pre-defined database. Structured threshold data is retrieved from a MySQL database using SQL statements, while unstructured threshold data is extracted from a MongoDB database using an aggregation pipeline. An EMI threshold table, including electric and magnetic field strength thresholds, is established based on material type and sensitivity. The threshold data is standardized using the Z-score method to convert the thresholds for different material types into a standard normal distribution, ensuring comparability. Real-time monitoring data from the warehouse is acquired, and ordinary kriging is used to spatially interpolate the discrete monitoring point data, generating continuous EMI distribution data. A theoretical model is fitted, selecting the optimal fit from a spherical, exponential, or Gaussian model. Cross-validation is used to evaluate the interpolation accuracy, calculating the root mean square error (RMSE) and average kriging variance. Uncertainty analysis is performed on the interpolation results, generating a standard error plot to provide input for subsequent Bayesian inference. The monitoring data and theoretical distribution are fused using the Markov chain Monte Carlo (MCMC) method to update the prior probability of the EMI distribution, obtaining the posterior probability distribution. The Metropol is-Hast ings algorithm was used for parameter sampling, with a combustion period of 1000 iterations and a total of 10000 iterations. The mean, variance, and 95% confidence interval of the posterior distribution were calculated to quantify the uncertainty of the fusion result. A fused electromagnetic field intensity distribution map and an uncertainty distribution map were generated. The fused electromagnetic field intensity distribution was compared with a standardized electromagnetic interference threshold table, and the C4.5 decision tree algorithm was used to identify areas exceeding the threshold and their corresponding material types. The minimum leaf node sample size of the decision tree was set to 5, the maximum depth to 10, and the Gini coefficient was used as the splitting criterion. Based on the decision tree output, the risk level was divided into low, medium, and high levels, and a risk heatmap was generated for visualization. Dynamic threshold adjustments were made, with historical data updated weekly using a sliding window method to adjust the electromagnetic interference threshold, ensuring the accuracy and timeliness of risk identification. SQL statements were used in the warehouse management system.

[0039] The query "SELECT material_type, em_field_type, threshold_value FROM em_thresholds WHERE sensitivity_level>3" extracts the electromagnetic interference threshold for highly sensitive materials from a MySQL database and aggregates it via a MongoDB pipeline.

[0040] Unstructured threshold data was obtained using the following method: [{$match:{sensitivity:{$gt:3}}},{$group:{_id:"$material_type",avg_threshold:{$avg:"$threshold_value"}}}]. Z-score standardization was applied to the extracted threshold data; for example, the electric field strength threshold of 5V / m for electronic components was converted to a standard score of 1.2. Spatial interpolation was performed on the data from 50 monitoring points using ordinary kriging, employing a spherical semi-variogram model with an effective range of 20 meters, a nugget value of 0.05, and a sill value of 0.8. The interpolation accuracy was evaluated using leave-one-out cross-validation, yielding an RMSE of 0.3V / m and a mean kriging variance of 0.25. A standard error was generated, ranging from 0.1 to 0.5V / m. MCMC sampling was performed using the Metropol is-Hastings algorithm, with a normal distribution N(μ,σ^2) chosen for the proposal distribution, where μ represents the current state and σ = 0.1. The combustion period was set to 1000 iterations, with a total of 10000 iterations. The posterior distribution mean was calculated to be 6.2 V / m, with a variance of 0.4 and a 95% confidence interval of [5.8 V / m, 6.6 V / m]. A fused electromagnetic field intensity distribution map was generated, with an intensity range of 2-8 V / m and a resolution of 0.5m × 0.5m. The C4.5 decision tree algorithm was applied, with a minimum leaf node sample size of 5, a maximum depth of 10, and the Gini coefficient used as the splitting criterion. Based on the decision tree output, the risk level was divided into three levels: low (<4 V / m), medium (4-6 V / m), and high (>6 V / m), generating a risk heatmap. Historical data was updated weekly using a 30-day sliding window, and the electromagnetic interference threshold was dynamically adjusted using the pre-set knowledge base rule "IF average field strength > historical 90th percentile THEN increase threshold by 10%". The final output included the risk area coordinates (x:15m, y:25m, z:2m) and the corresponding high-risk material type "precision electronic instruments".

[0041] S105. If the electromagnetic field strength in a certain area is found to exceed the electromagnetic sensitivity threshold of the stored materials in that area, an early warning message will be generated, and protective facilities will be configured according to the actual electromagnetic field, and corresponding suggestions for adjusting protective measures will be given.

[0042] The warehouse space is divided into regions using the K-means clustering algorithm, with a preset number of clusters and Euclidean distance as the similarity metric, to obtain an electromagnetic environment zoning map. Based on this map, the electromagnetic field strength of each region is compared with a preset electromagnetic sensitivity threshold for stored materials using the C4.5 decision tree algorithm to determine if any thresholds are exceeded. If the result indicates that the electromagnetic field strength of a certain region exceeds the sensitivity threshold, an early warning is triggered, generating a JSON-formatted warning message containing the location of the exceeding region, the electromagnetic field strength value, and the type of material. Based on the JSON-formatted warning message, information on existing electromagnetic protection facilities in the region is obtained, and recommendations for adjusting protection measures are generated. For the proposed adjustments, a cost-benefit analysis is performed on the protection measures to calculate the return on investment (ROI), and a priority list is generated by sorting the ROIs in descending order.

[0043] Specifically, based on real-time monitoring data, the K-means clustering algorithm is used to divide the warehouse space into regions, grouping areas with similar electromagnetic field strength characteristics into one category to obtain an electromagnetic environment zoning map. The clustering number K is set to 5, Euclidean distance is used as the similarity metric, and the maximum number of iterations is 100. The zoning results are evaluated using the silhouette coefficient, and the zoning scheme with the largest silhouette coefficient is selected. A heatmap of the zoning results is generated, using different colors to represent the electromagnetic field strength level of each region. The electromagnetic susceptibility thresholds of stored materials in each region are queried from the database, and the C4.5 decision tree algorithm is used to compare the electromagnetic field strength with the susceptibility thresholds to determine if any exceedances occur. The maximum depth of the decision tree is set to 5, the minimum number of leaf node samples is 10, and the information gain ratio is used as the splitting criterion. A decision tree visualization is generated to show the judgment process. The judgment results are overlaid on the zoning heatmap, and regions exceeding the threshold are highlighted with special markers. If the judgment results show that the electromagnetic field strength of a certain region exceeds the threshold, an early warning is triggered, generating a JSON-formatted early warning message containing information such as the location of the exceeding region, the electromagnetic field strength value, and the type of material. The structure of the early warning message is as follows:

[0044] {"warning_id":"W001","area":"A3","em_strength":8.5,"threshold":7.0,"material_type":"Precision Electronic Components","t imestamp":"2023-04-0110:30:00"}. Based on the degree to which the threshold is exceeded, the warning levels are divided into three levels: low, medium, and high, corresponding to yellow, orange, and red warnings, respectively. High-level warnings are automatically pushed to the mobile devices of relevant management personnel. Information on existing electromagnetic protection facilities in the area is extracted from the protection facility database. Combined with protection rules in a preset knowledge base, the Mamdani fuzzy inference system is used to generate suggestions for adjusting protection measures. Input variables are set as "degree exceeding the threshold" and "existing protection level," and output variable is "protection measure strength." A fuzzy rule set is defined, such as "IF degree exceeding the threshold = high AND existing protection level = low THEN protection measure strength = strong." Defuzzification is performed using the centroid method to obtain specific protection measure strength values. Based on the strength value, corresponding measures are selected from the protection measure library, such as replacing shielding materials or optimizing the grounding system. A cost-benefit analysis was conducted on the protective measures, and the return on investment (ROI) was calculated. A priority list was generated by sorting the data in descending order of ROI. Electromagnetic field strength data were collected from 500 monitoring points in a 100m × 50m × 10m warehouse space. K-means clustering was used for region division. The number of clusters K was set to 5, Euclidean distance was used as the similarity metric, and the maximum number of iterations was 100. After 50 iterations, convergence was achieved, resulting in 5 cluster centers located at (20, 15, 3), (40, 30, 5), (60, 10, 2), (80, 40, 7), and (90, 25, 4), in meters. The silhouette coefficient was calculated to be 0.72, indicating good clustering performance. A heatmap was generated, using red, orange, yellow, green, and blue colors to represent five regions with electromagnetic field strength ranging from high to low. The electromagnetic susceptibility thresholds for materials in each region are retrieved from a preset database. For example, the threshold for precision electronic components is 5V / m, and the threshold for ordinary electronic equipment is 10V / m. The C4.5 decision tree algorithm is used to compare the electromagnetic field strength with the thresholds, with a maximum depth of 5, a minimum leaf node sample size of 10, and an information gain ratio threshold of 0.1. A decision tree is generated, with the root node representing "electromagnetic field strength" and the leaf nodes representing "exceeds threshold" or "does not exceed threshold." The results show that the electromagnetic field strength in the region (40, 30, 5) is 8.5V / m, exceeding the 5V / m threshold for that region. An alert is triggered, generating a JSON-formatted alert message.

[0045] {"warning_id":"W001","area":"A2","em_strength":8.5,"threshold":5.0,"material_type":"Precision Electronic Components","t imestamp":"2023-04-0110:30:00","level":"High"}. This high-level warning is pushed to the administrator's mobile device via the WebSocket protocol. The protective facility database indicates that the area is currently shielded with fine metal materials including copper, aluminum, and steel, with a shielding effectiveness of 60dB. The Mamdani fuzzy inference system is used to generate protective measure suggestions. The fuzzy set for the input variable "threshold exceedance level" is set to {low, medium, high}, and the membership function is a Gaussian function; the fuzzy set for "existing protection level" is set to {weak, medium, strong}, and the membership function is a trapezoidal function. The fuzzy set for the output variable "protective measure strength" is set to {fine-tuning, strengthening, rebuilding}, and the membership function is a trigonometric function. Nine fuzzy rules were defined, such as "IF exceeding threshold level = high AND existing protection level = medium THEN protection measure strength = enhanced". The fuzziness was de-fuzzified using the centroid method, yielding a protection measure strength value of 0.75. From the preset protection measure library, "replace with 80dB shielding effectiveness metal shielding" was selected. The estimated cost was 50,000 yuan, with an expected benefit of 80,000 yuan, resulting in an ROI of 60%. This measure was then listed as the highest priority protection recommendation.

[0046] S106. If the overall electromagnetic environment of the warehouse deteriorates, the early warning information will be sent to the power grid dispatch center to coordinate the dynamic adjustment of the operating parameters of the high-voltage transmission lines and reduce the electromagnetic interference level generated by the transmission lines to a safe range that various materials can withstand.

[0047] The system acquires electromagnetic environment data from a pre-defined distributed sensor network monitoring warehouse. Based on this data, a sliding window algorithm is used to calculate the spatiotemporal variation trend of the electromagnetic field intensity, yielding electromagnetic field intensity variation data. This data is then quantitatively evaluated. If the average electromagnetic field intensity increases by a pre-defined range or the peak value exceeds a pre-defined multiple of the safety standard, the electromagnetic environment is deemed to have deteriorated, and a warning data packet is generated. This warning data packet is encrypted using the SSL / TLS protocol, resulting in an encrypted warning data packet, which is then sent to the power grid dispatch center. Upon receiving the encrypted warning data packet, the power grid dispatch center uses a weighted least squares method to estimate the power system state, identifying the key transmission lines affecting the warehouse's electromagnetic environment and their operating parameters. For the operating parameters of these key transmission lines, a non-dominated sorting genetic algorithm is used for multi-objective optimization. Based on the optimized operating parameters, the transmission lines are dynamically adjusted.

[0048] Specifically, a pre-set distributed sensor network monitors the overall electromagnetic environment of the warehouse in real time. A sliding window algorithm is used to calculate the spatiotemporal variation trend of electromagnetic field strength, with a window size of 60 minutes and a sliding step size of 5 minutes. The degree of electromagnetic environment deterioration is quantitatively assessed, and a threshold for triggering an early warning is set at a 20% increase in the average electromagnetic field strength or a peak exceeding 1.5 times the safety standard. If the assessment indicates overall electromagnetic environment deterioration, an early warning data packet containing information such as warehouse location, electromagnetic field strength distribution, and material type is generated. Early warning information is divided into low, medium, and high priorities based on the degree of deterioration, with high-priority information being processed first. The early warning information is encrypted using the SSL / TLS protocol to ensure communication security before being sent to the power grid dispatch center. Upon receiving the early warning information, the power grid dispatch center uses weighted least squares to estimate the power system state, and, combined with real-time transmission line operation data, calculates the key transmission lines affecting the warehouse's electromagnetic environment and their operating parameters. The impact of each key transmission line is quantitatively assessed, its contribution rate to the warehouse's electromagnetic environment is calculated, and the feasibility of parameter adjustments is evaluated considering power grid operation constraints. Based on the calculation results, the Non-Dominated Sorting Genetic Algorithm II (NSGA-II) was used to perform multi-objective optimization and dynamic adjustment of the transmission line operating parameters. The optimization objectives included minimizing electromagnetic interference levels and maximizing grid operational stability. After each parameter adjustment, the adjustment effect was verified through a sensor network within a warehouse, and the results were fed back to the optimization algorithm. Simultaneously, a dynamic relationship model between the electromagnetic environment and transmission line parameters was constructed using a Long Short-Term Memory (LSTM) network for continuous learning and updating to improve the accuracy and efficiency of future adjustments. Fifty electromagnetic field sensors were deployed within a 100m × 50m × 10m warehouse, with a sampling frequency of 1Hz. A sliding window algorithm was used, with a window size of 60 minutes and a step size of 5 minutes, to calculate changes in electromagnetic field strength. A warning threshold was set at an average increase of 20% or a peak value exceeding 5V / m. When a deterioration in the electromagnetic environment was detected, such as an increase in average field strength from 2.5V / m to 3.2V / m within the 10:00-11:00 window, a warning was triggered, generating a JSON-formatted warning data packet.

[0049] {"locat ion":[30.5,104.1],"em_strength":{"avg":3.2,"peak":4.8},"material_types":["electronics","metals"]}, with high priority. The TLS 1.3 protocol is used, with a 256-bit key length, to encrypt the early warning information. After receiving and decrypting the data, the power grid dispatch center applies weighted least squares for state estimation. The weight matrix W is set as a diagonal matrix, with diagonal elements being the reciprocal of the measurement accuracy. Five iterations are performed, with a convergence threshold of 0.001, yielding state estimates for four key transmission lines. The contribution rate of each line to the warehouse's electromagnetic environment is calculated, e.g., line 1 accounts for 40%, line 2 for 30%, line 3 for 20%, and line 4 for 10%. Considering voltage stability constraints, the load power factor is no less than 0.85. The NSGA-II algorithm is used to optimize the transmission line parameters, with a population size of 100, 50 iterations, a crossover probability of 0.9, and a mutation probability of 0.1. The optimization objectives are to minimize electromagnetic interference (weight 0.6) and maximize grid stability (weight 0.4). After each parameter adjustment, the sensor network is used to verify the effect, such as when the field strength decreases to 2.8 V / m. Simultaneously, an LSTM network is used to build a dynamic relationship model. The input layer contains voltage and current parameters of four transmission lines, the hidden layer has 64 neurons, and the output layer is the predicted electromagnetic field strength. The model is trained using historical data from the past 30 days and updated every 24 hours, achieving a prediction accuracy of 92%.

[0050] S107. Mine historical data on power grid load and electromagnetic environment, extract the correlation between electromagnetic interference characteristics and transmission line operating conditions, and establish a collaborative prediction model for electromagnetic security status of multiple warehouses.

[0051] Historical data from power grid monitoring and warehouse electromagnetic environment monitoring were acquired, including power grid load, transmission line operating parameters, and warehouse electromagnetic field strength. Outliers were processed using the moving median method, and missing values ​​were processed using interpolation to obtain cleaned data. The cleaned data was then standardized using a min-max normalization method to determine standardized data. Electromagnetic interference signals in the standardized data were decomposed using Daubechies wavelets at five levels to obtain time-domain and frequency-domain features. Principal component analysis was used to reduce the feature dimensionality based on the time-domain and frequency-domain features, and it was determined whether the cumulative contribution rate reached a preset range. If it did, the electromagnetic interference feature vector was determined. The Apriori algorithm was used to analyze the correlation between the electromagnetic interference feature vector and the transmission line operating parameters. Minimum support and minimum confidence were set, and it was determined whether the support and confidence met the conditions. If the conditions were met, a strong correlation rule was determined, resulting in the correlation between electromagnetic interference and transmission line operating conditions. A graph convolutional network (GCN) is constructed based on the aforementioned correlation model. The GCN comprises a two-layer GCN structure, with ReLU as the activation function. The Adam optimizer is used to train the GCN, resulting in a collaborative prediction model for electromagnetic security situations in multiple warehouses. The prediction error is calculated based on the prediction results. It is then determined whether the prediction error exceeds a preset threshold. If it does, the warning threshold is adjusted to obtain a dynamic warning threshold.

[0052] Specifically, historical data, including grid load, transmission line operating parameters, and warehouse electromagnetic field strength, was extracted from the power grid monitoring system and the warehouse electromagnetic environment monitoring system. Outliers were cleaned using the moving median method, and missing values ​​were cleaned using interpolation. Standardization was performed using the min-max normalization method to unify data of different dimensions into the [0,1] interval. Timestamps were used to align multi-source heterogeneous data, forming a structured multi-dimensional time series dataset to ensure data consistency and integrity. Daubechies wavelets were used to perform a 5-level multi-scale decomposition of the electromagnetic interference signal, extracting time-domain and frequency-domain features. Time-domain features included statistics such as mean, variance, skewness, and kurtosis, while frequency-domain features included the energy proportion of each frequency band and major harmonic components. Principal component analysis was used to reduce the feature dimensionality, retaining principal components with a cumulative contribution rate of 95%, resulting in the feature vector of the electromagnetic interference. The extracted features were interpreted physically; for example, the mean reflects the interference intensity, the variance represents the degree of fluctuation, and the frequency band energy proportion reflects the interference spectrum characteristics. The Apriori algorithm was used to analyze the correlation between electromagnetic interference characteristics and transmission line operating conditions, with a minimum support of 0.05 and a minimum confidence of 0.7. Considering the time lag effect, the correlation patterns at different time scales (0-24 hours) were analyzed. Support and confidence were calculated, and strongly correlated rules were selected to construct a correlation model between electromagnetic interference and transmission line operating conditions. The correlation rules were quantitatively evaluated, and the Lift and Kulczynski metrics were calculated to provide weight initialization for subsequent graph neural networks. A collaborative prediction model of electromagnetic safety status for multiple warehouses was constructed based on a graph convolutional network (GCN). Each warehouse was treated as a node in the graph, and the transmission line as an edge. Node features included warehouse electromagnetic environment parameters, and edge features included transmission line operating parameters. A two-layer GCN structure was designed, with a hidden layer dimension of 64 and ReLU activation function. Information interaction and collaborative prediction among multiple warehouses were achieved through message passing. Mean squared error was used as the loss function, and the Adam optimizer was used for model training with a learning rate of 0.001 and 200 training epochs. Mean absolute error (MAE) and root mean square error (RMSE) were selected as performance evaluation metrics. The prediction results were visualized using the Matplotlib library, and the warning threshold was dynamically adjusted based on the prediction error to improve the model's practicality and adaptability. Five warehouses were located within a power grid area. One year's worth of 15-minute sampling interval data was extracted from the power grid monitoring system, including parameters such as grid load, voltage, and current of 10 transmission lines. Simultaneously, electromagnetic field strength data was obtained from the warehouse electromagnetic environment monitoring system. Anomalies were processed using the 11-point moving median method, detecting and correcting 342 outliers. Linear interpolation was used to handle 567 missing values. Min-max normalization was used to map the data to the [0,1] interval, such as normalizing the 50Hz electromagnetic field strength from 0-10V / m.The multi-source data was aligned using timestamps to form a 35040×87 dimensional data matrix. A 5-level decomposition of the electromagnetic interference signal was performed using the db4 wavelet, extracting 12 time-domain features (mean 3.2 V / m, variance 0.8) and 20 frequency-domain features (50Hz component accounting for 85%). Dimensionality reduction was achieved through principal component analysis, retaining 14 principal components with a cumulative contribution rate of 95.3%. The Apriori algorithm, with a minimum support of 0.05 and a minimum confidence of 0.7, identified 78 strongly correlated rules, such as "A 10% increase in current on line 1 → an 8% increase in electromagnetic field strength in warehouse A after 2 hours" (support 0.06, confidence 0.75). The Lift value (1.8) and Kulczynski metric (0.72) were calculated to verify the significance of the rules. A graph structure with 5 nodes and 10 edges was constructed, with node feature dimensions of 32 (including electromagnetic field strength, temperature, etc.) and edge feature dimensions of 16 (including line voltage, current, etc.). A two-layer GCN was designed with 64 hidden layers and ReLU activation. The mean squared error loss function was used, with an Adam optimizer learning rate of 0.001, a batch size of 32, and 200 training epochs. On the test set, the performance achieved a mean squared error (MAE) of 0.15 V / m and an RMSE of 0.22 V / m. Matplotlib was used to plot the comparison between predicted and actual values. A dynamic warning threshold was set to 1.2 times the predicted value to achieve real-time monitoring and early warning of electromagnetic safety.

[0053] S108. Based on the prediction results of the prediction model, adjust the electromagnetic protection measures through the warehouse hardware facilities, dynamically optimize the storage area of ​​sensitive materials, and at the same time feed the early warning information back to the power grid dispatch center to coordinate and optimize the operation mode of the transmission lines.

[0054] The system acquires predicted electromagnetic field intensity distribution and electromagnetic susceptibility data of materials within the warehouse, and calculates the optimal storage plan based on these data. The operating parameters of the electromagnetic shielding equipment within the warehouse, including the shielding plate angle and absorbing material thickness, are adjusted via a PLC. Electromagnetic field sensors monitor the shielding effect in real time, and a closed-loop feedback mechanism is established based on the monitoring results. If the monitoring results indicate that the shielding effect is substandard, the PLC is triggered to adjust the control parameters according to the feedback data. Early warning information, including the predicted electromagnetic exceedance area, time, and degree, is sent to the power grid dispatch center via the IEC 61850 protocol. Upon receiving the early warning information, the power grid dispatch center uses the NSGA-II algorithm to generate an optimized operation plan for the transmission lines, including line load distribution and reactive power compensation device adjustment parameters.

[0055] Specifically, based on the prediction results of the forecasting model, the operating parameters of the electromagnetic shielding equipment in the warehouse are adjusted by a PLC (Programmable Logic Controller), such as adjusting the angle of the shielding plate and changing the thickness of the absorbing material, to achieve dynamic adjustment of electromagnetic protection measures. High-precision electromagnetic field sensors are used to monitor the shielding effect in real time, forming a closed-loop feedback to ensure the effectiveness of the protection measures. The PLC continuously optimizes the control parameters based on the feedback data to maintain the best shielding effect. Utilizing the automated logistics system in the warehouse, combined with the electromagnetic field intensity distribution prediction results and the electromagnetic susceptibility data of the materials, a genetic algorithm is used to calculate the optimal storage plan. The algorithm considers both the importance and urgency of the materials and optimizes the handling sequence. After the automatic handling and rearrangement of sensitive materials, the adjusted electromagnetic environment is evaluated through a distributed sensor network to verify the optimization effect. If the adjusted electromagnetic environment still poses a risk, a new round of optimization calculation and adjustment is triggered. The early warning information is sent to the power grid dispatch control center via the IEC61850 protocol, including detailed information such as the predicted electromagnetic exceedance area, time, and degree. According to the severity of the early warning, the information is divided into three levels: general, important, and urgent, and different response measures are adopted for each level. For emergency warnings, emergency plans are activated, prioritizing adjustments to the operating parameters of relevant transmission lines. Upon receiving the warning information, the power grid dispatch center uses the NSGA-II (Non-Dominated Sorting Genetic Algorithm II) algorithm to generate an optimized operation plan for the transmission lines, comprehensively considering factors such as power grid safety, economy, and electromagnetic interference. Corresponding adjustments are executed remotely, such as changing line load distribution and adjusting reactive power compensation devices. After implementing the optimized plan, changes in the electromagnetic environment are continuously monitored, and the adjustment effect is evaluated. Based on the evaluation results, optimization objectives and constraints are dynamically adjusted to achieve continuous improvement. Simultaneously, the adjusted power grid operation data is fed back to the warehouse management system, forming a collaborative optimization closed loop between the warehouse and the power grid. In a 100m×50m×10m smart warehouse, a predictive model shows that the electromagnetic field intensity in the northwest corner will exceed the standard by 20% within the next 4 hours. After receiving the warning information, the Siemens S7-1500 PLC controls the angle of the 12 adjustable shielding plates in the northwest corner from 45° to 60°, while simultaneously increasing the thickness of the absorbing material from 2cm to 3cm. Ten high-precision electromagnetic field sensors (sensitivity 0.1V / m) monitor the shielding effect in real time, sampling every 30 seconds. The PLC employs a PID control algorithm with a proportional gain Kp = 0.8, integral time Ti = 120s, and derivative time Td = 30s, continuously optimizing control parameters to maintain the shielding effect above 30dB. The automated logistics system uses a genetic algorithm to calculate the optimal storage scheme, with a population size of 100, a crossover probability of 0.8, a mutation probability of 0.1, and 500 iterations. The algorithm categorizes 1000 items into three classes based on electromagnetic sensitivity, considering importance (levels 1-5) and urgency (urgent if used within 24 hours). AGVs rearrange the items in the optimized order within 30 minutes.Fifty distributed sensors were used to evaluate the adjusted electromagnetic environment, and Kriging interpolation was used to generate a new electromagnetic field distribution map.

[0056] The IEC 61850 protocol encodes early warning information into MMS messages, which are sent to the power grid dispatch and control center via an encrypted VPN tunnel. Early warnings are categorized into three levels: general (exceeding limits <10%), important (10%-30%), and urgent (>30%), triggering responses after 5 minutes, 2 minutes, and immediately, respectively. The power grid dispatch center uses the NSGA-II algorithm to optimize transmission line operation schemes, with a population size of 200 and 1000 iterations. Optimization objectives include minimizing electromagnetic interference, network losses, and adjustment costs, while constraints include voltage stability and N-1 safety checks. The remote control system executes the optimization scheme, such as reducing the load on line 2 by 15% and adjusting the SVG reactive power compensation by +20 Mvar. The adjustment effect is evaluated every 5 minutes, and the weighting coefficients are dynamically adjusted. Ultimately, the warehouse electromagnetic environment compliance rate increased from 85% to 98%, while power grid losses decreased by 3%.

[0057] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for early warning of power supply in temporary construction warehouses, characterized in that, The method includes: establishing a three-dimensional electromagnetic field distribution model based on the relative positional relationship between temporary warehouses for highway construction and high-voltage transmission lines, thereby obtaining the theoretical distribution of the electromagnetic environment for each warehouse; within the temporary warehouses, deploying electromagnetic sensors for materials whose electronic components are highly sensitive to electromagnetic interference, and deploying large dynamic range electromagnetic sensors for materials whose metallic materials are susceptible to strong electromagnetic fields; transmitting monitoring data collected by various electromagnetic sensors in the warehouses back to the power grid dispatch center in real time via a low-power wide-area Internet of Things (IoT), and correlating this data with the material structure parameters of the warehouses to construct a database; extracting electromagnetic interference thresholds for various materials from the database, fusing the monitoring data with the theoretical distribution, determining the distribution of electromagnetic field intensity within the warehouses, and identifying the types of materials and specific areas with electromagnetic interference risks; if a certain area is found to be at risk... If the electromagnetic field strength exceeds the electromagnetic susceptibility threshold of the stored materials in the area, an early warning message is generated, and protective facilities are configured according to the actual electromagnetic field, with corresponding suggestions for adjusting protective measures. If the overall electromagnetic environment of the warehouse deteriorates, the early warning message is sent to the power grid dispatch center to coordinate the dynamic adjustment of the operating parameters of high-voltage transmission lines, reducing the electromagnetic interference level generated by the transmission lines to a safe range that various materials can withstand. Historical data on power grid load and electromagnetic environment are mined to extract the correlation between electromagnetic interference characteristics and transmission line operating conditions, establishing a collaborative prediction model for the electromagnetic security situation of multiple warehouses. Based on the prediction results of the model, electromagnetic protection measures are adjusted through warehouse hardware facilities to dynamically optimize the storage area of ​​sensitive materials, while the early warning message is fed back to the power grid dispatch center to coordinate and optimize the operation mode of transmission lines.

2. The method according to claim 1, wherein, The process involves establishing a three-dimensional electromagnetic field distribution model based on the relative positions of temporary highway construction warehouses and high-voltage transmission lines, taking into account the actual environmental influences. This model yields the theoretical electromagnetic environment distribution for each warehouse. The process includes: acquiring three-dimensional coordinate data of the temporary construction warehouses and high-voltage transmission lines, including warehouse locations and dimensions, and tower coordinates; determining the spatial distribution of electromagnetic field sources based on the three-dimensional coordinate data, determined by the voltage level, current load, phase line arrangement, and tower structure of the transmission lines; establishing a three-dimensional electromagnetic field distribution model using the finite element method, with transmission line parameters as boundary conditions; establishing a digital terrain model if topographic undulations exist; constructing an atmospheric propagation model if atmospheric influences exist; analyzing the spatial distribution of electromagnetic field intensity, which incorporates the shielding and reflection effects of topographic undulations on electromagnetic wave propagation, as well as the influence of atmospheric conditions on electromagnetic wave attenuation; and calculating the distribution of electromagnetic fields within each warehouse based on its specific location and building structure, determined by the electromagnetic shielding coefficient and reflection coefficient of the warehouse wall materials.

3. The method according to claim 1, wherein, Within the temporary warehouse, electromagnetic sensors are deployed for materials containing electronic components that are highly sensitive to electromagnetic interference, and large dynamic range electromagnetic sensors are deployed for materials containing metallic materials that are susceptible to strong electromagnetic fields. This includes: using a laser scanner to perform a three-dimensional spatial scan of the warehouse interior to acquire point cloud data; constructing a spatial model of the warehouse using 3D reconstruction software based on the point cloud data; assessing the electromagnetic sensitivity of the electronic component and metallic material areas within the warehouse spatial model to determine the deployment coordinates of the high-sensitivity and large dynamic range electromagnetic sensors; and deploying sensors in the electronic component areas based on the deployment coordinates. A Hall effect sensor and a capacitive sensor are used, and a Rogowski coil sensor and an electrostatic field meter are deployed in the metal material area. Sensor data from the Hall effect sensor, capacitive sensor, Rogowski coil sensor, and electrostatic field meter are collected. Kalman filter state equations and observation equations are established based on the sensor data. Optimal estimates are obtained through iterative calculation based on the state equations and observation equations. A three-dimensional model of the electromagnetic field distribution within the warehouse is constructed based on the optimal estimates and the coordinates of the deployment points. If electromagnetic interference exceeding a preset threshold is detected in the three-dimensional electromagnetic field distribution model, an alarm is triggered, and the location information of the interference source is provided through a triangulation algorithm.

4. The method according to claim 1, wherein, The process involves transmitting real-time monitoring data collected by various electromagnetic sensors in the warehouse to the power grid dispatch center via a low-power wide-area Internet of Things (LPIoT) network, and correlating this data with warehouse material structure parameters to construct a database. This includes: acquiring raw electromagnetic sensor data collected by LoRa node devices; calculating the difference between adjacent sampling points using the DPCM algorithm based on the raw data; generating encrypted data packets using the AES-128 encryption algorithm for the difference information; transmitting the encrypted data packets to the power grid dispatch center via the LoRaWAN network; decrypting the encrypted data packets using a preset key to obtain the raw monitoring data; and constructing the database, including acquiring material structure parameter information and converting the raw monitoring data and the material structure parameter information into JSON format data. If the JSON data is structured material parameter information, it is stored in a MySQL database; if the JSON data is semi-structured monitoring data, it is stored in a MongoDB database.

5. The method according to claim 1, wherein, The process of extracting electromagnetic interference thresholds for various materials from the database, fusing monitoring data with theoretical distributions, determining the distribution of electromagnetic field strength within the warehouse, and identifying the types of materials and specific areas at risk of electromagnetic interference includes: obtaining electromagnetic interference threshold data for various materials from the database; establishing an electromagnetic interference threshold table based on material type and sensitivity, the electromagnetic interference threshold table including electric field strength thresholds and magnetic field strength thresholds; obtaining real-time monitoring data within the warehouse; and if the real-time monitoring data consists of discrete monitoring points, using ordinary kriging to spatially interpolate the discrete monitoring point data to obtain continuous electromagnetic field strength data. The electromagnetic field intensity distribution data is obtained by fusing the continuous electromagnetic field intensity distribution data with the theoretical distribution using the Markov chain Monte Carlo method. The fused electromagnetic field intensity distribution is then compared with the electromagnetic interference threshold table, and the C4.5 decision tree algorithm is used to identify regions exceeding the threshold. The minimum number of leaf node samples in the C4.5 decision tree algorithm is a preset sample value, and the maximum depth is a preset depth value. Historical data is updated using the sliding window method, and the electromagnetic interference threshold in the electromagnetic interference threshold table is adjusted according to the updated historical data. The time window of the sliding window method is a preset window value.

6. The method according to claim 1, wherein, If the electromagnetic field strength in a certain area is found to exceed the electromagnetic susceptibility threshold for the stored materials in that area, an early warning message is generated, and protective facilities are configured according to the actual electromagnetic field strength. Corresponding suggestions for adjusting protective measures are provided, including: using a K-means clustering algorithm to divide the warehouse space into regions, where the number of clusters is set to a preset value and Euclidean distance is used as a similarity metric to obtain an electromagnetic environment zoning map; based on the electromagnetic environment zoning map, the C4.5 decision tree algorithm is used to compare the electromagnetic field strength of each region with the preset electromagnetic susceptibility threshold for the stored materials to determine whether there is a threshold exceeding situation. If the judgment result shows that the electromagnetic field strength in a certain area exceeds the sensitivity threshold, an early warning is triggered, and a JSON-formatted early warning message containing the location of the area exceeding the standard, the electromagnetic field strength value, and the type of material is generated; based on the JSON-formatted early warning message, information on existing electromagnetic protection facilities in the area is obtained, and suggestions for adjusting protection measures are generated; for the protection measures adjustment suggestions, a cost-benefit analysis is performed on the protection measures, the return on investment is calculated, and a priority list is generated by sorting the protection measures in descending order of the return on investment.

7. The method according to claim 1, wherein, If the overall electromagnetic environment of the warehouse deteriorates, an early warning message will be sent to the power grid dispatch center to coordinate the dynamic adjustment of the operating parameters of the high-voltage transmission lines and reduce the electromagnetic interference level generated by the transmission lines to a safe range that various materials can withstand. This includes: acquiring warehouse electromagnetic environment data monitored by a preset distributed sensor network; calculating the spatiotemporal variation trend of electromagnetic field strength using a sliding window algorithm based on the data to obtain electromagnetic field strength variation data; and performing a quantitative evaluation of the electromagnetic field strength variation data. If the average electromagnetic field strength increases by a preset range or the peak value exceeds a preset multiple of the safety standard, it is determined that the electromagnetic environment has deteriorated, and an early warning data packet is generated. The warning data packet is encrypted using the SSL / TLS protocol to obtain an encrypted warning data packet, which is then sent to the power grid dispatch center. The power grid dispatch center receives encrypted early warning data packets, uses weighted least squares to estimate the power system state, and determines the key transmission lines and their operating parameters that affect the electromagnetic environment of the warehouse. For the operating parameters of the key transmission lines, a non-dominated sorting genetic algorithm is used for multi-objective optimization. Based on the optimized transmission line operating parameters, the transmission lines are dynamically adjusted.

8. The method according to claim 1, wherein, The process involves mining historical data on power grid load and electromagnetic environment to extract the correlation between electromagnetic interference characteristics and transmission line operating conditions, and establishing a collaborative prediction model for the electromagnetic safety status of multiple warehouses. This includes: acquiring historical data from power grid monitoring and warehouse electromagnetic environment monitoring, including power grid load, transmission line operating parameters, and warehouse electromagnetic field strength; processing outliers using the moving median method and missing values ​​using interpolation to obtain cleaned data; standardizing the cleaned data using the min-max normalization method to determine standardized data; and performing a 5-level multi-scale decomposition of electromagnetic interference signals in the standardized data using Daubechies wavelet to obtain time-domain and frequency-domain features. Principal component analysis is used to reduce the feature dimensionality based on the time-domain and frequency-domain features. The cumulative contribution rate is then assessed to determine if it reaches a preset range. If it does, an electromagnetic interference (EMI) feature vector is determined. The Apriori algorithm is used to analyze the correlation between the EMI feature vector and the transmission line operating parameters. Minimum support and minimum confidence are set, and their conditions are assessed. If these conditions are met, a strong correlation rule is determined, resulting in a correlation between EMI and the transmission line operating conditions. A graph convolutional network (GCN) is constructed based on this correlation. The GCN consists of two layers and uses ReLU as its activation function. The Adam optimizer is used to train the GCN, resulting in a multi-warehouse electromagnetic security situation collaborative prediction model. The prediction error is calculated based on the prediction results, and it is determined whether the prediction error exceeds a preset threshold. If it does, the warning threshold is adjusted to obtain a dynamic warning threshold.

9. The method according to claim 1, wherein, The process involves adjusting electromagnetic protection measures based on the prediction results of the prediction model, dynamically optimizing the storage area of ​​sensitive materials by adjusting warehouse hardware facilities, and simultaneously feeding back early warning information to the power grid dispatch center to coordinate and optimize the operation mode of transmission lines. This includes: acquiring predicted electromagnetic field intensity distribution results and electromagnetic susceptibility data of materials within the warehouse; calculating the optimal storage scheme for materials based on the prediction results and susceptibility data; adjusting the operating parameters of the electromagnetic shielding equipment within the warehouse via PLC, including the shielding plate angle and the thickness of the absorbing material; using electromagnetic field sensors to monitor the shielding effect in real time and forming a closed-loop feedback based on the monitoring results; if the monitoring results show that the shielding effect is substandard, triggering the PLC to adjust control parameters based on the feedback data; and sending early warning information to the power grid dispatch center via the IEC61850 protocol, the early warning information including the predicted electromagnetic exceedance area, time, and degree. The power grid dispatch center receives the early warning information and uses the NSGA-II algorithm to generate an optimized operation plan for the transmission lines. The optimized operation plan includes line load allocation and reactive power compensation device adjustment parameters.

Images