Power grid thunder and lightning early warning analysis method and system based on AI

Abstract

The invention relates to the technical field of power grid thunder early warning, in particular to a power grid thunder early warning analysis method and system based on AI, and the method comprises the following steps: based on real-time meteorological data, through continuously monitoring weather mode changes, recognizing thunder frequency information at key time points, and correspondingly adjusting a time window of data collection for each time point according to a recognition result; and matching the lightning activity change to generate a lightning time dynamic adjustment result. According to the invention, through real-time monitoring of meteorological data and dynamic adjustment of a time window of data acquisition, lightning early warning accuracy and response speed are significantly improved, continuous time sequence analysis is adopted, and adaptive capacity to rapid change of lightning activity is enhanced, so that an early warning system can be adjusted in real time to match dynamic change of lightning, and lightning early warning accuracy and response speed are improved. Through refined thunder and lightning trend prediction, thunder and lightning events can be predicted more accurately, a power grid operation strategy is optimized accordingly, and the safety, stability, power grid guarantee and user safety of a power grid are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid lightning warning, and in particular to an AI-based power grid lightning warning analysis method and system. Background Art

[0002] The field of power grid lightning warning technology involves the use of monitoring equipment and analysis methods to predict and warn of potential threats related to lightning activities to protect the stability and safety of the power system, centrally detect lightning generation conditions in the atmosphere, monitor the progress of lightning activities, and evaluate the impact that lightning may have on the power grid. Technical applications include the analysis of meteorological radar data, the deployment of electromagnetic sensors, and the use of big data analysis technology to predict the intensity, location, and movement trend of lightning in real time, which is crucial to maintaining the normal operation of the power grid, reducing power outages, and improving the responsiveness of the power grid system.

[0003] Among them, the power grid lightning warning analysis method specifically refers to the use of artificial intelligence (AI) and data analysis technology to extract relevant information from multi-source data to predict the occurrence of lightning events and issue warnings to power grid operators in real time. Its main purpose is to identify potential lightning risks in advance so that power grid operators can take preventive measures, such as adjusting grid loads, switching the operating status of key equipment, or temporarily disconnecting power to susceptible areas to minimize the damage caused by lightning to the power grid.

[0004] Existing technologies mainly rely on traditional weather radars and electromagnetic sensors that update data periodically, which are limited in capturing rapidly changing weather conditions in real time, making lightning warning systems slow to respond to rapidly evolving lightning event predictions. In addition, existing technologies also show obvious deficiencies in the accuracy of lightning intensity and location predictions, resulting in the inability to accurately deploy protective measures for the power grid, increasing the risk of failures and equipment damage. For example, if power grid nodes fail to adjust protection settings based on accurate predictions, they will be severely damaged during lightning strikes, causing widespread power outages and delays in restoration work, limiting the power grid's protection capabilities and emergency response speed, and increasing unnecessary economic losses and safety risks. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and propose an AI-based power grid lightning warning analysis method and system.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: an AI-based power grid lightning early warning analysis method, comprising the following steps:

[0007] S1: Based on real-time meteorological data, by continuously monitoring changes in weather patterns, identify frequent lightning information at key time points, adjust the data collection time window for each time point according to the identification results, match the changes in lightning activity, and generate dynamic adjustment results for lightning time;

[0008] S2: According to the dynamic adjustment result of the lightning time, the result is input into the dynamic time elastic network, and the network connection weight is adjusted in real time to match the dynamic characteristics of the lightning data. According to the adjustment result and the warning sensing threshold, the trend of lightning activity in the short term in the future is predicted, and the lightning intensity is analyzed at the same time to generate a comprehensive prediction result of the lightning trend;

[0009] S3: using the comprehensive lightning trend prediction results, evaluating the voltage and current data corresponding to the grid nodes, calculating and predicting the potential impact of lightning activities on the grid node loads, comparing the current grid load status with the calculation results, performing grid load fluctuation analysis, and generating real-time grid load analysis results;

[0010] S4: According to the real-time analysis result of the power grid load, the power grid load balance is adjusted, an emergency maintenance plan is formulated, and a power grid lightning emergency treatment plan is generated.

[0011] As a further solution of the present invention, the step of identifying the frequent lightning information at the key time point is:

[0012] S111: Access to weather stations and meteorological satellites to collect real-time temperature, humidity, air pressure and electric field strength data, integrate information through data fusion, and generate comprehensive meteorological data;

[0013] S112: Using the comprehensive meteorological data, continuously monitoring weather patterns using time series analysis, predicting weather change trends and abnormal patterns, and obtaining dynamic analysis results of weather patterns;

[0014] S113: According to the dynamic analysis result of the weather pattern, using the formula:

[0015]

[0016] Calculate and predict the probability of lightning occurrence P lightning , generate the frequent lightning information at the key time point, where I T ,I H ,I P ,I E Represent the current temperature index, humidity index, air pressure index and electric field index respectively, T mean , H mean , P mean 、E mean Corresponding to the historical average values ​​of temperature index, humidity index, air pressure index and electric field index, T std , H std , P std 、E std is the historical standard deviation of the corresponding index.

[0017] As a further solution of the present invention, the steps for obtaining the result of the dynamic adjustment of the lightning time are:

[0018] S121: Based on the lightning frequency information at the key time point, adjusting the data collection time window for the predicted high lightning frequency time period, optimizing the data collection frequency and time alignment, and generating a matched data collection time window;

[0019] S122: Synchronize the matched data collection time window to the monitoring site, automatically adjust the data collection frequency, accurately capture all key meteorological data at key time points, and obtain an updated data collection plan;

[0020] S123: Based on the updated data collection plan, continuously monitor and record meteorological data at key time points, conduct in-depth analysis on the collected data, and generate dynamic adjustment results of lightning time corresponding to each predicted lightning-frequent time point.

[0021] As a further solution of the present invention, the real-time adjustment step of the network connection weight is:

[0022] S211: Based on the dynamic adjustment result of the lightning time, input it into the dynamic time elastic network to reveal the time variation characteristics of the lightning data and generate the adjustment data after inputting into the network;

[0023] S212: According to the adjusted data after inputting into the network, the formula is adopted:

[0024]

[0025] Calculate and output the new network weight W new , where W old is the old weight, ΔT is the time interval, λ is the time sensitivity factor, e represents the natural logarithm base, and π is the circumference of a circle;

[0026] S213: Apply the new network weight, update the network parameter configuration, confirm that the dynamic time elastic network matches the latest dynamic characteristics of the lightning data, and obtain the network connection weight configuration.

[0027] As a further solution of the present invention, the steps for obtaining the comprehensive prediction result of lightning trend are:

[0028] S221: Based on the network connection weight configuration, compare with the set warning sensing threshold to determine whether the potential conditions for lightning occurrence are met, and generate a comparison result;

[0029] S222: Based on the comparison result, predict the trend of lightning activity in the short term in the future to obtain a lightning activity trend analysis result;

[0030] S223: Based on the lightning activity trend analysis results, using the formula:

[0031]

[0032] Calculation of lightning intensity I lightning , generate comprehensive lightning trend prediction results, where F freq Represents the frequency of lightning, Q charge represents the charge of lightning, and R represents the radius of the affected area.

[0033] As a further solution of the present invention, the calculation steps for predicting the potential impact of lightning activity on grid node loads are:

[0034] S311: using the lightning trend comprehensive prediction result, selecting a target power grid node, acquiring voltage and current data of the target node in real time, and generating a power grid node electrical parameter reading result;

[0035] S312: Based on the reading result of the electrical parameters of the power grid nodes, perform electrical load analysis, analyze the expected changes of voltage and current of the target node during the lightning activity, and generate an estimated result of the electrical parameter changes;

[0036] S313: Using the electrical parameter change estimation result, adopting the formula:

[0037]

[0038] The load change ΔL of the target node caused by lightning activity is calculated to generate the calculation result of the potential impact of the grid node load, where ΔV represents the absolute value of the voltage change and ΔI represents the absolute value of the current change.

[0039] As a further solution of the present invention, the steps of obtaining the real-time analysis result of the power grid load are:

[0040] S321: Obtain current grid node load data, including voltage, current, and key power consumption parameters, analyze the operation status of the grid, and generate real-time grid node load data;

[0041] S322: Compare the real-time grid node load data with the calculation result of the potential impact of the grid node load, determine the actual impact of lightning activity on the grid node load, capture load peaks and fluctuations, and generate a grid load state comparison analysis result;

[0042] S323: Using the grid load status comparison analysis results, combined with the grid historical load data, potential risk assessment is performed to analyze the overall grid load stability and the ability to cope with lightning impacts, and generate real-time grid load results.

[0043] As a further solution of the present invention, the steps for obtaining the power grid lightning emergency treatment plan are:

[0044] S411: using the real-time analysis result of the power grid load, evaluating the load status and load distribution of the current power grid nodes, identifying key nodes with abnormal loads, and obtaining abnormal node identification results;

[0045] S412: Based on the abnormal node identification result, the operating parameters of the transmission line and the transformer are adjusted to optimize the load distribution, using the formula:

[0046]

[0047] Calculate the ideal power output value P of the node after adjustment adj , get the output power adjustment result, P ex Represents the current power, P req represents the required power, and N represents the total number of affected nodes;

[0048] S413: Analyze the emergency situation according to the output power adjustment result, formulate a power grid emergency maintenance and response plan, and obtain a power grid lightning emergency treatment plan.

[0049] AI-based power grid lightning warning analysis system, including:

[0050] The meteorological data access module accesses meteorological stations and meteorological satellites, collects real-time data, integrates information through data fusion, continuously monitors weather patterns, predicts weather trends and abnormal patterns, calculates and predicts the possibility of lightning, and generates information on frequent lightning at key time points;

[0051] The data collection and adjustment module adjusts the data collection time window for the predicted high lightning frequency time period based on the lightning frequency information at the key time point, automatically adjusts the data collection frequency, performs in-depth analysis on the collected data, and generates a dynamic adjustment result of the lightning time corresponding to each predicted lightning frequency time point;

[0052] The dynamic network adjustment module inputs the dynamic adjustment result of the lightning time into the dynamic time elastic network, reveals the time variation characteristics of the lightning data, calculates and outputs the new network weight, updates the parameter configuration of the network, confirms that the dynamic time elastic network matches the latest dynamic characteristics of the lightning data, and obtains the network connection weight configuration;

[0053] The lightning warning analysis module compares the network connection weight configuration with the set warning sensing threshold, predicts the trend of lightning activity in the short term, calculates the lightning intensity, and generates a comprehensive prediction result of the lightning trend;

[0054] The grid node analysis module uses the comprehensive prediction results of lightning trends to select target grid nodes and obtain target node data in real time, analyze the expected changes in voltage and current of the target nodes during lightning activity, calculate the change in load of the target nodes due to lightning activity, and generate calculation results of potential impact of grid node load;

[0055] The power grid emergency strategy module uses the real-time analysis results of the power grid load to evaluate the load status and load distribution of the current power grid nodes, adjust the operating parameters of the transmission lines and transformers, analyze emergency situations, formulate power grid emergency maintenance and response plans, and obtain power grid lightning emergency treatment plans.

[0056] Compared with the prior art, the advantages and positive effects of the present invention are:

[0057] In the present invention, by real-time monitoring of meteorological data and dynamically adjusting the time window for data collection, the accuracy and response speed of lightning warning are significantly improved. Continuous time series analysis is used to enhance the adaptability to rapid changes in lightning activities, so that the warning system can be adjusted in real time to match the dynamic changes of lightning. Through refined lightning trend prediction, lightning events can be foreseen more accurately, and the power grid operation strategy can be optimized accordingly, greatly improving the safety, stability, security and user safety of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a main step flow chart of the present invention;

[0059] Figure 2 This is a flow chart for identifying frequent lightning information at key time points of the present invention;

[0060] Figure 3 This is a flow chart for obtaining the result of dynamic adjustment of lightning time according to the present invention;

[0061] Figure 4 A flowchart of real-time adjustment of network connection weights of the present invention;

[0062] Figure 5 This is a flow chart for obtaining the comprehensive prediction results of lightning trends of the present invention;

[0063] Figure 6 A calculation flow chart for predicting the potential impact of lightning activity on power grid node loads according to the present invention;

[0064] Figure 7 A flowchart for obtaining real-time analysis results of power grid loads according to the present invention;

[0065] Figure 8 The present invention is a flowchart for obtaining the power grid lightning emergency processing solution. DETAILED DESCRIPTION

[0066] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0067] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0068] See also Figure 1 , the AI-based power grid lightning warning analysis method includes the following steps:

[0069] S1: Based on real-time meteorological data, by continuously monitoring changes in weather patterns, identify frequent lightning information at key time points, adjust the data collection time window for each time point according to the identification results, match the changes in lightning activity, and generate dynamic adjustment results for lightning time;

[0070] S2: According to the dynamic adjustment results of lightning time, the results are input into the dynamic time elastic network, and the network connection weights are adjusted in real time to match the dynamic characteristics of lightning data. According to the adjustment results and the warning sensing threshold, the trend of lightning activity in the short term in the future is predicted, and the lightning intensity is analyzed at the same time to generate a comprehensive prediction result of lightning trend;

[0071] S3: Use the comprehensive prediction results of lightning trends to evaluate the voltage and current data corresponding to the grid nodes, calculate and predict the potential impact of lightning activities on the grid node load, compare the current grid load status based on the calculation results, perform grid load fluctuation analysis, and generate real-time grid load analysis results;

[0072] S4: According to the real-time analysis results of the power grid load, adjust the power grid load balance, formulate an emergency maintenance plan, and generate a power grid lightning emergency treatment plan.

[0073] The dynamic adjustment results of lightning time include the adjusted time window size, the adjusted collection parameters, and the monitored lightning frequency index. The comprehensive prediction results of lightning trends include the lightning location coordinates, the predicted occurrence time, and the estimated lightning intensity level. The real-time analysis results of power grid load include the voltage change value, current change value, and overall load fluctuation degree of the power grid nodes. The power grid lightning emergency handling plan includes the load redistribution plan, the emergency response measures list, and the preventive maintenance plan.

[0074] See also Figure 2 ,The steps for identifying frequent lightning information at key time points are:

[0075] S111: Access to weather stations and meteorological satellites to collect real-time temperature, humidity, air pressure and electric field strength data, integrate information through data fusion, and generate comprehensive meteorological data;

[0076] After accessing the meteorological station and satellite system, data is received from multiple sensors and satellite channels. The data include but are not limited to real-time temperature, humidity, air pressure and electric field strength. The received data undergoes preliminary formatting and verification processing to ensure the integrity and validity of the data. The processing process involves data denoising, missing value processing and outlier detection. Through data fusion technology, data from different sources and types are reasonably integrated, such as using weighted averaging and other methods to fuse sensor data, to ensure that the final generated comprehensive meteorological data can accurately reflect the real-time meteorological conditions, which will be used for further meteorological analysis and model prediction to support subsequent meteorological monitoring and alarm systems.

[0077] S112: Using comprehensive meteorological data, adopting time series analysis to continuously monitor weather patterns, predict weather change trends and abnormal patterns, and obtain dynamic analysis results of weather patterns;

[0078] Utilizing comprehensive meteorological data, time series analysis techniques are applied to continuously monitor weather patterns and predict changing trends. The process mainly relies on advanced data analysis software and algorithms, including autoregressive models and moving average models. The models can effectively process and analyze large-scale time series data, and identify short-term and long-term trends in weather patterns. The analysis also includes seasonal adjustments and trend smoothing of weather data. Through careful analysis, weather trends in the next few days or weeks can be predicted. The dynamic analysis results are not only based on historical data, but also combined with real-time meteorological observations. The resulting dynamic analysis results of weather patterns provide a scientific basis for future meteorological warnings. The final analysis results can help meteorological departments issue weather warnings in a timely manner and reduce the potential impact of natural disasters.

[0079] S113: Based on the results of the dynamic analysis of weather patterns, the formula is used.

[0080]

[0081] Calculate and predict the probability of lightning occurrence P lightning , generate the frequent lightning information at the key time point, where I T ,I H ,I P ,I E Represent the current temperature index, humidity index, air pressure index and electric field index respectively, T mean , H mean , P mean 、E mean Corresponding to the historical average values ​​of temperature index, humidity index, air pressure index and electric field index, T std , H std , P std 、E std is the historical standard deviation of the corresponding index;

[0082] I T =30℃, I H =70%, I P =1013hPa, I E =500V / m represents the current temperature index, humidity index, air pressure index and electric field index respectively, T mean =25℃,H mean =60%, P mean =1010hPa, E mean =300V / m is the historical average value of the index, T std =5℃,H std =10%, P std =5hPa, E std =150V / m is the historical standard deviation of the corresponding index. The calculation formula is:

[0083]

[0084] The results show that under given conditions, the probability index of lightning occurrence is 1.98, which indicates the risk of lightning occurrence under current meteorological conditions (the larger the value, the higher the risk), which has a direct impact on the early warning system.

[0085] See also Figure 3 , the steps to obtain the results of dynamic adjustment of lightning time are:

[0086] S121: Based on the lightning frequency information at the key time points, the data collection time window for the predicted high lightning frequency time period is adjusted to optimize the data collection frequency and time alignment, and generate a matched data collection time window;

[0087] Based on the information on frequent lightning occurrence at key time points, the data collection time window is adjusted to adapt to the predicted high lightning frequency period. The adjustment process is carried out by analyzing the time distribution of historical lightning events. The purpose of optimization is to increase the frequency and accuracy of data collection during the predicted high incidence period. The adjustment includes reducing the data collection frequency during times with less lightning activity and increasing the collection frequency and the density of data collection points during periods of predicted high lightning activity. The adjustment strategy enables the data collection system to capture more key data at critical moments, thereby providing more accurate real-time data required for meteorological analysis and forecasting.

[0088] S122: Synchronize the matched data collection time window to the monitoring site, automatically adjust the data collection frequency, accurately capture all key meteorological data at key time points, and obtain an updated data collection plan;

[0089] Once the data collection time window plan is determined, the plan is synchronized to all relevant meteorological monitoring sites, and the monitoring system of the site is updated to automatically adjust the data collection frequency and parameters to ensure that all important meteorological data at key time points can be captured. During this process, the monitoring system will automatically adjust the settings of the collection equipment according to the latest lightning activity forecast, such as the sensitivity and collection frequency of the sensor, to adapt to the predicted high-incidence period of lightning. At the same time, the system ensures that all monitoring sites can receive the latest adjustment instructions and execute them accurately through the real-time communication network. This dynamic adjustment can maximize monitoring efficiency while reducing analysis delays that may be caused by data overload.

[0090] S123: Based on the updated data collection plan, continuously monitor and record meteorological data at key time points, conduct in-depth analysis on the collected data, and generate dynamic adjustment results of lightning time corresponding to each predicted lightning-frequent time point;

[0091] By implementing the updated data collection plan, the monitoring system continuously monitors and records detailed meteorological data at key time points. The data is processed through advanced trend analysis and pattern recognition tools to generate dynamic adjustment results for each predicted lightning-frequent time point. Data processing includes real-time stream analysis of data and historical data comparison to identify upcoming lightning patterns and predict their intensity and duration, which not only improves the response speed of the early warning system, but also enhances the reliability and accuracy of the system under extreme weather conditions.

[0092] See also Figure 4 , the real-time adjustment steps of network connection weights are:

[0093] S211: Based on the dynamic adjustment result of lightning time, input it into the dynamic time elastic network to reveal the time variation characteristics of lightning data and generate adjustment data after inputting into the network;

[0094] Receive the results of dynamic adjustment of lightning time, collect, correct and screen data, monitor lightning activities in real time through the sensor network installed in high-risk lightning areas, the sensor records the time and location of each lightning strike, and transmits it to the data center via wireless signals. The data center performs preliminary noise removal and outlier detection on the received raw data to ensure the accuracy and availability of the data, and establishes a preprocessing model based on the data history records of the past year. The newly collected data is corrected in real time through this model to correct data errors that may be caused by sensor deviations or environmental factors, and finally a set of corrected high-quality data is formed. These data will be directly input into the dynamic time elastic network for subsequent weight adjustment and network training.

[0095] S212: Based on the adjusted data after inputting the network, the formula is used.

[0096]

[0097] Calculate and output the new network weight W new , where W old is the old weight, ΔT is the time interval, λ is the time sensitivity factor, e represents the natural logarithm base, and π is the circumference of a circle;

[0098] W old represents the previous weight value of the network, which is set to 0.5. ΔT represents the time interval between two data updates, which is 5 seconds. λ is a constant that describes the time decay rate. It is obtained through the analysis of lightning activity data in the past year and its value is 0.1. e is the base of the natural logarithm, which is approximately equal to 2.718. π is the circumference of a circle, which is approximately 3.1416. The calculation process of the formula is as follows: First, calculate e -λ·ΔT The value of e -0.1·5 ≈e -0.5 ≈0.606, then calculate |W old ·e -λ·ΔT |, that is, |0.5×0.606|=0.303, then calculate the denominator Right now Finally, substituting these two results into the formula, we get This result shows that under the given parameter settings, the new weight W new The substantial reduction reflects the adjustment of the network’s time sensitivity and illustrates the time-decay property of the network weights, which helps the network adapt to the dynamically changing environment and thus improves the accuracy of predictions.

[0099] S213: Apply the new network weight, update the network parameter configuration, confirm that the dynamic time elastic network matches the latest dynamic characteristics of the lightning data, and obtain the network connection weight configuration;

[0100] The process of updating the network parameter configuration includes parameter optimization and model verification. According to the newly calculated network weights, the network's learning rate and the threshold of the activation function are adjusted. The back-propagation algorithm is used to fine-tune the weight of each connection in the network to ensure that each adjustment is made in the direction of reducing the prediction error. At the same time, the performance of the model on the new data set is tested by cross-validation. The verification method includes dividing the collected lightning data into training sets and test sets, training the network with the training set data, and then testing the network's prediction performance with the test set data to ensure the generalization ability of the model. In addition, it is necessary to monitor the performance of the model under different conditions, such as different seasons and different time periods. Through these detailed parameter settings and tests, a set of network parameter configurations that can accurately match the dynamic characteristics of lightning data is finally obtained. The configuration will be applied to the actual lightning prediction system to improve the system's prediction accuracy of lightning activity trends.

[0101] See also Figure 5 , the steps to obtain the comprehensive prediction results of lightning trends are:

[0102] S221: Based on the network connection weight configuration, compare with the set warning sensing threshold to determine whether the potential conditions for lightning occurrence are met and generate a comparison result;

[0103] In the monitoring system, the comparison of the real-time updated network weight results with the predetermined warning sensing threshold is a key operation. The threshold is set based on the statistical analysis of the lightning activity data of the past five years and represents the minimum network weight when lightning may occur. The implementation of this comparison involves a complex data processing process, including the real-time synchronization of lightning data collected from sensors in multiple geographical locations, as well as the preliminary cleaning and standardization of the data to ensure that the data input to the network is accurate and reliable. In addition, the system will adjust the sensitivity of the threshold based on the historical data model to adapt to climate change or changes in regional characteristics. If the current network weight exceeds the warning threshold, the system will automatically start the warning procedure. The response of the warning system includes sending immediate notifications to relevant departments and launching emergency plans to ensure that all relevant parties can be notified in time when the lightning risk increases, so that appropriate safety measures can be taken to prevent possible safety accidents.

[0104] S222: Based on the comparison result, the trend of lightning activity in the short term in the future is predicted to obtain the analysis result of the lightning activity trend;

[0105] Based on the comparison results, a special prediction for short-term lightning is made using a support vector machine. The input data processed by this model includes meteorological data collected in real time from meteorological stations and satellite systems, such as temperature, humidity, atmospheric pressure, and weather change data in the past 24 hours. In addition, the model also integrates real-time meteorological information from ground observation stations. All these data are pre-processed to filter out noise and outliers, and then used to train the model to predict the trend of lightning activity in the next 24 to 48 hours. The model accurately predicts the probability of future lightning by analyzing the correlation between short-term meteorological data changes and historical lightning activity. The prediction results include not only whether lightning will occur, but also the specific time and place where it is expected to occur. This will provide a scientific basis for emergency management departments so that they can more effectively formulate early warning strategies and prepare necessary safety measures to reduce property and personnel losses that may be caused by lightning.

[0106] S223: Based on the results of lightning activity trend analysis, the formula is used.

[0107]

[0108] Calculation of lightning intensity I lightning , generate comprehensive lightning trend prediction results, where F freq Represents the frequency of lightning, Q charge represents the charge of lightning, and R represents the radius of the affected area;

[0109] In a specific monitoring area, F freq 0.8 times / hour, Q charge is 30C (Coulomb), R is 5km. First calculate the molecule F freq Q charge The value is 0.8×30=24C / hour. Then calculate the denominator π·R 2 , that is 3.14159·(5000 2 )≈78539816.5m 2 Substitute the value into the formula to calculate the lightning intensity I lightning ,Right now The results show that the intensity of lightning is very low, which means that even if lightning occurs, its impact on the ground is extremely limited, which helps decision makers take appropriate preventive measures for lightning activities of different intensities.

[0110] See also Figure 6 , the calculation steps for predicting the potential impact of lightning activity on grid node loads are:

[0111] S311: using the lightning trend comprehensive prediction result, selecting the target grid node, acquiring the voltage and current data of the target node in real time, and generating the grid node electrical parameter reading result;

[0112] Using the comprehensive prediction results of lightning trends, the target power grid node is selected, and the voltage and current data of the node are obtained in real time. In this process, the comprehensive prediction results of lightning trends are provided by the meteorological monitoring system, which involves big data analysis and pattern recognition technology. By analyzing past lightning activity data, the probability and intensity of future lightning are predicted. The process of selecting power grid nodes depends on the real-time monitoring system of the power grid, which can capture the instantaneous load and operating status of each node to ensure that the selected node is the node most likely to be affected by lightning. Real-time acquisition of voltage and current data is completed through sensors installed on the power grid nodes. The sensors can accurately read the changes in current and voltage, and transmit the data to the data processing center in real time to generate the electrical parameter reading results of the power grid nodes.

[0113] S312: Based on the reading results of the electrical parameters of the power grid nodes, perform electrical load analysis, analyze the expected changes of the voltage and current of the target node during the lightning activity, and generate an estimated result of the electrical parameter changes;

[0114] Based on the reading results of the electrical parameters of the power grid nodes, electrical load analysis is performed to determine the expected changes in voltage and current of the power grid nodes during the period. During this process, the electrical load analysis is adjusted according to the characteristics of the power grid nodes and historical data to adapt to different types of power grid nodes and lightning conditions. The expected changes in voltage and current are calculated by simulating the response of the power grid during lightning activity. The data of historical lightning events are used as input, and the performance of the power grid nodes in future lightning activities is predicted through physical models and the principles of electrical engineering. This prediction includes estimating the peak values ​​that the current and voltage of the power grid nodes may reach, as well as the possible fluctuation range, and generating electrical parameter change prediction results to provide decision support for the final risk assessment and response measures.

[0115] S313: Using the estimated results of the electrical parameter changes, the formula is used.

[0116]

[0117] Calculate the load change ΔL of the target node caused by lightning activity, and generate the calculation result of the potential impact of the grid node load, where ΔV represents the absolute value of the voltage change and ΔI represents the absolute value of the current change;

[0118] In a certain monitoring, ΔV was measured as 15 volts and ΔI was measured as 10 amperes. According to the formula, the calculation process is as follows:

[0119]

[0120] The results show that under the influence of lightning, the load change of the power grid node is 30 units. This change reflects the stability and risk level of the node under current and voltage fluctuations, and has important reference value for the safety of power grid operation.

[0121] See also Figure 7 , the steps for obtaining the real-time analysis results of the power grid load are:

[0122] S321: Obtain current grid node load data, including voltage, current, and key power consumption parameters, analyze the operation status of the grid, and generate real-time grid node load data;

[0123] Real-time acquisition of load data of power grid nodes includes key indicators such as voltage, current and power consumption. In this process, high-precision sensors distributed at power grid nodes are used for data collection. The sensors are designed to work stably under various climatic conditions and monitor the load changes of the power grid in real time. The data collected by the sensors are sent to the central data processing center via an encrypted wireless network. The data processing center can analyze and record the voltage and current data of each node in real time. In addition, through real-time comparative analysis, any abnormal fluctuations or possible power grid failures can be detected to ensure the accuracy and reliability of the data. Through advanced monitoring and analysis technology, detailed real-time power grid node load data results can be generated, providing important decision-making support information for power grid operation and maintenance.

[0124] S322: Compare the real-time grid node load data with the calculation results of the potential impact of the grid node load, determine the actual impact of lightning activities on the grid node load, capture the load peak and fluctuation, and generate a grid load state comparison analysis result;

[0125] The comparison and analysis of the real-time load data results of the power grid nodes and the lightning impact prediction results mainly rely on customized data analysis software that can handle large-scale data sets and realize complex statistical analysis. During the analysis process, the collected real-time load data and historical data are first standardized, and then the lightning impact prediction model is used to predict the potential load changes under lightning weather. Through comparative analysis, it is possible to identify the voltage and current peaks that may be caused under lightning conditions, as well as potential threats to the stability of the power grid. In addition, the analysis also includes an assessment of the load response speed and recovery capacity of the power grid nodes so that the power grid operation and maintenance team can respond quickly to possible power outages or failures. The generated grid load status comparative analysis results provide a scientific basis for power grid risk assessment and the formulation of emergency response plans.

[0126] S323: Using the grid load status comparison analysis results, combined with the grid historical load data, potential risk assessment is performed to analyze the overall grid load stability and the ability to cope with lightning impacts, and generate real-time grid load results;

[0127] The results of the grid load status comparison analysis are comprehensively utilized, combined with historical load data, and in-depth analysis is performed through a comprehensive data analysis platform. The platform introduces artificial intelligence technology to improve the accuracy and efficiency of the analysis. In the analysis, the historical load data is first pattern recognized to determine the general trend of load changes, and then the trend is compared with the real-time data to identify fluctuation patterns that deviate from the general trend. In addition, the platform also uses machine learning models to predict future load fluctuations and provide early warning for the operation and maintenance of the power grid. This comprehensive analysis helps the power grid operation team optimize the load management strategy and improve the overall performance and reliability of the power grid. The real-time analysis results of the power grid load generated provide real-time and comprehensive data support for decision makers.

[0128] See also Figure 8 , the steps to obtain the power grid lightning emergency treatment plan are:

[0129] S411: using the real-time analysis results of the power grid load, evaluating the load status and load distribution of the current power grid nodes, identifying key nodes with abnormal loads, and obtaining abnormal node identification results;

[0130] When evaluating the load status of power grid nodes, we first collect the current and voltage data of each node to generate a real-time power grid status table, which involves preliminary processing and classification of the data. The data is transmitted in real time from each monitoring point of the power grid and quality controlled by the data processing center to ensure the accuracy and availability of the data. Next, the power grid nodes are partitioned, and each partition includes several power grid nodes. By analyzing the load index of each node, which reflects the current power usage of the node, the load comparison between nodes can find nodes with abnormal loads. The load capacity of the node is analyzed according to the current and voltage data of the node, and whether the node is overloaded or underloaded is analyzed, which provides a basis for the next step of load adjustment and optimization.

[0131] S412: Based on the abnormal node identification results, adjust the operating parameters of the transmission line and transformer, optimize the load distribution, and use the formula,

[0132]

[0133] Calculate the ideal power output value P of the node after adjustment adj , get the output power adjustment result, P ex Represents the current power, P req represents the required power, and N represents the total number of affected nodes;

[0134] In the power grid system, there are 5 nodes in an area, and the actual power and power demand of each node are P ex =

[0135] [100, 120, 90, 110, 105] kW, P req =[95, 115, 85, 105, 100] kilowatts. Substituting the values ​​into the formula, the calculation process is as follows:

[0136]

[0137] It means that after the adjustment, the output of each node will increase by 5 kilowatts on average to meet the power demand in the area, thereby achieving a more balanced load distribution.

[0138] S413: Analyze the emergency situation according to the output power adjustment result, formulate an emergency maintenance and response plan for the power grid, and obtain an emergency treatment plan for lightning in the power grid;

[0139] Based on the results of the load adjustment, the emergency maintenance and response plan of the power grid is updated, including formulating specific response strategies based on extreme weather conditions such as lightning, for example, switching to the emergency power supply system in advance, or adjusting the load of key nodes to prevent overload of power equipment. The emergency maintenance plan also includes regular inspections of key parts of the power grid, such as transformers and transmission lines, as well as strengthening monitoring of the overall security of the power grid, formulating detailed maintenance schedules and preventive measures to ensure the stable operation of the power grid when extreme weather occurs. These measures will be integrated into the daily operation and maintenance strategy of the power grid to enhance the power grid's resistance to potential threats in the future.

[0140] AI-based power grid lightning warning analysis system, including:

[0141] The meteorological data access module accesses meteorological stations and meteorological satellites, collects real-time data, integrates information through data fusion, continuously monitors weather patterns, predicts weather trends and abnormal patterns, calculates and predicts the possibility of lightning, and generates information on frequent lightning at key time points;

[0142] The data collection and adjustment module adjusts the data collection time window for the predicted high lightning frequency time period based on the lightning frequency information at key time points, automatically adjusts the data collection frequency, conducts in-depth analysis on the collected data, and generates dynamic adjustment results for lightning time corresponding to each predicted lightning frequency time point;

[0143] The dynamic network adjustment module is based on the dynamic adjustment results of lightning time, input into the dynamic time elastic network, reveal the time variation characteristics of lightning data, calculate and output new network weights, update the network parameter configuration, confirm that the dynamic time elastic network matches the latest dynamic characteristics of lightning data, and obtain the network connection weight configuration;

[0144] The lightning warning analysis module compares the network connection weight configuration with the set warning sensing threshold, predicts the trend of lightning activity in the short term, calculates the lightning intensity, and generates a comprehensive prediction result of the lightning trend;

[0145] The grid node analysis module uses the comprehensive prediction results of lightning trends to select target grid nodes and obtain target node data in real time, analyze the expected changes in voltage and current of target nodes during lightning activity, calculate the change in load of target nodes caused by lightning activity, and generate calculation results of potential impact of grid node load;

[0146] The power grid emergency strategy module uses the real-time analysis results of the power grid load to evaluate the load status and load distribution of the current power grid nodes, adjust the operating parameters of the transmission lines and transformers, analyze emergency situations, formulate power grid emergency maintenance and response plans, and obtain power grid lightning emergency treatment plans.

[0147] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. The AI-based power grid lightning warning analysis method is characterized by: The following steps are involved: Based on real-time meteorological data, by continuously monitoring changes in weather patterns, we can identify frequent lightning occurrences at key time points, adjust the data collection time window for each time point according to the identification results, match the changes in lightning activity, and generate dynamic adjustment results for lightning time; According to the dynamic adjustment result of the lightning time, the result is input into the dynamic time elastic network, the network connection weight is adjusted in real time to match the dynamic characteristics of the lightning data, and the trend of lightning activity in the short term in the future is predicted according to the adjustment result combined with the warning sensing threshold, and the lightning intensity is analyzed at the same time to generate a comprehensive prediction result of the lightning trend; Using the comprehensive lightning trend prediction results, evaluate the voltage and current data corresponding to the power grid nodes, calculate and predict the potential impact of lightning activities on the load of the power grid nodes, compare the current power grid load status according to the calculation results, perform power grid load fluctuation analysis, and generate real-time power grid load analysis results; According to the real-time analysis result of the power grid load, the power grid load balance is adjusted, an emergency maintenance plan is formulated, and a power grid lightning emergency treatment plan is generated.

2. The AI-based power grid lightning early warning analysis method according to claim 1 is characterized in that: The steps for identifying the frequent lightning information at the critical time point are: Access to weather stations and meteorological satellites to collect real-time temperature, humidity, air pressure and electric field strength data, integrate information through data fusion, and generate comprehensive meteorological data; Using the comprehensive meteorological data, continuously monitoring weather patterns using time series analysis, predicting weather change trends and abnormal patterns, and obtaining dynamic analysis results of weather patterns; According to the dynamic analysis results of the weather pattern, the formula is used: Calculate and predict the probability of lightning occurrence P lightning , generate the frequent lightning information at the key time point, where I T ,I H ,I P ,I E Represent the current temperature index, humidity index, air pressure index and electric field index respectively, T mean , H mean , P mean 、E mean Corresponding to the historical average values ​​of temperature index, humidity index, air pressure index and electric field index, T std , H std , P std 、E std is the historical standard deviation of the corresponding index.

3. The AI-based power grid lightning early warning analysis method according to claim 2 is characterized in that: The steps for obtaining the dynamic adjustment result of the lightning time are as follows: Based on the lightning frequency information at the key time points, the data collection time window for the predicted high lightning frequency time period is adjusted to optimize the data collection frequency and time alignment, and generate a matched data collection time window; Synchronize the matched data collection time window to the monitoring site, automatically adjust the data collection frequency, accurately capture all key meteorological data at key time points, and obtain an updated data collection plan; Based on the updated data collection plan, the meteorological data at key time points are continuously monitored and recorded, and the collected data are deeply analyzed to generate dynamic adjustment results of lightning time corresponding to each predicted lightning-frequent time point.

4. The AI-based power grid lightning early warning analysis method according to claim 3 is characterized in that: The real-time adjustment steps of the network connection weight are: Based on the dynamic adjustment result of lightning time, the result is input into the dynamic time elastic network to reveal the time variation characteristics of lightning data and generate adjustment data after input into the network; According to the adjusted data after inputting the network, the formula is adopted, Calculate and output the new network weight W new , where W old is the old weight, ΔT is the time interval, λ is the time sensitivity factor, e represents the natural logarithm base, and π is the circumference of a circle; The new network weight is applied to update the parameter configuration of the network, confirm that the dynamic time elastic network matches the latest dynamic characteristics of the lightning data, and obtain the network connection weight configuration.

5. The AI-based power grid lightning early warning analysis method according to claim 4 is characterized in that: The steps for obtaining the comprehensive prediction result of lightning trend are as follows: Based on the network connection weight configuration, the network connection weight configuration is compared with a set warning sensing threshold to determine whether a potential condition for lightning occurrence is met and generate a comparison result; Based on the comparison result, the trend of lightning activity in the short term in the future is predicted to obtain a lightning activity trend analysis result; According to the analysis results of the lightning activity trend, the formula is used: Calculation of lightning intensity I lightning , generate comprehensive lightning trend prediction results, where F freq Represents the frequency of lightning, Q charge represents the charge of lightning, and R represents the radius of the affected area.

6. The AI-based power grid lightning early warning analysis method according to claim 5 is characterized in that: The calculation steps for predicting the potential impact of lightning activity on grid node load are: Using the lightning trend comprehensive prediction results, select the target power grid node, obtain the voltage and current data of the target node in real time, and generate the power grid node electrical parameter reading results; Based on the electrical parameter reading results of the power grid nodes, an electrical load analysis is performed to analyze the expected changes in voltage and current of the target nodes during lightning activity, and an electrical parameter change prediction result is generated; Using the electrical parameter change estimation results, the formula is adopted: The load change ΔL of the target node caused by lightning activity is calculated to generate the calculation result of the potential impact of the grid node load, where ΔV represents the absolute value of the voltage change and ΔI represents the absolute value of the current change.

7. The AI-based power grid lightning early warning analysis method according to claim 6 is characterized in that: The steps for obtaining the real-time analysis result of the power grid load are as follows: Obtain current grid node load data, including voltage, current, and key power consumption parameters, analyze the operation status of the grid, and generate real-time grid node load data; Comparing the real-time grid node load data with the calculation result of the potential impact of the grid node load, determining the actual impact of lightning activity on the grid node load, capturing load peaks and fluctuations, and generating grid load status comparison analysis results; The grid load status comparison and analysis results are used in combination with the grid historical load data to perform potential risk assessment, analyze the overall grid load stability and the ability to cope with lightning impacts, and generate real-time grid load results.

8. The AI-based power grid lightning early warning analysis method according to claim 7 is characterized in that: The steps for obtaining the power grid lightning emergency treatment plan are: Using the real-time analysis results of the power grid load, evaluate the load status and load distribution of the current power grid nodes, identify key nodes with abnormal loads, and obtain abnormal node identification results; Based on the abnormal node identification results, the operating parameters of the transmission line and transformer are adjusted to optimize the load distribution. The formula is used. Calculate the ideal power output value P of the node after adjustment adj , get the output power adjustment result, P ex Represents the current power, P req represents the required power, and N represents the total number of affected nodes; According to the output power adjustment result, the emergency situation is analyzed, an emergency maintenance and response plan for the power grid is formulated, and an emergency treatment plan for lightning in the power grid is obtained.

9. The AI-based power grid lightning warning analysis system is characterized by: The system is used to execute the AI-based power grid lightning early warning analysis method according to any one of claims 1 to 7, comprising: The meteorological data access module accesses meteorological stations and meteorological satellites, collects real-time data, integrates information through data fusion, continuously monitors weather patterns, predicts weather trends and abnormal patterns, calculates and predicts the possibility of lightning, and generates information on frequent lightning at key time points; The data collection and adjustment module adjusts the data collection time window for the predicted high lightning frequency time period based on the lightning frequency information at the key time point, automatically adjusts the data collection frequency, performs in-depth analysis on the collected data, and generates a dynamic adjustment result of the lightning time corresponding to each predicted lightning frequency time point; The dynamic network adjustment module inputs the dynamic adjustment result of the lightning time into the dynamic time elastic network, reveals the time variation characteristics of the lightning data, calculates and outputs the new network weight, updates the parameter configuration of the network, confirms that the dynamic time elastic network matches the latest dynamic characteristics of the lightning data, and obtains the network connection weight configuration; The lightning warning analysis module compares the network connection weight configuration with the set warning sensing threshold, predicts the trend of lightning activity in the short term, calculates the lightning intensity, and generates a comprehensive prediction result of the lightning trend; The grid node analysis module uses the comprehensive prediction results of lightning trends to select target grid nodes and obtain target node data in real time, analyze the expected changes in voltage and current of the target nodes during lightning activity, calculate the change in load of the target nodes due to lightning activity, and generate calculation results of potential impact of grid node load; The power grid emergency strategy module uses the real-time analysis results of the power grid load to evaluate the load status and load distribution of the current power grid nodes, adjust the operating parameters of the transmission lines and transformers, analyze emergency situations, formulate power grid emergency maintenance and response plans, and obtain power grid lightning emergency treatment plans.

Images