Power grid abnormal event prediction and early warning method and device based on big data analysis

Abstract

The invention relates to a power grid abnormal event prediction and early warning method and device based on big data analysis. The method comprises the following steps: acquiring power grid operation data, external environment data and equipment state data of a target power grid; performing abnormal event analysis on the power grid operation data, the external environment data and the equipment state data, and identifying each abnormal event key factor of the target power grid; performing time sequence analysis on the power grid operation data, the external environment data and the equipment state data according to each abnormal event key factor, and predicting abnormal event data and an abnormal event triggering probability of the target power grid; and generating power grid early warning information according to the abnormal event data and the abnormal event triggering probability. By adopting the method, a prediction result obtained under the condition of predicting complex or atypical faults is more real-time and accurate, and operation and maintenance personnel are helped to pre-judge potential risks, so that the self-healing capability of a power grid is improved, and the reliability and efficiency of power grid operation are greatly improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of smart grid, in particular to a power grid abnormal event prediction and early warning method and device based on big data analysis. BACKGROUND

[0002] With the development of power grid technology, power grid abnormal event prediction technology has emerged, which refers to identifying and predicting possible abnormal events or faults in advance by analyzing various data in power grid operation. The purpose is to discover potential problems in time during power grid operation, prevent sudden power outages, equipment damage and other accidents, and improve the stability and efficiency of the power grid. This prediction can provide important reference for the maintenance and optimization of the power system, helping grid managers to better develop emergency plans and protection measures.

[0003] In traditional technology, power grid abnormal event prediction mainly relies on rules and threshold methods based on experience. Specifically, power grid operating parameters (such as voltage, current, frequency, etc.) will set a fixed normal range, and once the monitoring data exceeds these preset thresholds, the system will judge it as abnormal and issue an alarm. These thresholds are usually set by experts according to historical experience or equipment specifications, combined with simple statistical analysis tools, to monitor the operating state of the power grid. However, this method cannot dynamically adapt to changes in the power grid environment, resulting in inaccurate prediction of complex or atypical faults. SUMMARY

[0004] Therefore, it is necessary to provide a power grid abnormal event prediction and early warning method, device, computer equipment, computer readable storage medium and computer program product based on big data analysis, which can accurately predict complex or atypical faults.

[0005] In a first aspect, the present application provides a power grid abnormal event prediction and early warning method based on big data analysis, comprising:

[0006] Obtaining power grid operating data, external environment data and equipment state data of a target power grid;

[0007] Performing abnormal event analysis on the power grid operating data, the external environment data and the equipment state data to identify key factors of each abnormal event of the target power grid;

[0008] According to each abnormal event key factor, performing time series analysis on the power grid operating data, the external environment data and the equipment state data to predict abnormal event data and abnormal event trigger probability of the target power grid;

[0009] Generating power grid early warning information according to the abnormal event data and the abnormal event trigger probability.

[0010] Secondly, the present application also provides a power grid abnormal event prediction and early warning device based on big data analysis, including:

[0011] A data acquisition module is used to obtain grid operation data, external environment data, and equipment status data of the target grid;

[0012] a data analysis module, configured to perform abnormal event analysis on the power grid operation data, the external environment data, and the device status data, and identify key factors of abnormal events in the target power grid;

[0013] An abnormality prediction module is used to perform time series analysis on the power grid operation data, the external environment data and the device status data according to each of the key factors of the abnormal events, and predict the abnormal event data and the abnormal event triggering probability of the target power grid;

[0014] The power grid early warning module is used to generate power grid early warning information based on the abnormal event data and the abnormal event triggering probability.

[0015] In a third aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0016] Obtain grid operation data, external environment data, and equipment status data of the target grid;

[0017] Performing abnormal event analysis on the power grid operation data, the external environment data, and the device status data to identify key factors of abnormal events in the target power grid;

[0018] According to each of the key factors of the abnormal events, a time series analysis is performed on the power grid operation data, the external environment data, and the device status data to predict the abnormal event data of the target power grid and the probability of triggering the abnormal event;

[0019] Power grid warning information is generated based on the abnormal event data and the abnormal event triggering probability.

[0020] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0021] Obtain grid operation data, external environment data, and equipment status data of the target grid;

[0022] Performing abnormal event analysis on the power grid operation data, the external environment data, and the device status data to identify key factors of abnormal events in the target power grid;

[0023] According to each of the abnormal event key factors, time series analysis is performed on the power grid operation data, the external environment data and the equipment state data to predict abnormal event data and abnormal event triggering probability of the target power grid;

[0024] According to the abnormal event data and the abnormal event triggering probability, power grid early warning information is generated.

[0025] In a fifth aspect, the present application further provides a computer program product comprising a computer program which, when executed by a processor, implements the following steps:

[0026] Obtaining power grid operation data, external environment data and equipment state data of a target power grid;

[0027] Performing abnormal event analysis on the power grid operation data, the external environment data and the equipment state data to identify each abnormal event key factor of the target power grid;

[0028] According to each of the abnormal event key factors, time series analysis is performed on the power grid operation data, the external environment data and the equipment state data to predict abnormal event data and abnormal event triggering probability of the target power grid;

[0029] According to the abnormal event data and the abnormal event triggering probability, power grid early warning information is generated.

[0030] The above-mentioned power grid abnormal event prediction and early warning method, device, computer equipment, storage medium and computer program product based on big data analysis, by obtaining power grid operation data, external environment data and equipment state data of a target power grid, performing abnormal event analysis on the power grid operation data, the external environment data and the equipment state data to identify each abnormal event key factor of the target power grid, according to each of the abnormal event key factors, time series analysis is performed on the power grid operation data, the external environment data and the equipment state data to predict abnormal event data and abnormal event triggering probability of the target power grid, and according to the abnormal event data and the abnormal event triggering probability, power grid early warning information is generated.

[0031] By comprehensively acquiring power grid operation data, external environment data and equipment state data, combined with abnormal event analysis, the key factors affecting the safe operation of the power grid can be accurately identified. Through time series analysis, the system can not only predict abnormal events of the power grid, but also quantify the probability of event occurrence, dynamically considering the changing trend of the power grid under different external environments and equipment states. Based on these prediction results, the system can generate early warning information in advance, so that the prediction results obtained in the case of predicting complex or atypical faults are more real-time and accurate, and help operation and maintenance personnel to predict potential risks, thereby improving the self-healing ability of the power grid, optimizing maintenance and dispatching strategies, reducing the possibility of fault outage, and greatly improving the reliability and efficiency of power grid operation. BRIEF DESCRIPTION OF DRAWINGS

[0032] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the drawings needed to be used in the embodiments or related art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0033] Figure 1 An application environment diagram of the power grid abnormal event prediction and early warning method based on big data analysis in an embodiment;

[0034] Figure 2 A flowchart of the power grid abnormal event prediction and early warning method based on big data analysis in an embodiment;

[0035] Figure 3 A flowchart of the abnormal event key factor identification method in an embodiment;

[0036] Figure 4 A flowchart of the abnormal event key factor identification method in another embodiment;

[0037] Figure 5 A flowchart of the first abnormal event data and abnormal event trigger probability prediction method in an embodiment;

[0038] Figure 6 A flowchart of the second abnormal event data and abnormal event trigger probability prediction method in an embodiment;

[0039] Figure 7 A flowchart of the third abnormal event data and abnormal event trigger probability method in an embodiment;

[0040] Figure 8 A structural block diagram of the power grid abnormal event prediction and early warning device based on big data analysis in an embodiment;

[0041] Figure 9 Figure 1 is a schematic diagram of the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0042] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and should not be used to limit the present application.

[0043] The power grid abnormal event prediction and early warning method based on big data analysis provided by the embodiments of the present application can be applied to the application environment as shown in Figure 1 . The terminal 102 communicates with the server 104 through the network. The data storage system can store the data required by the server 104 for processing. The data storage system can be integrated on the server 104, or placed on the cloud or other network servers. The server 104 obtains the power grid operation data, external environment data and equipment state data of the target power grid through the terminal 102; performs abnormal event analysis on the power grid operation data, external environment data and equipment state data, and identifies the key factors of each abnormal event of the target power grid; performs time series analysis on the power grid operation data, external environment data and equipment state data according to the key factors of each abnormal event, and predicts the abnormal event data and the abnormal event triggering probability of the target power grid; and generates power grid early warning information according to the abnormal event data and the abnormal event triggering probability. The terminal 102 can be, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart televisions, smart air conditioners, smart vehicle devices, etc. The portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0044] In one exemplary embodiment, as shown in Figure 2 , a power grid abnormal event prediction and early warning method based on big data analysis is provided. Taking the server in Figure 1 as an example, the method includes the following steps 202 to 208. Wherein:

[0045] Step 202, obtaining the power grid operation data, external environment data and equipment state data of the target power grid.

[0046] The power grid operation data can be dynamic information reflecting various electrical parameters of the power grid in the daily work process. These data include voltage, current, power factor, frequency, line load, operating state of the transformer, etc.

[0047] The external environment data can be external natural conditions or climate information affecting power grid operation, including temperature, humidity, wind speed, rainfall, air pressure, thunderstorm, snow disaster, sandstorm, etc.

[0048] The device status data can be information reflecting the operation health status of various devices in the power grid, including the temperature, vibration, service life, fault record, maintenance record, etc. of devices such as transformers, switches, distribution boxes, and generators.

[0049] Specifically, through extensive sensor networks and data collection systems, power grid operation data covering power grid operation are obtained, including core electrical parameters such as voltage, current, power factor, frequency, line load, etc. In addition, external environment data is also obtained from meteorological monitoring stations, satellite data and environmental sensors, involving air temperature, humidity, wind speed, rainfall, air pressure, etc. Natural conditions also cover monitoring information of extreme environmental events such as geological disasters or lightning strikes. Device status data is obtained through Internet of Things (IoT) technology or regular inspection records, including the working status, service life, temperature, vibration, fault record, maintenance history, etc. of devices such as transformers, switches, distribution boxes, etc. to ensure comprehensive coverage of all key influencing factors in power grid operation.

[0050] Step 204, abnormal event analysis is performed on the power grid operation data, external environment data and device status data to identify the key factors of each abnormal event of the target power grid.

[0051] The abnormal event analysis can be a comprehensive analysis of power grid operation data, external environment data and device status data to identify abnormal situations or faults occurring during power grid operation. These abnormal events can include voltage fluctuations, device failures, line short circuits or overloads, etc.

[0052] The abnormal event key factor can be a core factor that directly or indirectly leads to the occurrence of power grid abnormal events. These factors can come from device aging, overload operation, external environment changes (such as high temperature, thunderstorm weather), and delayed maintenance, etc. For example, high internal temperature of the device can be a key factor of device failure, or long-term overload operation can be the main reason for voltage fluctuation.

[0053] Specifically, after collecting the power grid operation data, external environment data, and equipment state data, the system will send the power grid operation data, external environment data, and equipment state data into the anomaly detection algorithm for multi-dimensional analysis, and automatically detect potential anomalies in the power grid operation using statistical methods, machine learning, or artificial intelligence models. For example, by comparing the current and historical voltage fluctuation trends, abnormal rise in equipment temperature, or sudden changes in load, etc., the system can identify potential abnormal events. Then, the system will further analyze these anomalies to find the key factors of each abnormal event that caused the target power grid anomaly. For example, through correlation analysis, it may identify that high environmental temperature, equipment aging, or overloading operation are the main reasons for electrical equipment overheating, or identify that high humidity leads to poor electrical contact and other problems.

[0054] Step 206, according to each abnormal event key factor, time series analysis is performed on the power grid operation data, external environment data, and equipment state data to predict the abnormal event data and abnormal event trigger probability of the target power grid.

[0055] Among them, time series analysis can be used to analyze continuous data that changes over time, revealing its inherent trends, periodic fluctuations, or abnormal patterns.

[0056] Among them, abnormal event data can be specific information related to the abnormal events identified during the operation of the power grid, including the type of abnormal event, the time of occurrence, the scope of influence, the detailed record of equipment failure or voltage fluctuation, etc.

[0057] Among them, the abnormal event trigger probability can be the possibility of the power grid occurring abnormal events in a certain period of time in the future based on historical data and time series analysis.

[0058] Specifically, after identifying the key factors of each abnormal event, the system will use time series analysis models (such as ARIMA, LSTM, etc.) to predict possible abnormal events in the power grid in the future through continuous dynamic analysis of power grid operation history data, equipment aging trends, and environmental meteorological data. For example, the system may predict that the probability of power grid load fluctuation due to high equipment temperature will rise in the next three days, or estimate that the possibility of equipment short circuit due to environmental deterioration during thunderstorm weather will increase. In addition, the system will provide specific abnormal event occurrence time range, location, severity, and scope based on data modeling, and improve the accuracy of the prediction by combining external factors such as high temperature period, thunderstorm season, etc., to obtain the abnormal event data and abnormal event trigger probability of the target power grid.

[0059] Step 208, according to the abnormal event data and the abnormal event trigger probability, generate power grid warning information.

[0060] The power grid early warning information can be prompt information generated based on real-time analysis and prediction results of power grid operation data, equipment state data, and external environment data, and is used to remind power grid managers and operation and maintenance personnel of potential abnormal or fault risks.

[0061] Specifically, based on the abnormal event data and the abnormal event triggering probability obtained through time series analysis, the system generates early warning information for different abnormal events. The early warning information not only includes the predicted abnormal events and their triggering probabilities, but also lists in detail the power grid areas, specific equipment, and weak links of power grid operation that may be affected. The system also provides relevant preventive suggestions, such as suggesting operation and maintenance personnel to increase cooling measures, adjust load distribution, or conduct advance inspection of transformers and line equipment prone to lightning strikes before thunderstorm weather. At the same time, the early warning information is sent in real time to relevant power grid managers and maintenance teams through multiple channels such as SMS, APP, and email, ensuring that they can take timely action before the abnormal event occurs, minimize power grid operation risks, and improve the overall power grid response capability and stability.

[0062] In the above-mentioned power grid abnormal event prediction and early warning method based on big data analysis, power grid operation data, external environment data, and equipment state data of a target power grid are obtained. Abnormal event analysis is performed on the power grid operation data, external environment data, and equipment state data to identify key factors of each abnormal event of the target power grid. Time series analysis is performed on the power grid operation data, external environment data, and equipment state data according to the key factors of each abnormal event to predict abnormal event data and abnormal event triggering probability of the target power grid. Power grid early warning information is generated based on the abnormal event data and the abnormal event triggering probability.

[0063] By comprehensively obtaining power grid operation data, external environment data, and equipment state data, and combining abnormal event analysis, key factors affecting the safe operation of the power grid can be accurately identified. Through time series analysis, the system can not only predict abnormal events of the power grid, but also quantify the probability of event occurrence, dynamically considering the changing trend of the power grid under different external environments and equipment states. Based on these prediction results, the system can generate early warning information in advance, making the prediction results more real-time and accurate in the case of predicting complex or atypical faults, and helping operation and maintenance personnel to predict potential risks, thereby improving the self-healing capability of the power grid, optimizing maintenance and dispatching strategies, reducing the possibility of fault outages, and significantly improving the reliability and efficiency of power grid operation.

[0064] In an exemplary embodiment, as shown in Figure 3 Abnormal event analysis is performed on the power grid operation data, external environment data, and equipment state data to identify key factors of each abnormal event of the target power grid, including steps 302 to 306.

[0065] Step 302, feature extraction is performed on grid operation data, external environment data, and equipment state data to identify grid failure patterns.

[0066] Among them, the grid failure pattern can be the failure rule or abnormal behavior of the grid under certain conditions identified by analyzing the grid operation data, equipment state data and external environment data.

[0067] Specifically, the system will use feature engineering techniques to extract key features related to grid failures. These features may involve voltage and current fluctuation amplitude, frequency anomalies, equipment temperature rising trend, load change pattern, and the influence of external environment such as extreme weather conditions. Then, the system identifies the grid failure patterns hidden in these features through clustering analysis, anomaly detection or classification algorithms (such as K-means, SVM, etc.). Grid failure patterns can reflect different types of potential problems in the grid, such as voltage transient drop, overload-induced line failure, or equipment aging caused by external environmental changes. The goal is to patternize abnormal operation behavior of the grid through in-depth analysis of multi-dimensional data.

[0068] Step 304, determine the key factor screening method according to the grid failure pattern.

[0069] Among them, the key factor screening method can be a technical means for extracting core variables or parameters closely related to grid failure or abnormal events from a large amount of data.

[0070] Specifically, after identifying the grid failure pattern, the system needs to develop an effective screening method to extract key factors closely related to these failure patterns. First, the system will select appropriate screening algorithms according to the characteristics of different failure patterns, such as correlation analysis, principal component analysis (PCA), mutual information method or recursive feature elimination (RFE), etc. For specific grid failure patterns, the system will analyze the contribution of each feature to the occurrence of the failure, and evaluate which data features are closely related to the triggering and evolution of the failure. For example, in the failure pattern of equipment overload, the time series change of load fluctuation may be used to identify the feature of high load; while in the failure caused by external environment, environmental temperature, humidity, etc. may become key screening factors.

[0071] Step 306, according to the key factor screening method, identify the key factors of each abnormal event of the target grid from the grid operation data, external environment data and equipment state data.

[0072] Specifically, the system applies previously defined screening methods to comprehensively analyze power grid operation data, external environment data, and equipment state data. Through specific algorithms, the system gradually eliminates secondary factors with low contribution to fault prediction, and extracts key factors closely related to power grid abnormal events. For example, in equipment state data, factors such as equipment temperature, vibration amplitude, and service life may be identified as key parameters leading to faults; while in external environment data, factors such as temperature, humidity, and wind speed may become decisive factors affecting the stable operation of the power grid. The system not only identifies these key factors, but also quantitatively analyzes the degree of their influence to determine how much each factor contributes to the triggering probability of abnormal events. After optimization of the extracted key factors, the key factors of each abnormal event are obtained.

[0073] In this embodiment, by extracting features from power grid operation data, external environment data, and equipment state data, the power grid fault mode can be accurately identified, and then the key factor screening method can be customized according to the fault mode to efficiently screen out key factors that are crucial for abnormal event prediction from massive data. This process significantly improves the accuracy and efficiency of power grid abnormal event identification, enabling the system to accurately locate potential fault sources and provide a scientific basis for subsequent risk prediction and preventive maintenance, thereby improving the stability and security of power grid operation and reducing the occurrence of sudden failures.

[0074] In one exemplary embodiment, as shown in Figure 4 The key factor screening method includes correlation screening method, principal component screening method, feature screening method, and regularization screening method. According to the key factor screening method, the key factors of each abnormal event of the target power grid are identified from the power grid operation data, external environment data, and equipment state data, including steps 402 to 410. Among them:

[0075] Step 402, according to the correlation screening method, analyze the correlation of power grid operation data, external environment data and equipment state data, and get the correlation key factor.

[0076] The correlation screening method can be a method of screening out factors closely related to the target variable by analyzing the linear or nonlinear relationship between data features. By calculating the correlation coefficient between the features and the target variable (such as power grid abnormal events), the system can identify key features that have a greater impact on abnormal events.

[0077] The correlation key factor can be a factor screened out by analyzing the correlation between each feature in the power grid operation data, equipment state data, and external environment data and the target variable (such as abnormal events). These factors show strong correlation with abnormal events, indicating that they play an important role in predicting or explaining abnormal events.

[0078] Specifically, the system utilizes a correlation screening method to analyze the linear or nonlinear relationships between power grid operation data, external environment data, and equipment state data. By calculating the correlation coefficients (such as Pearson correlation, Spearman correlation, etc.) between different data features, the system identifies highly correlated data pairs. These data pairs include the correlation between voltage, current, equipment temperature, and external environmental factors such as humidity and air temperature. High correlation features indicate that these factors play an important role in abnormal events, such as simultaneous abnormality of equipment temperature and current fluctuations, indicating the risk of equipment overload. Finally, key factors with significant correlation are screened out as correlation key factors to explain and predict power grid abnormal events.

[0079] Step 404, according to the principal component screening method, the variance between different data of power grid operation data, external environment data and equipment state data is calculated, and the principal component key factor is obtained.

[0080] Among them, the principal component screening method can be to use principal component analysis (PCA) to reduce the dimension of high-dimensional data, and by calculating the variance contribution of different features, the principal component factors that can explain most of the changes in the data set are screened out. These principal components represent the main information in the original data, eliminating redundant features, which helps to simplify the model and improve computational efficiency.

[0081] Among them, the principal component key factor can be several comprehensive factors that can explain the main part of data changes extracted from high-dimensional data through principal component analysis (PCA). They are condensed versions of the most important information in the original data, which can reduce redundant information and noise. For example, multiple related voltage fluctuations and current change data can be compressed into a principal component key factor to describe the overall operation state of the power grid.

[0082] Specifically, the system uses principal component screening method (PCA) to reduce the dimension of large-scale, multi-dimensional power grid data. First, the system calculates the variance of each data feature, and identifies the principal components that contribute most to data changes through the covariance matrix. Through PCA, multiple original features are compressed into a few principal components, which explain most of the variance in the data set. For example, multiple parameters related to equipment aging can be simplified into a principal component variable. Finally, the system extracts the principal component key factor that best reflects the power grid anomaly to improve the efficiency of data processing and the accuracy of prediction.

[0083] Step 406, according to the feature screening method, factors with contribution greater than a threshold value to abnormal conditions are screened out from power grid operation data, external environment data and equipment state data, and feature key factors are obtained.

[0084] Among them, the feature screening method can be to evaluate the contribution of each data feature to the prediction target, and screen out those factors that have a significant impact on the target variable. Common evaluation methods include information gain, chi-square test, tree-based models, etc. The system will sort the features according to their importance, and screen out features whose contribution exceeds a certain threshold. These features are the most explanatory in predicting power grid abnormal events, such as the over-temperature phenomenon in the device state, which may be an important feature.

[0085] Among them, the abnormal situation contribution can refer to the contribution of each data feature to the prediction or explanation of the power grid abnormal event, measuring the influence of the feature in the abnormal event.

[0086] Among them, the feature key factor can be an important feature with a large contribution to predicting power grid abnormal events screened out by the feature screening method. These features are evaluated to have a significant impact on the target variable (such as failure, abnormal event), exceeding the set contribution threshold, and therefore selected as key factors. For example, significant changes in the aging state of the device and the environment temperature may be used as a feature key factor to predict future possible abnormal events.

[0087] Specifically, the system uses a feature screening method to evaluate power grid, environmental and device state data one by one. The system calculates the contribution of each feature to the abnormal event in the model, such as using information gain, chi-square test or tree-based models (such as random forest) to evaluate the importance of the feature. Features with a larger contribution mean that they have a significant effect on the prediction of power grid abnormal events, such as an increase in the probability of device failure when the power grid load is too high. The system sets a contribution threshold, and features exceeding the threshold are selected as feature key factors.

[0088] Step 408, according to the regularization screening method, eliminate the redundant information of power grid operation data, external environment data and device state data, and get the regularization key factor.

[0089] Among them, the regularization screening method can be to introduce a penalty term (such as Lasso or Ridge regression) in the regression model to eliminate redundant features that have a small contribution to model prediction, and prevent overfitting. Lasso regression can reduce the coefficients of some unimportant features to zero, effectively screening out a small number of useful features, and Ridge regression reduces the influence of multicollinearity by limiting the size of the coefficient.

[0090] The regularization key factor can be a feature extracted from the data by the regularization screening method. The system eliminates redundant or unimportant features through regularization techniques and only retains those factors that have strong explanatory power for abnormal event prediction. The regularization key factor effectively reduces model complexity and avoids multicollinearity problems, such as when multiple grid parameters are highly correlated, the regularization will screen out the most influential parameters and remove redundant factors.

[0091] Specifically, the system eliminates redundant and noisy features through regularization screening methods such as Lasso or Ridge regression. By introducing a penalty term in the regression model, the regularization method suppresses features that are useless or overly correlated to the model. Lasso regression can directly reduce the coefficient of some unimportant features to zero, thereby achieving automatic feature selection, while Ridge regression prevents the model from overfitting to a small number of features. This method is particularly suitable for high-dimensional data, as it eliminates redundant information, reduces model complexity and noise effects. For example, features with multicollinearity (such as multiple similar temperature readings in device status) are eliminated or reduced in weight by regularization, resulting in a regularization key factor.

[0092] Step 410: Obtain each abnormal event key factor according to the correlation key factor, the principal component key factor, the feature key factor, and the regularization key factor.

[0093] Specifically, cross-validation and model evaluation are used to further refine and integrate these factors. The system analyzes the performance of these factors under different fault modes to ensure the effectiveness and complementarity of each factor, avoiding redundancy and conflicts. For example, a factor may perform well in correlation analysis but be identified as a redundant feature by the regularization screening method. In this case, the system will prioritize factors that contribute most to prediction and have strong explanatory power. Ultimately, the system will integrate all screening results into an optimized set of abnormal event key factors.

[0094] In this embodiment, by comprehensively using correlation screening methods, principal component analysis, feature screening methods, and regularization screening methods, key factors can be effectively extracted from a large amount of grid operation data, external environment data, and device status data. Each screening method optimizes different aspects of the data, from correlation to variance, contribution, and elimination of redundant information, gradually refining key factors closely related to grid abnormal events. This multi-level screening method improves the accuracy of abnormal event prediction, reduces data noise and redundant information interference, and ultimately provides more reliable decision-making basis for grid early warning and maintenance, improving the safety and efficiency of grid operation.

[0095] In one exemplary embodiment, as Figure 5As shown, according to each abnormal event key factor, time series analysis is performed on power grid operation data, external environment data and equipment state data to predict abnormal event data and abnormal event triggering probability of the target power grid, including steps 502 to 504. Among them:

[0096] Step 502, according to the power grid operation data, external environment data and equipment state data, determine the analysis period of time series analysis, and the analysis content corresponding to each analysis period.

[0097] Among them, the analysis period can be a time interval or frequency set in time series analysis in order to better understand the time dependence and trend of the data. These periods can be short-term (such as minutes, hours), medium-term (such as days, weeks), or long-term (such as months, seasons) etc. to capture the change pattern of data on different time scales.

[0098] Among them, the analysis content can be the data features and specific problems to be solved for each determined analysis period in time series analysis. For example, short-term analysis content may include transient current fluctuations in power grid operation or rapid rise in equipment temperature, while long-term analysis may focus on equipment aging trend or seasonal load change.

[0099] Specifically, the system will first extract time-dependent features from power grid operation data, external environment data and equipment state data, and identify different data periodicity, i.e. analysis period, through time series analysis techniques (such as autocorrelation function ACF, Fourier transform, etc.). Power grid data may have multiple levels of time periods, for example: short-term period may include hourly data fluctuations, reflecting real-time load or rapid changes in equipment temperature; medium-term period such as day, week may reflect the trend of power demand or the day-night fluctuation of environmental temperature; long-term period may focus on the impact of seasonal changes on the power grid, such as equipment aging or long-term environmental deterioration. Further determine the analysis content corresponding to each period, each analysis content depends on the pattern it reveals, for example, short-term period analysis may focus on the rapid response of sudden events, while long-term period may be used to monitor the overall trend of equipment aging or load growth.

[0100] Step 504, based on the analysis content of each analysis period, cross-influence analysis is performed on each abnormal event key factor and corresponding power grid operation data, external environment data and equipment state data to obtain abnormal event data and abnormal event triggering probability.

[0101] The cross impact analysis can be to evaluate the interaction and influence between multiple data features or variables. In particular, in power grid monitoring, the system combines the key factors of abnormal events in power grid operation data, equipment status data, and external environment data through cross impact analysis, and analyzes how they jointly affect the operation of the power grid.

[0102] Specifically, based on the analysis content of each analysis period, the system uses time series analysis models (such as ARIMA, LSTM, or Prophet) to combine the analysis period and the corresponding key factors of abnormal events determined in the previous step for in-depth analysis. The system first models the interrelationships between the key factors of abnormal events and the power grid operation data, external environment data, and equipment status data within each period. Through cross impact analysis, the system can identify the influence of key factors on abnormal events in different periods, obtaining abnormal event data, such as in the short term, rapid temperature rise of equipment may have an important impact on load fluctuations, while in the long term, temperature and humidity changes in the external environment may have a greater impact on equipment aging and failure. The system further calculates the triggering probability of abnormal events through these models to predict the likelihood of abnormal event occurrence in a specific future time period. These triggering probabilities are dynamically adjusted by combining historical data, environmental conditions, and equipment status to obtain abnormal event triggering probabilities.

[0103] In this embodiment, by determining the time series analysis of different time periods and conducting in-depth analysis for the characteristics of each analysis period, the changing patterns of power grid operation, external environment, and equipment status at different time scales can be comprehensively captured. By cross impact analysis of each key factor of abnormal events and related data, the system can identify the interaction between different factors, and accurately predict the occurrence time and triggering probability of abnormal events. This method effectively improves the accuracy and timeliness of abnormal event prediction, helping power grid operators better understand potential risks and take preventive measures in a timely manner to ensure the continuous and stable operation of the power grid.

[0104] In one exemplary embodiment, as shown in Figure 6 the cross impact analysis of each key factor of abnormal events and the corresponding power grid operation data, external environment data, and equipment status data is performed to obtain abnormal event data and abnormal event triggering probability, including steps 602 to 604. Among them:

[0105] Step 602, covariance analysis is performed on each key factor of abnormal events to obtain key factor change data.

[0106] The key factor change data can be statistical results reflecting the changes of these factors over time or conditions by analyzing the interrelationships between key variables in power grid operation, equipment status, and external environment.

[0107] Specifically, the system will first perform a covariance analysis on the collected critical factors of abnormal events. Covariance analysis is a statistical method used to measure the linear relationship between two or more variables and determine whether they change simultaneously. The system will calculate the covariance value between each pair of critical factors to analyze their correlation and mutual influence. For example, the system may analyze the covariance between changes in grid load and increases in device temperature to assess their synchronous fluctuations. The size and sign of the covariance value reveal the direction (positive or negative) and strength of the association between these critical factors, which helps identify which factor combinations have a greater combined effect on abnormal events. Through this step, the system generates a set of critical factor change data representing the associated patterns of each factor and their potential contribution to grid abnormal events.

[0108] Step 604, according to the critical factor change data, time-dependent analysis is performed on the grid operation data, external environment data and device state data to obtain abnormal event data and abnormal event trigger probability.

[0109] Among them, time-dependent analysis can be an analysis method that studies the changes and patterns of data at different time points through time series models. It evaluates the rules of changes in grid operation data, device state data and external environment data over time, and analyzes the trends and periodicity of critical factors in the time dimension.

[0110] Specifically, the system performs time-dependent analysis based on the critical factor change data obtained from covariance analysis. Time-dependent analysis is performed through time series models to evaluate the trend of critical factors over time and their impact on grid abnormal events. The system uses time series algorithms (such as ARIMA, LSTM or Prophet) to combine historical information of grid operation data, external environment data and device state data with the time correlation of critical factors, and identify the fluctuation patterns of factors in different time periods. By observing the time dependence of these factors, the system can predict the likelihood of abnormal events in a specific future period. For example, the system may find that the cumulative effect of device temperature and external environment temperature significantly increases the risk of failure after a period of time. Based on these time-dependent relationships, the system can not only generate accurate abnormal event data, but also calculate the abnormal event trigger probability of each event in the future, helping grid operators take preventive measures or adjust operation strategies in advance.

[0111] In this embodiment, by performing covariance analysis between the key factors of each abnormal event, the mutual relationship and change trend between the key factors are identified, and the key factor change data is generated. Then, through time-dependent analysis, the dynamic influence of these key factors over time is further revealed, and the triggering time and probability of abnormal events are accurately predicted. This method effectively integrates the correlation between factors and the dependence on time, can more comprehensively capture the causes of power grid abnormal events, improve the accuracy of abnormal event prediction, help power grid managers discover potential problems earlier and take preventive measures, thereby enhancing the stability and reliability of the power grid.

[0112] In one exemplary embodiment, as shown in Figure 7 According to the key factor change data, time-dependent analysis is performed on the power grid operation data, external environment data and equipment state data to obtain abnormal event data and abnormal event triggering probability, including steps 702 to 706. Among them:

[0113] Step 702, according to the key factor change data, the time-dependent data between the key factors of each abnormal event is identified.

[0114] Specifically, the system analyzes the key factor change data in depth, including using time series models (such as autoregressive model AR or moving average model MA), to identify the dependence of each abnormal event key factor on time dimension, such as analyzing how the past values of each key factor affect the current and future state. For example, the system may find through analysis that the increase in equipment temperature in the past few hours will lead to an increase in future power grid load fluctuations. These time-dependent data reflect the dynamic relationship between key factors at different time points, helping the system understand how each key factor evolves over time and plays a predictive role in the occurrence of abnormal events.

[0115] Step 704, according to the key factor change data and the time-dependent data, the trend-dependent data between the key factors of each abnormal event is identified.

[0116] Specifically, the system further analyzes the long-term trends and cumulative effects between key factors by combining key factor change data and time-dependent data. Through trend analysis methods (such as trend decomposition or moving average methods), the system can reveal the change direction and dependence relationship of key factors over a longer period of time, and obtain trend-dependent data. For example, under the condition of continuous high temperature, the temperature and load of the equipment gradually increase, indicating that the risk of equipment failure will increase significantly at some time in the future. Trend-dependent data shows how these factors interact over a larger time frame and reveals the potential impact of long-term fluctuations or growth on the triggering of abnormal events, providing a longer-term predictive perspective.

[0117] At step 706, the power grid operation data, external environment data, and equipment state data are predicted according to the time-dependent data and the trend-dependent data to obtain abnormal event data and abnormal event triggering probability.

[0118] Specifically, the system utilizes the time-dependent data and the trend-dependent data, applies a complex time series prediction model (such as ARIMA, LSTM, or Prophet) to combine the power grid operation data, external environment data, and equipment state data, and predicts the abnormal events that may occur in the future by using time and trend information; by modeling the historical data, the system can generate abnormal event data, determine the type, occurrence time, and possible impact range of the abnormal event. Finally, the system calculates the abnormal event triggering probability of each abnormal event according to the abnormal event data, helping the operation and maintenance personnel to determine the possibility of future abnormal events in the power grid.

[0119] In this embodiment, by first identifying the time-dependent data between the key factors of each abnormal event, the short-term dynamic changes of the key factors over time are revealed, and then the trend-dependent data is further identified to analyze the change trend and cumulative effect of the key factors in the long term. By combining these two types of data, the system can comprehensively predict the power grid operation, external environment, and equipment state, accurately predict the type, time, and triggering probability of future abnormal events. This method integrates the analysis of short-term fluctuations and long-term trends, significantly improves the accuracy and timeliness of abnormal event prediction, helps operation and maintenance personnel to better develop preventive measures, reduces the failure rate, and ensures the continuous and reliable operation of the power grid.

[0120] It should be understood that although each step in the flowchart involved in the above embodiments is displayed in sequence according to the arrow, these steps are not necessarily executed in the order indicated by the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other orders. Moreover, at least part of the steps in the flowchart involved in the above embodiments can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of these steps or stages is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or steps or stages in other steps.

[0121] Based on the same inventive concept, the embodiments of the present application also provide a big data analysis based power grid abnormal event prediction and early warning device for implementing the above-mentioned big data analysis based power grid abnormal event prediction and early warning method. The implementation scheme for solving problems provided by the device is similar to the implementation scheme described in the above-mentioned method, so the specific limitations in one or more big data analysis based power grid abnormal event prediction and early warning device embodiments provided below can be referred to the limitations of the big data analysis based power grid abnormal event prediction and early warning method described above, and will not be repeated here.

[0122] In one exemplary embodiment, as shown in Figure 8 a big data analysis based power grid abnormal event prediction and early warning device is provided, comprising: a data acquisition module 802, a data analysis module 804, an abnormal event prediction module 806 and a power grid early warning module 808, wherein:

[0123] The data acquisition module 802 is configured to acquire power grid operation data, external environment data and equipment state data of a target power grid.

[0124] The data analysis module 804 is configured to perform abnormal event analysis on the power grid operation data, the external environment data and the equipment state data, and identify each abnormal event key factor of the target power grid.

[0125] The abnormal event prediction module 806 is configured to perform time series analysis on the power grid operation data, the external environment data and the equipment state data according to each abnormal event key factor, and predict abnormal event data and abnormal event trigger probability of the target power grid.

[0126] The power grid early warning module 808 is configured to generate power grid early warning information according to the abnormal event data and the abnormal event trigger probability.

[0127] In one embodiment, the data analysis module 804 is further configured to perform feature extraction on the power grid operation data, the external environment data and the equipment state data, identify a power grid failure mode, determine a key factor screening method according to the power grid failure mode, and identify each abnormal event key factor of the target power grid from the power grid operation data, the external environment data and the equipment state data according to the key factor screening method.

[0128] In one embodiment, the data analysis module 804 is further configured to analyze the correlation between the power grid operation data, the external environment data, and the equipment state data according to a correlation screening method to obtain correlation key factors; calculate the variance between different data of the power grid operation data, the external environment data, and the equipment state data according to a principal component screening method to obtain principal component key factors; screen out factors with a contribution degree greater than a threshold from the power grid operation data, the external environment data, and the equipment state data according to a feature screening method to obtain feature key factors; eliminate redundant information of the power grid operation data, the external environment data, and the equipment state data according to a regularization screening method to obtain regularization key factors; and obtain each abnormal event key factor according to the correlation key factors, the principal component key factors, the feature key factors, and the regularization key factors.

[0129] In one embodiment, the abnormal event prediction module 806 is further configured to determine each analysis period of time series analysis and analysis content corresponding to each analysis period according to the power grid operation data, the external environment data, and the equipment state data; and perform cross-influence analysis on each abnormal event key factor and corresponding power grid operation data, external environment data, and equipment state data based on the analysis content of each analysis period to obtain abnormal event data and abnormal event trigger probability.

[0130] In one embodiment, the abnormal event prediction module 806 is further configured to perform covariance analysis on each abnormal event key factor to obtain key factor change data; and perform time-dependent analysis on the power grid operation data, the external environment data, and the equipment state data according to the key factor change data to obtain abnormal event data and abnormal event trigger probability.

[0131] In one embodiment, the abnormal event prediction module 806 is further configured to identify time-dependent data between each abnormal event key factor according to the key factor change data; identify trend-dependent data between each abnormal event key factor according to the key factor change data and the time-dependent data; and perform prediction on the power grid operation data, the external environment data, and the equipment state data according to the time-dependent data and the trend-dependent data to obtain abnormal event data and abnormal event trigger probability.

[0132] Each module in the above-described power grid abnormal event prediction and early warning device based on big data analysis can be realized by software, hardware, and combinations thereof, in whole or in part. Each module can be embedded in or independent of a processor in a computer device in hardware form, or can be stored in a memory in a computer device in software form, so as to be called and executed by a processor to perform operations corresponding to each module.

[0133] In one exemplary embodiment, a computer device, which can be a server, is provided, and an internal structure diagram of the computer device can be as shown inFigure 9 The computer device shown in the figure includes a processor, a memory, an input / output interface (I / O for short) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store server data. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals through network connection. The computer program is executed by the processor to implement a power grid abnormal event prediction and early warning method based on big data analysis.

[0134] Those skilled in the art can understand that, Figure 9 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0135] In one embodiment, a computer device is also provided, including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps in each of the method embodiments.

[0136] In one embodiment, a computer readable storage medium is provided, storing a computer program, which is executed by a processor to implement the steps in each of the method embodiments.

[0137] In one embodiment, a computer program product or computer program is provided, which includes computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions, so that the computer device executes the steps in each of the method embodiments.

[0138] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant regulations.

[0139] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0140] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combination of the technical features does not exist, it should be considered as the scope of the present application.

[0141] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of protection of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method for predicting and warning abnormal power grid events based on big data analysis, characterized in that: The method comprises: Obtain grid operation data, external environment data, and equipment status data of the target grid; Performing abnormal event analysis on the power grid operation data, the external environment data, and the device status data to identify key factors of abnormal events in the target power grid; According to each of the key factors of the abnormal events, a time series analysis is performed on the power grid operation data, the external environment data, and the device status data to predict the abnormal event data of the target power grid and the probability of triggering the abnormal event; Power grid warning information is generated based on the abnormal event data and the abnormal event triggering probability.

2. The method according to claim 1, characterized in that The performing abnormal event analysis on the power grid operation data, the external environment data, and the device status data to identify key factors of abnormal events of the target power grid includes: Extracting features from the power grid operation data, the external environment data, and the device status data to identify power grid failure modes; Determining a key factor screening method according to the power grid failure mode; According to the key factor screening method, each of the abnormal event key factors of the target power grid is identified from the power grid operation data, the external environment data, and the equipment status data.

3. The method according to claim 2, characterized in that The key factor screening method includes a correlation screening method, a principal component screening method, a feature screening method, and a regularization screening method; identifying the key factors of each abnormal event of the target power grid from the power grid operation data, the external environment data, and the device status data according to the key factor screening method includes: Analyzing the correlation among the power grid operation data, the external environment data, and the device status data according to the correlation screening method to obtain correlation key factors; According to the principal component screening method, the variances among the power grid operation data, the external environment data, and the device status data are calculated to obtain principal component key factors; According to the feature screening method, factors whose contribution to abnormal conditions is greater than a threshold are screened from the power grid operation data, the external environment data, and the device status data to obtain feature key factors; Eliminating redundant information of the power grid operation data, the external environment data, and the device status data according to the regularized screening method to obtain regularized key factors; The key factors of each abnormal event are obtained according to the correlation key factors, the principal component key factors, the characteristic key factors and the regularization key factors.

4. The method according to claim 1, wherein The step of performing time series analysis on the power grid operation data, the external environment data, and the device status data based on each of the key factors of the abnormal events to predict abnormal event data and abnormal event triggering probability of the target power grid includes: Determining each analysis period of the time series analysis and analysis content corresponding to each analysis period according to the power grid operation data, the external environment data, and the device status data; Based on the analysis content of each analysis cycle, a cross-impact analysis is performed on each of the key factors of the abnormal events and the corresponding power grid operation data, the external environment data, and the equipment status data to obtain the abnormal event data and the abnormal event triggering probability.

5. The method according to claim 4, characterized in that The cross-impact analysis of each of the key factors of the abnormal events with the corresponding power grid operation data, the external environment data, and the device status data to obtain the abnormal event data and the abnormal event triggering probability includes: Conducting covariance analysis on the key factors of each abnormal event to obtain key factor change data; According to the key factor change data, a time-dependent analysis is performed on the power grid operation data, the external environment data, and the device status data to obtain the abnormal event data and the abnormal event triggering probability.

6. The method according to claim 5, characterized in that The step of performing a time-dependent analysis on the power grid operation data, the external environment data, and the device status data based on the key factor change data to obtain the abnormal event data and the abnormal event triggering probability includes: Identifying time dependency data between key factors of each of the abnormal events based on the key factor change data; Identifying trend dependency data between key factors of each of the abnormal events based on the key factor change data and the time dependency data; The power grid operation data, the external environment data, and the device status data are predicted based on the time-dependent data and the trend-dependent data to obtain the abnormal event data and the abnormal event triggering probability.

7. A power grid abnormal event prediction and early warning device based on big data analysis, characterized in that: The device comprises: A data acquisition module is used to obtain grid operation data, external environment data, and equipment status data of the target grid; a data analysis module, configured to perform abnormal event analysis on the power grid operation data, the external environment data, and the device status data, and identify key factors of abnormal events in the target power grid; An abnormality prediction module is used to perform time series analysis on the power grid operation data, the external environment data and the device status data according to each of the key factors of the abnormal events, and predict the abnormal event data and the abnormal event triggering probability of the target power grid; The power grid early warning module is used to generate power grid early warning information based on the abnormal event data and the abnormal event triggering probability.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images