Disaster prevention and reduction monitoring and early warning method and system for power grid equipment

Abstract

The invention discloses a disaster prevention and reduction monitoring and early warning method and system for power grid equipment, and the method comprises the steps: S1, obtaining the meteorological parameter data of the power grid equipment, carrying out the abnormal detection processing of the meteorological parameter data, and obtaining the normal meteorological parameter data; s2, performing data preprocessing on the normal meteorological parameter data, and constructing a meteorological parameter data set; s3, establishing a disaster prevention and reduction monitoring and early warning integrated model, and training, predicting and evaluating the disaster prevention and reduction monitoring and early warning integrated model according to the training data set and the test data set to obtain a final disaster prevention and reduction monitoring and early warning integrated model; and S4, collecting real-time meteorological parameter data of power grid equipment, inputting the final disaster prevention and reduction monitoring and early warning integrated model, and outputting a meteorological event classification result. The problems that all-weather and all-directional monitoring of the power grid equipment cannot be achieved through the current power grid equipment monitoring technology, and reliability and accuracy of classification and prediction of disaster risks of the power grid equipment are low are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power grid equipment monitoring, and in particular relates to a disaster prevention and mitigation monitoring and early warning method and system for power grid equipment. Background Art

[0002] Grid equipment refers to various equipment and devices used to transmit, distribute and control electric energy. It constitutes the infrastructure of the power system and is responsible for delivering the electric energy generated by power plants to various power consumption locations, and controlling and protecting the electric energy during the transmission and distribution process. At present, the main technologies used to detect whether there are natural disasters and man-made risks in power grid equipment include traditional inspection methods and monitoring systems. Traditional inspection methods rely on regular manual inspections and empirical judgments. Although they can detect whether there are disaster risks in equipment to a certain extent, they have the disadvantages of low efficiency, strong blindness, and insufficient data acquisition. The monitoring system mainly uses advanced sensor technology, data acquisition devices and remote monitoring technology to realize real-time monitoring of the status of power grid equipment and fault warning. Although the above monitoring and warning technologies can directly monitor the physical parameters of the equipment and instantly understand the working status and environmental conditions of the equipment, the data monitored by the sensors are often raw data, and it is difficult to directly discover potential abnormal patterns.

[0003] The Chinese patent with publication number CN111600392A discloses an intelligent power grid equipment monitoring system, including an equipment monitoring module, a central processing module and a user terminal. The equipment monitoring module collects real-time monitoring data of power grid equipment based on a wireless sensor network. The central processing module is used to receive, store and display real-time monitoring data of power grid equipment, and compare the real-time monitoring data of power grid equipment with the boundary value of a pre-set normal threshold range. If the normal threshold range is exceeded, an alarm signal is output; the user terminal is used to access the real-time monitoring data of power grid equipment in the central processing module in real time; the equipment monitoring module includes a sensor node and a convergence node. The sensor node is used to collect real-time monitoring data of power grid equipment, and the convergence node is used to send the real-time monitoring data of power grid equipment to the central processing module. This invention mainly relies on setting static threshold comparison for abnormal detection and alarm, lacks in-depth understanding of complex data patterns and abnormal situations, and has certain limitations. Summary of the invention

[0004] The present invention provides a disaster prevention and mitigation monitoring and early warning method and system for power grid equipment, aiming to solve the problems that the current power grid equipment monitoring technology cannot achieve all-weather and all-round monitoring of power grid equipment, and the reliability and accuracy of classifying and predicting disaster risks of power grid equipment are low.

[0005] In order to solve the above technical problems, the present invention provides a disaster prevention and mitigation monitoring and early warning method for power grid equipment, comprising the following steps:

[0006] S1: Acquire meteorological parameter data of power grid equipment, perform abnormal detection processing on the meteorological parameter data, and obtain normal meteorological parameter data.

[0007] S2: Perform data preprocessing on normal meteorological parameter data, construct a meteorological parameter data set, and divide the meteorological parameter data set into a training data set and a test data set.

[0008] S3: Establish a disaster prevention and mitigation monitoring and early warning integrated model, train and predictively evaluate the disaster prevention and mitigation monitoring and early warning integrated model based on the training data set and the test data set, and obtain the final disaster prevention and mitigation monitoring and early warning integrated model.

[0009] S4: Collect real-time meteorological parameter data from power grid equipment, input into the final disaster prevention and mitigation monitoring and early warning integrated model, and output meteorological event classification results.

[0010] Preferably, the abnormality detection process for the meteorological parameter data is specifically as follows:

[0011] S11: Perform density-based clustering on the acquired meteorological parameter data to obtain anomaly detection results of normal clusters and abnormal points.

[0012] S12: Calculate the local anomaly factors of all data in the meteorological parameter data by using the LOF algorithm to obtain an anomaly detection result of the LOF algorithm.

[0013] S13: Comprehensively calculate the abnormal detection results obtained by the analysis of steps S21 and S22 to obtain a comprehensive abnormal detection result, retain the meteorological parameter data whose abnormal detection results are normal, and eliminate the data whose abnormal detection results are abnormal.

[0014] Preferably, the step S13 is specifically as follows:

[0015] S131: The outlier detection results of the normal clusters and outliers are graded and set, and the grade values ​​are set for the data classified as the core points in the normal clusters, the data classified as the boundary points in the normal clusters, and the data classified as the outliers in turn.

[0016] S132: For each meteorological parameter data, the obtained anomaly detection result is comprehensively processed to obtain a comprehensive anomaly detection result, an anomaly detection threshold is set, the meteorological parameter data within the anomaly detection threshold range is retained, and the data exceeding the anomaly detection threshold range is eliminated. The formula for the comprehensive processing is specifically:

[0017] F(x i )=αg(x i )+βl(x i )

[0018] In the formula, Fi is the anomaly detection result of the i-th meteorological parameter data; x i is the i-th meteorological parameter data; α is the clustering weight factor; g(x i ) is the classification value of the clustering anomaly detection result of the i-th meteorological parameter data; β is the weight factor of the LOF algorithm; l(x i ) is the local anomaly factor value of the i-th meteorological parameter data.

[0019] Preferably, step S3 specifically comprises:

[0020] S31: Establish a disaster prevention and mitigation monitoring and early warning integrated model based on the random forest algorithm and the SVC algorithm, and train the random forest model and the SVC model in the disaster prevention and mitigation monitoring and early warning integrated model according to the training data set.

[0021] S32: Use the test data set to perform prediction and evaluation on the trained random forest model and SVC model respectively, use weighted voting to fuse the two prediction results of the disaster prevention and mitigation monitoring and early warning integrated model to obtain the final prediction result, and adjust the parameters of the disaster prevention and mitigation monitoring and early warning integrated model according to the final prediction result to obtain the final disaster prevention and mitigation monitoring and early warning integrated model.

[0022] On the other hand, the present invention provides a disaster prevention and mitigation monitoring and early warning system for power grid equipment, including a meteorological data acquisition module, a data set construction module, a model training and testing module, and a monitoring and early warning module.

[0023] The meteorological data acquisition module is used to obtain meteorological parameter data of power grid equipment, perform abnormal detection and processing on the meteorological parameter data, and obtain normal meteorological parameter data.

[0024] The data set construction module is used to perform data preprocessing on normal meteorological parameter data, construct a meteorological parameter data set, and divide the meteorological parameter data set into a training data set and a test data set.

[0025] The model training and testing module is used to establish an integrated model for disaster prevention and mitigation monitoring and early warning. The integrated model for disaster prevention and mitigation monitoring and early warning is trained and predicted and evaluated based on the training data set and the test data set to obtain the final integrated model for disaster prevention and mitigation monitoring and early warning.

[0026] The monitoring and early warning module is used to collect real-time meteorological parameter data from power grid equipment, input it into the final disaster prevention and mitigation monitoring and early warning integrated model, and output the meteorological event classification results.

[0027] Preferably, the meteorological data collection module includes a clustering anomaly detection module, a LOF anomaly detection module and a comprehensive anomaly detection module.

[0028] The clustering anomaly detection module is used to perform density-based clustering on the acquired meteorological parameter data to obtain anomaly detection results of normal clusters and abnormal points.

[0029] The LOF anomaly detection module is used to calculate the local anomaly factors of all data in the meteorological parameter data through the LOF algorithm to obtain the anomaly detection results of the LOF algorithm.

[0030] The comprehensive anomaly detection module is used to comprehensively calculate the anomaly detection results obtained by the above modules to obtain comprehensive anomaly detection results, retain the meteorological parameter data with normal anomaly detection results, and eliminate the data with abnormal anomaly detection results.

[0031] Preferably, the comprehensive anomaly detection module includes a grading setting module and a comprehensive processing module.

[0032] The classification setting module is used to classify the abnormal point detection results of normal clusters and abnormal points, and set the classification values ​​for the data classified as core points in the normal cluster, the data classified as boundary points in the normal cluster, and the data classified as abnormal points in turn.

[0033] The comprehensive processing module is used to comprehensively process the obtained anomaly detection results for each meteorological parameter data to obtain a comprehensive anomaly detection result, set an anomaly detection threshold, retain the meteorological parameter data within the anomaly detection threshold range, and remove the data exceeding the anomaly detection threshold range. The formula for the comprehensive processing is specifically:

[0034] F(x i )=αg(x i )+βl(x i )

[0035] In the formula, F i is the anomaly detection result of the i-th meteorological parameter data; x i is the i-th meteorological parameter data; α is the clustering weight factor; g(x i ) is the classification value of the clustering anomaly detection result of the i-th meteorological parameter data; β is the weight factor of the LOF algorithm; l(x i ) is the local anomaly factor value of the i-th meteorological parameter data.

[0036] Preferably, the model training and testing module includes a training module and a testing module.

[0037] The training module is used to establish an integrated model of disaster prevention and mitigation monitoring and early warning based on the random forest algorithm and the SVC algorithm, and to train the random forest model and the SVC model in the integrated model of disaster prevention and mitigation monitoring and early warning according to the training data set.

[0038] The testing module is used to use the test data set to perform prediction and evaluation on the trained random forest model and SVC model respectively, and to use weighted voting to fuse the two prediction results of the disaster prevention and mitigation monitoring and early warning integrated model to obtain the final prediction result. According to the final prediction result, the parameters of the disaster prevention and mitigation monitoring and early warning integrated model are tuned to obtain the final disaster prevention and mitigation monitoring and early warning integrated model.

[0039] On the other hand, the present invention also provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for monitoring and early warning of disaster prevention and mitigation of power grid equipment as described in any embodiment of the present invention is implemented.

[0040] On the other hand, the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the disaster prevention and mitigation monitoring and early warning method for threshold power grid equipment as described in any embodiment of the present invention.

[0041] Compared with the prior art, the present invention has the following technical effects:

[0042] 1. The present invention uses meteorological parameter data for anomaly detection and processing, and can quickly and accurately identify anomalies in the data while acquiring power grid equipment data, thereby ensuring that the quality of the data used is highly reliable. Compared with traditional manual inspections, this automated anomaly detection and processing method not only reduces labor costs, but also enables efficient processing of large-scale data, improving the real-time and accuracy of monitoring.

[0043] 2. The present invention adopts an integrated model based on the random forest algorithm and the SVC algorithm, which fully utilizes the advantages of the two algorithms in the training and prediction process, improves the robustness and applicability of the early warning model, and fuses the prediction results of the two algorithms through weighted voting, so as to more accurately classify and predict meteorological events, thereby improving the reliability and accuracy of monitoring and early warning, and can realize all-weather and all-round monitoring of power grid equipment, greatly improving the efficiency and coverage of monitoring and early warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is an overall flow chart of a disaster prevention and mitigation monitoring and early warning method for power grid equipment described in the present invention. DETAILED DESCRIPTION

[0045] In order to make the objectives, technical solutions and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in combination with specific embodiments of the present application and with reference to the accompanying drawings.

[0046] Embodiment 1

[0047] This embodiment provides a disaster prevention and mitigation monitoring and early warning method for power grid equipment. Figure 1 As shown, the following steps are included:

[0048] S1: Acquire meteorological parameter data of power grid equipment, wherein the meteorological parameter data includes wind speed data, temperature data, humidity data, rainfall data, air pressure data, radiation data, etc. collected by various sensors, perform abnormal detection and processing on the meteorological parameter data, and obtain normal meteorological parameter data.

[0049] As a preferred implementation of this embodiment, the abnormality detection process for meteorological parameter data is specifically as follows:

[0050] S11: Perform density-based clustering on the acquired meteorological parameter data to obtain abnormality detection results of normal clusters and abnormal points. Specifically:

[0051] Set the neighborhood radius and minimum neighborhood density, and for each data point, calculate the number of data points centered on it and within the neighborhood radius based on the Harmanton distance.

[0052] Identify core points according to the set minimum neighborhood density, merge core points and their directly density-reachable points into clusters, find the point set directly reachable by its density for each core point, mark the core point as a new cluster label, and add all points in the point set to the cluster, and assign each boundary point to the cluster to which its neighbors belong.

[0053] Data points that are not assigned to any cluster are marked as outliers, and data points are divided into clusters and outliers.

[0054] S12: Calculate the local anomaly factors of all data in the meteorological parameter data by using the LOF algorithm to obtain the anomaly detection result of the LOF algorithm. Specifically:

[0055] For each data point, calculate the distance between it and other data points and sort them by distance. Select the distance of the kth nearest neighbor of each data point as the neighbor distance of the point. This k value is a hyperparameter, usually selected by cross-validation or heuristic methods.

[0056] For each data point, calculate its reachable density, which is the minimum reachable distance between the point and the farthest point in its neighborhood. This distance can be a measure of the local density of the point, reflecting the distribution of data points around the point.

[0057] For each data point, calculate its local anomaly factor, which is the ratio of the average reachable density of the point to its neighboring points. The local anomaly factor reflects the density of the point relative to its neighboring points, thus measuring the degree of anomaly of the point. The larger the local anomaly factor, the more abnormal the point is relative to its neighboring points.

[0058] S13: Comprehensively calculate the abnormal detection results obtained by the analysis of steps S11 and S12 to obtain a comprehensive abnormal detection result, retain the meteorological parameter data whose abnormal detection results are normal, and eliminate the data whose abnormal detection results are abnormal.

[0059] As a preferred implementation of this embodiment, step S13 is specifically as follows:

[0060] S131: The outlier detection results of the normal clusters and outliers are graded and set, and the grade values ​​are set for the data classified as the core points in the normal clusters, the data classified as the boundary points in the normal clusters, and the data classified as the outliers in turn.

[0061] S132: For each meteorological parameter data, the obtained anomaly detection result is comprehensively processed to obtain a comprehensive anomaly detection result, an anomaly detection threshold is set, the meteorological parameter data within the anomaly detection threshold range is retained, and the data exceeding the anomaly detection threshold range is eliminated. The formula for the comprehensive processing is specifically:

[0062] F(x i )=αg(x i )+βl(x i )

[0063] In the formula, F i is the anomaly detection result of the i-th meteorological parameter data; x i is the i-th meteorological parameter data; α is the clustering weight factor; g(x i ) is the classification value of the clustering anomaly detection result of the i-th meteorological parameter data; β is the weight factor of the LOF algorithm; l(x i ) is the local anomaly factor value of the i-th meteorological parameter data.

[0064] S2: Perform data preprocessing on normal meteorological parameter data, including standardization and data integration, to ensure that the numerical ranges of different features are consistent and to integrate multiple data sources into a consistent and complete data set. A meteorological parameter data set is constructed based on the preprocessed meteorological parameter data, and the meteorological parameter data set is divided into a training data set and a test data set.

[0065] S3: Establish a disaster prevention and mitigation monitoring and early warning integrated model, train and predictively evaluate the disaster prevention and mitigation monitoring and early warning integrated model based on the training data set and the test data set, and obtain the final disaster prevention and mitigation monitoring and early warning integrated model.

[0066] As a preferred implementation of this embodiment, step S3 is specifically:

[0067] S31: Establish a disaster prevention and mitigation monitoring and early warning integrated model based on the random forest algorithm and the SVC algorithm, and train the random forest model and the SVC model in the disaster prevention and mitigation monitoring and early warning integrated model according to the training data set.

[0068] Specifically, random forest is an ensemble learning method used to solve classification and regression problems. By building multiple decision trees and combining them for prediction, the accuracy and generalization ability of the model are improved. Random forest consists of multiple decision trees, each of which is built independently. When building each decision tree, a sample subset is randomly selected from the training set (with replacement sampling), and a feature subset is randomly selected as a candidate feature. Using the selected features and sample subsets, the nodes are split based on the value of the feature until the predefined stopping condition is reached. Random forest uses the idea of ​​ensemble learning to build a strong learner by combining multiple weak learners (decision trees). Each decision tree predicts the input sample, and the final prediction result is obtained by voting by all decision trees.

[0069] The SVC algorithm can handle multi-category classification problems. The core of the SVC algorithm is the support vector, which is the key sample point in the training data. These support vectors play a key role in determining the classification boundary. The SVC algorithm performs classification by finding the maximum margin hyperplane, that is, finding the best decision boundary that can separate sample points of different categories. The SVC algorithm can use the kernel function to map data to a high-dimensional space, thereby making linearly inseparable data linearly separable. In practical applications, the RBF kernel function can be used to map data to a high-dimensional space.

[0070] S32: Use the test data set to perform prediction evaluation on the trained random forest model and SVC model respectively with accuracy as the evaluation indicator, and use weighted voting to fusion the two prediction results of the disaster prevention and mitigation monitoring and early warning integrated model. Specifically, for each test sample, according to the prediction results of each basic model and the corresponding weights, the sum of weighted votes is calculated to obtain the final prediction result. According to the final prediction result, the parameters of the disaster prevention and mitigation monitoring and early warning integrated model are tuned to improve the prediction performance and obtain the final disaster prevention and mitigation monitoring and early warning integrated model.

[0071] S4: Collect real-time meteorological parameter data of power grid equipment, input the final disaster prevention and mitigation monitoring and early warning integrated model, and output meteorological event classification results. According to the actual design and task requirements of the disaster prevention and mitigation monitoring and early warning integrated model, the meteorological event classification results can be monitoring and early warning results such as wind disasters, rainstorms, floods, droughts, typhoons, and man-made risks. Further, the output meteorological event classification results are interpreted and subsequently processed, which may include converting the classification results into specific action recommendations or warning levels to help decision makers and relevant departments take corresponding response measures.

[0072] In order to verify the effectiveness and superiority of the method provided in this embodiment, some specific cases are provided below:

[0073] The simulation test was carried out through the Ubuntu Linux operating system, the Python language environment was configured, Jupyter Notebook and Scikit-learn were used as the simulation software, and the meteorological data parameters were obtained as shown in Table 1. The algorithm comparison results were shown in Table 2. In terms of the overall classification results, the average classification accuracy of the algorithm described in this embodiment was higher than that of the KNN algorithm.

[0074] Table 1 Meteorological data parameter unit table

[0075]

[0076]

[0077] Table 2 Algorithm comparison results

[0078] Classification result accuracy The algorithm described in this embodiment KNN Algorithm Typhoon 0.854 0.815 rainstorm 0.910 0.922 flood 0.923 0.857 drought 0.875 0.851 typhoon 0.817 0.789 Human risk 0.861 0.894

[0079] Embodiment 2

[0080] Accordingly, this embodiment provides a disaster prevention and mitigation monitoring and early warning system for power grid equipment, including a meteorological data collection module, a data set construction module, a model training and testing module, and a monitoring and early warning module.

[0081] The meteorological data acquisition module is used to obtain meteorological parameter data of power grid equipment, perform abnormal detection and processing on the meteorological parameter data, and obtain normal meteorological parameter data.

[0082] The data set construction module is used to perform data preprocessing on normal meteorological parameter data, construct a meteorological parameter data set, and divide the meteorological parameter data set into a training data set and a test data set.

[0083] The model training and testing module is used to establish an integrated model for disaster prevention and mitigation monitoring and early warning. The integrated model for disaster prevention and mitigation monitoring and early warning is trained and predicted and evaluated based on the training data set and the test data set to obtain the final integrated model for disaster prevention and mitigation monitoring and early warning.

[0084] The monitoring and early warning module is used to collect real-time meteorological parameter data from power grid equipment, input it into the final disaster prevention and mitigation monitoring and early warning integrated model, and output the meteorological event classification results.

[0085] As a preferred implementation of this embodiment, the meteorological data acquisition module includes a clustering anomaly detection module, a LOF anomaly detection module and a comprehensive anomaly detection module.

[0086] The clustering anomaly detection module is used to perform density-based clustering on the acquired meteorological parameter data to obtain anomaly detection results of normal clusters and abnormal points.

[0087] The LOF anomaly detection module is used to calculate the local anomaly factors of all data in the meteorological parameter data through the LOF algorithm to obtain the anomaly detection results of the LOF algorithm.

[0088] The comprehensive anomaly detection module is used to comprehensively calculate the anomaly detection results obtained by the above modules to obtain comprehensive anomaly detection results, retain the meteorological parameter data with normal anomaly detection results, and eliminate the data with abnormal anomaly detection results.

[0089] As a preferred implementation of this embodiment, the comprehensive anomaly detection module includes a classification setting module and a comprehensive processing module.

[0090] The classification setting module is used to classify the abnormal point detection results of normal clusters and abnormal points, and set the classification values ​​for the data classified as core points in the normal cluster, the data classified as boundary points in the normal cluster, and the data classified as abnormal points in turn.

[0091] The comprehensive processing module is used to comprehensively process the obtained anomaly detection results for each meteorological parameter data to obtain a comprehensive anomaly detection result, set an anomaly detection threshold, retain the meteorological parameter data within the anomaly detection threshold range, and remove the data exceeding the anomaly detection threshold range. The formula for the comprehensive processing is specifically:

[0092] F(x i )=αg(x i )+βl(x i )

[0093] In the formula, F i is the anomaly detection result of the i-th meteorological parameter data; x i is the i-th meteorological parameter data; α is the clustering weight factor; g(x i ) is the classification value of the clustering anomaly detection result of the i-th meteorological parameter data; β is the weight factor of the LOF algorithm; l(x i ) is the local anomaly factor value of the i-th meteorological parameter data.

[0094] As a preferred implementation of this embodiment, the model training and testing module includes a training module and a testing module.

[0095] The training module is used to establish an integrated model of disaster prevention and mitigation monitoring and early warning based on the random forest algorithm and the SVC algorithm, and to train the random forest model and the SVC model in the integrated model of disaster prevention and mitigation monitoring and early warning according to the training data set.

[0096] The testing module is used to use the test data set to perform prediction and evaluation on the trained random forest model and SVC model respectively, and to use weighted voting to fuse the two prediction results of the disaster prevention and mitigation monitoring and early warning integrated model to obtain the final prediction result. According to the final prediction result, the parameters of the disaster prevention and mitigation monitoring and early warning integrated model are tuned to obtain the final disaster prevention and mitigation monitoring and early warning integrated model.

[0097] Embodiment 3

[0098] This embodiment provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, a disaster prevention and mitigation monitoring and early warning method for power grid equipment as described in any embodiment of the present invention is implemented.

[0099] Embodiment 4

[0100] This embodiment provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, a disaster prevention and mitigation monitoring and early warning method for power grid equipment as described in any embodiment of the present invention is implemented.

[0101] In the embodiments of the present application, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent the existence of A alone, the existence of A and B at the same time, and the existence of B alone. Among them, A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b and c can be represented by: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, c can be single or multiple.

[0102] Those of ordinary skill in the art will appreciate that the various units and algorithm steps described in the embodiments disclosed herein can be implemented in a combination of electronic hardware, computer software, and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0103] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0104] In several embodiments provided in the present application, if any function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory; hereinafter referred to as: ROM), random access memory (Random Access Memory; hereinafter referred to as: RAM), disk or optical disk, and other media that can store program codes.

[0105] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A disaster prevention and mitigation monitoring and early warning method for power grid equipment, characterized in that: The following steps are involved: S1: Acquire meteorological parameter data of power grid equipment, perform abnormal detection on the meteorological parameter data, and obtain normal meteorological parameter data; S2: preprocessing the normal meteorological parameter data, constructing a meteorological parameter data set, and dividing the meteorological parameter data set into a training data set and a test data set; S3: Establish a disaster prevention and mitigation monitoring and early warning integrated model, train and predict and evaluate the disaster prevention and mitigation monitoring and early warning integrated model based on the training data set and the test data set, and obtain the final disaster prevention and mitigation monitoring and early warning integrated model; S4: Collect real-time meteorological parameter data from power grid equipment, input into the final disaster prevention and mitigation monitoring and early warning integrated model, and output meteorological event classification results.

2. A disaster prevention and mitigation monitoring and early warning method for power grid equipment according to claim 1, characterized in that: The specific process of anomaly detection on meteorological parameter data is as follows: S21: performing density-based clustering on the acquired meteorological parameter data to obtain anomaly detection results of normal clusters and abnormal points; S22: Calculate the local anomaly factors of all data in the meteorological parameter data by using the LOF algorithm to obtain an anomaly detection result of the LOF algorithm; S23: Comprehensively calculate the anomaly detection results obtained by analyzing steps S21 and S22 to obtain a comprehensive anomaly detection result, retain the meteorological parameter data whose anomaly detection results are normal, and eliminate the data whose anomaly detection results are abnormal.

3. A disaster prevention and mitigation monitoring and early warning method for power grid equipment according to claim 2, characterized in that: The step S23 is specifically as follows: S231: Classifying the abnormal point detection results of the normal clusters and abnormal points, and sequentially setting the classification values ​​for the data classified as the core points in the normal clusters, the data classified as the boundary points in the normal clusters, and the data classified as the abnormal points; S232: For each meteorological parameter data, the obtained anomaly detection result is comprehensively processed to obtain a comprehensive anomaly detection result, an anomaly detection threshold is set, the meteorological parameter data within the anomaly detection threshold range is retained, and the data exceeding the anomaly detection threshold range is eliminated. The formula for the comprehensive processing is specifically: F(x i )=αg(x i )+βl(x i ) In the formula, F i is the anomaly detection result of the i-th meteorological parameter data; x i is the i-th meteorological parameter data; α is the clustering weight factor; g(x i ) is the classification value of the clustering anomaly detection result of the i-th meteorological parameter data; β is the weight factor of the LOF algorithm; l(x i ) is the local anomaly factor value of the i-th meteorological parameter data.

4. A disaster prevention and mitigation monitoring and early warning method for power grid equipment according to claim 1, characterized in that: Step S3 is specifically as follows: S31: Establish a disaster prevention and mitigation monitoring and early warning integrated model based on the random forest algorithm and the SVC algorithm, and train the random forest model and the SVC model in the disaster prevention and mitigation monitoring and early warning integrated model according to the training data set; S32: Use the test data set to perform prediction and evaluation on the trained random forest model and SVC model respectively, use weighted voting to fuse the two prediction results of the disaster prevention and mitigation monitoring and early warning integrated model to obtain the final prediction result, and adjust the parameters of the disaster prevention and mitigation monitoring and early warning integrated model according to the final prediction result to obtain the final disaster prevention and mitigation monitoring and early warning integrated model.

5. A disaster prevention and mitigation monitoring and early warning system for power grid equipment, characterized in that: It includes meteorological data collection module, data set construction module, model training and testing module and monitoring and early warning module; The meteorological data acquisition module is used to obtain meteorological parameter data of power grid equipment, perform abnormal detection and processing on the meteorological parameter data, and obtain normal meteorological parameter data; A data set construction module is used to perform data preprocessing on normal meteorological parameter data, construct a meteorological parameter data set, and divide the meteorological parameter data set into a training data set and a test data set; The model training and testing module is used to establish a disaster prevention and mitigation monitoring and early warning integrated model, train and predict the disaster prevention and mitigation monitoring and early warning integrated model based on the training data set and the test data set to obtain the final disaster prevention and mitigation monitoring and early warning integrated model; The monitoring and early warning module is used to collect real-time meteorological parameter data from power grid equipment, input it into the final disaster prevention and mitigation monitoring and early warning integrated model, and output the meteorological event classification results.

6. A disaster prevention and mitigation monitoring and early warning system for power grid equipment according to claim 5, characterized in that: The meteorological data collection module includes cluster anomaly detection module, LOF anomaly detection module and comprehensive anomaly detection module: The clustering anomaly detection module is used to perform density-based clustering on the acquired meteorological parameter data to obtain anomaly detection results of normal clusters and abnormal points; The LOF anomaly detection module is used to calculate the local anomaly factors of all data in the meteorological parameter data through the LOF algorithm to obtain the anomaly detection results of the LOF algorithm; The comprehensive anomaly detection module is used to comprehensively calculate the anomaly detection results obtained by the above modules to obtain comprehensive anomaly detection results, retain the meteorological parameter data with normal anomaly detection results, and eliminate the data with abnormal anomaly detection results.

7. A disaster prevention and mitigation monitoring and early warning system for power grid equipment according to claim 6, characterized in that: The comprehensive anomaly detection module includes a classification setting module and a comprehensive processing module; A classification setting module is used to classify the abnormal point detection results of normal clusters and abnormal points, and set the classification values ​​for the data classified as core points in the normal cluster, the data classified as boundary points in the normal cluster, and the data classified as abnormal points in turn; The comprehensive processing module is used to comprehensively process the obtained anomaly detection results for each meteorological parameter data to obtain a comprehensive anomaly detection result, set an anomaly detection threshold, retain the meteorological parameter data within the anomaly detection threshold range, and remove the data exceeding the anomaly detection threshold range. The formula for the comprehensive processing is specifically: F(x i )=αg(x i )+βl(x i ) In the formula, F i is the anomaly detection result of the i-th meteorological parameter data; x i is the i-th meteorological parameter data; α is the clustering weight factor; g(x i ) is the classification value of the clustering anomaly detection result of the i-th meteorological parameter data; β is the weight factor of the LOF algorithm; l(x i ) is the local anomaly factor value of the i-th meteorological parameter data.

8. A disaster prevention and mitigation monitoring and early warning system for power grid equipment according to claim 5, characterized in that: The model training and testing module includes a training module and a testing module; A training module is used to establish a disaster prevention and mitigation monitoring and early warning integrated model based on the random forest algorithm and the SVC algorithm, and to train the random forest model and the SVC model in the disaster prevention and mitigation monitoring and early warning integrated model according to the training data set; The testing module is used to use the test data set to perform prediction and evaluation on the trained random forest model and SVC model respectively, and to use weighted voting to fuse the two prediction results of the disaster prevention and mitigation monitoring and early warning integrated model to obtain the final prediction result. According to the final prediction result, the parameters of the disaster prevention and mitigation monitoring and early warning integrated model are tuned to obtain the final disaster prevention and mitigation monitoring and early warning integrated model.

9. An electronic device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements a disaster prevention and mitigation monitoring and early warning method for power grid equipment as described in any one of claims 1 to 4 when executing the computer program.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, a disaster prevention and mitigation monitoring and early warning method for power grid equipment according to any one of claims 1 to 4 is implemented.

Images