Power grid inspection optimization method and system based on digital twin

Abstract

The digital twin-based power grid inspection optimization method and system provided in the present application relate to the field of data processing technology. The temperature information of normal positions in the target power grid is collected through a sensor array to obtain a missing temperature information array and a missing position; the inspection data accuracy analysis is performed to obtain the overall accuracy; the temperature information of the missing position is supplemented and optimized to obtain a temperature information array; the temperature information array is optimized for dimensionality reduction according to the overall accuracy to obtain the optimal dimensionality reduction ratio, and dimensionality reduction processing is performed to obtain a reduced-dimensional temperature information array; a temperature inspection twin model of the target power grid is constructed based on the digital twin for inspection judgment, which solves the technical problem that it is impossible to monitor the data information of the power grid in real time and comprehensively, and it is difficult to ensure the accuracy and comprehensiveness of the inspection data, improves the accuracy and response speed of power grid data monitoring, and effectively improves the data processing efficiency and the intelligence level of the inspection.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a power grid inspection optimization method and system based on digital twins. Background Art

[0002] With the continuous growth of global energy demand and the continuous improvement of industrialization level, the scale and complexity of power grid systems are also expanding, which not only increases the difficulty of power grid maintenance, but also puts forward higher requirements for the operation efficiency and safety of power grids. Traditional power grid inspection methods mainly rely on manual inspections or basic automated monitoring systems. These methods have defects such as slow response speed, insufficient monitoring accuracy and poor adaptability, and it is difficult to effectively cope with the challenges of current power grid operation and maintenance. In the existing technology, the inspection and maintenance of power grids are often completed by periodic manual inspections. This method is not only time-consuming, but also unable to achieve real-time monitoring of the power grid status. In addition, although some automated systems are capable of performing basic monitoring tasks, these systems often lack the ability to predict and respond quickly to complex events and sudden failures. Therefore, how to improve the real-time, accuracy and intelligence level of power grid inspection has become a technical problem that needs to be solved in the field of power grid management.

[0003] In summary, existing technologies often have the problem of being unable to comprehensively monitor data information at various locations in the power grid in real time. Especially when data collection is incomplete or lost, it is difficult to ensure the accuracy and comprehensiveness of inspection data, resulting in technical problems such as low response speed and reliability. Summary of the invention

[0004] The present application provides a power grid inspection optimization method and system based on digital twins, which is used to solve the technical problem that the prior art cannot comprehensively monitor the data information of each location in the power grid in real time, especially when data collection is incomplete or lost, it is difficult to ensure the accuracy and comprehensiveness of the inspection data, resulting in low response speed and reliability.

[0005] In view of the above problems, the present application provides a power grid inspection optimization method and system based on digital twins.

[0006] In a first aspect, the present application provides a power grid inspection optimization method based on digital twins, the method comprising:

[0007] Based on the inspection sensor array arranged in the target power grid, the temperature information of multiple normal positions detected in the target power grid is collected to obtain the missing temperature information array, and the multiple missing positions where the temperature information has not been collected are determined; based on the missing temperature information array, the inspection data accuracy analysis is performed to obtain the overall accuracy; based on the missing temperature information array and the overall accuracy, the temperature information of the multiple missing positions is supplemented and optimized to obtain multiple optimal supplementary temperature information, the missing temperature information array is supplemented to obtain a temperature information array; based on the overall accuracy, the temperature information array is optimized by dimensionality reduction to obtain the optimal dimensionality reduction ratio, the temperature information array is processed by dimensionality reduction to obtain a reduced dimensionality temperature information array; based on the digital twin, a temperature inspection twin model of the target power grid is constructed according to the reduced dimensionality temperature information array to perform inspection judgment.

[0008] In a second aspect, the present application provides a power grid inspection optimization system based on digital twins, the system comprising:

[0009] A temperature information acquisition module is used to collect temperature information of multiple normal positions detected in the target power grid based on the patrol sensor array deployed in the target power grid, obtain a missing temperature information array, and determine multiple missing positions where temperature information has not been collected; a data accuracy analysis module is used to perform patrol data accuracy analysis based on the missing temperature information array to obtain overall accuracy; a supplementary optimization module is used to supplement and optimize the temperature information of the multiple missing positions based on the missing temperature information array and the overall accuracy, obtain multiple optimal supplementary temperature information, supplement the missing temperature information array, and obtain a temperature information array; a dimensionality reduction processing module is used to perform dimensionality reduction optimization of the temperature information array based on the overall accuracy, obtain an optimal dimensionality reduction ratio, perform dimensionality reduction processing on the temperature information array, and obtain a reduced dimensionality temperature information array; an inspection judgment module is used to construct a temperature inspection twin model of the target power grid based on the reduced dimensionality temperature information array based on digital twins, and perform inspection judgment.

[0010] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0011] The power grid inspection optimization method based on digital twin provided in the present application collects temperature information of multiple normal positions detected in the target power grid based on the inspection sensor array arranged in the target power grid, obtains a missing temperature information array, and determines multiple missing positions where temperature information has not been collected; performs inspection data accuracy analysis based on the missing temperature information array to obtain overall accuracy; performs supplementary optimization on the temperature information of the multiple missing positions based on the missing temperature information array and the overall accuracy to obtain multiple optimal supplementary temperature information, supplements the missing temperature information array to obtain a temperature information array; performs dimensionality reduction optimization on the temperature information array based on the overall accuracy, An optimal dimensionality reduction ratio is obtained, and the temperature information array is subjected to dimensionality reduction processing to obtain a reduced-dimensional temperature information array; based on the digital twin, a temperature inspection twin model of the target power grid is constructed according to the reduced-dimensional temperature information array to perform inspection judgment, thereby solving the technical problem that the data information of each location in the power grid cannot be fully monitored in real time, especially when data collection is incomplete or lost, it is difficult to ensure the accuracy and comprehensiveness of the inspection data, resulting in low response speed and reliability. Through real-time monitoring, data analysis and supplementary optimization, the accuracy and response speed of power grid data monitoring are significantly improved. At the same time, the use of dimensionality reduction processing and digital twin models effectively improves the data processing efficiency and the intelligence level of inspection. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 A flow chart of a power grid inspection optimization method based on digital twin is provided for this application.

[0013] Figure 2 A structural schematic diagram of a power grid inspection and optimization system based on digital twins is provided for this application.

[0014] Explanation of the reference numerals: temperature information acquisition module 11 , data accuracy analysis module 12 , supplementary optimization module 13 , dimensionality reduction processing module 14 , inspection and discrimination module 15 . DETAILED DESCRIPTION

[0015] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0016] Embodiment 1, as Figure 1 As shown, the present application provides a power grid inspection optimization method based on digital twins, the method comprising:

[0017] Step S100: Based on the inspection sensor array arranged in the target power grid, temperature information of multiple normal positions detected in the target power grid is collected, an array of missing temperature information is obtained, and multiple missing positions where temperature information is not collected are determined.

[0018] Power grid inspection refers to the regular or irregular inspection, testing and evaluation of various equipment, lines and related facilities of the power system to ensure the continuity, reliability and safety of power supply, prevent power facility failures, reduce power outages, and extend the service life of equipment. Since temperature abnormalities usually indicate equipment failure, overload or insulation loss, they reflect more inspection characteristics and accurate power grid conditions, which is crucial for early identification and prevention of power grid failures to ensure stable operation of the system. Therefore, in this embodiment, the inspection results are obtained mainly based on analysis based on temperature data.

[0019] Specifically, a series of patrol sensors are first arranged at key locations within the target power grid, which are used to capture temperature data during the operation of the power grid. All patrol sensors together form a patrol sensor array, which covers various parts of the power grid, such as substations, transmission lines, and distribution facilities to ensure comprehensive monitoring. Among them, each sensor automatically monitors and records the temperature information of its location, which is sent back to the central monitoring system in real time. For example, the sensor can record temperature data once a minute to ensure the timeliness and accuracy of the data. At the same time, during the data collection process, the system automatically analyzes the received temperature information to identify which locations have failed to collect data successfully, which may be due to sensor failure, communication interruption or external environmental factors. The missing data location is marked by the system and stored in a missing temperature information array, which records the specific location and time of each unsuccessful data collection in detail. Furthermore, once the specific location of the missing temperature information is identified and recorded, the system will further analyze the pattern and frequency of these locations to determine the cause of the missing data and optimize the sensor network. For example, if data is frequently missing in a specific area, it may indicate that the sensors in that area need maintenance or new sensors need to be added.

[0020] Through the above steps, the temperature information in the power grid can be effectively monitored and managed, and the data missing problem can be discovered and handled in time, thereby improving the accuracy and reliability of power grid inspection. In addition, this real-time data collection and analysis based on the sensor array is a key means to improve the efficiency of power grid operation monitoring.

[0021] Step S200: performing inspection data accuracy analysis based on the missing temperature information array to obtain overall accuracy.

[0022] Optionally, an array of missing temperature information that has been collected by the sensor array but some data has not been successfully transmitted is determined. The array contains the detailed locations of all temperature data points that have not been successfully recorded and their corresponding timestamps. This information is essential for analyzing the integrity and accuracy of the overall data collection. Further, based on the collected temperature data, the consistency and logic of the temperature readings at adjacent locations are analyzed. For example, assuming that the temperature at one location of the transformer is recorded as A, because the physical location and environmental conditions are similar, the temperature at the adjacent location is expected to be close to A. If the temperature reading at the adjacent location is too different from A, it may indicate inaccurate data collection or abnormal sensor function. Next, a statistical analysis method, such as standard deviation or variance analysis, is used to evaluate all collected temperature data as a whole. The consistency and reliability of the data can be quantified through overall evaluation and analysis. The overall accuracy is determined by calculating the average and distribution of the temperature difference between each data point and its adjacent data point. The overall accuracy directly affects the reliability of the inspection results.

[0023] Through the above steps, not only can the problem areas in data collection be effectively identified, but also the high accuracy of the data can be ensured through continuous monitoring and analysis, thereby improving the overall efficiency and reliability of power grid inspections. This process ensures the power grid monitoring system's ability to respond quickly to abnormal conditions, greatly enhancing the operational safety and stability of the power grid.

[0024] Step S300: According to the missing temperature information array and the overall accuracy, the temperature information of the multiple missing positions is supplemented and optimized to obtain multiple optimal supplementary temperature information, and the missing temperature information array is supplemented to obtain a temperature information array.

[0025] Further, all missing positions marked in the missing temperature information array are classified and analyzed, and possible temperature values ​​are predicted based on the existing temperature data and the geographical and environmental characteristics of the missing positions. This prediction is based on machine learning models or statistical inferences and aims to provide a preliminary estimate for the missing data. During the supplementary optimization process, a series of possible temperature values ​​are randomly generated for supplementation. These temperature values ​​are obtained based on the prediction model and the measured data of adjacent positions. For example, if the normal temperature range of a transformer location is 35°C to 40°C, a temperature value may be randomly generated within this range as supplementary information. After supplementation, the overall accuracy is recalculated and the consistency of the newly supplemented temperature information with the existing data is evaluated. This step is completed by comparing the consistency of the data before and after supplementation and the change in the overall temperature distribution. If the supplemented data causes a significant decrease in overall accuracy, other temperature values ​​may need to be reselected for supplementation. After multiple rounds of supplementation and evaluation, the supplementary temperature information that minimizes the overall accuracy deviation is selected. This process involves iterative testing of different supplementary values ​​until the most suitable data point is found to ensure that the supplemented temperature information array not only completes the data, but also does not destroy the accuracy and consistency of the original data. Once the optimal supplementary temperature information is obtained, this data is formally integrated into the existing temperature information array, thereby ensuring the integrity and high accuracy of the data array, and providing a reliable basis for subsequent data processing and analysis, such as dimensionality reduction.

[0026] Through the above steps, the problem of data incompleteness is effectively solved, the quality of the overall data set is optimized, and the consistency and accuracy of subsequent data processing are ensured, thereby greatly improving the efficiency and reliability of power grid inspection. At the same time, it not only improves the practicality of data, but also ensures the efficient operation of the power grid monitoring system.

[0027] Step S400: According to the overall accuracy, perform dimensionality reduction optimization on the temperature information array to obtain an optimal dimensionality reduction ratio, perform dimensionality reduction processing on the temperature information array to obtain a reduced-dimensional temperature information array.

[0028] Exemplarily, the dimensionality reduction optimization of the temperature information array is performed according to the overall accuracy, that is, the obtained data accuracy is used to guide how to reduce (reduce) the size of the temperature data set while maintaining the integrity of key information, thereby optimizing the subsequent processing and analysis efficiency. In this process, the overall accuracy is used as an evaluation indicator to help determine the degree of dimensionality reduction processing to ensure that despite the reduction in the amount of data, the retained data can still effectively reflect the actual operating status of the power grid and support reliable decision-making and monitoring. Specifically, if the overall accuracy of the data is high, it means that the information in the original data is relatively accurate and reliable, so a higher dimensionality reduction ratio (i.e., less data is retained) can be adopted, because even with a significant reduction in the amount of data, the remaining data is still sufficient to maintain the effectiveness of analysis and monitoring. On the contrary, if the data accuracy is low, then more caution is needed when performing dimensionality reduction, and more data may need to be retained to ensure sufficient information to avoid losing important monitoring signals. This method effectively balances the efficiency and quality of data processing, ensuring that the necessary analysis and decision accuracy is not sacrificed while reducing the burden of data processing.

[0029] The dimensionality reduction ratio is the ratio of the number of temperature information in the temperature information array after dimensionality reduction to the number of temperature information in the original temperature information array. For example, if the original data contains 1,000 data points, and 600 data points are retained after dimensionality reduction processing, the dimensionality reduction ratio is 60%, which is directly related to the complexity and computational efficiency of data processing. According to the overall accuracy, the corresponding dimensionality reduction ratio is selected for dimensionality reduction processing, and data dimensionality reduction techniques such as principal component analysis (PCA) or autoencoders are used to systematically reduce the dimension of temperature information while retaining the most critical information as much as possible. This step is to improve data processing speed and efficiency while reducing storage requirements.

[0030] Furthermore, simulations were performed using different overall sample accuracy and dimensionality reduction ratios to analyze the accuracy of digital twin inspections at different dimensionality reduction ratios. For example, the accuracy probability of judging that the temperature of power grid equipment is too high was evaluated at a dimensionality reduction ratio of 60%. This analysis helps determine the optimal dimensionality reduction ratio to ensure that both high accuracy and efficiency can be maintained. By comparing the inspection accuracy at different dimensionality reduction ratios, an optimal dimensionality reduction ratio is selected. This ratio should be the ratio that maximizes efficiency and reduces data processing requirements while ensuring accuracy. After determining the optimal dimensionality reduction ratio, the entire temperature information array is formally reduced in dimensionality to generate the final reduced dimensionality temperature information array, which will be used for subsequent digital twin model construction and inspection tasks.

[0031] Through the above steps, a method for dynamically adjusting data processing depth according to data accuracy is realized, which optimizes the data processing process, improves the efficiency and accuracy of power grid inspection, and effectively supports power grid health management and fault prevention.

[0032] Step S500: Based on the digital twin, a temperature inspection twin model of the target power grid is constructed according to the reduced-dimensional temperature information array to perform inspection judgment.

[0033] Specifically, digital twin technology creates virtual copies of physical entities, making it possible to simulate and analyze the behavior of entities in digital space. In this embodiment, the temperature inspection digital twin model of the target power grid is first constructed using the temperature information array after dimensionality reduction. This model not only contains the structural information of the power grid, such as the location and connection method of wires and transformers, but also integrates the data of temperature monitoring points, which are collected from the actual power grid through the sensor array and optimized. Further, the temperature inspection twin model of the target power grid is constructed based on the reduced-dimensional temperature information array. The use of these selected data can effectively simulate the temperature performance of the power grid under different operating conditions.

[0034] Next, the constructed digital twin model is used for inspection and judgment, that is, to evaluate the current and future status of the power grid. Specifically, the trend of temperature data changes in time series is analyzed to identify possible abnormal patterns, such as a sudden increase in temperature, which may indicate equipment failure or overload. In addition, by comparing historical data with real-time data, problems that may occur in specific areas can be predicted, so that warnings can be given in advance.

[0035] Through the above steps, not only the efficiency and accuracy of power grid inspection are improved, but also the reliability and safety of the power grid can be improved through real-time monitoring and prediction. This digital twin-based method can effectively reduce power outages and equipment failures, and provide strong technical support for power grid operation and maintenance.

[0036] Furthermore, based on the inspection sensor array arranged in the target power grid, the temperature information of multiple normal positions detected in the target power grid is collected to obtain the missing temperature information array. Step S100 of the present application further includes:

[0037] Step S110: deploying inspection sensors at multiple locations in the target power grid to obtain the inspection sensor array.

[0038] Step S120: Acquire temperature information of multiple normal positions detected and successfully transmitted by the inspection sensor array, and construct a missing temperature information array by combining the position information of the multiple normal positions.

[0039] Step S130: treating the multiple positions where the temperature information is not obtained as multiple missing positions.

[0040] Optionally, patrol sensors are first deployed at multiple key locations of the target power grid. These sensors are specifically designed to monitor and record temperature data from various parts of the power grid. The selection and layout of sensors take into account the geographical layout of the power grid, the location of key equipment, and historical fault data to ensure comprehensive coverage of important parts of the power grid. Each sensor regularly monitors the temperature at its location and transmits the data back to the central monitoring system in real time. During this process, the sensor records the temperature reading with the corresponding timestamp and location information. For example, a sensor on a transformer records the temperature every 10 minutes to ensure that any critical temperature changes can be captured. At the same time, during the data collection process, the system automatically detects the data returned by all sensors. If the data from some locations cannot be successfully received (possibly due to sensor failure, communication failure or other technical problems), these locations will be marked as missing locations. These missing locations are recorded in the missing temperature information array, which is a database that details the specific location and scheduled collection time of each missing data.

[0041] Through the above steps, comprehensive temperature monitoring of the target power grid is ensured, and problems in data collection are discovered and solved in a timely manner, thereby improving the accuracy and reliability of power grid monitoring, helping to identify potential equipment failures and abnormal conditions in advance, and then taking appropriate preventive measures to reduce failure rates and maintenance costs.

[0042] Furthermore, according to the missing temperature information array, the inspection data accuracy analysis is performed to obtain the overall accuracy. The step S200 of the present application further includes:

[0043] Step S210: According to the historical records of temperature inspection of the target power grid, a set of sample missing temperature information arrays is collected, and according to the matching degree of temperature information of multiple normal positions in each sample missing temperature information array, an overall accuracy set of samples is obtained by marking.

[0044] Step S220: using the sample missing temperature information array set and the sample overall accuracy set to train an inspection accuracy analyzer, wherein the inspection accuracy analyzer includes a plurality of inspection accuracy analysis branches corresponding to a plurality of missing position combinations.

[0045] Step S230: According to the missing position combination of the multiple missing positions, the missing temperature information array is input into the corresponding inspection accuracy analysis branch in the inspection accuracy analyzer to analyze and obtain the overall accuracy.

[0046] Furthermore, historical temperature inspection records of the target power grid are collected, with special attention paid to the missing temperature information in the records. These historical records form a set of sample missing temperature information arrays. For each sample, the degree of match between the temperature information in the record and the temperature information of the surrounding (or related) locations is analyzed as an important criterion for evaluating data accuracy. For example, if the normal operating temperature of a transformer is very different from that of an adjacent transformer, this may indicate a data error or equipment abnormality. Based on the above analysis, an overall accuracy is labeled for each sample. This accuracy reflects the reliability and consistency of the temperature information, that is, the degree of conformity of the sample data with the expected model or data in a similar environment. The higher the overall accuracy, the more reliable the data, and vice versa.

[0047] The obtained sample missing temperature information array set and the corresponding overall accuracy set are used to train a complex inspection accuracy analyzer, which contains multiple inspection accuracy analysis branches for different missing position combinations. Each branch specializes in processing a specific type of missing pattern and recovers or estimates the missing data from it. For example, if temperature data is frequently missing in a specific area, the relevant analysis branch will focus on inferring the most likely temperature of these missing points based on adjacent data. In actual inspections, the currently acquired missing temperature information array is input into the corresponding branch of the inspection accuracy analyzer, which uses the model established during training to evaluate the overall accuracy of the current data. In this way, not only can data quality be monitored in real time, but potential errors can also be predicted and corrected, thereby maintaining high accuracy in data monitoring.

[0048] Furthermore, the sample missing temperature information array set and the sample overall accuracy set are used to train the inspection accuracy analyzer, and step S220 of the present application further includes:

[0049] Step S221: acquiring a plurality of missing position combinations according to the historical inspection records of the target power grid.

[0050] Step S222: dividing the sample missing temperature information array set and the sample overall accuracy set according to the multiple missing position combinations, and training multiple inspection accuracy analysis branches respectively, wherein the multiple missing position combinations include different numbers of missing position combinations.

[0051] Step S223: Based on the multiple inspection accurate analysis branches, obtain the inspection accurate analyzer.

[0052] Specifically, data is extracted from the historical inspection records of the target power grid. These records contain temperature information at various detection points at different time points, as well as the specific location and situation of missing data in the records. These historical data provide the basis for training the inspection accuracy analyzer. Then, multiple combinations of locations where missing temperature information occurs are analyzed and identified. These combinations reflect specific areas or patterns in the power grid where data may be missing due to equipment failure, environmental factors or other reasons. For example, a specific transformer area may often have missing data due to its environmental sensitivity. Next, the sample missing temperature information array and the sample overall accuracy set are divided according to the different missing location combinations identified. Each specific missing location combination corresponds to an inspection accuracy analysis branch, and these branches are trained separately according to their corresponding data missing patterns. During the training process, each branch learns how to estimate and correct the missing data based on the surrounding normal temperature data and known missing patterns to improve the accuracy and completeness of the data set. After the training is completed, all inspection accuracy analysis branches are integrated into a comprehensive inspection accuracy analyzer, which can quickly determine the applicable analysis branch based on the input current missing temperature information array, and use the branch to evaluate and correct the data missing problem. For example, if the current missingness pattern matches a certain historical pattern, the corresponding analysis branch will be used to predict the missing data and estimate the overall accuracy.

[0053] Through the above steps, not only the historical inspection data is effectively utilized to enhance the monitoring and analysis capabilities of the power grid, but also the response capability to specific missing patterns is improved through the dedicated inspection accurate analysis branch. This method greatly improves the data quality and reliability of power grid inspection, thereby ensuring the stable operation and efficient management of the power grid.

[0054] Further, according to the missing temperature information array and the overall accuracy, the temperature information of the multiple missing positions is supplemented and optimized to obtain multiple optimal supplementary temperature information, and the missing temperature information array is supplemented. Step S300 of the present application also includes:

[0055] Step S310: constructing a supplementary optimization space according to the historical inspection temperature intervals of the plurality of missing positions.

[0056] Step S320: randomly generating a plurality of first supplementary temperature information of the plurality of missing positions in the supplementary optimization space.

[0057] Step S330: adding the plurality of first supplementary temperature information to the missing temperature information array for supplementation to obtain a first supplementary temperature information array.

[0058] Step S340: performing inspection data accuracy analysis on the first supplementary temperature information array to obtain a first overall accuracy, and calculating a deviation from the overall accuracy to obtain a first accuracy deviation.

[0059] Step S350: Continue to iteratively supplement and optimize the temperature information of the multiple missing positions in the supplementary optimization space until convergence, and output multiple supplementary temperature information with the smallest accuracy deviation as multiple optimal supplementary temperature information to supplement the missing temperature information array to obtain a temperature information array.

[0060] Further, based on the historical inspection temperature data of multiple missing locations known in the target power grid, a supplementary optimization space is constructed, which defines the range of possible temperature values, which should be statistically consistent with the historical data. For example, if the temperature displayed in the past records of a missing location is usually between 20°C and 25°C, the supplementary optimization space will be limited to this range or slightly expanded. In the supplementary optimization space, a set of initial supplementary temperature information is randomly generated, which is selected to fill the current missing temperature data points. The purpose of this step is to provide a reasonable temperature estimate based on historical data for the missing location to restore the consistency of the data. Then, the generated first supplementary temperature information is added to the original missing temperature information array to form an updated temperature information array as the first supplementary temperature information array. Next, the accuracy of the inspection data is analyzed on this updated array to evaluate the overall accuracy of the supplemented data set and calculate the deviation from the original overall accuracy. If there is a large deviation between the overall accuracy of the first supplement and the accuracy of the original data, the system will continue to iteratively supplement the temperature information in the supplementary optimization space, and each iteration will adjust the supplementary temperature information to reduce the deviation of the overall accuracy. This process is repeated until the minimum value of the accuracy deviation is reached, or until the accuracy improvement is no longer significant. Finally, when the accuracy deviation is minimized, the obtained set of supplementary temperature information is regarded as the optimal solution, and this information is formally integrated into the missing temperature information array, thus completing the supplement of the entire temperature information array and obtaining a more complete and accurate data set for subsequent power grid monitoring and analysis.

[0061] Through the above steps, the problem of missing data in power grid temperature monitoring can be effectively solved, and the accuracy and reliability of data can be improved, making power grid operation and maintenance more efficient and safer.

[0062] Further, according to the overall accuracy, the temperature information array is optimized for dimensionality reduction to obtain an optimal dimensionality reduction ratio, and the temperature information array is subjected to dimensionality reduction processing to obtain a reduced-dimensional temperature information array. Step S400 of the present application further includes:

[0063] Step S410: Obtain a dimension reduction ratio space.

[0064] Step S420: randomly selecting a first dimensionality reduction ratio in the dimensionality reduction ratio space, and performing inspection discrimination accuracy analysis on the temperature information array after dimensionality reduction in combination with the overall accuracy to obtain a first inspection accuracy.

[0065] Step S430: Continue to perform dimensionality reduction optimization in the dimensionality reduction ratio space until convergence, and output the dimensionality reduction ratio with the maximum inspection accuracy as the optimal dimensionality reduction ratio.

[0066] Step S440: performing dimension reduction processing on the temperature information array according to the optimal dimension reduction ratio, retaining the temperature information with the optimal dimension reduction ratio, and obtaining a dimension-reduced temperature information array.

[0067] Optionally, define a dimensionality reduction ratio space, which is a set of all possible dimensionality reduction ratios. The dimensionality reduction ratio refers to the ratio of the amount of data retained to the amount of original data after dimensionality reduction. For example, a range from 90% to 10% may be considered a reasonable dimensionality reduction ratio space, allowing to minimize information loss while retaining less data. Randomly select a preliminary dimensionality reduction ratio in the dimensionality reduction ratio space, such as 50% as the first attempt. Combined with the known overall accuracy, perform preliminary dimensionality reduction on the temperature information array. This step uses statistical methods such as principal component analysis (PCA) to reduce the data dimension while retaining the most important information as much as possible.

[0068] After the dimensionality reduction process, the updated temperature information array is used for patrol judgment. The patrol judgment accuracy refers to the impact of the data after dimensionality reduction on the ability to distinguish the power grid state, that is, analyzing the performance of the reduced dimensionality data set under simulated power grid operation conditions. In this way, the effect of the initial dimensionality reduction ratio is evaluated. Then, iterative optimization is continued in the dimensionality reduction ratio space, different dimensionality reduction ratios are tested and the patrol judgment accuracy analysis is repeated, and the ratio is adjusted each iteration to find a better solution. During the iteration process, the system monitors the changes in patrol accuracy and searches for the dimensionality reduction ratio that maximizes the patrol accuracy. When the iterative process converges, that is, a maximum patrol accuracy value that cannot be surpassed by other ratios within the test range is found, the dimensionality reduction ratio is determined as the optimal dimensionality reduction ratio. Subsequently, the original temperature information array is subjected to final dimensionality reduction processing according to this ratio to generate a final reduced dimensionality temperature information array.

[0069] Through the above steps, data dimensionality reduction is effectively achieved while ensuring the quality and accuracy of power grid inspection data, improving the efficiency of data processing, and reducing the consumption of computing resources, making the power grid monitoring system more efficient and reliable.

[0070] Furthermore, a first dimensionality reduction ratio is randomly selected in the dimensionality reduction ratio space, and the inspection discrimination accuracy analysis after the dimensionality reduction of the temperature information array is performed in combination with the overall accuracy. Step S420 of the present application further includes:

[0071] Step S421: randomly selecting a first dimensionality reduction ratio in the dimensionality reduction ratio space.

[0072] Step S422: According to the temperature inspection history of the target power grid, a sample overall accuracy set and a sample dimensionality reduction ratio set are collected, and the accuracy of inspection judgment is performed according to the temperature information arrays with different sample overall accuracies and different sample dimensionality reduction ratios, and the sample inspection accuracy set is marked.

[0073] Step S423: using the sample overall accuracy set and the sample dimension reduction ratio set as input data, using the sample inspection accuracy set as supervision data, and training the dimension reduction inspection accuracy analyzer.

[0074] Step S424: using the dimension reduction inspection accuracy analyzer, performing inspection discrimination accuracy analysis on the first dimension reduction ratio and the overall accuracy after the dimension reduction of the temperature information array, to obtain a first inspection accuracy.

[0075] Specifically, in the defined dimensionality reduction ratio space, a dimensionality reduction ratio is randomly selected as the starting point. This starting point is the first dimensionality reduction ratio, which represents the proportion of the original temperature information array retained during the dimensionality reduction process. For example, if the selected dimensionality reduction ratio is 50%, it means that half of the original data will be retained during the dimensionality reduction process. Further, using the historical temperature inspection records of the target power grid, the sample overall accuracy set and the sample dimensionality reduction ratio set are collected and analyzed. These sets record the accuracy of the historical data and the dimensionality reduction ratio applied to these data, respectively. These sample data are analyzed to evaluate the inspection discrimination accuracy of temperature information arrays with different overall accuracies after applying different dimensionality reduction ratios, and the sample inspection accuracy set is labeled accordingly.

[0076] Next, the collected sample overall accuracy set and sample dimensionality reduction ratio set are used as input data, and the sample inspection accuracy set is used as supervision data to train the dimensionality reduction inspection accuracy analyzer. The purpose of this analyzer is to learn how to accurately predict the state of the power grid based on the dimensionality reduced data to achieve efficient and accurate inspections. Finally, the trained dimensionality reduction inspection accuracy analyzer is used to evaluate the randomly selected first dimensionality reduction ratio. This ratio is applied to the current temperature information array, and the inspection discrimination accuracy of the dimensionality reduced data is evaluated by the analyzer. This step will determine whether the initially selected dimensionality reduction ratio is effective, or whether further adjustment is needed to achieve a better inspection effect.

[0077] Through the above steps, not only the processing efficiency of power grid temperature information is improved, but also the accuracy of inspection is optimized by accurately adjusting the proportion of data retention.

[0078] Through the technical solutions of the above-mentioned embodiments, the digital twin-based power grid inspection optimization method provided in the present application solves the technical problems existing in the prior art that the data information of each location in the power grid cannot be monitored comprehensively and in real time, especially when data collection is incomplete or lost, it is difficult to ensure the accuracy and comprehensiveness of the inspection data, resulting in low response speed and reliability. Through real-time monitoring, data analysis and supplementary optimization, the accuracy and response speed of power grid data monitoring are significantly improved. At the same time, the use of dimensionality reduction processing and digital twin models effectively improves the data processing efficiency and the intelligence level of inspection.

[0079] Embodiment 2 is based on the same inventive concept as the power grid inspection optimization method based on digital twin in the previous embodiment. Figure 2 As shown, the present application provides a power grid inspection optimization system based on digital twins, the system comprising:

[0080] The temperature information collection module 11 is used to collect temperature information of multiple normal positions detected in the target power grid based on the inspection sensor array deployed in the target power grid, obtain a missing temperature information array, and determine multiple missing positions where no temperature information is collected.

[0081] The data accuracy analysis module 12 is used to perform inspection data accuracy analysis based on the missing temperature information array to obtain overall accuracy.

[0082] The supplementary optimization module 13 is used to supplement and optimize the temperature information of the multiple missing positions according to the missing temperature information array and the overall accuracy, obtain multiple optimal supplementary temperature information, supplement the missing temperature information array, and obtain a temperature information array.

[0083] The dimension reduction processing module 14 is used to perform dimension reduction optimization of the temperature information array according to the overall accuracy, obtain an optimal dimension reduction ratio, perform dimension reduction processing on the temperature information array, and obtain a reduced-dimensional temperature information array.

[0084] The inspection judgment module 15 is used to construct a temperature inspection twin model of the target power grid based on the digital twin and according to the reduced-dimensional temperature information array to perform inspection judgment.

[0085] Furthermore, the temperature information acquisition module 11 is further configured to perform the following steps:

[0086] Patrol sensors are arranged at multiple locations in the target power grid to obtain the patrol sensor array.

[0087] The temperature information of multiple normal positions detected and successfully transmitted by the inspection sensor array is obtained, and the missing temperature information array is constructed by combining the position information of the multiple normal positions.

[0088] A plurality of positions for which temperature information is not obtained are regarded as a plurality of missing positions.

[0089] Furthermore, the data accuracy analysis module 12 is further configured to perform the following steps:

[0090] According to the historical records of temperature inspection of the target power grid, a set of sample missing temperature information arrays is collected, and according to the matching degree of temperature information of multiple normal positions in each sample missing temperature information array, an overall sample accuracy set is obtained by marking.

[0091] The sample missing temperature information array set and the sample overall accuracy set are used to train an inspection accuracy analyzer, wherein the inspection accuracy analyzer includes multiple inspection accuracy analysis branches corresponding to multiple missing position combinations.

[0092] According to the missing position combination of the multiple missing positions, the missing temperature information array is input into the corresponding inspection accuracy analysis branch in the inspection accuracy analyzer, and the overall accuracy is obtained by analysis.

[0093] Furthermore, the data accuracy analysis module 12 is further configured to perform the following steps:

[0094] According to the historical inspection records of the target power grid, a plurality of missing position combinations are obtained.

[0095] According to the multiple missing position combinations, the sample missing temperature information array set and the sample overall accuracy set are divided, and multiple inspection accuracy analysis branches are trained respectively, wherein the multiple missing position combinations include different numbers of missing position combinations.

[0096] Based on the multiple inspection accurate analysis branches, the inspection accurate analyzer is obtained.

[0097] Furthermore, the supplementary optimization module 13 is further configured to perform the following steps:

[0098] A supplementary optimization space is constructed based on the historical inspection temperature intervals of the multiple missing positions.

[0099] A plurality of first supplementary temperature information of the plurality of missing positions are randomly generated in the supplementary optimization space.

[0100] The plurality of first supplementary temperature information are added to the missing temperature information array for supplementation to obtain a first supplementary temperature information array.

[0101] Performing an inspection data accuracy analysis on the first supplementary temperature information array to obtain a first overall accuracy, and calculating a deviation from the overall accuracy to obtain a first accuracy deviation.

[0102] Continue to iteratively supplement and optimize the temperature information of the multiple missing positions in the supplementary optimization space until convergence, output multiple supplementary temperature information with the smallest accuracy deviation as multiple optimal supplementary temperature information, supplement the missing temperature information array, and obtain a temperature information array.

[0103] Furthermore, the dimension reduction processing module 14 is further configured to perform the following steps:

[0104] Get the dimension reduction ratio space.

[0105] A first dimensionality reduction ratio is randomly selected in the dimensionality reduction ratio space, and an inspection discrimination accuracy analysis is performed on the temperature information array after dimensionality reduction in combination with the overall accuracy to obtain a first inspection accuracy.

[0106] Continue to perform dimensionality reduction optimization in the dimensionality reduction ratio space until convergence, and output the dimensionality reduction ratio with the maximum inspection accuracy as the optimal dimensionality reduction ratio.

[0107] According to the optimal dimensionality reduction ratio, the temperature information array is subjected to dimensionality reduction processing, the temperature information of the optimal dimensionality reduction ratio is retained, and a reduced-dimensional temperature information array is obtained.

[0108] Furthermore, the dimension reduction processing module 14 is further configured to perform the following steps:

[0109] A first dimensionality reduction ratio is obtained by randomly selecting in the dimensionality reduction ratio space.

[0110] According to the temperature inspection history of the target power grid, a sample overall accuracy set and a sample dimensionality reduction ratio set are collected, and the accuracy of inspection judgment is performed according to different sample overall accuracy temperature information arrays after different sample dimensionality reduction ratios, and the sample inspection accuracy set is marked.

[0111] The sample overall accuracy set and the sample dimension reduction ratio set are used as input data, and the sample inspection accuracy set is used as supervision data to train a dimension reduction inspection accuracy analyzer.

[0112] The dimension reduction inspection accuracy analyzer is used to perform inspection discrimination accuracy analysis on the first dimension reduction ratio and the overall accuracy after the temperature information array is dimensionally reduced to obtain a first inspection accuracy.

[0113] Through the above-mentioned detailed description of the power grid inspection optimization method based on digital twins in this specification, those skilled in the art can clearly understand the power grid inspection optimization system based on digital twins in this embodiment. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0114] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A power grid inspection optimization method based on digital twins, characterized in that: The method comprises: Based on the inspection sensor array arranged in the target power grid, temperature information of multiple normal positions detected in the target power grid is collected to obtain a missing temperature information array, and multiple missing positions where temperature information is not collected are determined; According to the missing temperature information array, performing inspection data accuracy analysis to obtain overall accuracy; According to the missing temperature information array and the overall accuracy, the temperature information of the multiple missing positions is supplemented and optimized to obtain multiple optimal supplementary temperature information, and the missing temperature information array is supplemented to obtain a temperature information array; According to the overall accuracy, perform dimensionality reduction optimization on the temperature information array to obtain an optimal dimensionality reduction ratio, perform dimensionality reduction processing on the temperature information array to obtain a reduced-dimensional temperature information array; Based on the digital twin, a temperature inspection twin model of the target power grid is constructed according to the reduced-dimensional temperature information array to perform inspection discrimination; According to the missing temperature information array, the inspection data accuracy analysis is performed to obtain the overall accuracy, including: According to the historical records of temperature inspection of the target power grid, a set of sample missing temperature information arrays is collected, and according to the matching degree of temperature information of multiple normal positions in each sample missing temperature information array, an overall accuracy set of samples is obtained by marking; The sample missing temperature information array set and the sample overall accuracy set are used to train an inspection accuracy analyzer, wherein the inspection accuracy analyzer includes a plurality of inspection accuracy analysis branches corresponding to a plurality of missing position combinations; According to the missing position combination of the multiple missing positions, the missing temperature information array is input into the corresponding inspection accuracy analysis branch in the inspection accuracy analyzer to analyze and obtain the overall accuracy; The sample missing temperature information array set and the sample overall accuracy set are used to train the inspection accuracy analyzer, including: Acquire multiple missing position combinations according to historical inspection records of the target power grid; According to the multiple missing position combinations, the sample missing temperature information array set and the sample overall accuracy set are divided, and multiple inspection accuracy analysis branches are trained respectively, wherein the multiple missing position combinations include different numbers of missing position combinations; Based on the multiple inspection accurate analysis branches, obtaining the inspection accurate analyzer; According to the missing temperature information array and the overall accuracy, the temperature information of the multiple missing positions is supplemented and optimized to obtain multiple optimal supplementary temperature information, and the missing temperature information array is supplemented, including: Constructing a supplementary optimization space according to the historical inspection temperature intervals of the plurality of missing positions; Randomly generating a plurality of first supplementary temperature information of the plurality of missing positions in the supplementary optimization space; Adding the plurality of first supplementary temperature information to the missing temperature information array for supplementation to obtain a first supplementary temperature information array; Performing an inspection data accuracy analysis on the first supplementary temperature information array to obtain a first overall accuracy, and calculating a deviation from the overall accuracy to obtain a first accuracy deviation; Continue to iteratively supplement and optimize the temperature information of the multiple missing positions in the supplement optimization space until convergence, output multiple supplementary temperature information with the smallest accuracy deviation as multiple optimal supplementary temperature information, supplement the missing temperature information array, and obtain a temperature information array; According to the overall accuracy, the temperature information array is optimized for dimensionality reduction to obtain an optimal dimensionality reduction ratio, and the temperature information array is processed for dimensionality reduction to obtain a dimensionality-reduced temperature information array, including: Get the dimension reduction ratio space; Randomly select a first dimensionality reduction ratio in the dimensionality reduction ratio space, and perform inspection discrimination accuracy analysis on the temperature information array after dimensionality reduction in combination with the overall accuracy to obtain a first inspection accuracy; Continue to perform dimensionality reduction optimization in the dimensionality reduction ratio space until convergence, and output the dimensionality reduction ratio with the maximum inspection accuracy as the optimal dimensionality reduction ratio; According to the optimal dimensionality reduction ratio, the temperature information array is subjected to dimensionality reduction processing, the temperature information of the optimal dimensionality reduction ratio is retained, and a dimensionality reduction temperature information array is obtained; Among them, randomly selecting a first dimension reduction ratio in the dimension reduction ratio space, and combining the overall accuracy to perform inspection discrimination accuracy analysis after the dimension reduction of the temperature information array, including: Randomly selecting a first dimensionality reduction ratio in the dimensionality reduction ratio space; According to the temperature inspection history of the target power grid, a sample overall accuracy set and a sample dimension reduction ratio set are collected, and the accuracy of inspection judgment is performed according to the temperature information arrays of different sample overall accuracy and different sample dimension reduction ratios, and the sample inspection accuracy set is marked; The sample overall accuracy set and the sample dimension reduction ratio set are used as input data, and the sample inspection accuracy set is used as supervision data to train a dimension reduction inspection accuracy analyzer; The dimension reduction inspection accuracy analyzer is used to perform inspection discrimination accuracy analysis on the first dimension reduction ratio and the overall accuracy after the temperature information array is dimensionally reduced to obtain a first inspection accuracy.

2. The power grid inspection optimization method based on digital twin according to claim 1 is characterized in that: Based on the inspection sensor array arranged in the target power grid, temperature information of multiple normal positions detected in the target power grid is collected to obtain a missing temperature information array, including: Deploying inspection sensors at multiple locations within the target power grid to obtain the inspection sensor array; Acquire temperature information of multiple normal positions detected and successfully transmitted by the inspection sensor array, and construct a missing temperature information array by combining the position information of the multiple normal positions; A plurality of positions for which temperature information is not obtained are regarded as a plurality of missing positions.

3. The power grid inspection and optimization system based on digital twin is characterized by: For implementing the digital twin-based power grid inspection optimization method according to any one of claims 1 to 2, the system comprises: A temperature information acquisition module is used to acquire temperature information of multiple normal positions detected in the target power grid based on a patrol sensor array arranged in the target power grid, obtain a missing temperature information array, and determine multiple missing positions where temperature information has not been acquired; A data accuracy analysis module, used to perform inspection data accuracy analysis based on the missing temperature information array to obtain overall accuracy; A supplementary optimization module, used for supplementing and optimizing the temperature information of the multiple missing positions according to the missing temperature information array and the overall accuracy, obtaining multiple optimal supplementary temperature information, supplementing the missing temperature information array, and obtaining a temperature information array; A dimensionality reduction processing module, used to perform dimensionality reduction optimization of the temperature information array according to the overall accuracy, obtain an optimal dimensionality reduction ratio, perform dimensionality reduction processing on the temperature information array, and obtain a reduced-dimensional temperature information array; The inspection judgment module is used to construct a temperature inspection twin model of the target power grid based on the digital twin and the reduced-dimensional temperature information array to perform inspection judgment.

Images