Power Fault Calculation and Prediction Method and System Based on Digital Twin

Abstract

The present invention provides a method and system for calculating and predicting power faults based on digital twins, which relates to the technical field of data processing. The method includes: obtaining real-time operation parameters of a power system through a data sensing cloud platform to establish a digital power twin model, simulating and emulating historical fault information through the digital power twin model, generating a power fault simulation data set to calculate and obtain multiple power fault prediction information, using the multiple power fault prediction information for estimating and analyzing operation faults of the power system, and formulating a power operation scheduling strategy to perform operation and maintenance on the power system. The present invention solves the technical problem in the prior art that the prediction accuracy of power system faults is low, resulting in difficulty in timely maintaining the power system when faults occur, and realizes reasonable and accurate prediction of faults through digital twin technology, and better performs operation and maintenance on the power system.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly relates to a method and system for calculating and predicting power failures based on digital twins. Background Art

[0002] With the rapid development of the power system, as well as the continuous expansion and update of the power grid, the scale of the power system has increased year by year, and the operation difficulty has also increased accordingly. The power system is an important support for the development and production of modern society. However, the stable operation of the power system faces many challenges. Some external factors such as natural disasters will directly cause power system failures, and there are also certain internal out-of-control factors. Moreover, in the prior art, the prediction accuracy of power system failures is low, resulting in the technical problem that it is difficult to maintain the power system in time when a failure occurs. Summary of the Invention

[0003] This application provides a method and system for calculating and predicting power failures based on digital twins, which are used to solve the technical problem in the prior art that the prediction accuracy of power system failures is low, resulting in the difficulty of maintaining the power system in time when a failure occurs.

[0004] In view of the above problems, this application provides a method and system for calculating and predicting power failures based on digital twins.

[0005] In a first aspect, this application provides a method for calculating and predicting power failures based on digital twins. The method includes: obtaining real-time operation parameters of the power system through the data sensing cloud platform; establishing a digital power twin model based on the real-time operation parameters; extracting historical fault information of the power system, and performing simulation on the historical fault information based on the digital power twin model to generate a power failure simulation data set; calculating and obtaining multiple power failure prediction information according to the power failure simulation data set; using the multiple power failure prediction information to perform an estimation and analysis of the operation failure of the power system, and formulating a power operation scheduling strategy; and performing operation and maintenance on the power system through the power operation scheduling strategy.

[0006] In a second aspect, the present application provides a power fault calculation and prediction system based on digital twin. The system includes: a first parameter acquisition module for acquiring real-time operation parameters of a power system through the data sensing cloud platform; a model establishment module for establishing a digital power twin model based on the real-time operation parameters; a first simulation module for extracting historical fault information of the power system and performing simulation on the historical fault information based on the digital power twin model to generate a power fault simulation data set; a first calculation module for calculating and obtaining multiple power fault prediction information according to the power fault simulation data set; a first analysis module for using the multiple power fault prediction information to perform estimation and analysis on operation faults of the power system and formulate a power operation scheduling strategy; and an operation and maintenance module for performing operation and maintenance on the power system through the power operation scheduling strategy.

[0007] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0008] The power fault calculation and prediction method and system based on digital twin provided in the present application relate to the technical field of data processing, solve the technical problem in the prior art that the prediction accuracy of power system faults is low, resulting in difficulties in timely maintenance of power system faults, and realize rational and accurate prediction of faults through digital twin technology, and better perform operation and maintenance on the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 It is a schematic flow chart of the power fault calculation and prediction method based on digital twin provided by the present application;

[0010] Figure 2 It is a schematic structural diagram of the power fault calculation and prediction system based on digital twin provided by the present application.

[0011] Description of the reference numerals: First parameter acquisition module 1, model establishment module 2, first simulation module 3, first calculation module 4, first analysis module 5, operation and maintenance module 6. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0012] The present application provides a power fault calculation and prediction method and system based on digital twin to solve the technical problem in the prior art that the prediction accuracy of power system faults is low, resulting in difficulties in timely maintenance of power system faults.

[0013] Embodiment 1

[0014] As Figure 1As shown in the figure, an embodiment of the present application provides a method for calculating and predicting power faults based on digital twins. The method is applied to a system for calculating and predicting power faults based on digital twins. The system for calculating and predicting power faults based on digital twins is communicatively connected to a data sensing cloud platform. The method includes:

[0015] Step A100: Obtain the real-time operation parameters of the power system through the data sensing cloud platform;

[0016] Furthermore, step A100 of the present application further includes:

[0017] Step A110: Select multiple sensors and monitoring devices based on the power system structure to construct the data sensing cloud platform;

[0018] Step A120: Collect data from the power system in real time through the data sensing cloud platform to obtain real-time power sensing monitoring data. Among them, the real-time power sensing monitoring data includes real-time current monitoring parameters, real-time voltage monitoring parameters, and real-time temperature monitoring parameters;

[0019] Step A130: Clean the data of the real-time current monitoring parameters, the real-time voltage monitoring parameters, and the real-time temperature monitoring parameters in sequence, and obtain the real-time operation parameters of the power system through aggregation.

[0020] In this application, the power fault calculation and prediction method based on digital twin provided by the embodiments of this application is applied to a power fault calculation and prediction system based on digital twin. The power fault calculation and prediction system based on digital twin is communicatively connected to a data sensing cloud platform. The data sensing cloud platform is used for collecting power parameters. The data sensing cloud platform selects appropriate sensors and monitoring devices on the basis of the power system structure, which may include current transformers, voltage transformers, temperature transformers, etc., to perform real-time sensing and monitoring of real-time current parameters, real-time voltage parameters, and real-time temperature parameters in the power system, so as to obtain the power sensing and monitoring data of the power system. And the power sensing and monitoring data includes real-time current monitoring parameters, real-time voltage monitoring parameters, and real-time temperature monitoring parameters. Further, data cleaning is performed on the real-time current monitoring parameters, real-time voltage monitoring parameters, and real-time temperature monitoring parameters in sequence. It means that the real-time current monitoring parameters, real-time voltage monitoring parameters, and real-time temperature monitoring parameters all contain errors, missing values, duplicates, and inconsistencies. The above problems are solved through data cleaning, and the real-time current monitoring parameters, real-time voltage monitoring parameters, and real-time temperature monitoring parameters are identified and repaired to remove useless data and error data to ensure the quality and consistency of the data. Finally, the multiple cleaned data are aggregated, which means integrating the data from multiple data sources into one data source, so as to facilitate later data analysis and decision-making for the power system, and at the same time obtain the real-time operating parameters of the power system for output, which serves as an important reference basis for later realizing the calculation and prediction of power faults based on digital twin technology.

[0021] Step A200: Establish a digital power twin model based on the real-time operating parameters;

[0022] Furthermore, step A200 of this application further includes:

[0023] Step A210: Extract features from the real-time current monitoring parameters, the real-time voltage monitoring parameters, and the real-time temperature monitoring parameters to generate a feature vector set, where the feature vector set includes a current feature vector, a voltage feature vector, and a temperature feature vector;

[0024] Step A220: Use the current feature vector, the voltage feature vector, and the temperature feature vector as data labels to perform corresponding identification on the real-time current monitoring parameters, the real-time voltage monitoring parameters, and the real-time temperature monitoring parameters, and obtain multiple identification results;

[0025] Step A230: Based on the multiple identification results, use deep learning for supervised training. When the training data converges, the construction of the digital power twin model is completed.

[0026] Furthermore, step A343 of this application includes:

[0027] Step A221: Update data according to the real-time state of the power system to obtain actual operation parameters;

[0028] Step A222: Obtain simulation operation parameters through the digital power twin model;

[0029] Step A223: Match the actual operation parameters and the simulation operation parameters respectively, perform fitness analysis according to the matching results, and obtain parameter fitness;

[0030] Step A224: Determine whether the parameter fitness is greater than or equal to a preset threshold. If it is greater than or equal to, consider the operation state of the power system to be normal and add a first data label, where the first data label is 0;

[0031] Step A225: If it is less than, consider the operation state of the power system to be abnormal and add a second data label, where the second data label is 1;

[0032] Step A226: Update the multiple identification results based on the first data label and the second data label.

[0033] In this application, in order to better calculate and predict power system faults, it is necessary to construct a digital power twin model through the real-time operation parameters obtained above. It means that first, the principal component analysis method is used to extract effective features from the real-time current monitoring parameters, real-time voltage monitoring parameters, and real-time temperature monitoring parameters. To reduce the later calculation complexity, it is necessary to perform data dimensionality reduction on the extracted features, convert the high-dimensional feature vectors into low-dimensional feature vectors, so as to obtain a feature vector set, and the feature vector set includes current feature vectors, voltage feature vectors, and temperature feature vectors. Further, the values of the current feature vectors, voltage feature vectors, and temperature feature vectors are respectively corresponding to a data label, and each data label contains the feature values of the corresponding feature vectors, so as to perform corresponding identification on the real-time current monitoring parameters, real-time voltage monitoring parameters, and real-time temperature monitoring parameters. On this basis, multiple identification results are obtained, and the multiple identification results include the correlation relationships between the real-time current monitoring parameters, real-time voltage monitoring parameters, real-time temperature monitoring parameters and the current feature vectors, voltage feature vectors, and temperature feature vectors. And according to the multiple identification results, supervised training is carried out using deep learning to construct a digital power twin model. And the digital power twin model includes an input layer, a hidden layer, and an output layer. The input layer is the layer for data input, the hidden layer is for better separating data features, and the output layer is for result output. The digital power twin model is obtained through training using a training data set and a supervision data set. Among them, each set of training data in the training data set includes multiple identification results, and the supervision data set is the supervision data corresponding one-to-one to the training data set.

[0034] Furthermore, the process of constructing the digital power twin model is as follows: Input each set of training data in the training dataset into the digital power twin model, and adjust the output of the digital power twin model through the corresponding supervision data of this set of training data. When the output result of the digital power twin model is consistent with the supervision data, the training of the current set is completed. When all the training data in the training dataset have been trained, the training of the digital power twin model is completed.

[0035] To ensure the convergence and accuracy of the digital power twin model, its convergence process can be that when the output data in the digital power twin model converges to a point and approaches a certain value, it is considered convergent. Its accuracy can be tested by the test dataset for the digital power twin model. For example, the test accuracy rate can be set to 80%. When the test accuracy rate of the test dataset meets 80%, the construction of the digital power twin model is completed.

[0036] In order to more accurately monitor faults in real-time current monitoring parameters, real-time voltage monitoring parameters, and real-time temperature monitoring parameters through multiple identification results, it is necessary to update the multiple identification results regularly. First, synchronously update the operation data in the power system based on the real-time operation state of the current power system to obtain the real-time operation parameters of the power system. Then, output the simulation operation parameters of the power system through the digital power twin model, and respectively match the degree of coincidence of the operation data between the actual operation parameters and the simulation operation parameters. Conduct fitness analysis based on the matching results. The higher the degree of data coincidence, the better the fitness, so as to determine the parameter fitness of the power system. Further, compare the parameter fitness with a preset threshold. The preset threshold is obtained by setting based on the average operation fitness of the power system in the historical period, so as to judge whether the parameter fitness is greater than or equal to the preset threshold. If the parameter fitness is greater than or equal to the preset threshold, it is considered that the operation state of the power system is normal, and add a data label to the operation state in the power system at this time, which is recorded as the first data label, and the first data label is 0. If the parameter fitness is less than the preset threshold, it is considered that the operation state of the power system is abnormal, and add a data label to the operation state in the power system at this time, which is recorded as the second data label, and the second data label is 1. Finally, update and replace the corresponding values of the data labels in the multiple identification results based on the first data label and the second data label to ensure its matching degree with the actual power system and improve the accuracy of the digital power twin model.

[0037] Step A300: Extract the historical fault information of the power system, and perform simulation on the historical fault information based on the digital power twin model to generate a power fault simulation dataset;

[0038] Furthermore, step A300 of the present application further includes:

[0039] Step A310: Extract historical fault information by traversing the historical operation logs of the power system;

[0040] Step A320: Set fault conditions according to the historical fault information;

[0041] Step A330: Introduce the fault conditions to perform simulation on the digital power twin model, generating multiple simulation results;

[0042] Step A340: Analyze the multiple simulation results to generate the power fault simulation dataset.

[0043] In this application, by extracting the historical operation logs within the power system, which are used to record various operation parameters of the power system during a historical time period, based on this, the fault information recorded in the historical operation logs is screened to obtain the historical fault information of the power system. At the same time, according to each power fault situation included in the historical fault information, the fault conditions of the power system are set. Further, performing simulation on the above - constructed digital power twin model based on the set fault conditions means introducing various fault situations, such as equipment failure, line fault, etc., into the digital power twin model to simulate the impact of faults under different situations on the power system, conduct fault simulation and prediction, generating multiple simulation results, where each simulation result corresponds to a fault situation. Finally, perform fault analysis on the multiple simulation results, predict the potential fault occurrence probability and fault influence range, and record the analysis results after integration as the power fault simulation dataset for output, thereby providing a guarantee for realizing the calculation and prediction of power faults based on digital twin technology.

[0044] Step A400: Calculate and obtain multiple power fault prediction information according to the power fault simulation dataset;

[0045] Furthermore, step A400 of this application further includes:

[0046] Step A410: Decompose the power fault simulation dataset according to time sequence to determine the operation time sequence of the power fault simulation dataset;

[0047] Step A420: Extract multiple fault simulation data from the power fault simulation dataset in sequence according to the operation time sequence, where each fault simulation data in the multiple fault simulation data includes fault node data and fault probability data;

[0048] Step A430: Traverse the correlation between each fault node and other nodes in the multiple fault simulation data to determine multiple fault - related nodes;

[0049] Step A440: Determine whether there is abnormal data in the multiple fault-related nodes. If so, locate the node with abnormal data and extract the power abnormal data;

[0050] Step A450: Perform fault prediction based on the power abnormal data and the fault probability data of the nodes with abnormal data, and generate the multiple power fault prediction information.

[0051] In this application, in order to more accurately predict the faults of the power system in the later stage, fault prediction calculations are performed through the generated power fault simulation data set. First, the power fault simulation data set is decomposed according to the operating time sequence of the power system to obtain the power fault simulation data corresponding to each operating moment of the power system. At the same time, the operating time sequence of the power fault simulation data set is determined. Further, according to the operating time sequence, a plurality of fault simulation data included in the power fault simulation data set are sequentially extracted. Each fault simulation data in the plurality of fault simulation data includes fault node data and fault probability data. The fault nodes of each fault simulation data in the plurality of fault simulation data are the node positions corresponding to the fault simulation data in the power system. The fault probability data has a one-to-one correspondence with the fault node data, and the fault probability data is the probability of a fault occurring in the corresponding fault node. Further, for

[0052] Step A430: Each fault node in the multiple fault simulation data is sequentially accessed and traversed with other nodes. According to the relevance between the nodes, multiple fault-related nodes are determined, where the relevance between the multiple fault-related nodes and the fault nodes is greater than 80%. Further, determining whether there is abnormal data in the multiple fault-related nodes means that the multiple fault-related nodes related to the fault nodes may or may not have occurred faults. According to the judgment result, it is determined whether there are faults in the multiple fault-related nodes. If there is abnormal data in the multiple fault-related nodes, node positioning is performed according to the abnormal data, and the power parameters included in the node are extracted and recorded as power abnormal data. Finally, based on the power abnormal data and the fault probability data of the nodes with abnormal data, a prediction and evaluation of the fault type and the probability of fault occurrence of the power system are performed, and the evaluation result is output as multiple power fault prediction information, laying a solid foundation for subsequent realization of power fault calculation and prediction based on digital twin technology.

[0053] Step A500: Use the multiple power fault prediction information to perform an estimation and analysis of the operating faults of the power system and formulate a power operation scheduling strategy;

[0054] Furthermore, step A500 of this application further includes:

[0055] Step A510: Match the multiple power failure prediction information with the power failure data in big data to generate a failure risk level;

[0056] Step A520: Perform weighted calculation on the multiple power failure prediction information based on the failure risk level to obtain a failure prediction calculation result;

[0057] Step A530: Perform risk prediction of operation failures on the power system based on the failure prediction calculation result to generate power prediction risk information;

[0058] Step A540: Formulate the power operation scheduling strategy of the power system according to the power prediction risk information.

[0059] In this application, the estimation and analysis of power system operation failures by using multiple power failure prediction information determined through the above calculations refers to matching the multiple power failure prediction information with power failure data in big data to generate a failure risk level. The failure risk level can be divided according to an index for evaluating and classifying the likelihood of potential failures occurring in the power system and the degree of impact on system operation, and can include high risk, medium risk, and low risk. High risk means that the likelihood of a potential failure occurring is very high, and once a failure occurs, it will have a serious impact on the operation of the power system, even leading to system collapse or power outage. Medium risk means that the likelihood of a potential failure occurring is relatively high, and it has a certain impact on the operation of the power system, but will not cause a complete system collapse or power outage. Low risk means that the likelihood of a potential failure occurring is relatively low, and it has a small impact on the operation of the power system and can be handled through routine maintenance and monitoring. Further, weighted calculations are performed on the multiple power failure prediction information based on the failure risk level. Performing weighted calculations on the multiple power failure prediction information based on the failure risk level can help determine the priorities and importance of different failures, thereby providing guidance for failure handling and maintenance plans. First, the weights can be set according to actual situations and requirements. For example, higher weights can be assigned to high risks, moderate weights to medium risks, and lower weights to low risks, so as to determine the weights for each failure risk level to reflect its importance in the overall failure prediction. Multiply each failure prediction information within the multiple power failure prediction information by its corresponding failure risk level and the corresponding weight to obtain a weighted value denoted as the failure prediction calculation result.Then add up all the weighted values to obtain the final weighted calculation result. Multiple power fault prediction information can include the health status monitoring of the power system, fault history records, environmental sensor data, etc. At the same time, based on the weights in the fault prediction calculation result, the risk prediction of operation faults in the power system is carried out, which means that based on the fault prediction calculation result, by combining the fault prediction calculation result with information such as historical data and equipment status monitoring, the occurrence probability of each fault type in the power system is evaluated. A higher fault probability indicates that the fault is more likely to occur in the system. For each fault type, evaluate the degree of its impact on the operation of the power system, and a higher fault impact indicates that the fault may have a significant impact on the power system. Based on statistical analysis, expert judgment or risk matrix, etc., the evaluation and classification of fault probability and fault impact are carried out, and the risks of different fault types in the power system are predicted to generate power prediction risk information. Further, according to the power prediction risk information, a power operation scheduling strategy for the power system is formulated, which means that by analyzing historical fault data, equipment status monitoring data, etc., the power prediction risk information is analyzed to determine the risks and probabilities of different power fault types, and then based on the power prediction risk information, a power operation scheduling strategy is formulated by adjusting the generator output, changing the operation mode of transmission lines, regulating the load, etc., so as to make the best use of existing resources and ensure the safe and stable operation of the power system.

[0060] Step A600: Carry out operation and maintenance on the power system through the power operation scheduling strategy.

[0061] In this application, according to the power operation scheduling strategy formulated above, the power system can be regularly inspected and maintained for key equipment such as generators, transformers, and switchgear, including cleaning, lubrication, fastening, etc., to complete the equipment inspection and maintenance within the power system. It is also possible to formulate a preventive maintenance plan for preventive maintenance by regularly overhauling key equipment and replacing aging components according to the service life of the equipment, historical fault data, etc. Use the monitoring technology deployed in the power system to monitor the operation status and health status of power monitoring equipment in real time, timely detect abnormal situations in the power system, and take corresponding maintenance measures. At the same time, monitor and record the operation parameters of the power system, such as voltage, frequency, power factor, etc., and compare them with the rated values of the equipment to timely detect and solve potential operation problems. Further, use historical data and data analysis technology to analyze and optimize the operation and maintenance of the power system to improve the accuracy of subsequent power fault calculation and prediction based on digital twin technology.

[0062] In summary, the power fault calculation and prediction method based on digital twin provided by the embodiment of this application has at least the following technical effects: realizing reasonable and accurate prediction of faults through digital twin technology, and better performing operation and maintenance on the power system.

[0063] Embodiment 2

[0064] Based on the same inventive concept as the digital twin-based power fault calculation and prediction method in the foregoing embodiment, as Figure 2 shown, this application provides a digital twin-based power fault calculation and prediction system, which includes:

[0065] The first parameter acquisition module 1, which is used to acquire the real-time operation parameters of the power system through the data sensing cloud platform;

[0066] The model establishment module 2, which is used to establish a digital power twin model based on the real-time operation parameters;

[0067] The first simulation module 3, which is used to extract the historical fault information of the power system, and perform simulation on the historical fault information based on the digital power twin model to generate a power fault simulation data set;

[0068] The first calculation module 4, which is used to calculate and obtain multiple power fault prediction information according to the power fault simulation data set;

[0069] The first analysis module 5, which is used to perform pre-estimation analysis of the operation faults of the power system by using the multiple power fault prediction information and formulate a power operation scheduling strategy;

[0070] The operation and maintenance module 6, which is used to perform operation and maintenance on the power system through the power operation scheduling strategy.

[0071] Furthermore, the system further includes:

[0072] The platform construction module, which is used to construct the data sensing cloud platform by selecting multiple sensors and monitoring devices based on the power system structure;

[0073] The data acquisition module, which is used to perform real-time data acquisition on the power system through the data sensing cloud platform to obtain real-time power sensing monitoring data, where the real-time power sensing monitoring data includes real-time current monitoring parameters, real-time voltage monitoring parameters, and real-time temperature monitoring parameters;

[0074] The second parameter acquisition module, which is used to perform data cleaning on the real-time current monitoring parameters, the real-time voltage monitoring parameters, and the real-time temperature monitoring parameters in sequence, and obtain the real-time operation parameters of the power system through aggregation.

[0075] Furthermore, the system further includes:

[0076] A vector module, which is used to extract features from the real-time current monitoring parameters, the real-time voltage monitoring parameters, and the real-time temperature monitoring parameters to generate a feature vector set. Among them, the feature vector set includes a current feature vector, a voltage feature vector, and a temperature feature vector;

[0077] An identification module, which is used to use the current feature vector, the voltage feature vector, and the temperature feature vector as data labels to correspondingly identify the real-time current monitoring parameters, the real-time voltage monitoring parameters, and the real-time temperature monitoring parameters to obtain multiple identification results;

[0078] A construction module, which is used to perform supervised training using deep learning based on the multiple identification results. When the training data converges, the construction of the digital power twin model is completed.

[0079] Furthermore, the system further includes:

[0080] A first update module, which is used to update data according to the real-time state of the power system to obtain actual operation parameters;

[0081] A third parameter acquisition module, which is used to obtain simulation operation parameters through the digital power twin model;

[0082] A first matching module, which is used to respectively match the actual operation parameters and the simulation operation parameters, perform fitness analysis according to the matching results, and obtain parameter fitness;

[0083] A first judgment module, which is used to judge whether the parameter fitness is greater than or equal to a preset threshold. If it is greater than or equal to, it is considered that the operation state of the power system is normal, and a first data label is added, where the first data label is 0;

[0084] A second judgment module, which is used to consider that the operation state of the power system is abnormal and add a second data label if it is less than, where the second data label is 1;

[0085] A second update module, which is used to update the multiple identification results based on the first data label and the second data label.

[0086] Furthermore, the system further includes:

[0087] A traversal module, which is used to extract historical fault information by traversing the historical operation logs of the power system;

[0088] A condition setting module, which is used to set fault conditions according to the historical fault information;

[0089] A second simulation module, which is used to introduce the fault conditions to simulate and imitate the digital power twin model, and generate a plurality of simulation results;

[0090] A second analysis module, which is used to analyze the plurality of simulation results to generate the power fault simulation data set.

[0091] Furthermore, the system further includes:

[0092] A decomposition module, which is used to decompose the power fault simulation data set according to time sequence to determine the running time sequence of the power fault simulation data set;

[0093] A data acquisition module, which is used to sequentially extract a plurality of fault simulation data from the power fault simulation data set according to the running time sequence, wherein each fault simulation data in the plurality of fault simulation data includes fault node data and fault probability data;

[0094] A node determination module, which is used to traverse the correlation between each fault node in the plurality of fault simulation data and other nodes to determine a plurality of fault-related nodes;

[0095] A third judgment module, which is used to judge whether there is abnormal data in the plurality of fault-related nodes. If so, locate the node of the abnormal data and extract the power abnormal data;

[0096] A fault prediction module, which is used to perform fault prediction based on the power abnormal data and the fault probability data of the node with abnormal data to generate a plurality of power fault prediction information.

[0097] Furthermore, the system further includes:

[0098] A second matching module, which is used to match the plurality of power fault prediction information according to the power fault data in big data to generate a fault risk level;

[0099] A second calculation module, which is used to perform weighted calculation on the plurality of power fault prediction information based on the fault risk level to obtain a fault prediction calculation result;

[0100] A risk prediction module, which is used to perform risk prediction of running faults on the power system based on the fault prediction calculation result to generate power prediction risk information;

[0101] A strategy formulation module, which is used to formulate the power operation scheduling strategy of the power system according to the power prediction risk information.

[0102] Through the foregoing detailed description of the power fault calculation prediction method based on digital twin in this specification, those skilled in the art can clearly know the power fault calculation prediction system based on digital twin in this embodiment. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple. For the relevant parts, refer to the description in the method part.

[0103] 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 obvious to those skilled in the art, and the general principles defined herein can 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 these embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A power fault calculation and prediction method based on digital twins, characterized in that: The method is applied to a power fault calculation prediction system based on digital twins, wherein the power fault calculation prediction system based on digital twins is communicatively connected to a data sensing cloud platform, and the method comprises: Acquire real-time operating parameters of the power system through the data sensing cloud platform; Establishing a digital power twin model based on the real-time operating parameters; Extracting historical fault information of the power system, simulating the historical fault information based on the digital power twin model, and generating a power fault simulation data set; Calculate and obtain multiple power failure prediction information according to the power failure simulation data set; Using the plurality of power fault prediction information to perform prediction analysis of power system operation faults and formulate power operation dispatching strategies; Performing operation and maintenance on the power system through the power operation and dispatching strategy; A digital power twin model is established based on the real-time operating parameters, and the method includes: Performing feature extraction on real-time current monitoring parameters, real-time voltage monitoring parameters, and real-time temperature monitoring parameters to generate a feature vector set, wherein the feature vector set includes a current feature vector, a voltage feature vector, and a temperature feature vector; The current characteristic vector, the voltage characteristic vector, and the temperature characteristic vector are used as data labels to correspondingly identify the real-time current monitoring parameters, the real-time voltage monitoring parameters, and the real-time temperature monitoring parameters to obtain multiple identification results; Based on the multiple identification results, supervised training is performed using deep learning, and when the training data converges, the construction of the digital power twin model is completed; The method for calculating and obtaining a plurality of power failure prediction information according to the power failure simulation data set includes: Decomposing the power failure simulation data set according to time sequence to determine the running time sequence of the power failure simulation data set; According to the running time sequence, sequentially extract multiple fault simulation data in the power fault simulation data set, wherein each fault simulation data in the multiple fault simulation data includes fault node data and fault probability data; Traversing the correlation between each fault node and other nodes in the plurality of fault simulation data to determine a plurality of fault-related nodes; Determine whether there is abnormal data in the multiple fault-related nodes, and if so, locate the node with abnormal data and extract the abnormal power data; Fault prediction is performed based on the power abnormality data and the fault probability data of the node with abnormal data to generate the plurality of power fault prediction information.

2. The method according to claim 1, characterized in that The real-time operating parameters of the power system are obtained through the data sensing cloud platform, and the method includes: Select multiple sensors and monitoring equipment based on the power system structure to build the data sensing cloud platform; The data sensing cloud platform is used to collect data of the power system in real time to obtain real-time power sensing monitoring data, wherein the real-time power sensing monitoring data includes real-time current monitoring parameters, real-time voltage monitoring parameters, and real-time temperature monitoring parameters; The real-time current monitoring parameters, the real-time voltage monitoring parameters, and the real-time temperature monitoring parameters are cleaned in turn, and the real-time operation parameters of the power system are obtained through aggregation.

3. The method according to claim 2, characterized in that A digital power twin model is established based on the real-time operating parameters, and the method includes: Performing feature extraction on the real-time current monitoring parameter, the real-time voltage monitoring parameter, and the real-time temperature monitoring parameter to generate a feature vector set, wherein the feature vector set includes a current feature vector, a voltage feature vector, and a temperature feature vector; The current characteristic vector, the voltage characteristic vector, and the temperature characteristic vector are used as data labels to correspondingly identify the real-time current monitoring parameters, the real-time voltage monitoring parameters, and the real-time temperature monitoring parameters to obtain multiple identification results; Based on the multiple identification results, deep learning is used for supervised training, and when the training data converges, the construction of the digital power twin model is completed.

4. The method according to claim 3, characterized in that Methods include: Update data based on the real-time status of the power system to obtain actual operating parameters; Obtaining simulation operation parameters through the digital power twin model; Matching the actual operating parameters and the simulation operating parameters respectively, performing fitness analysis according to the matching results, and obtaining parameter fitness; Determine whether the parameter fitness is greater than or equal to a preset threshold value. If so, the operation state of the power system is considered to be normal, and a first data label is added, wherein the first data label is 0; If it is less than, the operation state of the power system is considered to be abnormal, and a second data tag is added, wherein the second data tag is 1; The plurality of identification results are updated based on the first data tag and the second data tag.

5. The method according to claim 1, characterized in that Extracting historical fault information of the power system, simulating the historical fault information based on the digital power twin model, and generating a power fault simulation data set, the method comprising: By traversing the historical operation logs of the power system, historical fault information can be extracted; Setting a fault condition according to the historical fault information; Introducing the fault condition to simulate the digital power twin model and generate multiple simulation results; The multiple simulation results are analyzed to generate the power fault simulation data set.

6. The method according to claim 1, characterized in that The plurality of power fault prediction information are used to perform an estimation analysis of power system operation faults and formulate a power operation dispatching strategy, the method comprising: Perform fault matching on the plurality of power fault prediction information according to the power fault data in the big data to generate a fault risk level; Performing weighted calculation on the plurality of power fault prediction information based on the fault risk level to obtain a fault prediction calculation result; Based on the fault prediction calculation results, the risk of operating faults in the power system is predicted to generate power prediction risk information; The power operation dispatching strategy of the power system is formulated according to the power forecast risk information.

7. The power fault calculation and prediction system based on digital twin is characterized by: Used to implement the power fault calculation prediction method based on digital twins according to any one of claims 1 to 6, the power fault calculation prediction system based on digital twins is communicatively connected with a data sensing cloud platform, and the system comprises: A first parameter acquisition module, the first parameter acquisition module is used to obtain real-time operating parameters of the power system through the data sensing cloud platform; A model building module, wherein the model building module is used to build a digital power twin model based on the real-time operating parameters; A first simulation module, wherein the first simulation module is used to extract historical fault information of the power system, simulate the historical fault information based on the digital power twin model, and generate a power fault simulation data set; A first calculation module, the first calculation module is used to calculate and obtain multiple power fault prediction information according to the power fault simulation data set; A first analysis module, the first analysis module is used to use the plurality of power fault prediction information to perform an estimation analysis of power system operation faults and formulate a power operation dispatching strategy; An operation and maintenance module, wherein the operation and maintenance module is used to perform operation and maintenance on the power system through the power operation and scheduling strategy.

Images