A Quasi-Real-Time Fault Identification Method and System for Distribution Networks Based on Line Topology Analysis

Abstract

This invention discloses a quasi-real-time fault identification method for distribution networks based on line topology analysis, comprising: acquiring basic line information, line topology map information, sign information, switch action information, power outage record information from the data acquisition system, power outage record information from the fusion terminal, line fault reclosing success record information, and line maintenance plan information; generating a line action set based on the acquired switch action information, power outage record information from the data acquisition system, and power outage record information from the fusion terminal; filtering the line action set based on the sign information, line fault reclosing success record information, and line maintenance plan information; generating a feature sample set based on the line topology map information and the filtered line action set; training a fault identification model using the feature sample set; and identifying distribution network faults using the trained fault identification model. This invention can accurately identify distribution network faults.

Description

Technical Field

[0001] This invention belongs to the field of power grid fault identification technology, and in particular relates to a quasi-real-time fault identification method and system for distribution networks based on line topology analysis. Background Technology

[0002] The current distribution network fault statistics and reporting system has basically achieved fault assessment and statistics for Jiangsu distribution networks at a time of T-1. In order to further improve the distribution network fault management capability and guide rapid emergency repairs, it is necessary to strengthen the ability to assess faults in real time. However, in the context of real-time faults, the problem of data incompleteness becomes prominent. How to accurately diagnose real-time faults using limited data has become an urgent problem to be solved. Summary of the Invention

[0003] To address the problems existing in the prior art, this invention provides a quasi-real-time fault identification method and system for distribution networks based on line topology analysis, which can accurately identify faults in distribution networks.

[0004] The technical problem to be solved by the present invention is achieved through the following technical solution:

[0005] Firstly, a quasi-real-time fault identification method for distribution networks based on line topology analysis is provided, including:

[0006] Acquire basic line information, line topology map information, sign information, switch action information, power outage record information of the data acquisition system, power outage record information of the converged terminal, line fault reclosing success record information, and line maintenance plan information;

[0007] A set of line actions is generated based on the obtained switch action information, power outage record information from the data acquisition system, and power outage record information from the fusion terminal.

[0008] The line action set is filtered based on the sign information, the line fault overlap success record information, and the line maintenance plan information.

[0009] A feature sample set is generated based on the route topology map information and the filtered route action set;

[0010] The fault identification model is trained using a feature sample set;

[0011] The fault identification model is used to identify faults in the distribution network.

[0012] In conjunction with the first aspect, the switch action information further includes: fault tripping information of quasi-real-time substations, remote signaling change information of substation and distribution network switches, protection SOE information, and FA action information.

[0013] In conjunction with the first aspect, further, the filtering of the line action set includes: determining whether each line action in the line action set matches the substation and distribution network switch placement information, the line fault overlap success record information, and the line maintenance plan information. If they match, the corresponding line action time period is considered not to be within the fault identification range, and it is filtered out from the line action set.

[0014] In conjunction with the first aspect, further, the step of generating a feature sample set based on the filtered set of line actions includes:

[0015] Based on the information from the line action set and line topology map, examine the matching results of each line within the filtered line action set with the fault tripping information of the near real-time substation, the remote signaling change information of the substation and distribution network switches, the protection SOE information, the FA action information, the power outage record information of the data acquisition system, and the power outage record information of the fusion terminal to generate a feature sample set.

[0016] In conjunction with the first aspect, further, the feature sample set includes the following feature values:

[0017] Matching status of substation fault tripping information, substation remote signaling change information, substation protection SOE information, line maintenance plan, distribution network switch remote signaling change information, distribution network switch protection SOE information, FA action information, line-substation transformer and integrated terminal power outage information, line-substation transformer and user power outage information, distribution network switch remote signaling change information and its downstream transformer integrated terminal power outage information, distribution network switch remote signaling change information and its downstream transformer user power outage information, distribution network switch protection SOE information and its downstream transformer integrated terminal power outage information, distribution network switch protection SOE information and its downstream transformer user power outage information, FA action information and its downstream transformer integrated terminal power outage information, FA action information and its downstream transformer user power outage information.

[0018] In conjunction with the first aspect, furthermore, 70% of the feature sample set is used as training samples and 30% as test samples.

[0019] In conjunction with the first aspect, the fault identification model is further described as a BP neural network model, and the F1 score is used to evaluate the model.

[0020] In conjunction with the first aspect, further, the identification of distribution network faults includes:

[0021] Feature extraction is performed on the near real-time actions of lines in the actual power distribution network, and the extracted features are input into the trained fault identification model for fault identification.

[0022] Secondly, a near real-time fault identification system for distribution networks based on line topology analysis is provided, including:

[0023] The information acquisition module is used to acquire basic line information, line topology map information, sign information, switch action information, power outage record information of the data acquisition system, power outage record information of the converged terminal, line fault reclosing success record information, and line maintenance plan information.

[0024] The set generation module is used to generate a set of line actions based on the acquired switch action information, power outage record information from the data acquisition system, and power outage record information from the fusion terminal.

[0025] The filtering module is used to filter the set of line actions based on the tag information, the line fault re-closing success record information, and the line maintenance plan information;

[0026] The sample generation module is used to generate a feature sample set based on the line topology map information and the filtered line action set.

[0027] The model training module is used to train the fault identification model using a feature sample set;

[0028] The fault identification module is used to identify faults in the distribution network using a trained fault identification model.

[0029] The beneficial effects of this invention include the following:

[0030] First, this invention learns the changing trends of massive historical fault data in a data-driven manner, collects 15 feature values ​​related to line switch action information and topology information, and uses the F1 coefficient to accurately determine whether a line fault event has occurred, thus providing timely and accurate decision support.

[0031] Secondly, this invention comprehensively considers multiple information combinations such as line markings, maintenance plans, successful reclosing, topology diagrams, and switch actions to generate relevant feature values ​​for the matching status of each switch information, gradually eliminating and accurately identifying fault events, resulting in a high accuracy rate in judgment.

[0032] Furthermore, this invention is simple to calculate and has a clear principle, which can help distribution network operators detect fault events in a timely manner and has good application prospects. Attached Figure Description

[0033] Figure 1 This is a flowchart of the quasi-real-time fault identification method for distribution networks based on line topology analysis according to the present invention. Detailed Implementation

[0034] To further describe the technical features and effects of the present invention, the present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0035] Example 1

[0036] like Figure 1 As shown, a quasi-real-time fault identification method for distribution networks based on line topology analysis is provided, including the following steps:

[0037] Step 1: Obtain Information

[0038] To subsequently diagnose the fault, it is first necessary to collect some information, including basic line information, line topology diagram information, signage information, switch action information, power outage records from the data acquisition system, power outage records from the converged terminal, successful reclosing records of line faults, and line maintenance plan information. The specific implementation process is as follows:

[0039] In the energy management system (which is a system that exists in the prior art), select the medium-voltage distribution network to be processed, read the basic data of the lines and corresponding distribution transformers and the line topology information stored in the existing system, extract the planned power outage data in the production management system, the power consumption and collection power outage record data in the power consumption information collection management system, the power outage data of the fusion terminal of the distribution cloud master station system, the accident tripping record data, FA (feeder automation) action record data, remote signaling change action record data and tag data in the dispatch management system, the distribution automation master station system's distribution network switch action record data and protection SOE (event sequence record) signal data, and the fault record data of the fault management system.

[0040] Step 2: Generate a set of line actions

[0041] The line action set C is generated based on the obtained switch action information (fault tripping information of near real-time substations, remote signaling change information of substation and distribution network switches, protection SOE information and FA action information), power outage record information of the data acquisition system and power outage record information of the fusion terminal.

[0042] Suppose that during a certain period of time, a connection to line f... i The switch on T k If there is one action record within the above information, then the generated set C = [(f1,T1),...,(f i ,T k ...,(f m ,T n )], where m is the total number of lines with action records in the medium-voltage distribution network within a certain time period, and n is the total number of all action records within a certain time period.

[0043] From the obtained data of the lines and corresponding action switches, generate a set C of lines and corresponding time points in units of line and action time.

[0044] Step 3: Set Filtering

[0045] In set C, some information represents normal switching actions, so it needs to be filtered out. Specifically, this includes:

[0046] View the lines f within set C i In T k Matching information with real-time and near real-time substation and distribution network switchboard information, line maintenance plan information, and line reclosing success fault records. The matching method is based on the action record information within set C, specifically the T value. k If a match is found within a time period offset of ±30 minutes, then the line f is considered to be valid. i In T k Actions occurring during periods not within the scope of fault identification are excluded, and related records in set C are removed.

[0047] Step 4: Generate a feature sample set

[0048] Based on the generated set C of all routes and their corresponding time points, and combined with the route topology information, we can examine the route f within set C. i In T k The matching status of fault tripping information from substations and near real-time substations, remote signaling change information from substations and distribution network switches, protection SOE information, FA action information, power outage record information from the data acquisition system, and power outage record information from the fusion terminal is determined by matching the action record information within the filtered set C'. i In T k Within the time period after a time offset of plus or minus 30 minutes.

[0049] Finally, the feature values ​​are generated:

[0050] One is the line f within the filtered set C'. i In T k Matching status of time with substation fault tripping information;

[0051] Second, the lines f within the filtered set C' i In T k Matching status of time with substation remote signaling change information;

[0052] The third is the line f within the filtered set C'. i In T k Matching status of time with substation protection SOE information;

[0053] Fourthly, the lines f within the filtered set C' i In T k Matching of time with line maintenance plan;

[0054] The fifth is the line f within the filtered set C'. i In T kMatching status of time with remote signaling change information of distribution network switches;

[0055] Six are the lines f within the filtered set C'. i In T k Matching status of time with SOE information of distribution network switch protection;

[0056] Seven is the line f within the filtered set C'. i In T k Matching status between time and FA action information;

[0057] Eight is the line f within the filtered set C'. i In T k Matching status of time and line-based power outage information between distribution transformers and integrated terminals;

[0058] Nine represents the lines f within the filtered set C'. i In T k Matching of time and line-related power outage information between distribution transformers and users;

[0059] Ten are the lines f within the filtered set C' i In T k Matching of time and distribution network switch remote signaling change information with the integrated terminal power outage information of the switch and its downstream distribution transformer;

[0060] Eleven refers to the lines f within the filtered set C'. i In T k Matching of power outage information between the switch and its downstream distribution transformer in time and remote signaling change information of the distribution network switch;

[0061] Twelve are the lines f within the filtered set C'. i In T k Matching status of power outage information between the switch and its downstream distribution transformer in the SOE information of the time and distribution network switch protection;

[0062] Thirteen are the lines f within the filtered set C'. i In T k Matching of time with the outage information of the switch and its downstream distribution transformer in the SOE information of the distribution network switch protection;

[0063] Fourteen is the line f within the filtered set C'. i In T k Matching status of power outage information between the switch and its downstream distribution transformer in time and FA action information;

[0064] Fifteen represents the line f within the filtered set C'. i In T k Matching of power outage information between the switch and its downstream distribution transformer in time and FA action information;

[0065] After generating a total of 15 feature values, the feature data is finally normalized to obtain the feature sample set.

[0066] Step 5: Model Training

[0067] The fault identification model is trained using a feature sample set.

[0068] Before training, the training samples need to be divided, that is, the feature sample set is divided into a 7:3 ratio, with 70% used to train the fault identification model and 30% used as the validation set to test the model training effect and model tuning; the fault identification model adopts the BP neural network model, and the F1 coefficient is used to evaluate the trained model.

[0069] The specific formula for calculating the F1 coefficient is as follows:

[0070]

[0071] In the formula, P is precision, which refers to the model's ability not to label negative samples as positive samples, and R is recall, which refers to the model's ability to find all positive samples.

[0072] The final evaluation coefficient F1 of the model is 0.96, which proves that the model has a high accuracy in prediction and good training effect.

[0073] Step Six: Fault Identification

[0074] The trained model is used to perform batch real-time fault identification on the action records of line-related power distribution switches and substations.

[0075] After the model predicts the evaluation coefficients and meets the requirements, the trained model is used to perform batch fault identification and judgment on the near real-time action record data of the lines in the actual medium-voltage distribution network. That is, the required feature values ​​are generated and extracted for each line in batches, and then fed into the model for calculation. Finally, the fault diagnosis results for each line are output.

[0076] Example 2

[0077] This invention also provides a quasi-real-time fault identification system for distribution networks based on line topology analysis, comprising:

[0078] The information acquisition module is used to acquire basic line information, line topology map information, sign information, switch action information, power outage record information of the data acquisition system, power outage record information of the converged terminal, line fault reclosing success record information, and line maintenance plan information.

[0079] The set generation module is used to generate a set of line actions based on the acquired switch action information, power outage record information from the data acquisition system, and power outage record information from the fusion terminal.

[0080] The filtering module is used to filter the set of line actions based on the tag information, the line fault re-closing success record information, and the line maintenance plan information;

[0081] The sample generation module is used to generate a feature sample set based on the line topology map information and the filtered line action set.

[0082] The model training module is used to train the fault identification model using a feature sample set;

[0083] The fault identification module is used to identify faults in the distribution network using a trained fault identification model.

[0084] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0085] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0086] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0087] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0088] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A quasi-real-time fault identification method for distribution networks based on line topology analysis, characterized in that, include: Acquire basic line information, line topology map information, sign information, switch action information, power outage record information of the data acquisition system, power outage record information of the converged terminal, line fault reclosing success record information, and line maintenance plan information; A set of line actions is generated based on the obtained switch action information, power outage record information from the data acquisition system, and power outage record information from the fusion terminal. The line action set is filtered based on the sign information, the line fault overlap success record information, and the line maintenance plan information. A feature sample set is generated based on the route topology map information and the filtered route action set; The fault identification model is trained using a feature sample set; The fault identification model is used to identify faults in the distribution network. The step of generating a feature sample set based on the filtered set of line actions includes: Based on the line action set and line topology information, examine the matching of each line with the near real-time substation's fault tripping information, the substation and distribution network switch's remote signaling change information, protection SOE information, FA action information, power outage record information of the user acquisition system, and power outage record information of the fusion terminal within the filtered line action set to generate a feature sample set. The feature sample set contains the following feature values: Matching status of substation fault tripping information, substation remote signaling change information, substation protection SOE information, line maintenance plan, distribution network switch remote signaling change information, distribution network switch protection SOE information, FA action information, line-substation transformer and integrated terminal power outage information, line-substation transformer and user power outage information, distribution network switch remote signaling change information and its downstream transformer integrated terminal power outage information, distribution network switch remote signaling change information and its downstream transformer user power outage information, distribution network switch protection SOE information and its downstream transformer integrated terminal power outage information, distribution network switch protection SOE information and its downstream transformer user power outage information, FA action information and its downstream transformer integrated terminal power outage information, FA action information and its downstream transformer user power outage information.

2. The method for quasi-real-time fault identification in distribution networks based on line topology analysis according to claim 1, characterized in that, The switch action information includes: fault tripping information of quasi-real-time substations, remote signaling change information of substation and distribution network switches, protection SOE information, and FA action information.

3. The method for quasi-real-time fault identification in distribution networks based on line topology analysis according to claim 1, characterized in that, The filtering of the line action set includes: determining whether each line action in the line action set matches the substation and distribution network switch labeling information, the line fault overlap success record information, and the line maintenance plan information. If they match, the corresponding line action time period is considered not to be within the fault identification range, and it is filtered out from the line action set.

4. The method for quasi-real-time fault identification in a distribution network based on line topology analysis according to claim 1, characterized in that, 70% of the feature sample set is used as training samples and 30% is used as test samples.

5. The method for quasi-real-time fault identification in distribution networks based on line topology analysis according to claim 1, characterized in that, The fault identification model is a BP neural network model, and the F1 score is used to evaluate the model.

6. The method for quasi-real-time fault identification in distribution networks based on line topology analysis according to claim 1, characterized in that, The identification of distribution network faults includes: Feature extraction is performed on the near real-time actions of lines in the actual power distribution network, and the extracted features are input into the trained fault identification model for fault identification.

7. A quasi-real-time fault identification method for distribution networks based on line topology analysis, characterized in that, include: The information acquisition module is used to acquire basic line information, line topology map information, sign information, switch action information, power outage record information of the data acquisition system, power outage record information of the converged terminal, line fault reclosing success record information, and line maintenance plan information. The set generation module is used to generate a set of line actions based on the acquired switch action information, power outage record information from the data acquisition system, and power outage record information from the fusion terminal. The filtering module is used to filter the set of line actions based on the tag information, the line fault re-closing success record information, and the line maintenance plan information; A sample generation module is used to generate a feature sample set based on the line topology map information and the filtered line action set; wherein, generating the feature sample set based on the filtered line action set includes: Based on the line action set and line topology information, examine the matching of each line with the near real-time substation's fault tripping information, the substation and distribution network switch's remote signaling change information, protection SOE information, FA action information, power outage record information of the user acquisition system, and power outage record information of the fusion terminal within the filtered line action set to generate a feature sample set. The feature sample set contains the following feature values: Matching status of substation fault tripping information, substation remote signaling change information, substation protection SOE information, line maintenance plan, distribution network switch remote signaling change information, distribution network switch protection SOE information, FA action information, line-substation transformer and integrated terminal power outage information, line-substation transformer and user power outage information, distribution network switch remote signaling change information and its downstream transformer integrated terminal power outage information, distribution network switch remote signaling change information and its downstream transformer user power outage information, distribution network switch protection SOE information and its downstream transformer integrated terminal power outage information, distribution network switch protection SOE information and its downstream transformer user power outage information, distribution network switch protection SOE information and its downstream transformer user power outage information, FA action information and its downstream transformer integrated terminal power outage information, FA action information and its downstream transformer user power outage information. The model training module is used to train the fault identification model using a feature sample set; The fault identification module is used to identify faults in the distribution network using a trained fault identification model.

Images