Low-voltage user power failure event identification and analysis method, system and device based on deep neural network, and storage medium

Abstract

The invention discloses a low-voltage user power failure event identification and analysis method, system and device based on a deep neural network, and a storage medium. The method comprises the following steps: obtaining measurement, ledger, topology, power supply and collection power failure events, customer service fault records and switch action signal data of a low-voltage user and a distribution transformer to which the low-voltage user belongs; generating an initial power outage event according to the user voltage, the power measurement and the uploaded power utilization and collection power outage event; generating a characteristic value according to the acquired data and the initial power failure event; forming a sample data set in combination with each user characteristic value and historical power failure data, dividing the sample data set, training and optimizing a deep neural network model, and evaluating the deep neural network model; and performing batch identification on the initial power failure events of the low-voltage users by using the deep neural network model, and finally outputting true and false power failure results corresponding to the low-voltage power failure events. According to the invention, low-voltage user operation data change characteristics are learned through the data model, whether the low-voltage user has real power failure is judged, and accurate collection of power failure events of the low-voltage user is realized.

Description

Technical Field

[0001] The present invention belongs to the field of low-voltage distribution network power outage event recognition and analysis, and specifically relates to a low-voltage user power outage event recognition and analysis method, system, device and storage medium based on a deep neural network. Background Art

[0002] As a key energy source in modern society, electricity is widely used in various fields such as industry, commerce, and residential life. The low-voltage distribution system is the last link in power transmission and is directly facing end users. Its reliability is crucial. At present, the low-voltage power outage detection method is still in a relatively traditional stage and has great limitations. For example, it mainly relies on manual inspections, which is inefficient and cannot detect power outages in time. Inspection personnel need to check each household or determine the scope and cause of the power outage based on user reports, which consumes a lot of time and manpower; another example is a simple alarm device, which can only provide basic power outage signals and cannot accurately determine the specific location of the power outage. It is also likely to have false alarms or missed alarms.

[0003] With the continuous development of smart grid technology, higher requirements are placed on the rapid and accurate diagnosis and identification of power outages for low-voltage users. The diagnosis method of power outages in low-voltage distribution networks is an important research topic. The research results can help operators detect power outages in a timely manner and identify weak risk points in the grid. Summary of the invention

[0004] Purpose of the invention: The purpose of the present invention is to provide a method, system, device and storage medium for identifying and analyzing power outage events for low-voltage users based on deep neural networks, which can be used to identify power outage events for low-voltage users of any scale. The calculation is simple and the principle is clear, which can help distribution network operators to detect real power outage events in a timely manner.

[0005] Technical solution: A method for identifying and analyzing low-voltage user power outage events based on a deep neural network of the present invention comprises:

[0006] Obtain the measurement, ledger, topology, power outage events, customer service fault records, and switch action signal data of low-voltage users and their distribution transformers; among them, the measurements include voltage, power, and current;

[0007] Generate an initial power outage event based on the voltage and power measurement of low-voltage users and the power outage events sent up;

[0008] Generate characteristic values ​​based on the acquired data and the initial power outage event, including the distribution transformer current drop ratio, the user power consumption drop ratio, the matching power outage event, the user power outage ratio under the distribution transformer, the meter box power outage ratio under the distribution transformer, the topology segment power outage ratio, the power outage duration of the user's initial power outage event, the matching low-voltage fault record, whether it is a photovoltaic user, the existence of an upstream switch action signal, and the existence of a contact switch downstream;

[0009] Combine the characteristic values ​​of each user and the historical power outage situation to form a sample data set and divide it, train and optimize the deep neural network model, and finally use the F1 coefficient to evaluate the deep neural network model;

[0010] The trained deep neural network model is used to batch identify the initial power outage events of each low-voltage user, and finally outputs the true and false power outage results corresponding to the power outage events of each low-voltage user.

[0011] Furthermore, the measurement, ledger, topology, power outage events, customer service fault records, and switch action signal data of low-voltage users and their distribution transformers are obtained, among which,

[0012] The measurement, ledger, topology, and power outage event data of the distribution transformer include the basic information maintained by the distribution transformer in the PMS system, i.e., the power information collection system, the drawing topology path data, and the power outage event data sent by the equipment;

[0013] The switch action signal data includes the protection signals, opening and closing signals, FA test and switch listing information data sent by all automatic switches in the distribution network of the distribution automation system;

[0014] Customer service fault records include fault records of distribution transformers related to the release of customer service power outage information.

[0015] Furthermore, an initial power outage event is generated based on the acquired user voltage, power measurement, and the sent power outage event, including:

[0016] Assume low voltage user C k Connected to distribution transformer T, low voltage user C k The active power measurement data is P k =[p k,1 ,p k,2 ,…,p k,i ,…,p k,n ], low voltage user C k The voltage measurement data is V k =[v k,1 ,v k,2 ,…,v k,i ,…,v k,n ], n represents the number of sampling points of active power and voltage measurement data of low-voltage users;

[0017] If low voltage user C k Active power measurement data P k In the continuous data segment [p k,i ,…,p k,j ] is 0 or empty, and the voltage measurement data V k In the continuous data segment [v k,i ,…,vk,j ] is 0 or empty, a low voltage user C is generated k The data segment [i,…,j] generates a power outage event set for the data;

[0018] Low voltage user C k If a power outage event is collected, and the power outage time sent is earlier than the power restoration time, and the power outage duration during the power outage period is within two natural days, a power outage event is generated as a low-voltage user C k The user adopts the power outage event set sent up;

[0019] From the obtained low-voltage user measurement data and the collected power outage events, a power outage event set and a power outage event set for uploading are generated for each user on a user-by-user basis to form the initial power outage event for each user.

[0020] Furthermore, characteristic values ​​are generated based on the acquired data and the initial power outage event. The characteristic values ​​include the distribution transformer current drop ratio, the user power consumption drop ratio, the matching power outage event, the user power outage ratio under the distribution transformer, the meter box power outage ratio under the distribution transformer, the topology segment power outage ratio, the power outage duration of the user's initial power outage event, the matching low-voltage fault record, whether it is a photovoltaic user, the existence of an upstream switch action signal, and the existence of a tie switch downstream, including:

[0021] Based on the ledger and topology information, obtain the low-voltage user C k The number of low-voltage users in operation under the corresponding distribution transformer D is used as the denominator, and the number of users with initial power outage events under the same distribution transformer is used as the numerator to calculate the power outage ratio of users under the distribution transformer;

[0022] Based on the ledger and topology information, obtain the low-voltage user C k The number of meter boxes connected to the corresponding distribution transformer D is used as the denominator, and the number of meter boxes with initial power outage events in the meter boxes of the distribution transformer is used as the numerator to calculate the power outage ratio of the meter boxes under the distribution transformer;

[0023] Based on the ledger and topology information, the nearest upstream switch in the user's area is obtained as the start switch, and the nearest downstream switch in the user's area is obtained as the end switch to form a minimum topology segment. The number of users in the segment is counted as the denominator, and the number of users with initial power outage events in the segment is counted as the numerator to calculate the power outage ratio of the topology segment.

[0024] Based on the obtained user current data under the distribution transformer, the distribution transformer current drop ratio DP is calculated, and the calculation formula is as follows:

[0025]

[0026] Among them, m is the number of power outage time points of the power outage event, n is the total number of users in operation under the distribution transformer to which the user belongs, I x,iis the user current value at a certain time point during the power outage period on a similar power outage day, I i is the current value of the power outage user at a certain time during the power outage period on the power outage day, l x is the current value of the distribution transformer of the user at a certain time during the power outage period on the same day of power outage, and l is the power value of the distribution transformer of the power outage user at a certain time during the power outage period on the same day of power outage;

[0027] The user power consumption reduction ratio EP is calculated as follows:

[0028]

[0029] Among them, m is the number of power outage time points of the power outage event, E x is the power consumption value of users during the power outage period on similar power outage days, and E is the power consumption value of users during the power outage period on the power outage day;

[0030] According to the divided minimum topological segment, check whether there is a tie switch downstream of the switch. If any switch exists, it is yes, otherwise it is no, and it is obtained that there is a tie switch downstream of the user.

[0031] Furthermore, characteristic values ​​are generated based on the acquired data and the initial power outage event. The characteristic values ​​include the distribution transformer current drop ratio, the user power consumption drop ratio, the matching power outage event, the user power outage ratio under the distribution transformer, the meter box power outage ratio under the distribution transformer, the topology segment power outage ratio, the power outage duration of the user's initial power outage event, the matching low-voltage fault record, whether it is a photovoltaic user, the existence of an upstream switch action signal, and the existence of a tie switch downstream, including:

[0032] According to the power outage start time of the user's power outage event, check whether the user sends the power outage event within the sampling point density of the time difference from the power outage start time. If so, it is yes, otherwise it is no, and the matching power outage event is obtained;

[0033] According to the power outage start time of the user's power outage event, check whether the user has a customer service low-voltage fault record within the time difference sampling point density from the power outage start time. If so, it is yes, otherwise it is no, and the matching low-voltage fault record is obtained;

[0034] According to the power outage start time of the user's power outage event, check whether the user's upstream switch in the topology sends a switch opening action signal within the time difference sampling point density from the power outage start time. If yes, it is yes, otherwise it is no, and the upstream switch action signal is obtained;

[0035] According to the power outage start time and end time of the user's power outage event, the time difference is calculated to obtain the power outage duration of the user's initial power outage event;

[0036] According to the user account information, check whether the user is a photovoltaic user and obtain whether the user is a photovoltaic user.

[0037] Furthermore, the characteristic values ​​of each user and the historical power outages are combined to form a sample data set, and the deep neural network model is trained and optimized. Finally, the deep neural network model is evaluated, including:

[0038] A sample data set is formed by combining the characteristic values ​​of each user and the historical power outages, and then divided into a training set and a validation set in proportion. The training set is used to train the deep neural network model, and the validation set is used to test the training effect of the deep neural network model and to tune the deep neural network model. Finally, the F1 coefficient is calculated and used to evaluate the deep neural network model.

[0039] Furthermore, the trained deep neural network model is used to batch identify the initial power outage events of each low-voltage user, and finally outputs the true and false power outage results corresponding to the power outage events of each low-voltage user, including:

[0040] The initial power outage events of each low-voltage user are identified in batches, the required feature values ​​are generated and extracted, and sent to the deep neural network model for calculation, and finally the true and false power outage results corresponding to the power outage events of each low-voltage user are output.

[0041] Based on the same inventive concept, a low-voltage user power outage event recognition and analysis system based on a deep neural network of the present invention comprises:

[0042] The data acquisition module is used to obtain the measurement, ledger, topology, power outage events, customer service fault records, and switch action signal data of low-voltage users and their distribution transformers; the measurements include voltage, power, and current;

[0043] An initial power outage event generation module is used to generate an initial power outage event based on the voltage and power measurement of low-voltage users and the sent power outage events;

[0044] The characteristic value generation module is used to generate characteristic values ​​according to the acquired data and the initial power outage event. The characteristic values ​​include the distribution transformer current reduction ratio, the user power consumption reduction ratio, the matching power outage event, the user power outage ratio under the distribution transformer, the meter box power outage ratio under the distribution transformer, the topological segment power outage ratio, the power outage duration of the user's initial power outage event, the matching low-voltage fault record, whether it is a photovoltaic user, the existence of an upstream switch action signal, and the existence of a contact switch downstream;

[0045] The model building and training module is used to combine the characteristic values ​​of each user and the historical power outage situation to form and divide the sample data set, train and optimize the deep neural network model, and finally evaluate the deep neural network model;

[0046] The power outage event recognition module is used to use the trained deep neural network model to batch identify the initial power outage events of each low-voltage user, and finally output the true and false power outage results corresponding to the power outage events of each low-voltage user.

[0047] Based on the same inventive concept, a low-voltage user power outage event identification and analysis device based on a deep neural network of the present invention includes a processor and a memory, wherein the memory stores computer instructions, and the processor is used to execute the computer instructions stored in the memory. When the computer instructions are executed by the processor, the electronic device implements the steps of the above-mentioned low-voltage user power outage event identification and analysis method based on a deep neural network.

[0048] Based on the same inventive concept, a computer-readable storage medium of the present invention stores a computer program thereon, which, when executed by a processor, implements the steps of the above-mentioned method for identifying and analyzing low-voltage user power outage events based on a deep neural network.

[0049] Beneficial effects: Compared with the prior art, the significant technical effects of the present invention are:

[0050] The present invention learns the changing trends of massive historical data in a data-driven manner, collects a total of 11 relevant characteristic values ​​related to power outage events of low-voltage users, and uses the F1 coefficient to accurately evaluate and determine whether the power outage event model is highly accurate, thereby avoiding manual entry and device collection errors.

[0051] The present invention comprehensively considers various data features such as the measurement drop ratio of each low-voltage user, topological aggregation relationship, topological communication status, and uplink signal, and gradually eliminates false power outage events with high accuracy.

[0052] The invention has simple calculation and clear principle, can help distribution network operators to timely discover real power outage events of low-voltage users, and has good application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a flow chart of a method for identifying and analyzing power outage events of low-voltage users based on a deep neural network disclosed in an embodiment of the present invention;

[0054] Figure 2 It is a structural schematic diagram of a low-voltage user power outage event recognition and analysis system based on a deep neural network disclosed in an embodiment of the present invention;

[0055] Figure 3 It is a structural schematic diagram of a low-voltage user power outage event recognition and analysis device based on a deep neural network disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0056] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. It will be understood by those skilled in the art that the purposes and advantages that can be achieved with the present invention are not limited to the specific description of the above beneficial effects, and the above and other purposes that can be achieved by the present invention will be more clearly understood based on the following detailed description.

[0057] It should be understood by those skilled in the art that the exemplary components, systems and methods described in conjunction with the embodiments disclosed in the present invention can be implemented in hardware, software or a combination of the two. Whether it is implemented in hardware or software depends on the specific application and design of the technical solution. Professional and technical personnel can use different methods to implement the functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0058] Reference to "embodiments" in the present invention means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0059] Example 1

[0060] See also Figure 1 , Figure 1 1 is a flow chart of a method for identifying and analyzing power outage events for low-voltage users based on a deep neural network disclosed in an embodiment of the present invention. Figure 1 The described low-voltage user power outage event identification and analysis method is applied to the power system, such as for low-voltage user power outage event identification, etc., which is not limited in the embodiments of the present invention. Figure 1 As shown, the method for identifying and analyzing low-voltage user power outage events based on a deep neural network may include the following operations:

[0061] S1. Obtain the measurement, ledger, topology, power outage events, customer service fault records, and switch action signal data of low-voltage users and their distribution transformers. The measurement includes voltage, power, and current.

[0062] Measurement, ledger, topology, power outage event data of distribution transformer: basic information maintained by the distribution transformer in the PMS system (power consumption information collection system), drawing topology path data, and power outage event data sent by the equipment.

[0063] Switch action signal data: protection signals, opening and closing signals, FA (feeder automation) tests, switch listing information and other data sent by all automated switches in the distribution network of the distribution automation system.

[0064] Customer service fault record: Customer service power outage information releases the fault record of the relevant distribution transformer.

[0065] In this embodiment, a low-voltage user to be processed is selected in the electricity consumption information collection system (i.e., the PMS system is a system existing in the prior art), and the basic data of the low-voltage users under the distribution transformer stored in the existing system is read, including basic ledger data, various equipment connection relationship data, and uploaded user power outage event data.

[0066] The power consumption information collection system exports the active power, electricity meter bottom, and voltage measurement data of low-voltage users sampled every 15 minutes. The aforementioned sampling frequency of sampling every 15 minutes can be modified according to actual conditions.

[0067] Switch action signal data: protection signals, opening and closing signals, FA test and switch listing information sent by all automatic switches in the distribution network of the D5200 system (distribution automation system).

[0068] Distribution transformer fault record: Customer service power outage information releases the fault record of the relevant distribution transformer.

[0069] S2. Generate an initial power outage event based on the voltage and power measurement of the low-voltage users and the uploaded power outage events.

[0070] Considering that when the active power and voltage measurement data of low-voltage users drop to 0 or are empty at the same time, it may be a real power outage, or it may be a false power outage caused by signal transmission, data storage, etc., it is necessary to effectively and reasonably distinguish between real power outages and false power outages through data analysis methods. Step S2 specifically includes the following:

[0071] Assume a low voltage user C k Connected to distribution transformer T, low voltage user C k The active power measurement data is P k =[p k,1 ,p k,2 ,…,p k,i ,…,p k,n ], low voltage user C k The voltage measurement data is V k =[v k,1 ,v k,2 ,…,v k,i ,…,v k,n ], n represents the number of sampling points of active power and voltage measurement data of low-voltage users;

[0072] If low voltage user C k Active power measurement data P k In the continuous data segment [p k,i ,…,p k,j ] is 0 or empty, and the voltage measurement data Vk In the continuous data segment [v k,i ,…,v k,j ] is 0 or empty, a low voltage user C is generated k The data segment [i,…,j] generates a power outage event set for the data;

[0073] A low voltage user C k If a power outage event is collected, and the power outage time sent is earlier than the power restoration time, and the power outage duration during the power outage period is within two natural days, a power outage event is generated as a low-voltage user C k The user adopts the power outage event set sent up;

[0074] From the obtained low-voltage user measurement data and the collected power outage events, a power outage event set and a power outage event set for uploading are generated for each user on a user-by-user basis to form the initial power outage event for each user.

[0075] S3. Generate characteristic values ​​based on the acquired data and the initial power outage event. The characteristic values ​​include distribution transformer current reduction ratio, user power consumption reduction ratio, matching power outage events, user power outage ratio under distribution transformer, meter box power outage ratio under distribution transformer, topology segment power outage ratio, power outage duration of user initial power outage event, matching low-voltage fault record, whether it is a photovoltaic user, existence of upstream switch action signal, and existence of downstream contact switch, a total of 11 characteristic values.

[0076] According to the measured data during the initial power outage event period of low-voltage users, two characteristic values, the current drop ratio of the distribution transformer and the power consumption drop ratio of users, are calculated; according to the low-voltage user ledger, the distribution transformer topology and the initial power outage event, three characteristic values, the power outage ratio of users under the distribution transformer to which the user belongs, the power outage ratio of the meter box under the distribution transformer, and the power outage ratio of the topological segment, are generated; according to the distribution transformer topology, the characteristic value of the existence of a connecting switch downstream of the user is generated; according to the power outage signal sent by the low-voltage user, the characteristic value of the matching power outage event is generated; according to the sent low-voltage fault record, the characteristic value of the matching low-voltage fault record is generated; according to the switch action signal sent and the topology information, the characteristic value of the existence of the upstream switch action signal is generated; according to the initial power outage event, the characteristic value of the power outage duration of the user's initial power outage event is generated; according to the low-voltage user ledger, the characteristic value of whether it is a photovoltaic user is generated. Finally, a total of 11 characteristic values ​​are generated. Specifically, the following steps are included:

[0077] S3.1. Based on the ledger and topology information, obtain the low-voltage user C k The number of low-voltage users in operation under the corresponding distribution transformer D is taken as the denominator, and the number of users with initial power outage events under the same distribution transformer is taken as the numerator, and the power outage ratio of users under the distribution transformer is calculated.

[0078] S3.2. Based on the ledger and topology information, obtain the low-voltage user C kThe number of meter boxes connected to the corresponding distribution transformer D is used as the denominator, and the number of meter boxes with initial power outage events in the meter boxes of the distribution transformer is used as the numerator to calculate the power outage ratio of the meter boxes under the distribution transformer.

[0079] S3.3. Based on the ledger and topology information, obtain the nearest upstream switch in the user's area as the starting switch, and the nearest downstream switch in the user's area as the ending switch to form a minimum topological segment. Count the number of users in the segment as the denominator, and the number of users with initial power outage events in the segment as the numerator to calculate the power outage ratio of the topological segment.

[0080] S3.4. Based on the obtained user current data of the distribution transformer, calculate the distribution transformer current drop ratio DP, and the calculation formula is as follows:

[0081]

[0082] Among them, m is the number of power outage time points of the power outage event, n is the total number of users in operation under the distribution transformer to which the user belongs, I x,i The current value of the user at a certain time during the power outage period on a similar day (the default is 7 days before the power outage), I i is the current value of the power outage user at a certain time during the power outage period on the power outage day, l x It is the current value of the distribution transformer of the user at a certain time during the power outage period on a similar day (the default is 7 days before) and l is the power value of the distribution transformer of the power outage user at a certain time during the power outage period on the day of the power outage.

[0083] S3.5, user power consumption reduction ratio EP, the calculation formula is as follows:

[0084]

[0085] Among them, m is the number of power outage time points of the power outage event, E x is the power consumption value of users during the power outage period on a similar day of power outage (the default is the 7th day before), and E is the power consumption value of users during the power outage period on the day of the power outage.

[0086] S3.6. According to the minimum topological segment divided in step S3.3, check whether there is a tie switch downstream of the switch. If any switch exists, it is yes, otherwise it is no, and it is obtained that there is a tie switch downstream of the user.

[0087] S3.7. According to the power outage start time of the user's power outage event, check whether the user is within the sampling point density of the time difference from the power outage start time to submit the power outage event for use. If so, it is yes, otherwise it is no, and the matching power outage event is obtained.

[0088] S3.8. According to the power outage start time of the user's power outage event, check whether the user has a customer service low-voltage fault record within the time difference sampling point density from the power outage start time. If so, it is yes, otherwise it is no, and the matching low-voltage fault record characteristics are obtained.

[0089] S3.9. According to the power outage start time of the user's power outage event, check whether the user's upstream switch in the topology sends a switch opening action signal within the time difference sampling point density from the power outage start time. If so, it is yes, otherwise it is no, and the upstream switch action signal is obtained.

[0090] S3.10. Calculate the time difference based on the power outage start time and end time of the user's power outage event to obtain the power outage duration of the user's initial power outage event.

[0091] S3.11. Check whether the user is a photovoltaic user based on the user account information to obtain whether the user is a photovoltaic user.

[0092] S4. Combine the characteristic values ​​of each user and the historical power outage situation to form a sample data set and divide it, train and optimize the deep neural network model (DNN), and finally use the F1 coefficient to evaluate the deep neural network model.

[0093] The sample data set is formed by combining the characteristic values ​​of each user and the historical power outage situation, and is divided into a training set and a validation set at a ratio of 7:3, with 70% as the training set and 30% as the validation set. The deep neural network model is trained with the training set, and the training effect of the deep neural network model is tested and the deep neural network model is tuned with the validation set. Finally, the F1 coefficient is calculated and the deep neural network model is evaluated using the F1 coefficient.

[0094] In this embodiment, the calculation formula of the F1 coefficient is as follows:

[0095]

[0096] Table 1 Model evaluation results. It can be seen from the table below that the precision rate of model training is close to 1, indicating that the accuracy of the model classification result is relatively high; the recall rate of model training is between 0.9 and 1, indicating that the model can identify power outage events for almost all users; the F1 score of the evaluation index calculated based on the precision rate and recall rate of model training is close to 1, indicating that the model classification effect is good and can effectively output whether the user actually has a power outage based on the input user feature value.

[0097] Precision Recall F1 score 0.983 0.912 0.946

[0098] It is worth mentioning that the deep neural network model is an open source mature model, also known as a fully connected network. It is the basic framework of deep learning and a specific model in the neural network model. It can be called directly.

[0099] S5. Use the trained deep neural network model to batch identify the initial power outage events of each low-voltage user, and finally output the true and false power outage results corresponding to the power outage events of each low-voltage user.

[0100] After the model prediction accuracy meets the requirements, the trained model is used to batch identify the initial power outage events of each low-voltage user, and finally the true and false power outage results corresponding to the power outage events of each low-voltage user are output. That is, the initial power outage events of each low-voltage user are batch identified, the required feature values ​​are generated and extracted, and sent to the deep neural network model for calculation, and finally the true and false power outage results corresponding to the power outage events of each low-voltage user are output.

[0101] The present invention uses a data model to learn the changing characteristics of low-voltage user operating data, and determines whether a real power outage occurs to the low-voltage user, thereby achieving accurate collection of power outage events of low-voltage users.

[0102] The present invention uses data analysis capabilities, big data analysis and artificial intelligence algorithms to process and analyze large amounts of monitoring data, quickly and accurately identify power outage events, and infer the location of power outages, thereby achieving real-time monitoring of the operating status of low-voltage users in the power system, providing a new means for diagnosing power outage events for low-voltage users.

[0103] Example 2

[0104] See also Figure 2 , Figure 2 The present invention discloses a structural diagram of a system for identifying and analyzing power outage events for low-voltage users based on a deep neural network. The system can identify power outage events for low-voltage users, and specifically includes:

[0105] Data acquisition module, used to obtain measurement, ledger, topology, power outage events, customer service fault records, and switch action signal data of low-voltage users and their distribution transformers;

[0106] An initial power outage event generation module is used to generate an initial power outage event based on the acquired user voltage, power measurement, and the sent user power outage event;

[0107] The characteristic value generation module is used to generate characteristic values ​​according to the acquired data and the initial power outage event. The characteristic values ​​include the distribution transformer current reduction ratio, the user power consumption reduction ratio, the matching power outage event, the user power outage ratio under the distribution transformer, the meter box power outage ratio under the distribution transformer, the topological segment power outage ratio, the power outage duration of the user's initial power outage event, the matching low-voltage fault record, whether it is a photovoltaic user, the existence of an upstream switch action signal, and the existence of a contact switch downstream;

[0108] The model building and training module is used to combine the characteristic values ​​of each user and the historical power outage situation to form and divide the sample data set, train and optimize the deep neural network model, and finally evaluate the deep neural network model;

[0109] The power outage event recognition module is used to use the trained deep neural network model to batch identify the initial power outage events of each low-voltage user, and finally output the true and false power outage results corresponding to the power outage events of each low-voltage user.

[0110] In an optional embodiment, a method for identifying and analyzing power outage events for low-voltage users based on a deep neural network includes: a) obtaining measurements, ledgers, topology, power outage events, customer service fault records, and switch action signal data of low-voltage users and their distribution transformers; b) generating an initial power outage event based on the acquired user voltage and power measurement and the uploaded power outage event; c) generating 11 characteristic values ​​such as the distribution transformer current drop ratio, the user power consumption drop ratio, the matching power outage event, and the power outage ratio of users under the distribution transformer; c) combining the characteristic values ​​of each user and the historical power outage data to form a sample data set and divide it, train and optimize the deep neural network model, and evaluate the model; d) using the deep neural network model to batch identify the initial power outage events of each low-voltage user, and finally output the true and false power outage results corresponding to the power outage events of each low-voltage user.

[0111] Example 3

[0112] See also Figure 3 , Figure 3 1 is a schematic diagram of a structure of a low-voltage user power outage event recognition and analysis device based on a deep neural network disclosed in an embodiment of the present invention. Figure 3 The described device can be applied to a power system, such as for identifying power outage events of low-voltage users, etc., which is not limited in the embodiments of the present invention.

[0113] like Figure 3 As shown, the device may include a processor and a memory, wherein the memory stores computer instructions, and the processor is used to execute the computer instructions stored in the memory. When the computer instructions are executed by the processor, the electronic device implements the steps of the method described in the above embodiment and can achieve the technical effect consistent with the above method.

[0114] The memory may include computer system readable media in the form of volatile memory, such as random access memory (RAM) and / or cache memory. The device may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the memory may be used to read and write non-removable, non-volatile magnetic media (commonly referred to as a "hard drive"). A program / utility having a set (at least one) of program modules may be stored in, for example, the memory, such program modules including but not limited to an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include the implementation of a network environment. The program modules typically perform the functions and / or methods in the embodiments described herein.

[0115] The processor executes various functional applications and data processing by running the program stored in the memory, such as implementing the method provided in the first embodiment of the present invention.

[0116] Example 4

[0117] Embodiment 4 of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps of the method described in the above embodiment are implemented and the technical effect consistent with the above method can be achieved.

[0118] The computer storage medium of the embodiment of the present invention can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, a device or a device or used in combination with it.

[0119] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0120] The program code embodied on the computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0121] Computer program code for performing the operation of the present invention may be written in one or more programming languages ​​or combinations thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0122] Of course, the storage medium containing computer executable instructions provided by an embodiment of the present invention, whose computer executable instructions are not limited to the above method operations, can also execute related operations in the method provided by any embodiment of the present invention.

[0123] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for identifying and analyzing low-voltage user power outage events based on deep neural networks, characterized in that: include: Obtain the measurement, ledger, topology, power outage events, customer service fault records, and switch action signal data of low-voltage users and their distribution transformers; among them, the measurements include voltage, power, and current; Generate an initial power outage event based on the voltage and power of the low-voltage user and the power outage event sent up; Generate characteristic values ​​based on the acquired data and the initial power outage event. The characteristic values ​​include the distribution transformer current drop ratio, the user power consumption drop ratio, the matching power outage event, the user power outage ratio under the distribution transformer, the meter box power outage ratio under the distribution transformer, the topology segment power outage ratio, the power outage duration of the user's initial power outage event, the matching low-voltage fault record, whether it is a photovoltaic user, the existence of an upstream switch action signal, and the existence of a contact switch downstream; Combine the characteristic values ​​of each user and historical power outage data to form a sample data set and divide it, train and optimize the deep neural network model, and finally use the F1 coefficient to evaluate the deep neural network model; The trained deep neural network model is used to batch identify the initial power outage events of each low-voltage user, and finally outputs the true and false power outage results corresponding to the power outage events of each low-voltage user.

2. The method for identifying and analyzing low-voltage user power outage events based on a deep neural network according to claim 1 is characterized in that: Obtain the measurement, ledger, topology, power outage events, customer service fault records, and switch action signal data of low-voltage users and their distribution transformers, including: The measurement, ledger, topology, and power outage event data of the distribution transformer include the basic information maintained by the distribution transformer in the PMS system, the drawing topology path data, and the power outage event data sent by the equipment; The switch action signal data includes the protection signals, opening and closing signals, FA test and switch listing information data sent by all automatic switches in the distribution network of the distribution automation system; Customer service fault records include fault records of distribution transformers related to the release of customer service power outage information.

3. The method for identifying and analyzing low-voltage user power outage events based on deep neural networks according to claim 1 is characterized in that: Generate an initial power outage event based on the acquired user voltage, power measurement, and the sent power outage event, including: Assume low voltage user C k Connected to distribution transformer T, low voltage user C k The active power measurement data is P k =[p k,1 ,p k,2 ,…,p k,i ,…,p k,n ], low voltage user C k The voltage measurement data is V k =[v k,1 ,v k,2 ,…,v k,i ,…,v k,n ], n represents the number of sampling points of active power and voltage measurement data of low-voltage users; If low voltage user C k Active power measurement data P k In the continuous data segment [p k,i ,…,p k,j ] is 0 or empty, and the voltage measurement data V k In the continuous data segment [v k,i ,…,v k,j ] is 0 or empty, a low voltage user C is generated k The data segment [i,…,j] generates a power outage event set for the data; Low voltage user C k If a power outage event is collected, and the power outage time sent is earlier than the power restoration time, and the power outage duration during the power outage period is within two natural days, a power outage event is generated as a low-voltage user C k The user adopts the power outage event set sent up; From the obtained low-voltage user measurement data and the collected power outage events, a power outage event set and a power outage event set for uploading are generated for each user on a user-by-user basis to form the initial power outage event for each user.

4. The method for identifying and analyzing low-voltage user power outage events based on deep neural networks according to claim 1 is characterized in that: Generate characteristic values ​​based on the acquired data and the initial power outage event. The characteristic values ​​include the distribution transformer current drop ratio, the user power consumption drop ratio, the matching power outage event, the user power outage ratio under the distribution transformer, the meter box power outage ratio under the distribution transformer, the topology segment power outage ratio, the power outage duration of the user's initial power outage event, the matching low-voltage fault record, whether it is a photovoltaic user, the existence of an upstream switch action signal, and the existence of a contact switch downstream, including: Based on the ledger and topology information, obtain the low-voltage user C k The number of low-voltage users in operation under the corresponding distribution transformer D is used as the denominator, and the number of users with initial power outage events under the same distribution transformer is used as the numerator to calculate the power outage ratio of users under the distribution transformer; Based on the ledger and topology information, obtain the low-voltage user C k The number of meter boxes connected to the corresponding distribution transformer D is used as the denominator, and the number of meter boxes with initial power outage events in the meter boxes of the distribution transformer is used as the numerator to calculate the power outage ratio of the meter boxes under the distribution transformer; Based on the ledger and topology information, the nearest upstream switch in the user's area is obtained as the start switch, and the nearest downstream switch in the user's area is obtained as the end switch to form a minimum topology segment. The number of users in the segment is counted as the denominator, and the number of users with initial power outage events in the segment is counted as the numerator to calculate the power outage ratio of the topology segment. Based on the obtained user current data under the distribution transformer, the distribution transformer current drop ratio DP is calculated, and the calculation formula is as follows: Among them, m is the number of power outage time points of the power outage event, n is the total number of users in operation under the distribution transformer to which the user belongs, I x,i is the user current value at a certain time point during the power outage period on a similar power outage day, I i is the current value of the power outage user at a certain time during the power outage period on the power outage day, l x is the current value of the distribution transformer of the user at a certain time during the power outage period on the same day of power outage, and l is the power value of the distribution transformer of the power outage user at a certain time during the power outage period on the same day of power outage; The user power consumption reduction ratio EP is calculated as follows: Among them, m is the number of power outage time points of the power outage event, E x is the power consumption value of users during the power outage period on similar power outage days, and E is the power consumption value of users during the power outage period on the power outage day; According to the divided minimum topological segment, check whether there is a tie switch downstream of the switch. If any switch exists, it is yes, otherwise it is no, and it is obtained that there is a tie switch downstream of the user.

5. The method for identifying and analyzing low-voltage user power outage events based on deep neural networks according to claim 1 is characterized in that: Generate characteristic values ​​based on the acquired data and the initial power outage event. The characteristic values ​​include the distribution transformer current drop ratio, the user power consumption drop ratio, the matching power outage event, the user power outage ratio under the distribution transformer, the meter box power outage ratio under the distribution transformer, the topology segment power outage ratio, the power outage duration of the user's initial power outage event, the matching low-voltage fault record, whether it is a photovoltaic user, the existence of an upstream switch action signal, and the existence of a contact switch downstream, including: According to the power outage start time of the user's power outage event, check whether the user sends the power outage event within the sampling point density of the time difference from the power outage start time. If so, it is yes, otherwise it is no, and the matching power outage event is obtained; According to the power outage start time of the user's power outage event, check whether the user has a customer service low-voltage fault record within the time difference sampling point density from the power outage start time. If so, it is yes, otherwise it is no, and the matching low-voltage fault record is obtained; According to the power outage start time of the user's power outage event, check whether the user's upstream switch in the topology sends a switch opening action signal within the time difference sampling point density from the power outage start time. If yes, it is yes, otherwise it is no, and the upstream switch action signal is obtained; According to the power outage start time and end time of the user's power outage event, the time difference is calculated to obtain the power outage duration of the user's initial power outage event; According to the user account information, check whether the user is a photovoltaic user and obtain whether the user is a photovoltaic user.

6. The method for identifying and analyzing low-voltage user power outage events based on deep neural networks according to claim 1 is characterized in that: Combine the characteristic values ​​of each user and the historical power outage situation to form a sample data set, train and optimize the deep neural network model, and finally evaluate the deep neural network model, including: A sample data set is formed by combining the characteristic values ​​of each user and the historical power outages, and then divided into a training set and a validation set in proportion. The training set is used to train the deep neural network model, and the validation set is used to test the training effect of the deep neural network model and to tune the deep neural network model. Finally, the F1 coefficient is calculated and used to evaluate the deep neural network model.

7. The method for identifying and analyzing low-voltage user power outage events based on deep neural networks according to claim 1 is characterized in that: The trained deep neural network model is used to batch identify the initial power outage events of each low-voltage user, and finally outputs the true and false power outage results corresponding to the power outage events of each low-voltage user, including: The initial power outage events of each low-voltage user are identified in batches, the required feature values ​​are generated and extracted, and sent to the deep neural network model for calculation, and finally the true and false power outage results corresponding to the power outage events of each low-voltage user are output.

8. A low-voltage user power outage event recognition and analysis system based on deep neural network, characterized in that: include: The data acquisition module is used to obtain the measurement, ledger, topology, power outage events, customer service fault records, and switch action signal data of low-voltage users and their distribution transformers; the measurements include voltage, power, and current; An initial power outage event generation module is used to generate an initial power outage event based on the voltage and power measurement of low-voltage users and the sent power outage events; The characteristic value generation module is used to generate characteristic values ​​according to the acquired data and the initial power outage event. The characteristic values ​​include the distribution transformer current reduction ratio, the user power consumption reduction ratio, the matching power outage event, the user power outage ratio under the distribution transformer, the meter box power outage ratio under the distribution transformer, the topological segment power outage ratio, the power outage duration of the user's initial power outage event, the matching low-voltage fault record, whether it is a photovoltaic user, the existence of an upstream switch action signal, and the existence of a contact switch downstream; The model building and training module is used to combine the characteristic values ​​of each user and the historical power outage situation to form and divide the sample data set, train and optimize the deep neural network model, and finally evaluate the deep neural network model; The power outage event recognition module is used to use the trained deep neural network model to batch identify the initial power outage events of each low-voltage user, and finally output the true and false power outage results corresponding to the power outage events of each low-voltage user.

9. A low-voltage user power outage event recognition and analysis device based on deep neural network, characterized in that: The electronic device comprises a processor and a memory, wherein the memory stores computer instructions, and the processor is used to execute the computer instructions stored in the memory. When the computer instructions are executed by the processor, the electronic device implements the steps of the method for identifying and analyzing power outage events for low-voltage users based on a deep neural network as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a method for identifying and analyzing low-voltage user power outage events based on a deep neural network as described in any one of claims 1 to 7.

Images