CN117251809A - Power grid time sequence data anomaly detection method, device, equipment and storage medium - Google Patents
Power grid time sequence data anomaly detection method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN117251809A CN117251809A CN202311101772.XA CN202311101772A CN117251809A CN 117251809 A CN117251809 A CN 117251809A CN 202311101772 A CN202311101772 A CN 202311101772A CN 117251809 A CN117251809 A CN 117251809A
- Authority
- CN
- China
- Prior art keywords
- power grid
- time sequence
- grid time
- encoder
- subsequence
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 153
- 238000003860 storage Methods 0.000 title claims abstract description 18
- 230000005856 abnormality Effects 0.000 claims abstract description 95
- 238000000034 method Methods 0.000 claims abstract description 63
- 238000012549 training Methods 0.000 claims abstract description 63
- 230000002159 abnormal effect Effects 0.000 claims abstract description 48
- 230000006870 function Effects 0.000 claims description 36
- 238000004590 computer program Methods 0.000 claims description 16
- 238000011478 gradient descent method Methods 0.000 claims description 12
- 230000008569 process Effects 0.000 claims description 9
- 238000013507 mapping Methods 0.000 claims description 6
- 230000000875 corresponding effect Effects 0.000 description 59
- 238000004422 calculation algorithm Methods 0.000 description 10
- 238000012545 processing Methods 0.000 description 10
- 238000004891 communication Methods 0.000 description 9
- 239000000306 component Substances 0.000 description 9
- 230000011218 segmentation Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 238000007781 pre-processing Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000009467 reduction Effects 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 239000008358 core component Substances 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 239000000178 monomer Substances 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
- G06N3/0455—Auto-encoder networks; Encoder-decoder networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/088—Non-supervised learning, e.g. competitive learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2123/00—Data types
- G06F2123/02—Data types in the time domain, e.g. time-series data
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a method, a device, equipment and a storage medium for detecting power grid time sequence data abnormality. Wherein the method comprises the following steps: acquiring power grid time sequence data to be detected, and determining at least one power grid time sequence subsequence of the power grid time sequence data; determining a reconstruction result of the power grid time sequence subsequence based on a preset anomaly detection network model, wherein the preset anomaly detection network model is obtained after serial training of at least two self-encoders; and determining an abnormal detection result of the power grid time sequence subsequence according to the abnormal score of the reconstruction result. According to the embodiment of the invention, the acquired power grid time sequence subsequence of the power grid time sequence data is input into the preset abnormality detection network model to obtain the corresponding reconstruction result, so that the abnormality detection result corresponding to the power grid time sequence subsequence is determined according to the abnormality score of the reconstruction result, the problem of low accuracy of power grid time sequence data abnormality detection is solved, and an accurate and efficient power grid time sequence data abnormality detection method is realized.
Description
Technical Field
The present invention relates to the field of data processing technologies, and in particular, to a method, an apparatus, a device, and a storage medium for detecting power grid time sequence data anomalies.
Background
The real-time measurement center of the power grid resource business center supports real-time business application construction such as power grid operation analysis and equipment state monitoring by converging power-saving quantity and non-electric quantity measurement data of each ring of the power grid in real time and integrating data of the primary grid frame and the secondary device in a correlated manner, and comprehensively improves calculation and analysis capacity of the power grid and observability and descriptability of the distribution network. However, due to various communication faults, equipment faults, power grid fluctuation, abnormal user behaviors and other factors, a large amount of abnormal data are inevitably generated. These anomaly data have an important impact on the accuracy and integrity of the grid timing data. Therefore, how to identify abnormal data from a large amount of grid time sequence data has great significance for subsequent data analysis and data application.
At present, detection of abnormal data generally adopts the following modes: firstly, determining a reasonable fluctuation range of target detection data through the distribution of historical data based on a statistical abnormality detection algorithm, such as an abnormality detection algorithm based on a 3sigma criterion; secondly, an abnormality detection algorithm based on clustering is used for judging the boundary between an abnormal value and a normal value by clustering normal data; and thirdly, an abnormality detection algorithm based on classification, such as One-Class SVM, judges by finding a hyperplane between a normal value and an abnormal value. However, for multi-dimensional grid time sequence data, the anomaly detection algorithm generally has the problem of poor detection accuracy.
Disclosure of Invention
The invention provides a method, a device, equipment and a storage medium for detecting power grid time sequence data abnormality, which are used for solving the problem of poor accuracy of power grid time sequence data abnormality detection in the related technology.
According to an aspect of the present invention, there is provided a method for detecting abnormality of time-series data of a power grid, the method comprising:
acquiring power grid time sequence data to be detected, and determining at least one power grid time sequence subsequence of the power grid time sequence data;
determining a reconstruction result of the power grid time sequence subsequence based on a preset anomaly detection network model, wherein the preset anomaly detection network model is obtained after serial training of at least two self-encoders;
and determining an abnormal detection result of the power grid time sequence subsequence according to the abnormal score of the reconstruction result.
According to another aspect of the present invention, there is provided a power grid time series data anomaly detection device, the device comprising:
the data acquisition module is used for acquiring the power grid time sequence data to be detected and determining at least one power grid time sequence subsequence of the power grid time sequence data;
the reconstruction result determining module is used for determining a reconstruction result of the power grid time sequence subsequence based on a preset anomaly detection network model, wherein the preset anomaly detection network model is obtained after serial training of at least two self-encoders;
And the abnormality detection result determining module is used for determining an abnormality detection result of the power grid time sequence subsequence according to the abnormality score of the reconstruction result.
According to another aspect of the present invention, there is provided an electronic apparatus including:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,
the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the grid time series data anomaly detection method according to any one of the embodiments of the present invention.
According to another aspect of the present invention, there is provided a computer readable storage medium storing computer instructions for causing a processor to implement the grid time series data anomaly detection method according to any one of the embodiments of the present invention when executed.
According to the technical scheme, the power grid time sequence data to be detected are obtained, and at least one power grid time sequence subsequence of the power grid time sequence data is determined; determining a reconstruction result of the power grid time sequence subsequence based on a preset anomaly detection network model, wherein the preset anomaly detection network model is obtained after serial training of at least two self-encoders; and determining an abnormal detection result of the power grid time sequence subsequence according to the abnormal score of the reconstruction result. According to the embodiment of the invention, the acquired power grid time sequence subsequence of the power grid time sequence data is input into the preset abnormality detection network model to obtain the corresponding reconstruction result, so that the abnormality detection result corresponding to the power grid time sequence subsequence is determined according to the abnormality score of the reconstruction result, the problem of low accuracy of power grid time sequence data abnormality detection is solved, and an accurate and efficient power grid time sequence data abnormality detection method is realized.
It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the invention or to delineate the scope of the invention. Other features of the present invention will become apparent from the description that follows.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a method for detecting power grid time series data anomalies according to a first embodiment of the present invention;
fig. 2 is a flowchart of a method for detecting power grid time series data anomalies according to a second embodiment of the present invention;
FIG. 3 is a flowchart of a training process of a preset anomaly detection network model according to a second embodiment of the present invention;
fig. 4 is a flowchart of a method for detecting power grid time series data anomalies according to a third embodiment of the present invention;
fig. 5 is a schematic structural diagram of a power grid time-series data anomaly detection device according to a fourth embodiment of the present invention;
Fig. 6 is a schematic structural diagram of an electronic device for implementing the method for detecting power grid time-series data anomalies according to an embodiment of the present invention.
Detailed Description
In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and the claims of the present invention and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the invention described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
Example 1
Fig. 1 is a flowchart of a method for detecting an abnormality of power grid time series data according to an embodiment of the present invention, where the method may be performed by a power grid time series data abnormality detection device, and the power grid time series data abnormality detection device may be implemented in hardware and/or software, and the power grid time series data abnormality detection device may be configured in an electronic device, and the electronic device may be, for example, a computer device or a server. As shown in fig. 1, the method for detecting abnormal power grid time sequence data provided in the first embodiment specifically includes the following steps:
s110, acquiring power grid time sequence data to be detected, and determining at least one power grid time sequence subsequence of the power grid time sequence data.
The power grid time sequence data may refer to time sequence data collected by each link in the power grid, and the power grid time sequence data may include, but is not limited to: voltage, current, active power and reactive power, etc. The power grid time sequence subsequence may refer to a plurality of subsequences obtained by performing time sequence segmentation on power grid time sequence data.
In the embodiment of the present invention, the manner of acquiring the power grid time sequence data to be detected may include, but is not limited to, the following: the power grid time sequence data to be detected can be obtained from a database of a local or remote server; the power grid time sequence data can also be obtained by calling a service interface of a real-time measurement center of a power grid resource business, and the obtained power grid time sequence data has high density and large data volume, and is inconvenient for subsequent data processing operation, so that a pre-configured time sequence segmentation algorithm can be called to segment the power grid time sequence data, and a plurality of power grid time sequence subsequences are obtained, wherein the time sequence segmentation algorithm can comprise, but is not limited to: sliding window methods, K-means clustering methods, segmentation methods based on variable point detection, etc., the grid time series data may include, but are not limited to: voltage, current, active power and reactive power, etc. It may be understood that the grid time sequence data may include all the dimensional information, or may include only part of the dimensional information, or may further include information with more dimensions, which is not limited in the embodiment of the present invention.
S120, determining a reconstruction result of the power grid time sequence subsequence based on a preset anomaly detection network model, wherein the preset anomaly detection network model is obtained after serial training of at least two self-encoders.
The preset anomaly detection network model may be a preconfigured network model for anomaly detection of a power grid time sequence subsequence, and the preset anomaly detection network model may be obtained after serial training of at least two self-encoders. The reconstruction result may be the reconstruction data, which is obtained by performing the dimension reduction processing on the input power grid time sequence subsequence and then reconstructing the input power grid time sequence subsequence by using a preset anomaly detection network model.
In the embodiment of the present invention, the preset anomaly detection network model may be a self encoder (AE), which may be obtained after serial training by at least two self encoders, and the obtained power grid time sequence subsequence may be input to the trained preset anomaly detection network model, and the power grid time sequence subsequence is mapped to a low-dimensional space through encoding by using the preset anomaly detection network model to obtain a low-dimensional representation, and then mapped to a reconstruction result close to the power grid time sequence subsequence through decoding according to the low-dimensional representation.
It can be appreciated that the principle of the self-encoder for data anomaly detection is mainly: the self-encoder may encode the normal data into a low-dimensional representation, thereby capturing features of the normal data; when abnormal data is encountered, the self-encoder encodes according to the encoding mode of normal data, so that the reconstruction result obtained after the abnormal data is processed by the self-encoder has larger difference from the original data, and whether the input data is abnormal or not can be detected.
S130, determining an abnormality detection result of the power grid time sequence subsequence according to the abnormality score of the reconstruction result.
The anomaly score can be used for representing a reconstruction error of the power grid time sequence subsequence and a corresponding reconstruction result.
In the embodiment of the invention, the obtained reconstruction result and the reconstruction error of the corresponding power grid time sequence subsequence can be determined, for example, the abnormal score corresponding to the power grid time sequence subsequence can be determined by calculating the mean square error (Mean Square Error, MSE), the L1 norm, the L2 norm and the like between the reconstruction result and the corresponding power grid time sequence subsequence, whether the abnormal score is larger than a preset abnormal score threshold value is judged, if yes, the abnormal detection result corresponding to the power grid time sequence subsequence is determined to be abnormal, otherwise, the abnormal detection result is determined to be abnormal.
Further, on the basis of the embodiment of the invention, after the abnormality detection result corresponding to the power grid time sequence sub-sequence is determined, the abnormality detection result is the corresponding power grid time sequence sub-sequence with abnormality and the reconstruction result are visually displayed, so that a worker can quickly locate a time point or a time interval when the power grid time sequence data is abnormal.
According to the technical scheme, the power grid time sequence data to be detected are obtained, and at least one power grid time sequence subsequence of the power grid time sequence data is determined; determining a reconstruction result of the power grid time sequence subsequence based on a preset anomaly detection network model, wherein the preset anomaly detection network model is obtained after serial training of at least two self-encoders; and determining an abnormal detection result of the power grid time sequence subsequence according to the abnormal score of the reconstruction result. According to the embodiment of the invention, the acquired power grid time sequence subsequence of the power grid time sequence data is input into the preset abnormality detection network model to obtain the corresponding reconstruction result, so that the abnormality detection result corresponding to the power grid time sequence subsequence is determined according to the abnormality score of the reconstruction result, the problem of low accuracy of power grid time sequence data abnormality detection is solved, and an accurate and efficient power grid time sequence data abnormality detection method is realized.
Example two
Fig. 2 is a flowchart of a method for detecting abnormal power grid time sequence data according to a second embodiment of the present invention, which is further optimized and expanded based on the foregoing embodiments, and may be combined with each of the optional technical solutions in the foregoing embodiments. As shown in fig. 2, the method for detecting the abnormal power grid time sequence data provided in the second embodiment specifically includes the following steps:
s210, a service interface of a real-time measurement center of the power grid resource business center is called to acquire power grid time sequence data.
The power grid resource business center can be an aggregate with standardized model data, standardized service interfaces and basic commonality business capability, is a novel infrastructure platform different from a traditional monomer business system, is a platform for sharing and converging power grid resource data and equipment management service, and provides the capabilities of resource asset management, graphic topology management, equipment operation analysis and the like. The real-time measurement center is an important component of the power grid resource business center, and provides the capabilities of data real-time aggregation, high-speed forwarding sharing and the like. The service interface can be a data service interface provided by the real-time measurement center for the business center and the information system which cooperate with the real-time measurement center, and time sequence data such as voltage, current, power and the like of each power grid device such as a transformer, a bus, a circuit breaker and the like can be obtained in real time by utilizing the service interface.
In the embodiment of the invention, the real-time measurement center of the power grid resource business center serves as a core component of the novel power system, and the power grid time sequence data collected in each link in the power grid can be obtained by calling one or more service interfaces configured by the center.
S220, a sliding window method with a preset time window length is called to divide the power grid time sequence data so as to generate a power grid time sequence subsequence.
The sliding window method is a time sequence segmentation algorithm, and the preset time window length can be understood as the length of a time window in the pre-configured sliding window method, and the preset time window length can be set correspondingly according to actual power grid time sequence data, for example, the method is not limited to 12 hours, 24 hours and the like.
In the embodiment of the invention, because the acquired power grid time sequence data has high density and large data volume, the subsequent data processing operation is inconvenient, a pre-configured sliding window method can be called, the power grid time sequence data is segmented based on the length of a pre-set time window, and a plurality of power grid time sequence subsequences are obtained.
S230, inputting the power grid time sequence subsequence into a trained preset anomaly detection network model, mapping the power grid time sequence subsequence to a hidden layer by using an encoder of the preset anomaly detection network model to obtain coding features corresponding to the power grid time sequence subsequence, and decoding and restoring the coding features by using a decoder of the preset anomaly detection network model to obtain a reconstruction result.
The coding feature may be a low-dimensional representation obtained by mapping the input power grid time sequence subsequence to the hidden layer by using an encoder of a preset anomaly detection network model.
In the embodiment of the present invention, the preset anomaly detection network model may be a self-encoder (AE), which may be obtained after serial training by at least two self-encoders, where the preset anomaly detection network model includes an encoder and a decoder, where the encoder is configured to map an input power grid time sequence sub-sequence to a hidden layer to obtain a low-dimensional representation, i.e., an encoded feature, corresponding to the power grid time sequence sub-sequence, and the decoder is configured to decode, restore, i.e., reconstruct, the encoded feature to obtain a corresponding reconstruction result close to the input power grid time sequence sub-sequence.
S240, taking the L2 norm of the difference value between the reconstruction result and the corresponding element of the power grid time sequence sub-sequence as an anomaly score.
In the embodiment of the invention, the anomaly score can be used for representing the reconstruction error of the power grid time sequence subsequence and the corresponding reconstruction result, and specifically, the L2 norm of the difference value between the reconstruction result and the corresponding element of the power grid time sequence subsequence can be used as the anomaly score.
S250, judging whether the abnormality score is larger than a preset abnormality threshold, if so, determining that the abnormality detection result of the power grid time sequence subsequence is abnormal, and if not, determining that the abnormality detection result is not abnormal.
The preset anomaly threshold value may be understood as a threshold value for anomaly score that is configured in advance.
In the embodiment of the invention, after determining the abnormality score corresponding to the power grid time sequence sub-sequence, whether the abnormality score is larger than the preset abnormality threshold value or not can be judged, if yes, the abnormality detection result corresponding to the power grid time sequence sub-sequence is determined to be abnormal, otherwise, the abnormality detection result is determined to be abnormal.
Further, on the basis of the above embodiment of the present invention, the method for detecting the anomaly of the time-series data of the power grid provided in the second embodiment of the present invention further includes:
and visually displaying the abnormal detection result which is the corresponding power grid time sequence subsequence with the abnormality and the reconstruction result.
In the embodiment of the invention, if the abnormality detection result corresponding to the power grid time sequence subsequence is abnormal, the corresponding power grid time sequence subsequence and the reconstruction result are visually displayed in a time sequence chart mode and the like, so that a worker can quickly locate the time point or time interval when the power grid time sequence data is abnormal.
Further, on the basis of the above embodiment of the present invention, as shown in fig. 3, the training process of the preset anomaly detection network model in the embodiment of the present invention may include the following steps:
S310, acquiring power grid time sequence training data, and calling a sliding window method with a preset time window length to determine a power grid time sequence training sub-sequence set of the power grid time sequence training data.
The grid time sequence training data may refer to grid time sequence data for training a network model. The power grid time sequence training subsequence set may refer to a set containing a plurality of subsequences obtained by dividing power grid time sequence training data by a sliding window method.
In the embodiment of the invention, historical power grid time sequence data of a plurality of data volumes can be extracted from a database of a local or remote server to be used as power grid time sequence training data required by model training, and the acquired power grid time sequence training data is segmented by adopting the same data preprocessing strategy as that in the step S220, so that a corresponding power grid time sequence training subsequence set is obtained.
S320, constructing a first self-encoder and a second self-encoder, wherein the first self-encoder is composed of a first encoder of three full-connection layers and a first decoder of three full-connection layers, and the second self-encoder is composed of a second encoder of three full-connection layers and a second decoder of three full-connection layers.
In the embodiment of the present invention, two self encoders, namely, the first self encoder AE, may be respectively constructed 1 And a second self-encoder AE 2 Wherein, the first self-encoder AE 1 Consists of three full-connection-layer first encoders E1 and three full-connection-layer first decoders D1, and a second self-encoder AE 2 Consists of three fully connected layer second encoders E2 and three fully connected layer second decoders D2. It will be appreciated that the network structure and model parameters of the encoder and decoder in the two self-encoders may be the same or different, and embodiments of the invention are not limited in this respect.
S330, taking the power grid time sequence training subsequence set as the input of a first self-encoder, taking a first reconstruction result output by the first self-encoder as the input of a second self-encoder, taking an output result of the second self-encoder as a second reconstruction result, and calling a preset gradient descent method to update corresponding weight parameters and bias parameters in the first encoder, the first decoder, the second encoder and the second decoder based on a first loss function corresponding to the first self-encoder and a second loss function corresponding to the second self-encoder until the first loss function and the second loss function converge.
The first reconstruction result and the second reconstruction result may refer to reconstruction results output by the first self-encoder and the second self-encoder, respectively. The first and second loss functions may refer to corresponding loss functions of the first and second self-encoders, respectively, which may be represented using a minimized reconstruction error of the corresponding self-encoder. The preset gradient descent method may refer to an iterative optimization algorithm for solving optimization update of corresponding weight parameters and bias parameters in a first encoder and a first decoder in a first self-encoder and a second decoder in a second self-encoder, and the preset gradient descent method may include: random gradient descent method, random average gradient descent method, batch gradient descent method, etc.
In the embodiment of the invention, for the first self-encoder AE 1 The input is a power grid time sequence training subsequence set X, and the first coding characteristic mapped to the corresponding hidden layer is h 1 ＝σ(W E1 X+b E1 ) For h 1 The first reconstruction result obtained after decoding reduction is y=σ (W D1 X+b D1 ) Wherein W is E1 And b E1 Respectively represent the weight parameter and the bias parameter, W, in the first encoder E1 D1 And b D1 Representing the weight parameters and the bias parameters, respectively, in the first decoder D1, σ represents the activation function for implementing the nonlinear mapping, and the activation function σ may include, but is not limited to: sigmoid function, tanh function, relu function, etc.; similarly, for the second self-encoder AE 2 Its input is a first self-encoder AE 1 The second coding feature mapped to the corresponding hidden layer is h as the first reconstruction result Y of (1) 2 ＝σ(W E2 X+b E2 ) For h 2 The second reconstruction result obtained after the decoding reduction is z=σ (W D2 Y+b D2 ) Wherein W is E2 And b E2 Respectively represent the weight parameter and the bias parameter, W, in the second encoder E2 D2 And b D2 Respectively representing the weight parameters and the bias parameters in the second decoder D2.
At the first self-encoder AE 1 And a second self-encoder AE 2 In the serial training process of (2), the reconstruction errors of the minimized power grid time sequence training sub-sequence set X and the first reconstruction result Y and the reconstruction errors of the minimized power grid time sequence training sub-sequence set X and the second reconstruction result Z can be respectively used as a first loss function L corresponding to the first self-encoder AE1 And a second loss function L corresponding to the second self-encoder AE2 And the network parameters W are controlled by a preset gradient descent method such as a random gradient descent method and a batch gradient descent method E1 、b E1 、W D1 、b D1 、W E2 、b E2 、W D2 And b D2 Iteratively updating to achieve a first loss function L AE1 A second loss function L AE2 Is described.
Further, on the basis of the above embodiment of the invention, the first loss function L AE1 May include at least: l (L) AE1 ＝min║Y-X║ 2 Second loss function L AE2 May include at least: l (L) AE2 ＝min║Z-X║ 2 。
S340, constructing a third self-encoder by using the trained first encoder and the trained second decoder, and taking the third self-encoder as a trained preset anomaly detection network model.
In the embodiment of the invention, in the first self-encoder AE 1 And a second self-encoder AE 2 After serial training, the first self-encoder AE may be used 1 First encoder E1 and second self encoder AE of (a) 2 The second decoder D2 of (a) takes out and constructs a new self-encoder, namely a third self-encoder AE 3 And the third self-encoder AE 3 And the training is used as a preset anomaly detection network model.
According to the technical scheme, the power grid time sequence data are obtained by calling a service interface of a power grid resource business center station real-time measurement center; calling a sliding window method with a preset time window length to divide the power grid time sequence data so as to generate a power grid time sequence subsequence; inputting the power grid time sequence subsequence into a trained preset anomaly detection network model, mapping the power grid time sequence subsequence to a hidden layer by using an encoder of the preset anomaly detection network model to obtain coding features corresponding to the power grid time sequence subsequence, and decoding and restoring the coding features by using a decoder of the preset anomaly detection network model to obtain a reconstruction result; taking the L2 norm of the difference value between the reconstruction result and the corresponding element of the power grid time sequence sub-sequence as an abnormal score; judging whether the abnormality score is larger than a preset abnormality threshold, if so, determining that the abnormality detection result of the power grid time sequence subsequence is abnormal, and if not, determining that the abnormality detection result is not abnormal. The embodiment of the invention solves the problem of lower accuracy of power grid time sequence data anomaly detection, realizes an accurate and efficient power grid time sequence data anomaly detection method, and can provide a data anomaly early warning function for a power grid resource business center.
Example III
Fig. 4 is a flowchart of a method for detecting an anomaly of power grid time series data according to a third embodiment of the present invention, where the present embodiment provides an implementation manner of the method for detecting an anomaly of power grid time series data based on the foregoing embodiment, so as to enable anomaly detection of power grid time series data, and the anomaly detection accuracy is higher. As shown in fig. 4, the method for detecting the power grid time sequence data abnormality provided in the third embodiment of the present invention specifically includes the following steps:
s410, acquiring power grid time sequence training data, and determining a power grid time sequence training sub-sequence set after preprocessing the power grid time sequence training data.
In the embodiment of the invention, a group of power grid time sequence training data, such as voltage, current, active power, reactive power and the like, can be obtained from a database of a local or remote server, and then the power grid time sequence training data is segmented by utilizing a sliding window method with a time window length of k to obtain a power grid time sequence training subsequence set X= { X containing a plurality of time subsequences 1 ,x 2 ,……,x n And n represents the number of time sub-sequences.
S420, constructing a first self-encoder and a second self-encoder, and utilizing a power grid time sequence training subsequence set to train the first self-encoder and the second self-encoder in series, so as to generate a trained preset anomaly detection network model.
In the embodiment of the present invention, two self-encoders, namely, a first self-encoder AE, are respectively constructed 1 And a second self-encoder AE 2 Wherein, the first self-encoder AE 1 Is composed of three full-connection layer encoders E1 and three full-connection layer decoders D1, and a second self-encoder AE 2 Is composed of three full-connection layer encoders E2 and three full-connection layer decoders D2. First self-encoder AE 1 And a second self-encoder AE 2 The generating process of the serial training of the preset anomaly detection network model can comprise the following steps:
(1) taking the power grid time sequence training subsequence set X as a first self-encoder AE 1 Is input to the first self-encoder AE 1 The output first reconstruction result Y is taken as a second self-encoder AE 2 To obtain a second self-encoder AE 1 Outputting a second reconstruction result Z;
(2) correspondingly constructing a first self-encoder AE according to the reconstruction errors of the power grid time sequence training subsequence set X and the first reconstruction result Y and the second reconstruction result Z respectively 1 Corresponding first loss function L AE1 Second self-encoder AE 2 Corresponding second loss function L AE2 I.e. L AE1 ＝min║Y-X║ 2 ，L AE2 ＝min║Z-X║ 2 ；
(3) Invoking a random gradient descent method to update corresponding weight parameters and bias parameters in the first encoder E1, the first decoder D1, the second encoder E2 and the second decoder D2 until a first loss function L AE1 And a second loss function L AE2 Convergence, i.e. the first self-encoder AE is completed 1 And a second self-encoder AE 2 Is a serial training of (a);
(4) first self-encoder AE 1 First encoder E1 and second self encoder AE of (a) 2 The second decoder D2 of (a) takes out and constructs a new self-encoder, namely a third self-encoder AE 3 And the third self-encoder AE 3 And the training is used as a preset anomaly detection network model.
S430, inputting at least one power grid time sequence sub-sequence of the power grid time sequence data to be detected into a preset anomaly detection network model to obtain a reconstruction result corresponding to the power grid time sequence sub-sequence.
In the embodiment of the invention, the power grid time sequence data to be detected can be obtained by calling the service interface of the real-time measuring center of the power grid resource business center, and the power grid time sequence data is divided into m pieces of electricity by adopting the data preprocessing method which is the same as the model training stageNetwork timing subsequence p i (1.ltoreq.i.ltoreq.m); sub-sequence p of power grid time sequence i Sequentially inputting the obtained information into a trained preset anomaly detection network model to obtain a corresponding reconstruction result q i (1≤i≤m)。
S440, determining an abnormality score according to the power grid time sequence subsequence and the corresponding reconstruction result, and determining an abnormality detection result corresponding to the power grid time sequence subsequence according to the abnormality score.
In embodiments of the invention, the anomaly score may be used to characterize the grid timing subsequence p i And the corresponding reconstruction result q i Reconstruction errors of (a), i.e. grid timing sub-sequence p i Corresponding anomaly score (p i )＝║q i -p i ║ 2 The method comprises the steps of carrying out a first treatment on the surface of the Further judging the abnormality score (p) i ) Whether the power grid time sequence sub-sequence p is larger than a preset abnormal threshold value or not, if yes, determining the power grid time sequence sub-sequence p i And if the corresponding abnormality detection result is abnormal, otherwise, determining that the abnormality detection result is not abnormal.
Further, on the basis of the embodiment of the invention, after the abnormality detection result corresponding to the power grid time sequence sub-sequence is determined, the abnormality detection result is the corresponding power grid time sequence sub-sequence with abnormality and the reconstruction result are visually displayed, so that a worker can quickly locate a time point or a time interval when the power grid time sequence data is abnormal.
According to the technical scheme, the power grid time sequence training sub-sequence set after the preprocessing of the power grid time sequence training data is determined by acquiring the power grid time sequence training data; constructing a first self-encoder and a second self-encoder, and performing serial training on the first self-encoder and the second self-encoder by utilizing a power grid time sequence training subsequence set, so as to generate a trained preset anomaly detection network model; inputting at least one power grid time sequence subsequence of power grid time sequence data to be detected into a preset anomaly detection network model to obtain a reconstruction result corresponding to the power grid time sequence subsequence; and determining an abnormality score according to the power grid time sequence subsequence and the corresponding reconstruction result, and determining an abnormality detection result corresponding to the power grid time sequence subsequence according to the abnormality score. According to the embodiment of the invention, the problem of low accuracy of power grid time sequence data abnormality detection is solved by introducing the time sequence data abnormality detection technology based on the serial self-encoder into the power grid time sequence data abnormality early warning of the service center station, and the accurate and efficient power grid time sequence data abnormality detection method is realized.
Example IV
Fig. 5 is a schematic structural diagram of a power grid time-series data anomaly detection device according to a fourth embodiment of the present invention. As shown in fig. 5, the apparatus includes:
the data acquisition module 51 is configured to acquire power grid time sequence data to be detected, and determine at least one power grid time sequence sub-sequence of the power grid time sequence data.
The reconstruction result determining module 52 is configured to determine a reconstruction result of the power grid time sequence subsequence based on a preset anomaly detection network model, where the preset anomaly detection network model is obtained after serial training of at least two self-encoders.
The anomaly detection result determining module 53 is configured to determine an anomaly detection result of the power grid time sequence subsequence according to an anomaly score of the reconstruction result.
According to the technical scheme, the data acquisition module is used for acquiring the power grid time sequence data to be detected, and determining at least one power grid time sequence subsequence of the power grid time sequence data; the reconstruction result determining module determines a reconstruction result of the power grid time sequence subsequence based on a preset anomaly detection network model, wherein the preset anomaly detection network model is obtained after serial training of at least two self-encoders; and the abnormality detection result determining module determines an abnormality detection result of the power grid time sequence subsequence according to the abnormality score of the reconstruction result. According to the embodiment of the invention, the acquired power grid time sequence subsequence of the power grid time sequence data is input into the preset abnormality detection network model to obtain the corresponding reconstruction result, so that the abnormality detection result corresponding to the power grid time sequence subsequence is determined according to the abnormality score of the reconstruction result, the problem of low accuracy of power grid time sequence data abnormality detection is solved, and an accurate and efficient power grid time sequence data abnormality detection method is realized.
Further, on the basis of the above embodiment of the invention, the data acquisition module 51 includes:
the data acquisition unit is used for calling a service interface of the real-time measurement center of the power grid resource business center to acquire power grid time sequence data.
The data segmentation unit is used for calling a sliding window method with a preset time window length to segment the power grid time sequence data so as to generate a power grid time sequence subsequence.
Further, on the basis of the above embodiment of the invention, the reconstruction result determination module 52 includes:
the reconstruction result determining unit is used for inputting the power grid time sequence subsequence into the trained preset anomaly detection network model, mapping the power grid time sequence subsequence to the hidden layer by using an encoder of the preset anomaly detection network model to obtain coding features corresponding to the power grid time sequence subsequence, and decoding and restoring the coding features by using a decoder of the preset anomaly detection network model to obtain a reconstruction result.
Further, on the basis of the above embodiment of the invention, the abnormality detection result determination module 53 includes:
and the anomaly score determining unit is used for taking the L2 norm of the difference value between the reconstruction result and the corresponding element of the power grid time sequence sub-sequence as an anomaly score.
The abnormality detection result determining unit is used for judging whether the abnormality score is larger than a preset abnormality threshold value, if yes, determining that the abnormality detection result of the power grid time sequence subsequence is abnormal, and if not, determining that the abnormality detection result is not abnormal.
Further, on the basis of the above embodiment of the present invention, a training process of the anomaly detection network model is preset, including:
acquiring power grid time sequence training data, and calling a sliding window method with a preset time window length to determine a power grid time sequence training sub-sequence set of the power grid time sequence training data;
constructing a first self-encoder and a second self-encoder, wherein the first self-encoder is composed of a first encoder of three full-connection layers and a first decoder of three full-connection layers, and the second self-encoder is composed of a second encoder of three full-connection layers and a second decoder of three full-connection layers;
taking the power grid time sequence training subsequence set as the input of a first self-encoder, taking a first reconstruction result output by the first self-encoder as the input of a second self-encoder, taking an output result of the second self-encoder as a second reconstruction result, and calling a preset gradient descent method to update corresponding weight parameters and bias parameters in the first encoder, the first decoder, the second encoder and the second decoder based on a first loss function corresponding to the first self-encoder and a second loss function corresponding to the second self-encoder until the first loss function and the second loss function are converged;
And constructing a third self-encoder by using the trained first encoder and the trained second decoder, and taking the third self-encoder as a trained preset anomaly detection network model.
Further, on the basis of the above embodiment of the present invention, the first loss function includes at least: l (L) AE1 ＝min║Y-X║ 2 Wherein Y represents a first reconstruction result, and X represents a power grid time sequence training subsequence set;
the second loss function includes at least: l (L) AE2 ＝min║Z-X║ 2 Wherein Z represents the second reconstruction result.
Further, on the basis of the embodiment of the invention, the power grid time sequence data abnormality detection device further includes:
and the data display module is used for visually displaying the abnormal detection result which is the corresponding power grid time sequence subsequence with the abnormality and the reconstruction result.
The power grid time sequence data abnormality detection device provided by the embodiment of the invention can execute the power grid time sequence data abnormality detection method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.
Example five
Fig. 6 shows a schematic diagram of an electronic device 60 that may be used to implement an embodiment of the invention. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. Electronic equipment may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the inventions described and/or claimed herein.
As shown in fig. 6, the electronic device 60 includes at least one processor 61, and a memory, such as a Read Only Memory (ROM) 62, a Random Access Memory (RAM) 63, etc., communicatively connected to the at least one processor 61, in which the memory stores a computer program executable by the at least one processor, and the processor 61 may perform various appropriate actions and processes according to the computer program stored in the Read Only Memory (ROM) 62 or the computer program loaded from the storage unit 68 into the Random Access Memory (RAM) 63. In the RAM 63, various programs and data required for the operation of the electronic device 60 may also be stored. The processor 61, the ROM 62 and the RAM 63 are connected to each other via a bus 64. An input/output (I/O) interface 65 is also connected to bus 64.
Various components in the electronic device 60 are connected to the I/O interface 65, including: an input unit 66 such as a keyboard, a mouse, etc.; an output unit 67 such as various types of displays, speakers, and the like; a storage unit 68 such as a magnetic disk, an optical disk, or the like; and a communication unit 69 such as a network card, modem, wireless communication transceiver, etc. The communication unit 69 allows the electronic device 60 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunication networks.
Processor 61 can be a variety of general-purpose and/or special-purpose processing components having processing and computing capabilities. Some examples of processor 61 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various processors running machine learning model algorithms, digital Signal Processors (DSPs), and any suitable processor, controller, microcontroller, etc. The processor 61 performs the respective methods and processes described above, such as the grid time series data anomaly detection method.
In some embodiments, the grid time series data anomaly detection method may be implemented as a computer program tangibly embodied on a computer readable storage medium, such as storage unit 68. In some embodiments, part or all of the computer program may be loaded and/or installed onto the electronic device 60 via the ROM 62 and/or the communication unit 69. When the computer program is loaded into RAM 63 and executed by processor 61, one or more steps of the grid time series data anomaly detection method described above may be performed. Alternatively, in other embodiments, the processor 61 may be configured to perform the grid time series data anomaly detection method by any other suitable means (e.g., by means of firmware).
Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.
A computer program for carrying out methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions/acts specified in the flowchart and/or block diagram block or blocks to be implemented. The computer program may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.
In the context of the present invention, a computer-readable storage medium may be a tangible medium that can contain, or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, the computer readable storage medium may be a machine readable signal medium. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, 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 disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) through which a user can provide input to the electronic device. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), blockchain networks, and the internet.
The computing system may include clients and servers. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical hosts and VPS service are overcome.
It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps described in the present invention may be performed in parallel, sequentially, or in a different order, so long as the desired results of the technical solution of the present invention are achieved, and the present invention is not limited herein.
The above embodiments do not limit the scope of the present invention. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of the present invention.
Claims (10)
1. A method for detecting anomalies in time-series data of a power grid, the method comprising:
acquiring power grid time sequence data to be detected, and determining at least one power grid time sequence subsequence of the power grid time sequence data;
determining a reconstruction result of the power grid time sequence subsequence based on a preset anomaly detection network model, wherein the preset anomaly detection network model is obtained after serial training of at least two self-encoders;
And determining an abnormality detection result of the power grid time sequence subsequence according to the abnormality score of the reconstruction result.
2. The method of claim 1, wherein the acquiring grid-timing data to be detected and determining at least one grid-timing subsequence of the grid-timing data comprises:
calling a service interface of a real-time measurement center of a power grid resource business center to acquire the power grid time sequence data;
and calling a sliding window method with a preset time window length to divide the power grid time sequence data so as to generate the power grid time sequence subsequence.
3. The method of claim 1, wherein the determining the reconstruction result of the grid timing subsequence based on a preset anomaly detection network model comprises:
inputting the power grid time sequence subsequence into the trained preset anomaly detection network model, mapping the power grid time sequence subsequence to a hidden layer by using an encoder of the preset anomaly detection network model to obtain coding features corresponding to the power grid time sequence subsequence, and decoding and restoring the coding features by using a decoder of the preset anomaly detection network model to obtain the reconstruction result.
4. The method according to claim 1, wherein determining the anomaly detection result of the grid-timing subsequence according to the anomaly score of the reconstruction result comprises:
taking the L2 norm of the difference value between the reconstruction result and the corresponding element of the power grid time sequence sub-sequence as the anomaly score;
judging whether the abnormality score is larger than a preset abnormality threshold, if so, determining that the abnormality detection result of the power grid time sequence subsequence is abnormal, and if not, determining that the abnormality detection result is not abnormal.
5. The method of claim 1, wherein the training process of the preset anomaly detection network model comprises:
acquiring power grid time sequence training data, and calling a sliding window method with a preset time window length to determine a power grid time sequence training subsequence set of the power grid time sequence training data;
constructing a first self-encoder and a second self-encoder, wherein the first self-encoder is composed of a first encoder of three full-connection layers and a first decoder of three full-connection layers, and the second self-encoder is composed of a second encoder of three full-connection layers and a second decoder of three full-connection layers;
Taking the power grid time sequence training subsequence set as the input of the first self-encoder, taking a first reconstruction result output by the first self-encoder as the input of the second self-encoder, taking an output result of the second self-encoder as a second reconstruction result, and calling a preset gradient descent method to update corresponding weight parameters and bias parameters in the first encoder, the first decoder, the second encoder and the second decoder until the first loss function and the second loss function converge based on a first loss function corresponding to the first self-encoder and a second loss function corresponding to the second self-encoder;
and constructing a third self-encoder by using the trained first encoder and the trained second decoder, and taking the third self-encoder as the trained preset anomaly detection network model.
6. The method of claim 5, wherein the first loss function comprises at least: l (L) AE1 ＝min║Y-X║ 2 Wherein Y represents the first weightConstructing a result, wherein X represents the power grid time sequence training subsequence set;
the second loss function includes at least: l (L) AE2 ＝min║Z-X║ 2 Wherein Z represents the second reconstruction result.
7. The method as recited in claim 1, further comprising:
and visually displaying the power grid time sequence subsequence corresponding to the abnormality detection result as the abnormal power grid time sequence subsequence and the reconstruction result.
8. A power grid time series data anomaly detection device, the device comprising:
the data acquisition module is used for acquiring the power grid time sequence data to be detected and determining at least one power grid time sequence subsequence of the power grid time sequence data;
the reconstruction result determining module is used for determining a reconstruction result of the power grid time sequence subsequence based on a preset anomaly detection network model, wherein the preset anomaly detection network model is obtained after serial training of at least two self-encoders;
and the abnormality detection result determining module is used for determining an abnormality detection result of the power grid time sequence subsequence according to the abnormality score of the reconstruction result.
9. An electronic device, the electronic device comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,
the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the grid time series data anomaly detection method of any one of claims 1-7.
10. A computer readable storage medium storing computer instructions for causing a processor to implement the grid time series data anomaly detection method of any one of claims 1-7 when executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311101772.XA CN117251809A (en) | 2023-08-29 | 2023-08-29 | Power grid time sequence data anomaly detection method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311101772.XA CN117251809A (en) | 2023-08-29 | 2023-08-29 | Power grid time sequence data anomaly detection method, device, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117251809A true CN117251809A (en) | 2023-12-19 |
Family
ID=89135967
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311101772.XA Pending CN117251809A (en) | 2023-08-29 | 2023-08-29 | Power grid time sequence data anomaly detection method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117251809A (en) |
-
2023
- 2023-08-29 CN CN202311101772.XA patent/CN117251809A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112415331B (en) | Power grid secondary system fault diagnosis method based on multi-source fault information | |
| CN116049146B (en) | Database fault processing method, device, equipment and storage medium | |
| CN116021981A (en) | Method, device, equipment and storage medium for predicting ice coating faults of power distribution network line | |
| CN112528159B (en) | Feature quality assessment method and device, electronic equipment and storage medium | |
| CN113642727A (en) | Training method of neural network model and processing method and device of multimedia information | |
| CN117332898A (en) | New energy small time scale power time sequence rolling prediction method based on machine learning | |
| CN117332896A (en) | New energy small time scale power prediction method and system for multilayer integrated learning | |
| CN117034149A (en) | Fault processing strategy determining method and device, electronic equipment and storage medium | |
| CN117009903A (en) | Data anomaly detection method, device, equipment and storage medium | |
| CN115392715A (en) | Power utilization data risk assessment method, device, equipment and storage medium | |
| CN117251809A (en) | Power grid time sequence data anomaly detection method, device, equipment and storage medium | |
| CN113408808A (en) | Training method, data generation method, device, electronic device and storage medium | |
| CN117591983B (en) | Multi-index anomaly detection method and device, electronic equipment and storage medium | |
| CN117131315B (en) | Out-of-tolerance electric energy meter determining method and medium based on solving multi-element quadratic function extremum | |
| CN116070138B (en) | State monitoring method, device, equipment and medium for pumped storage unit | |
| CN117370914A (en) | Power grid attack detection model test method, device, equipment and storage medium | |
| CN117705178A (en) | Wind power bolt information detection method and device, electronic equipment and storage medium | |
| CN116111562A (en) | Method, device, equipment and storage medium for determining reliability of power distribution network | |
| CN116307159A (en) | Load prediction method and device, electronic equipment and storage medium | |
| CN117707899A (en) | Micro-service abnormality detection method, device, equipment and storage medium | |
| CN116842837A (en) | Transformer fault diagnosis method and device, electronic equipment and storage medium | |
| CN117667587A (en) | Abnormality detection method and device, electronic equipment and storage medium | |
| CN115081703A (en) | Gas quantity prediction method, device, equipment and storage medium | |
| CN117554747A (en) | Power grid fault diagnosis method and device, electronic equipment and storage medium | |
| CN116258167A (en) | Data detection method, device, electronic equipment and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |