CN113255775A - Method and device for identifying abnormal data of power system and intelligent chip - Google Patents
Method and device for identifying abnormal data of power system and intelligent chip Download PDFInfo
- Publication number
- CN113255775A CN113255775A CN202110587768.3A CN202110587768A CN113255775A CN 113255775 A CN113255775 A CN 113255775A CN 202110587768 A CN202110587768 A CN 202110587768A CN 113255775 A CN113255775 A CN 113255775A
- Authority
- CN
- China
- Prior art keywords
- data
- state estimation
- estimation value
- power system
- sample
- 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.)
- Granted
Links
- 230000002159 abnormal effect Effects 0.000 title claims abstract description 65
- 238000000034 method Methods 0.000 title claims abstract description 59
- 238000012549 training Methods 0.000 claims abstract description 29
- 238000013528 artificial neural network Methods 0.000 claims abstract description 17
- 125000004122 cyclic group Chemical group 0.000 claims abstract description 6
- 239000011159 matrix material Substances 0.000 claims description 27
- 238000005259 measurement Methods 0.000 claims description 27
- 238000012545 processing Methods 0.000 claims description 18
- 230000009467 reduction Effects 0.000 claims description 14
- 238000004364 calculation method Methods 0.000 claims description 13
- 230000015654 memory Effects 0.000 claims description 13
- 230000000306 recurrent effect Effects 0.000 claims description 13
- 238000004422 calculation algorithm Methods 0.000 claims description 9
- 238000007621 cluster analysis Methods 0.000 claims description 7
- 238000006243 chemical reaction Methods 0.000 claims description 5
- 238000013507 mapping Methods 0.000 claims description 5
- 230000007787 long-term memory Effects 0.000 claims description 4
- 230000006403 short-term memory Effects 0.000 claims description 4
- 230000008569 process Effects 0.000 description 7
- 230000006870 function Effects 0.000 description 6
- 238000000605 extraction Methods 0.000 description 4
- 239000013598 vector Substances 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 239000006185 dispersion Substances 0.000 description 3
- 238000004590 computer program Methods 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 230000002349 favourable effect Effects 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 238000000513 principal component analysis Methods 0.000 description 2
- 238000012896 Statistical algorithm Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000005206 flow analysis Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Economics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention relates to a method, a device and an intelligent chip for identifying abnormal data of an electric power system, wherein the method comprises the following steps: acquiring data to be tested of a power system and an actual state estimation value corresponding to the data to be tested; acquiring a state estimation value prediction model; the state estimation value prediction model is obtained by training a cyclic neural network by adopting at least one group of sample data groups of the power system, wherein each group of sample data groups comprises sample data and a pre-labeled state estimation value; calling the state estimation value prediction model to output to obtain a prediction state estimation value according to the data to be detected; and identifying abnormal data of the power system based on the actual state estimation value and the predicted state estimation value. The scheme of the invention can improve the accuracy of identifying the abnormal data of the power system.
Description
Technical Field
The invention relates to the technical field of computers, in particular to a method and a device for identifying abnormal data of a power system and an intelligent chip.
Background
Abnormal data existing in the power system can affect the functions of load flow analysis calculation, state estimation and online analysis software, and normal operation of the power system can be affected by the interference of the abnormal data on the decision of a dispatcher.
In the prior art, the abnormal data identification method analyzes the measurement residual error obtained by state estimation and calculation, and in the case of a plurality of abnormal data with strong correlation, residual error pollution and residual error inundation often occur, which affects the accuracy of abnormal data identification.
Therefore, in view of the above disadvantages, it is desirable to provide a method, an apparatus and an intelligent chip for identifying abnormal data of an electric power system.
Disclosure of Invention
The invention aims to solve the technical problem that the accuracy of identifying abnormal data of an electric power system is not high, and provides a method and a device for identifying abnormal data of the electric power system and an intelligent chip aiming at the defects in the prior art.
In order to solve the technical problem, the invention provides a method for identifying abnormal data of an electric power system, which comprises the following steps:
acquiring data to be tested of a power system and an actual state estimation value corresponding to the data to be tested;
acquiring a state estimation value prediction model; the state estimation value prediction model is obtained by training a cyclic neural network by adopting at least one group of sample data groups of the power system, wherein each group of sample data groups comprises sample data and a pre-labeled state estimation value;
calling the state estimation value prediction model to output to obtain a prediction state estimation value according to the data to be detected;
and identifying abnormal data of the power system based on the actual state estimation value and the predicted state estimation value.
In a possible implementation manner, the invoking the state estimation value prediction model to output a predicted state estimation value according to the data to be measured to obtain a predicted state estimation value includes:
extracting data characteristics from the data to be detected; wherein the data characteristics comprise current fluctuation characteristics and/or voltage fluctuation characteristics;
and inputting the data characteristics into the state estimation value prediction model, and outputting to obtain the prediction state estimation value.
In a possible implementation manner, before the state estimation value prediction model, the method further includes:
obtaining a training sample set, wherein the training sample set comprises at least one group of the sample data groups;
for each set of the sample data sets, extracting sample data features from the sample data;
inputting the sample data characteristics into the recurrent neural network, and outputting to obtain a sample prediction state estimation value; wherein the recurrent neural network is a long-term and short-term memory model;
comparing the sample prediction state estimation value with a pre-labeled state estimation value to obtain a calculation loss; wherein the computational loss is indicative of an error between the sample prediction state estimate and a pre-labeled state estimate;
and training by adopting an error back propagation algorithm according to the respective calculation loss of each group of the sample data group to obtain the state estimation value prediction model.
In a possible implementation manner, the invoking the state estimation value prediction model to output a predicted state estimation value according to the data to be measured to obtain a predicted state estimation value includes:
performing dimensionality reduction processing on the data to be detected to obtain low-dimensional data to be detected;
and inputting the low-dimensional data to be measured into the state estimation value prediction model, and outputting to obtain a prediction state estimation value.
In a possible implementation manner, the performing dimension reduction processing on the data to be measured to obtain low-dimensional data to be measured includes:
mapping the data to be detected to a high dimension, and carrying out centralized operation to form a data matrix;
performing linear conversion on the data matrix to obtain a covariance matrix of the data matrix;
and obtaining low-dimensional data to be measured based on the eigenvalue and the eigenvector of the covariance matrix.
In one possible implementation, the identifying abnormal data of the power system based on the actual state estimation value and the predicted state estimation value includes:
obtaining a measurement residual error based on the square error of the actual state estimation value and the prediction state estimation value;
and performing cluster analysis on the measurement residual error, and identifying abnormal data of the power system.
In one possible implementation, the identifying abnormal data of the power system based on the actual state estimation value and the predicted state estimation value includes:
determining an estimation threshold according to the following formula; wherein the formula is:
and if the absolute value of the difference value between the actual state estimation value and the predicted state estimation value is greater than the estimation threshold, determining that abnormal data exists in the data to be detected.
The invention also provides a device for identifying the abnormal data of the power system, which comprises the following components:
the system comprises a first acquisition module, a second acquisition module and a third acquisition module, wherein the first acquisition module is used for acquiring data to be detected of the power system and an actual state estimation value corresponding to the data to be detected;
the second acquisition module is used for acquiring a state estimation value prediction model; the state estimation value prediction model is obtained by training a cyclic neural network by adopting at least one group of sample data groups of the power system, wherein each group of sample data groups comprises sample data and a pre-labeled state estimation value;
the calling module is used for calling the state estimation value prediction model to output to obtain a prediction state estimation value according to the data to be detected;
and the identification module is used for identifying abnormal data of the power system based on the actual state estimation value and the predicted state estimation value.
The invention also provides an intelligent chip, comprising: at least one memory and at least one processor;
the at least one memory to store a machine readable program;
the at least one processor is configured to invoke the machine readable program to perform the method as described above.
The invention also provides a computer readable medium having stored thereon computer instructions which, when executed by a processor, cause the processor to perform the method as described above.
The implementation of the method, the device and the intelligent chip for identifying the abnormal data of the power system has the following beneficial effects:
according to the technical scheme provided by the invention, the data to be measured of the power system is input into a pre-trained state estimation value prediction model, a prediction state estimation value is obtained through output, and the abnormal data of the power system is identified according to the actual state estimation value corresponding to the data to be measured and the prediction state estimation value, wherein the state estimation value prediction model is obtained through training by utilizing massive operation data of the power system. The technical scheme avoids the problems of residual pollution and residual inundation of the traditional identification method, can accurately predict the power system with abnormal data, and improves the accuracy of identifying the abnormal data of the power system; and the operation time for identification by using the pre-trained state estimation value prediction model is short, so that the method is suitable for a power system with large scale and large data volume, and the rapidity of online identification application is greatly improved.
Drawings
Fig. 1 is a flowchart of a method for identifying abnormal data of an electric power system according to an embodiment of the present invention;
fig. 2 is a flowchart of a method for identifying abnormal data of an electric power system according to another embodiment of the present invention;
FIG. 3 is a schematic diagram of a smart chip provided by one embodiment of the present invention;
fig. 4 is a schematic diagram of an apparatus for identifying abnormal data of an electrical power system according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, are within the scope of the present invention.
Fig. 1 shows a flow chart of a method of identification of power system anomaly data according to one embodiment. It is to be appreciated that the method can be performed by any apparatus, device, platform, cluster of devices having computing and processing capabilities.
Referring to fig. 1, the method includes:
step 101: the method comprises the steps of obtaining data to be measured of the power system and an actual state estimation value corresponding to the data to be measured.
In step 101, the Data to be measured is obtained from the measurement values of each device of the power system on the time section to be identified, And may be obtained, for example, by an SCADA (Supervisory Control And Data Acquisition) system, specifically, the measurement Data acquired from each plant station device is affected by the accuracy of the Acquisition instrument, the Data transmission channel, the transmission mode, the transmission delay, And other factors, And measurement errors inevitably exist, And the Data with large errors is called as abnormal Data.
The environment noise makes the ideal motion equation unable to be solved accurately, and the random error of the measuring system makes the measuring vector unable to directly find the true value of the state through the ideal measuring equation. The estimation of the state vector is processed by statistical methods, called state estimation. In some embodiments, the actual State estimate corresponding to the data under test may be determined by a State Estimator (SE).
Step 102: and acquiring a state estimation value prediction model.
In step 102, the state estimation value prediction model is obtained by training the recurrent neural network by using at least one group of sample data sets of the power system, and each group of sample data sets includes sample data and a pre-labeled state estimation value.
The computer device obtains a pre-trained state estimate prediction model. In a possible implementation manner, when the computer device is a terminal, the terminal obtains a trained state estimation value prediction model stored in the terminal, or obtains the trained state estimation value prediction model from a server. In another possible implementation, when the computer device is a server, the server obtains a trained state estimation value prediction model stored in the server.
The state estimation value prediction model is a recurrent neural network model having a function of recognizing sample data. The state estimation value prediction model is a preset mathematical model and comprises model coefficients between the sample data and the state estimation value. The model coefficients may be fixed values, may be values dynamically modified over time, or may be values dynamically modified with the usage scenario.
In some embodiments, the state estimate prediction model is a hybrid model of a Convolutional Neural Network (CNN) and a recurrent Neural Network, or the state estimate prediction model is a recurrent Neural Network model. For example, the recurrent neural network is a Long-Short Term Memory (LSTM) network.
Step 103: and calling a state estimation value prediction model to output to obtain a prediction state estimation value according to the data to be detected.
In some embodiments, step 103 specifically includes the following steps:
extracting data characteristics from the data to be detected; wherein the data characteristics comprise current fluctuation characteristics and/or voltage fluctuation characteristics;
and inputting the data characteristics into a state estimation value prediction model, and outputting to obtain a predicted state estimation value.
In this embodiment, the feature extraction is a process of extracting features from data to be measured and converting the extracted features into structured data, and a prediction state estimation value is obtained by inputting the data features extracted by the features into a state estimation value prediction model and outputting the data features. In addition, since the data to be measured includes voltage measurement data and current measurement data, the data characteristics include a current fluctuation characteristic and/or a voltage fluctuation characteristic.
It should be noted that before the computer device obtains the state estimation value prediction model, it needs to train the training sample set to obtain the state estimation value prediction model. The following describes a training process of the state estimation value prediction model.
In one possible implementation, the training process for the state estimation prediction model includes the following steps:
step A1, the computer device obtains a training sample set, the training sample set includes at least one group of sample data group.
In step a1, the state estimation prediction model is trained from at least one set of sample data sets, where each set of sample data sets includes sample data and a pre-labeled state estimation value.
Step A2, for each group of sample data, the computer device extracts the sample data features from the sample data.
In step a2, the computer device performs feature extraction on the sample data, and determines the data after feature extraction as the sample data features.
Schematically, feature extraction is a process of extracting features from data to be measured and converting the extracted features into structured data.
Step A3, inputting the sample data characteristics into a recurrent neural network by the computer equipment to obtain a training result, wherein the recurrent neural network is a long-term and short-term memory model.
Illustratively, for each group of sample data groups, the computer device creates an input-output pair corresponding to the group of sample data groups, wherein input parameters of the input-output pair are sample data characteristics in the group of sample data groups, and output parameters are pre-labeled state estimation values in the group of sample data groups; and inputting the input parameters into the prediction model by the computer equipment to obtain a training result.
Alternatively, the input-output pairs are represented by feature vectors.
Step A4, the computer equipment compares the sample prediction state estimation value with the pre-labeled state estimation value to obtain the calculation loss, and the calculation loss is used for indicating the error between the sample prediction state estimation value and the pre-labeled state estimation value.
Optionally, the calculation loss is represented by cross entropy, which is not described herein.
And A5, training by the computer equipment according to the respective calculation loss of each group of sample data groups by adopting an error back propagation algorithm to obtain a state estimation value prediction model.
Optionally, the computer device determines a gradient direction of the state estimation value prediction model according to the calculation loss through a back propagation algorithm, and updates the model parameters in the state estimation value prediction model layer by layer from an output layer of the state estimation value prediction model.
In other embodiments, since the data to be detected may be a waveform continuous signal, the efficiency of data processing on the waveform continuous signal is low, which is not favorable for fast identification of abnormal data. Based on this, it can be considered to perform dimensionality reduction processing on the data to be measured.
Specifically, step 103 may include the steps of:
performing dimensionality reduction on the data to be detected to obtain low-dimensional data to be detected;
and inputting the low-dimensional data to be measured into the state estimation value prediction model, and outputting to obtain a prediction state estimation value.
In this embodiment, the data features are preliminarily extracted after the dimension reduction processing is performed on the data to be detected, so that the data features of the retained important information are used as the input of further processing, the processing efficiency of the data can be effectively improved under the background of large-data-volume processing, the training result is improved, and the accuracy of abnormal data identification is improved.
Further, the dimension reduction processing of the data can be performed by:
mapping the data to be detected to a high dimension, and carrying out centralized operation to form a data matrix;
performing linear conversion on the data matrix to obtain a covariance matrix of the data matrix;
and obtaining low-dimensional data to be measured based on the eigenvalue and the eigenvector of the covariance matrix.
In this embodiment, the data to be measured may be subjected to dimensionality reduction processing by a Principal Component Analysis (PCA) algorithm, so as to retain important data features and remove unimportant data features such as noise, thereby improving the processing speed of subsequent data. Each piece of data to be tested is processed. In the processing process, a row of fields is regarded as each data to be detected, and a plurality of data to be detected are simultaneously selected as a processing matrix unit, and the specific dimension reduction mode can be carried out through the following steps:
(1) mapping the data to a high dimension, performing centralization operation, centering m data to be detected, and forming a data matrix X with m rows and n columns by taking each data to be detected as one row and m data to be detected;
(2) performing linear conversion on the data matrix X, and calculating a covariance matrix C of the X matrix;
(3) calculating an eigenvalue and an eigenvector of the covariance matrix C;
(4) low-dimensional projection, namely taking k e (0, n) corresponding characteristic row vectors according to the size of the characteristic value to form a projection matrix P; in the selection of the k value, a cross validation method is adopted, and a lower limit threshold t is set to be 0.96;
(5) and Y is PX, namely the data (namely the low-dimensional data to be measured) from the dimensionality reduction to the dimensionality k.
Step 104: and identifying abnormal data of the power system based on the actual state estimation value and the predicted state estimation value.
In step 104, the following steps may be specifically included:
obtaining a measurement residual error based on the square error of the actual state estimation value and the prediction state estimation value;
and performing cluster analysis on the measurement residual error, and identifying abnormal data of the power system.
In this embodiment, the measured residuals are used as input data sets for the clustering algorithm in the subsequent step.
Specifically, the step of performing cluster analysis on the measurement residual error to identify abnormal data of the power system comprises the following steps:
and step B1, setting the measurement residual errors into a plurality of clusters.
And step B2, obtaining the optimal clustering number according to the set clustering number, the clustering gap value and the dispersion.
In step B2, the optimal cluster number is calculated, for example, using the improved GSA gap statistics algorithm. The GSA gap statistical algorithm is a data mining algorithm for enhancing the clustering effect, determines the optimal clustering number by comparing the dispersion index of the clustering result with the reference value thereof, and accurately distinguishes the clusters of normal data and abnormal data.
And B3, clustering the measurement residual errors based on the optimal clustering number to obtain a clustering result.
Specifically, step B3 includes the following steps:
step B31, according to the optimal clustering number k, optionally clustering k measurement residual data in the data set to be used as an initial clustering center;
step B32, calculating the distance between each measured residual data in the cluster data set and k cluster center data, and selecting the closest one to be classified into the cluster data set;
b33, after the measurement residual data of all the devices in the cluster data set are classified, calculating the average value of the measurement residual data of k clusters, taking the average value as a new cluster center, and repeating the step B32;
and step B34, when the clustering results of two continuous clustering are completely consistent, stopping the circulation of the steps B32 and B33, and determining the final clustering result.
And step B4, determining the quantity and the position of the abnormal data of the power system in the clustering result.
Specifically, the method for determining the number and the position of abnormal data of the power system comprises the following steps:
1) if the optimal clustering number is 1, abnormal data does not exist;
2) if the optimal clustering number is larger than 1, abnormal data exist; and in the clustering result, calculating the average value of the measurement error data in each cluster, wherein the cluster with the minimum average value is normal data, and all the measurements corresponding to the measurement error data in the other clusters are abnormal data.
In step 104, the following steps may be specifically included:
determining an estimation threshold according to the following formula; wherein the formula is:
and if the absolute value of the difference value between the actual state estimation value and the predicted state estimation value is greater than the estimation threshold, determining that abnormal data exists in the data to be detected.
In this embodiment, the estimation threshold is determined by unbiased estimation of the sample standard deviation of the sample data of n times, and the estimation threshold determined in this way is more favorable for determining that abnormal data exists in the data to be measured.
It can be seen that, in the process shown in fig. 1, the data to be measured of the power system is input into the pre-trained state estimation value prediction model, the predicted state estimation value is obtained through output, and then the abnormal data of the power system is identified according to the actual state estimation value corresponding to the data to be measured and the predicted state estimation value, wherein the state estimation value prediction model is obtained through training by using massive operation data of the power system. The technical scheme avoids the problems of residual pollution and residual inundation of the traditional identification method, can accurately predict the power system with abnormal data, and improves the accuracy of identifying the abnormal data of the power system; and the operation time for identification by using the pre-trained state estimation value prediction model is short, so that the method is suitable for a power system with large scale and large data volume, and the rapidity of online identification application is greatly improved.
Fig. 2 shows a flow chart of a method of identification of power system anomaly data according to another embodiment. Referring to fig. 2, the method includes:
step 201: the method comprises the steps of obtaining data to be measured of the power system and an actual state estimation value corresponding to the data to be measured.
Step 202: and acquiring a state estimation value prediction model.
Step 203: and performing dimensionality reduction on the data to be detected to obtain low-dimensional data to be detected.
Step 204: and extracting data characteristics from the low-dimensional data to be detected.
Step 205: and inputting the data characteristics extracted from the low-dimensional data to be measured into a state estimation value prediction model, and outputting to obtain a prediction state estimation value.
Step 206: and obtaining a measurement residual error based on the square error of the actual state estimation value and the prediction state estimation value.
Step 207: and performing cluster analysis on the measurement residual error, and identifying abnormal data of the power system.
As shown in fig. 3 and 4, the embodiment of the present invention provides an intelligent chip and an apparatus for identifying abnormal data of an electric power system. The device embodiments may be implemented by software, or by hardware, or by a combination of hardware and software. From a hardware level, as shown in fig. 3, a hardware structure diagram of an intelligent chip provided in the embodiment of the present invention is that, in addition to the processor, the memory, the network interface, and the nonvolatile memory shown in fig. 3, a device in which the apparatus is located in the embodiment may also generally include other hardware, such as a forwarding chip responsible for processing a packet, and the like. Taking a software implementation as an example, as shown in fig. 4, as a logical apparatus, the apparatus is formed by reading a corresponding computer program instruction in a non-volatile memory into a memory by a CPU of a device in which the apparatus is located and running the computer program instruction.
As shown in fig. 4, the apparatus for identifying abnormal data in an electrical power system according to this embodiment includes:
a first obtaining module 401, configured to obtain data to be measured of an electric power system and an actual state estimation value corresponding to the data to be measured;
a second obtaining module 402, configured to obtain a state estimation value prediction model; the state estimation value prediction model is obtained by training a cyclic neural network by adopting at least one group of sample data groups of the power system, wherein each group of sample data groups comprises sample data and a pre-labeled state estimation value;
a calling module 403, configured to call the state estimation value prediction model to output a predicted state estimation value according to the data to be measured;
an identifying module 404 configured to identify abnormal data of the power system based on the actual state estimation value and the predicted state estimation value.
In the embodiment of the present invention, the first obtaining module 401 may be configured to execute step 101 in the above-described method embodiment, the second obtaining module 402 may be configured to execute step 102 in the above-described method embodiment, the invoking module 403 may be configured to execute step 103 in the above-described method embodiment, and the identifying module 404 may be configured to execute step 104 in the above-described method embodiment.
In an embodiment of the present invention, the calling module 403 is configured to perform the following operations:
extracting data characteristics from the data to be detected; wherein the data characteristics comprise current fluctuation characteristics and/or voltage fluctuation characteristics;
and inputting the data characteristics into the state estimation value prediction model, and outputting to obtain the prediction state estimation value.
In one embodiment of the present invention, further comprising: a training module;
the training module is used for executing the following operations:
obtaining a training sample set, wherein the training sample set comprises at least one group of the sample data groups;
for each set of the sample data sets, extracting sample data features from the sample data;
inputting the sample data characteristics into the recurrent neural network, and outputting to obtain a sample prediction state estimation value; wherein the recurrent neural network is a long-term and short-term memory model;
comparing the sample prediction state estimation value with a pre-labeled state estimation value to obtain a calculation loss; wherein the computational loss is indicative of an error between the sample prediction state estimate and a pre-labeled state estimate;
and training by adopting an error back propagation algorithm according to the respective calculation loss of each group of the sample data group to obtain the state estimation value prediction model.
In an embodiment of the present invention, the calling module 403 is configured to perform the following operations:
performing dimensionality reduction processing on the data to be detected to obtain low-dimensional data to be detected;
and inputting the low-dimensional data to be measured into the state estimation value prediction model, and outputting to obtain a prediction state estimation value.
In an embodiment of the present invention, when performing dimension reduction processing on the data to be tested to obtain low-dimensional data to be tested, the calling module 403 is configured to perform the following operations:
mapping the data to be detected to a high dimension, and carrying out centralized operation to form a data matrix;
performing linear conversion on the data matrix to obtain a covariance matrix of the data matrix;
and obtaining low-dimensional data to be measured based on the eigenvalue and the eigenvector of the covariance matrix.
In an embodiment of the present invention, the identifying module 404 is configured to perform the following operations:
obtaining a measurement residual error based on the square error of the actual state estimation value and the prediction state estimation value;
and performing cluster analysis on the measurement residual error, and identifying abnormal data of the power system.
In an embodiment of the present invention, the identifying module 404, when performing the cluster analysis on the measured residuals and identifying abnormal data of the power system, is configured to:
setting the measurement residuals into a plurality of clusters;
obtaining the optimal clustering number according to the set clustering number, the clustering gap value and the dispersion;
clustering the measurement residual errors based on the optimal clustering number to obtain a clustering result;
and determining the quantity and the position of abnormal data of the power system in the clustering result.
In an embodiment of the present invention, the identifying module 404 is configured to perform the following operations:
determining an estimation threshold according to the following formula; wherein the formula is:
and if the absolute value of the difference value between the actual state estimation value and the predicted state estimation value is greater than the estimation threshold, determining that abnormal data exists in the data to be detected.
It is to be understood that the illustrated structure of the embodiment of the present invention does not constitute a specific limitation on the identification device of the abnormal data of the power system. In other embodiments of the invention, the means for identifying power system anomaly data may include more or fewer components than shown, or some components may be combined, some components may be split, or a different arrangement of components. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.
Because the content of information interaction, execution process, and the like among the modules in the device is based on the same concept as the method embodiment of the present invention, specific content can be referred to the description in the method embodiment of the present invention, and is not described herein again.
The embodiment of the invention also provides a device for identifying abnormal data of the power system, which comprises: at least one memory and at least one processor;
at least one memory for storing a machine readable program;
at least one processor for invoking a machine readable program to perform a method for identifying power system anomaly data in any embodiment of the present invention.
Embodiments of the present invention also provide a computer-readable medium storing instructions for causing a computer to perform a method of identifying abnormal data of an electric power system as described herein. Specifically, a method or an apparatus equipped with a storage medium on which a software program code that realizes the functions of any of the above-described embodiments is stored may be provided, and a computer (or a CPU or MPU) of the method or the apparatus is caused to read out and execute the program code stored in the storage medium.
In this case, the program code itself read from the storage medium can realize the functions of any of the above-described embodiments, and thus the program code and the storage medium storing the program code constitute a part of the present invention.
Examples of the storage medium for supplying the program code include a floppy disk, a hard disk, a magneto-optical disk, an optical disk (e.g., CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD + RW), a magnetic tape, a nonvolatile memory card, and a ROM. Alternatively, the program code may be downloaded from a server computer via a communications network.
Further, it should be clear that the functions of any one of the above-described embodiments can be implemented not only by executing the program code read out by the computer, but also by performing a part or all of the actual operations by an operation method or the like operating on the computer based on instructions of the program code.
Further, it is to be understood that the program code read out from the storage medium is written to a memory provided in an expansion board inserted into the computer or to a memory provided in an expansion unit connected to the computer, and then causes a CPU or the like mounted on the expansion board or the expansion unit to perform part or all of the actual operations based on instructions of the program code, thereby realizing the functions of any of the above-described embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments can still be repaired, or some technical features thereof can be equivalently replaced; and such repair or replacement does not depart from the spirit and scope of the corresponding technical solutions.
Claims (10)
1. A method for identifying abnormal data of an electric power system is characterized by comprising the following steps:
acquiring data to be tested of a power system and an actual state estimation value corresponding to the data to be tested;
acquiring a state estimation value prediction model; the state estimation value prediction model is obtained by training a cyclic neural network by adopting at least one group of sample data groups of the power system, wherein each group of sample data groups comprises sample data and a pre-labeled state estimation value;
calling the state estimation value prediction model to output to obtain a prediction state estimation value according to the data to be detected;
and identifying abnormal data of the power system based on the actual state estimation value and the predicted state estimation value.
2. The method of claim 1, wherein the invoking the state estimation prediction model output to obtain a predicted state estimation value according to the data to be tested comprises:
extracting data characteristics from the data to be detected; wherein the data characteristics comprise current fluctuation characteristics and/or voltage fluctuation characteristics;
and inputting the data characteristics into the state estimation value prediction model, and outputting to obtain the prediction state estimation value.
3. The method of claim 1, further comprising, prior to the state estimate prediction model:
obtaining a training sample set, wherein the training sample set comprises at least one group of the sample data groups;
for each set of the sample data sets, extracting sample data features from the sample data;
inputting the sample data characteristics into the recurrent neural network, and outputting to obtain a sample prediction state estimation value; wherein the recurrent neural network is a long-term and short-term memory model;
comparing the sample prediction state estimation value with a pre-labeled state estimation value to obtain a calculation loss; wherein the computational loss is indicative of an error between the sample prediction state estimate and a pre-labeled state estimate;
and training by adopting an error back propagation algorithm according to the respective calculation loss of each group of the sample data group to obtain the state estimation value prediction model.
4. The method according to any one of claims 1-3, wherein the invoking the state estimation prediction model output to obtain a predicted state estimation value according to the data to be tested comprises:
performing dimensionality reduction processing on the data to be detected to obtain low-dimensional data to be detected;
and inputting the low-dimensional data to be measured into the state estimation value prediction model, and outputting to obtain a prediction state estimation value.
5. The method according to claim 4, wherein the performing dimension reduction on the data to be measured to obtain low-dimensional data to be measured includes:
mapping the data to be detected to a high dimension, and carrying out centralized operation to form a data matrix;
performing linear conversion on the data matrix to obtain a covariance matrix of the data matrix;
and obtaining low-dimensional data to be measured based on the eigenvalue and the eigenvector of the covariance matrix.
6. The method according to any one of claims 1-3, wherein the identifying abnormal data of the power system based on the actual state estimate and the predicted state estimate comprises:
obtaining a measurement residual error based on the square error of the actual state estimation value and the prediction state estimation value;
and performing cluster analysis on the measurement residual error, and identifying abnormal data of the power system.
7. The method according to any one of claims 1-3, wherein the identifying abnormal data of the power system based on the actual state estimate and the predicted state estimate comprises:
determining an estimation threshold according to the following formula; wherein the formula is:
and if the absolute value of the difference value between the actual state estimation value and the predicted state estimation value is greater than the estimation threshold, determining that abnormal data exists in the data to be detected.
8. An apparatus for recognizing abnormal data of an electric power system, comprising:
the system comprises a first acquisition module, a second acquisition module and a third acquisition module, wherein the first acquisition module is used for acquiring data to be detected of the power system and an actual state estimation value corresponding to the data to be detected;
the second acquisition module is used for acquiring a state estimation value prediction model; the state estimation value prediction model is obtained by training a cyclic neural network by adopting at least one group of sample data groups of the power system, wherein each group of sample data groups comprises sample data and a pre-labeled state estimation value;
the calling module is used for calling the state estimation value prediction model to output to obtain a prediction state estimation value according to the data to be detected;
and the identification module is used for identifying abnormal data of the power system based on the actual state estimation value and the predicted state estimation value.
9. A smart chip, comprising: at least one memory and at least one processor;
the at least one memory to store a machine readable program;
the at least one processor, configured to invoke the machine readable program to perform the method of any of claims 1-7.
10. A computer readable medium having stored thereon computer instructions which, when executed by a processor, cause the processor to perform the method of any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110587768.3A CN113255775B (en) | 2021-05-28 | 2021-05-28 | Method and device for identifying abnormal data of power system and intelligent chip |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110587768.3A CN113255775B (en) | 2021-05-28 | 2021-05-28 | Method and device for identifying abnormal data of power system and intelligent chip |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113255775A true CN113255775A (en) | 2021-08-13 |
| CN113255775B CN113255775B (en) | 2021-09-24 |
Family
ID=77185013
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110587768.3A Active CN113255775B (en) | 2021-05-28 | 2021-05-28 | Method and device for identifying abnormal data of power system and intelligent chip |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113255775B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114580467A (en) * | 2022-02-22 | 2022-06-03 | 国网山东省电力公司信息通信公司 | Power data anomaly detection method and system based on data enhancement and Tri-tracing |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101893430A (en) * | 2010-07-20 | 2010-11-24 | 哈尔滨工业大学 | Processing method of abnormal measured values based on CNC gear measuring center |
| CN108629371A (en) * | 2018-05-02 | 2018-10-09 | 电子科技大学 | A kind of Method of Data with Adding Windows to two-dimentional time-frequency data |
| CN108900546A (en) * | 2018-08-13 | 2018-11-27 | 杭州安恒信息技术股份有限公司 | The method and apparatus of time series Network anomaly detection based on LSTM |
| CN110298369A (en) * | 2018-03-21 | 2019-10-01 | 中国电力科学研究院有限公司 | A kind of discrimination method and system of electric system bad data |
| US20200104224A1 (en) * | 2018-09-27 | 2020-04-02 | Kabushiki Kaisha Toshiba | Anomaly detection device, anomaly detection method and storage medium |
-
2021
- 2021-05-28 CN CN202110587768.3A patent/CN113255775B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101893430A (en) * | 2010-07-20 | 2010-11-24 | 哈尔滨工业大学 | Processing method of abnormal measured values based on CNC gear measuring center |
| CN110298369A (en) * | 2018-03-21 | 2019-10-01 | 中国电力科学研究院有限公司 | A kind of discrimination method and system of electric system bad data |
| CN108629371A (en) * | 2018-05-02 | 2018-10-09 | 电子科技大学 | A kind of Method of Data with Adding Windows to two-dimentional time-frequency data |
| CN108900546A (en) * | 2018-08-13 | 2018-11-27 | 杭州安恒信息技术股份有限公司 | The method and apparatus of time series Network anomaly detection based on LSTM |
| US20200104224A1 (en) * | 2018-09-27 | 2020-04-02 | Kabushiki Kaisha Toshiba | Anomaly detection device, anomaly detection method and storage medium |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114580467A (en) * | 2022-02-22 | 2022-06-03 | 国网山东省电力公司信息通信公司 | Power data anomaly detection method and system based on data enhancement and Tri-tracing |
| CN114580467B (en) * | 2022-02-22 | 2023-11-17 | 国网山东省电力公司信息通信公司 | Power data anomaly detection method and system based on data enhancement and Tri-Training |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113255775B (en) | 2021-09-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112731260B (en) | Online evaluation method for error state of voltage transformer based on concept drift recognition | |
| CN112101765A (en) | Abnormal data processing method and system for operation index data of power distribution network | |
| CN113255775B (en) | Method and device for identifying abnormal data of power system and intelligent chip | |
| CN114386537A (en) | Lithium battery fault diagnosis method and device based on Catboost and electronic equipment | |
| CN111768034A (en) | Method for interpolating and supplementing missing value based on neighbor algorithm in power load prediction | |
| CN113484817A (en) | Intelligent electric energy meter automatic verification system abnormity detection method based on TSVM model | |
| CN111104241A (en) | Server memory anomaly detection method, system and equipment based on self-encoder | |
| CN116070302A (en) | Cable insulation state prediction method and device | |
| CN115526258A (en) | Power system transient stability evaluation method based on Spearman correlation coefficient feature extraction | |
| CN107274025B (en) | System and method for realizing intelligent identification and management of power consumption mode | |
| CN112949201A (en) | Wind speed prediction method and device, electronic equipment and storage medium | |
| CN117407665A (en) | Retired battery time sequence data missing value filling method based on generation countermeasure network | |
| CN112734201A (en) | Multi-equipment overall quality evaluation method based on expected failure probability | |
| CN116957534A (en) | Method for predicting replacement number of intelligent electric meter | |
| CN116630989A (en) | Visual fault detection method and system for intelligent ammeter, electronic equipment and storage medium | |
| Ke et al. | A model for degradation prediction with change point based on Wiener process | |
| CN116400266A (en) | Transformer fault detection method, device and medium based on digital twin model | |
| CN115239971A (en) | GIS partial discharge type recognition model training method, recognition method and system | |
| CN110334125A (en) | A kind of power distribution network measurement anomalous data identification method and device | |
| CN116079498A (en) | Method for identifying abnormal signals of cutter | |
| CN112416709A (en) | Chip dynamic power consumption estimation method and device, processor chip and server | |
| CN116796213B (en) | Power distribution network line transformation relation identification method based on clustering algorithm | |
| CN117494588B (en) | Method, equipment and medium for optimizing residual effective life of fan bearing | |
| CN113239236B (en) | Video processing method and device, electronic equipment and storage medium | |
| CN117277422B (en) | Method, system, terminal and medium for evaluating stability of direct-drive wind farm |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20221215 Address after: 807-3, floor 8, block F, No. 9, Shangdi Third Street, Haidian District, Beijing 100080 Patentee after: Zhongcheng Hualong Computer Technology Co.,Ltd. Address before: No.114, 14th floor, block B, building 1, No.38, Zhongguancun Street, Haidian District, Beijing 100082 Patentee before: Shenwei Super Computing (Beijing) Technology Co.,Ltd. |
|
| TR01 | Transfer of patent right |