CN116539994A - Substation main equipment operation state detection method based on multi-source time sequence data - Google Patents
Substation main equipment operation state detection method based on multi-source time sequence data Download PDFInfo
- Publication number
- CN116539994A CN116539994A CN202310520060.5A CN202310520060A CN116539994A CN 116539994 A CN116539994 A CN 116539994A CN 202310520060 A CN202310520060 A CN 202310520060A CN 116539994 A CN116539994 A CN 116539994A
- Authority
- CN
- China
- Prior art keywords
- density
- target
- indexes
- data
- running state
- 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 30
- 238000000034 method Methods 0.000 claims abstract description 31
- 230000002159 abnormal effect Effects 0.000 claims abstract description 30
- 238000013528 artificial neural network Methods 0.000 claims abstract description 21
- 238000012216 screening Methods 0.000 claims abstract description 9
- 230000015654 memory Effects 0.000 claims description 27
- 235000014435 Mentha Nutrition 0.000 claims description 18
- 241001072983 Mentha Species 0.000 claims description 18
- 235000014569 mints Nutrition 0.000 claims description 18
- 238000012549 training Methods 0.000 claims description 12
- 230000008569 process Effects 0.000 claims description 9
- 230000002596 correlated effect Effects 0.000 claims description 8
- 230000004927 fusion Effects 0.000 claims description 8
- 230000005856 abnormality Effects 0.000 claims description 7
- 238000012300 Sequence Analysis Methods 0.000 claims description 4
- 238000004891 communication Methods 0.000 claims description 4
- 230000006870 function Effects 0.000 description 7
- 238000004364 calculation method Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 230000004913 activation Effects 0.000 description 3
- 230000000875 corresponding effect Effects 0.000 description 3
- 239000013598 vector Substances 0.000 description 3
- 101001095088 Homo sapiens Melanoma antigen preferentially expressed in tumors Proteins 0.000 description 2
- 102100037020 Melanoma antigen preferentially expressed in tumors Human genes 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000001174 ascending effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
-
- 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
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- 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
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of power grid operation prediction, and discloses a method for detecting the operation state of main equipment of a transformer substation based on multi-source time sequence data, which comprises the steps of firstly, obtaining a plurality of time sequence characteristics related to the abnormal operation height of the equipment of the transformer substation as target indexes according to the correlation among a plurality of state indexes contained in the operation state information of the equipment of the transformer substation; taking a plurality of target indexes as input and taking a single target index as output to train the LSTM neural network to obtain a prediction model; sequentially inputting a plurality of historical target indexes into a prediction model according to a preset time step to obtain target indexes of a plurality of future time sequences of the running state of the transformer substation; and clustering and screening out abnormal data by using a spatial clustering algorithm based on noise of density to perform detection of the running state of the main equipment of the transformer substation based on the multi-source time sequence data, and early warning that the running state of the equipment possibly appears abnormal in the future.
Description
Technical Field
The invention relates to the field of power grid running state prediction, in particular to a substation main equipment running state detection method based on multi-source time sequence data.
Background
With the rapid development of modern science and technology, the number and types of information systems used in smart grids are gradually increased, and complicated power services make early warning of abnormal operation states of smart grid devices more difficult. And the intelligent fusion terminal is utilized to count and analyze the running state information of the main flow equipment in the intelligent substation, and early warning is given when the equipment is abnormal, so that the intelligent fusion terminal has important significance for ensuring the stable running of the intelligent power grid.
The existing anomaly detection method mainly directly sets the operation parameter range of main stream equipment of an intelligent substation, and cannot meet the data anomaly detection of multi-index complex power service (such as transformer control service). Meanwhile, the existing intelligent substation equipment abnormality detection method is slow in response. Most of the current methods only can find the abnormality under the condition that the power grid abnormality has occurred, and cannot effectively pre-judge and pre-process the abnormal condition and possible abnormality of the transformer substation equipment.
Disclosure of Invention
In view of the above, the invention provides a method for detecting the running state of a transformer substation main device based on multi-source time sequence data, so as to solve the problem that the abnormal condition and possible abnormality of the transformer substation device cannot be effectively prejudged and processed in advance in the prior art.
In a first aspect, the present invention provides a method for detecting an operation state of a substation main device based on multi-source time series data, including:
collecting operation state information of substation equipment, analyzing correlations among a plurality of state indexes contained in the operation state information by utilizing a spearman correlation coefficient, and acquiring a plurality of time sequence characteristics which are highly correlated with the abnormal operation of the substation equipment as target indexes;
taking a plurality of target indexes with time sequences as input and taking a single target index as output, training the LSTM neural network, and obtaining a prediction model of taking the trained LSTM neural network as the single target index of the running state of the substation equipment;
sequentially inputting a plurality of historical target indexes into the prediction model according to a preset time step to obtain target indexes of a plurality of future time sequences of the running state of the transformer substation;
and clustering and screening out abnormal data by using a spatial clustering algorithm based on the noise of density to perform clustering on target indexes of the current time sequence and the future time sequence of the running state of the substation equipment, thereby completing the running state detection of the substation main equipment based on the multi-source time sequence data.
In an alternative embodiment, the spearman correlation coefficient between the plurality of state indexes is ranked a preset number or more as the index having time series characteristics highly correlated with the abnormal operation as the target index.
In an alternative embodiment, the plurality of target indicators with time series characteristics includes: the power index and other indexes of which the spearman correlation coefficient with the power index is larger than a preset threshold value.
In an alternative embodiment, the other indicators having a higher correlation with the power indicator include: voltage index, current index and impedance index.
In an optional implementation manner, the process of sequentially inputting a plurality of historical target indexes into the prediction model according to a preset time step to obtain target indexes of a plurality of future time sequences of the operation state of the transformer substation includes:
inputting a plurality of target indexes of a preset historical day into the prediction model to obtain a single target index of the running state of the transformer substation in a future day;
adding the target index obtained through prediction to preset historical data, and updating a plurality of target indexes of the preset historical days of the next step length to be input into the prediction model to obtain a single target index of the running state of the transformer substation of the next future day;
and in this cycle, obtaining target indexes of a plurality of future time sequences of the running state of the transformer substation.
In an optional implementation manner, the process of clustering and screening the abnormal data by using the density-based noise to apply a spatial clustering algorithm to target indexes of the current time sequence and the future time sequence of the running state of the substation equipment includes:
calculating the number of data contained in a circle with Eps as a radius by taking each target index as a data point and each data point as a circle center as a point density value;
selecting a density threshold value Mints, wherein the central points with the number of points smaller than the Mints in the circle are low-density points, recording the central points larger than or equal to the Mints as high-density points, and if one high-density point exists in the circle of the other high-density point, connecting the two high-density points;
the data points are connected continuously, if there are low density points in the high density circle, then connect to the nearest high density point as boundary point, all connected points form one cluster, the low density points in no cluster are marked as outliers.
In an alternative embodiment, the method further comprises: based on the detected abnormal data, early warning is provided for possible abnormal conditions in the future.
In a second aspect, the present invention provides a substation main device operation state detection device based on multi-source time series data, the device comprising:
the system comprises a plurality of target index acquisition modules, a plurality of time sequence analysis modules and a plurality of time sequence analysis modules, wherein the target index acquisition modules are used for acquiring the running state information of the substation equipment, analyzing the correlation among a plurality of state indexes contained in the running state information by utilizing the spearman correlation coefficient, and acquiring a plurality of time sequence characteristics which are highly correlated with the running abnormality of the substation equipment as target indexes;
the prediction model training module is used for training the LSTM neural network by taking a plurality of target indexes with time sequences as input and taking a single target index as output to obtain a trained prediction model with the LSTM neural network as a single target index of the running state of the substation equipment;
the target index prediction module is used for sequentially inputting a plurality of historical target indexes into the prediction model according to a preset time step to obtain target indexes of a plurality of future time sequences of the running state of the transformer substation;
and the anomaly detection module is used for clustering and screening out anomaly data by utilizing a density-based noise application spatial clustering algorithm to target indexes of the current time sequence and the future time sequence of the running state of the substation equipment, and finishing the running state detection of the substation main equipment based on the multi-source time sequence data.
In an alternative embodiment, the target index prediction module includes:
the single target index prediction unit is used for inputting a plurality of target indexes of a preset historical day into the prediction model to obtain a single target index of the running state of the transformer substation in a future day;
the target index prediction units are used for adding target indexes obtained through prediction into preset historical data, updating a plurality of target indexes of the preset historical days of the next step length, inputting the target indexes into the prediction model, and obtaining a single target index of the running state of the transformer substation of the next future day; and in this cycle, obtaining target indexes of a plurality of future time sequences of the running state of the transformer substation.
In an alternative embodiment, the anomaly detection module includes:
a point density value acquisition unit for calculating the number of data contained in a circle with Eps as a radius as a point density value with each target index as a data point and each data point as a circle center;
a high-density point obtaining unit, configured to select a density threshold value mints, wherein a center point with a number of points in a circle smaller than the mints is a low-density point, a center point with a number greater than or equal to the mints is recorded as a high-density point, and if one high-density point exists in a circle with another high-density point, the two high-density points are connected;
an outlier acquisition unit for continuously connecting the data points, connecting to the nearest high-density point as a boundary point if there is a low-density point in the high-density point circle, all the connected points forming one cluster, the low-density points not in any cluster being marked as outliers.
In a third aspect, the present invention provides an intelligent fusion terminal, including: the system comprises a memory and a processor, wherein the memory and the processor are in communication connection, the memory stores computer instructions, and the processor executes the computer instructions so as to execute the method for detecting the running state of the substation main equipment based on the multi-source time series data according to the first aspect or any corresponding embodiment of the first aspect.
In a fourth aspect, the present invention provides a computer readable storage medium, on which computer instructions are stored, the computer instructions being configured to cause a computer to perform the method for detecting an operating state of a transformer substation main device based on multi-source time series data according to the first aspect or any one of the embodiments corresponding thereto.
The technical scheme of the invention has the following advantages:
according to the method for detecting the running state of the transformer substation main equipment based on the multi-source time sequence data, firstly, correlation among a plurality of state indexes contained in running state information of the transformer substation equipment is well obtained by utilizing the spearman correlation coefficient, and a plurality of time sequence characteristics which are related to abnormal running height of the transformer substation equipment are obtained to serve as target indexes; training the LSTM neural network by taking a plurality of target indexes as input and taking a single target index as output to obtain a prediction model, and sequentially inputting a plurality of historical target indexes into the prediction model according to a preset time step length, so that a plurality of target indexes of future time sequences of the running state of the transformer substation can be accurately obtained; and clustering and screening out abnormal data by using a spatial clustering algorithm based on noise of density to perform detection of the running state of the main equipment of the transformer substation based on the multi-source time sequence data, and early warning that the running state of the future equipment is possibly abnormal.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are needed in the description of the embodiments or the prior art will be briefly described, and it is obvious that the drawings in the description below are some embodiments of the present invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flow chart of a method for detecting an operation state of a main device of a transformer substation based on multi-source time series data according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of the structure of an LSTM neural network in accordance with an embodiment of the invention;
FIG. 3 is a schematic diagram of identifying outliers based on a density-based noise application spatial clustering algorithm according to an embodiment of the present invention;
fig. 4 is a block diagram of a configuration of a substation main equipment operation state detection apparatus based on multi-source time series data according to an embodiment of the present invention;
fig. 5 is a schematic hardware structure diagram of an intelligent fusion terminal according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
According to an embodiment of the present invention, there is provided an embodiment of a method for detecting an operation state of a substation main device based on multi-source time series data, it should be noted that the steps shown in the flowchart of the drawings may be performed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in an order different from that herein.
In this embodiment, a method for detecting an operation state of a substation main device based on multi-source time series data is provided, which may be used for an intelligent fusion terminal, and fig. 1 is a flowchart of a method for detecting an operation state of a substation main device based on multi-source time series data according to an embodiment of the present invention, as shown in fig. 1, where the flowchart includes the following steps:
step S101, collecting operation state information of substation equipment, analyzing correlations among a plurality of state indexes contained in the operation state information by utilizing a Szelman correlation coefficient, and acquiring a plurality of time sequence characteristics which are highly correlated with abnormal operation of the substation equipment as target indexes.
The spearman correlation coefficient adopted in the embodiment of the invention can well measure the non-parameter index of the dependency between two variables, so that the target index with the correlation of the running state of the substation equipment can be well found out, and the calculation formula is as follows:
wherein X, Y is a data set containing N elements, X i ,Y i Respectively two sets of element values, x i ,y i Respectively X i ,Y i Is arranged in ascending or descending order. If the N lines of data are different integers, using d i Represents x i And y i Wherein d is the difference of i ＝x i -y i The correlation coefficient between the two indices can be calculated using the following indices:
when the calculation result of the correlation coefficient is positive, the monotonicity among the indexes is the same; when the correlation coefficient calculation result is negative, the monotonicity between the indexes is reversed, and the closer the coefficient is to 1, the stronger the correlation between the two indexes is indicated.
According to the embodiment of the invention, the abnormal condition of the power grid is detected more accurately by comprehensively considering the abnormal condition of the intelligent power grid equipment in the operation process and selecting the power index and other indexes with higher correlation with the power index. The correlation between different indexes is calculated through the spearman correlation coefficient to obtain the spearman correlation coefficients between the power and voltage indexes and between the current index and the impedance index which are respectively 0.88, 0.96 and 0.88. By way of example only, and not by way of limitation, an index that ranks the spearman correlation coefficient among a plurality of state indexes top 3 may be set as an index that has a higher correlation with power in practice.
Step S102, training the LSTM neural network by taking a plurality of target indexes with time sequences as input and taking a single target index as output to obtain a prediction model of taking the trained LSTM neural network as a single target index of the running state of substation equipment.
In an embodiment of the invention, the data set division is to divide the generated data set into a training set, a verification set and a test set. Wherein the training set is 80% of the total data set, the validation set is 10% of the total data set, and the test set is 10% of the total data set, by way of example only, and not by way of limitation.
Because the target index has stronger time sequence characteristics, the LSTM neural network has high prediction precision on the time sequence, and can automatically adjust and calculate the related weight coefficient of the hidden layer and learn the implicit relation between the time sequences, the embodiment of the invention adopts the LSTM neural network as a prediction model. As shown in fig. 2, the standard structure of the LSTM neural network adopts a forgetting gate, an input gate and an output gate structure to control the transmission process of time series information, and the state of the forgetting gate is used for determining how much information can be transmitted from the state of a cell at the previous moment to the state of the cell at the current moment. If the output is 0, the information of the previous moment is abandoned; if the output is 1, this indicates that the previous time information was retained. The forgetting door state calculation formula is as follows:
f t ＝σ(W xf *x t +W hf *h t-1 +b f )
wherein f t Indicating the state of the forgotten gate, sigma being the activation function, x t Is an input value, h t-1 Is the output value of the last hidden layer, W xf ,W hf Respectively forgetting the weight between the gate and the input layer, and forgetting the weight between the gate and the hidden layer, b f Is a bias vector.
The input gate is made up of two parts, which can determine which new input information can be added to the cell structure. The specific calculation formula is as follows:
i t ＝σ(W xi *x t +W hi *h t-1 +b i )
wherein i is c Indicating which new input information can be added to the cell structure, generating a new candidate cell state in the current state using the hyperbolic tangent function, σ being the activation function, x t Is an input value, h t-1 Is the output value of the last hidden layer, W xi ,W hi Representing the weights of the input gate and the input layer, respectively, and the weights between the input gate and the hidden layer, b i ,b c Are all bias vectors, W xc ,W hc Is the weight between the candidate and the input gate and the weight between the candidate and the hidden layer.
By forgetting the gate and the input gate, we can update the cell state. The new state consists of two parts, one of which is forgetting the gate output f t With the last old cell state C t-1 Is a product of (2); and the other is input gate input i t And candidate stateIs a product of (a) and (b). The two parts are added together to get a new cell state. The new cell state formula is calculated as follows:
output door O t Controlling the current output value h of the whole LSTM network t It determines how much information the previously updated cell state can output. The specific calculation formula is as follows:
O t ＝σ(W xo *x t +W ho *h t-1 +b o )
h t ＝O t *tanh(C t )
where σ is the activation function, x t Is an input value, h t-1 Is the output value of the last hidden layer, W xo ,W ho The weights between the output gate and the hidden layer and between the input layer and the hidden layer, b o Is the bias vector.
In one embodiment, the training process includes: the actual power, voltage, current and impedance index data of the history are selected and input into the LSTM neural network together, and the LSTM neural network is trained by taking one index of the power, the voltage, the current or the impedance as output to obtain a trained model. The prediction result adopts average absolute percentage error (MAPE) to evaluate the prediction precision, and the smaller the MAPE is, the higher the prediction precision is.
Step S103, sequentially inputting a plurality of historical target indexes into the prediction model according to a preset time step to obtain a plurality of future time series target indexes of the running state of the transformer substation.
In an embodiment, the actual power, voltage, current and impedance data of the historical operation of 1-10 days of the month may be input into the LSTM neural network together, and the actual power, voltage, current and impedance data of 11 days of the month may be obtained by respectively inputting the actual power, voltage, current and impedance data into the prediction model, and the data obtained by 11 days may be added to the historical data, and the operation data of 2-11 days of the month may be input into the prediction model to predict the operation data of 12 days, so that the cycle may be used, for example, the prediction index data of 11 to 16 days of the month may be obtained for seven days.
And step S104, clustering and screening out abnormal data by using a spatial clustering algorithm based on the noise of the density to perform clustering on target indexes of the current time sequence and the future time sequence of the running state of the substation equipment, thereby completing the running state detection of the substation main equipment based on the multi-source time sequence data.
The density-based unsupervised clustering algorithm can find clusters of any shape, can effectively find noise points and outliers, and is suitable for processing irregular data samples. The step S104 includes:
and a step a1, calculating the number of data contained in a circle with the Eps as a radius as a point density value by taking each target index as a data point and taking each data point as a circle center.
And a2, selecting a density threshold value Mints, wherein the central points with the number smaller than the Mints in the circle are low-density points, recording the central points with the number larger than or equal to the Mints as high-density points, and if one high-density point exists in the circle of the other high-density point, connecting the two high-density points.
Step a3, continuously connecting the data points, if there is a low-density point in the high-density point circle, connecting to the nearest high-density point as a boundary point, wherein all connected points form a cluster, and the low-density points not in any cluster are marked as abnormal values.
As shown in fig. 3, the A, B, C, D, E, F points are high-density points, P, Q is a boundary point, and M is an outlier. And taking the Minpts as 3, taking the high-density points as the circle centers, and the number of data points contained in the circle with the radius of Eps being more than or equal to 3.
Based on the detected abnormal data, early warning is provided for possible abnormal conditions in the future.
The embodiment also provides a substation main equipment operation state detection device based on the multi-source time sequence data, which is used for realizing the embodiment and the preferred implementation mode, and is not described again. As used below, the term "module" may be a combination of software and/or hardware that implements a predetermined function. While the means described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware, is also possible and contemplated.
The embodiment provides a substation main equipment operation state detection device based on multi-source time sequence data, as shown in fig. 4, including:
the plurality of target index obtaining modules 401 are configured to collect operation state information of the substation equipment, analyze correlations between a plurality of state indexes included in the operation state information by using spearman correlation coefficients, and obtain a plurality of time-series characteristics highly correlated with an abnormal operation of the substation equipment as target indexes.
The prediction model training module 402 is configured to train the LSTM neural network with a plurality of target indexes having time sequences as input and a single target index as output, so as to obtain a trained prediction model with the LSTM neural network as a single target index of an operation state of the substation equipment.
The target index prediction module 403 is configured to sequentially input a plurality of historical target indexes into the prediction model according to a preset time step, so as to obtain target indexes of a plurality of future time sequences of the operating state of the substation.
The anomaly detection module 404 is configured to perform clustering on target indexes of a current time sequence and a future time sequence of an operation state of substation equipment by using a density-based noise application spatial clustering algorithm to screen out anomaly data, thereby completing detection of the operation state of the substation main equipment based on multi-source time sequence data.
In this embodiment, the target index prediction module 403 includes:
the single target index prediction unit is used for inputting a plurality of target indexes of a preset historical day into the prediction model to obtain a single target index of the running state of the transformer substation in a future day;
the target index prediction units are used for adding target indexes obtained through prediction into preset historical data, updating a plurality of target indexes of the preset historical days of the next step length, inputting the target indexes into the prediction model, and obtaining a single target index of the running state of the transformer substation of the next future day; and in this cycle, obtaining target indexes of a plurality of future time sequences of the running state of the transformer substation.
In this embodiment, the anomaly detection module 404 includes:
a point density value acquisition unit for calculating the number of data contained in a circle with Eps as a radius as a point density value with each target index as a data point and each data point as a circle center;
a high-density point obtaining unit, configured to select a density threshold value mints, wherein a center point with a number of points in a circle smaller than the mints is a low-density point, a center point with a number greater than or equal to the mints is recorded as a high-density point, and if one high-density point exists in a circle with another high-density point, the two high-density points are connected;
an outlier acquisition unit for continuously connecting the data points, connecting to the nearest high-density point as a boundary point if there is a low-density point in the high-density point circle, all the connected points forming one cluster, the low-density points not in any cluster being marked as outliers.
The substation main device operation state detection device based on the multi-source time series data in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC circuit, a processor and a memory that execute one or more software or fixed programs, and/or other devices that can provide the above functions.
Further functional descriptions of the above respective modules and units are the same as those of the above corresponding embodiments, and are not repeated here.
The embodiment of the invention also provides an intelligent fusion terminal which is provided with the substation main equipment running state detection device based on the multi-source time sequence data shown in the figure 5.
Referring to fig. 5, fig. 5 is a schematic structural diagram of a computer device according to an alternative embodiment of the present invention, and as shown in fig. 5, the intelligent fusion terminal includes: one or more processors 10, memory 20, and interfaces for connecting the various components, including high-speed interfaces and low-speed interfaces. The various components are communicatively coupled to each other using different buses and may be mounted on a common motherboard or in other manners as desired. The processor may process instructions executing within the computer device, including instructions stored in or on memory to display graphical information of the GUI on an external input/output device, such as a display device coupled to the interface. In some alternative embodiments, multiple processors and/or multiple buses may be used, if desired, along with multiple memories and multiple memories. Also, multiple computer devices may be connected, each providing a portion of the necessary operations (e.g., as a server array, a set of blade servers, or a multiprocessor system). One processor 10 is illustrated in fig. 5.
The processor 10 may be a central processor, a network processor, or a combination thereof. The processor 10 may further include a hardware chip, among others. The hardware chip may be an application specific integrated circuit, a programmable logic device, or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general-purpose array logic, or any combination thereof.
Wherein the memory 20 stores instructions executable by the at least one processor 10 to cause the at least one processor 10 to perform the methods shown in implementing the above embodiments.
The memory 20 may include a storage program area that may store an operating system, at least one application program required for functions, and a storage data area; the storage data area may store data created from the use of the computer device of the presentation of a sort of applet landing page, and the like. In addition, the memory 20 may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, memory 20 may optionally include memory located remotely from processor 10, which may be connected to the computer device via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
Memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk, or solid state disk; the memory 20 may also comprise a combination of the above types of memories.
The computer device also includes a communication interface 30 for the computer device to communicate with other devices or communication networks.
The embodiments of the present invention also provide a computer readable storage medium, and the method according to the embodiments of the present invention described above may be implemented in hardware, firmware, or as a computer code which may be recorded on a storage medium, or as original stored in a remote storage medium or a non-transitory machine readable storage medium downloaded through a network and to be stored in a local storage medium, so that the method described herein may be stored on such software process on a storage medium using a general purpose computer, a special purpose processor, or programmable or special purpose hardware. The storage medium can be a magnetic disk, an optical disk, a read-only memory, a random access memory, a flash memory, a hard disk, a solid state disk or the like; further, the storage medium may also comprise a combination of memories of the kind described above. It will be appreciated that a computer, processor, microprocessor controller or programmable hardware includes a storage element that can store or receive software or computer code that, when accessed and executed by the computer, processor or hardware, implements the methods illustrated by the above embodiments.
Although embodiments of the present invention have been described in connection with the accompanying drawings, various modifications and variations may be made by those skilled in the art without departing from the spirit and scope of the invention, and such modifications and variations fall within the scope of the invention as defined by the appended claims.
Claims (12)
1. The utility model provides a transformer substation master equipment running state detection method based on multisource time series data which is characterized in that the method includes:
collecting operation state information of substation equipment, analyzing correlations among a plurality of state indexes contained in the operation state information by utilizing a spearman correlation coefficient, and acquiring a plurality of time sequence characteristics which are highly correlated with the abnormal operation of the substation equipment as target indexes;
taking a plurality of target indexes with time sequences as input and taking a single target index as output, training the LSTM neural network, and obtaining a prediction model of taking the trained LSTM neural network as the single target index of the running state of the substation equipment;
sequentially inputting a plurality of historical target indexes into the prediction model according to a preset time step to obtain target indexes of a plurality of future time sequences of the running state of the transformer substation;
and clustering and screening out abnormal data by using a spatial clustering algorithm based on the noise of density to perform clustering on target indexes of the current time sequence and the future time sequence of the running state of the substation equipment, thereby completing the running state detection of the substation main equipment based on the multi-source time sequence data.
2. The method according to claim 1, wherein the spearman correlation coefficient between the plurality of state indexes is ranked a preset number or more as the index having time series characteristics highly correlated with the abnormal operation as the target index.
3. The method of claim 2, wherein the plurality of target metrics having time series characteristics comprises: the power index and other indexes of which the spearman correlation coefficient with the power index is larger than a preset threshold value.
4. A method according to claim 3, wherein the other indicators having a higher correlation with the power indicator comprise: voltage index, current index and impedance index.
5. The method according to claim 1, wherein the process of sequentially inputting the plurality of historical target indexes into the prediction model according to a preset time step to obtain the target indexes of the plurality of future time sequences of the operation state of the substation comprises:
inputting a plurality of target indexes of a preset historical day into the prediction model to obtain a single target index of the running state of the transformer substation in a future day;
adding the target index obtained through prediction to preset historical data, and updating a plurality of target indexes of the preset historical days of the next step length to be input into the prediction model to obtain a single target index of the running state of the transformer substation of the next future day;
and in this cycle, obtaining target indexes of a plurality of future time sequences of the running state of the transformer substation.
6. The method according to claim 1, wherein the process of clustering and screening the abnormal data from the target indexes of the current time sequence and the future time sequence of the operation state of the substation equipment by using the density-based noise application spatial clustering algorithm comprises the following steps:
calculating the number of data contained in a circle with Eps as a radius by taking each target index as a data point and each data point as a circle center as a point density value;
selecting a density threshold value Mints, wherein the central points with the number of points smaller than the Mints in the circle are low-density points, recording the central points larger than or equal to the Mints as high-density points, and if one high-density point exists in the circle of the other high-density point, connecting the two high-density points;
the data points are connected continuously, if there are low density points in the high density circle, then connect to the nearest high density point as boundary point, all connected points form one cluster, the low density points in no cluster are marked as outliers.
7. The method as recited in claim 6, further comprising: based on the detected abnormal data, early warning is provided for possible abnormal conditions in the future.
8. A substation master device operation state detection device based on multi-source time series data, the device comprising:
the system comprises a plurality of target index acquisition modules, a plurality of time sequence analysis modules and a plurality of time sequence analysis modules, wherein the target index acquisition modules are used for acquiring the running state information of the substation equipment, analyzing the correlation among a plurality of state indexes contained in the running state information by utilizing the spearman correlation coefficient, and acquiring a plurality of time sequence characteristics which are highly correlated with the running abnormality of the substation equipment as target indexes;
the prediction model training module is used for training the LSTM neural network by taking a plurality of target indexes with time sequences as input and taking a single target index as output to obtain a trained prediction model with the LSTM neural network as a single target index of the running state of the substation equipment;
the target index prediction module is used for sequentially inputting a plurality of historical target indexes into the prediction model according to a preset time step to obtain target indexes of a plurality of future time sequences of the running state of the transformer substation;
and the anomaly detection module is used for clustering and screening out anomaly data by utilizing a density-based noise application spatial clustering algorithm to target indexes of the current time sequence and the future time sequence of the running state of the substation equipment, and finishing the running state detection of the substation main equipment based on the multi-source time sequence data.
9. The apparatus of claim 8, wherein the target index prediction module comprises:
the single target index prediction unit is used for inputting a plurality of target indexes of a preset historical day into the prediction model to obtain a single target index of the running state of the transformer substation in a future day;
the target index prediction units are used for adding target indexes obtained through prediction into preset historical data, updating a plurality of target indexes of the preset historical days of the next step length, inputting the target indexes into the prediction model, and obtaining a single target index of the running state of the transformer substation of the next future day; and in this cycle, obtaining target indexes of a plurality of future time sequences of the running state of the transformer substation.
10. The apparatus of claim 8, wherein the anomaly detection module comprises:
a point density value acquisition unit for calculating the number of data contained in a circle with Eps as a radius as a point density value with each target index as a data point and each data point as a circle center;
a high-density point obtaining unit, configured to select a density threshold value mints, wherein a center point with a number of points in a circle smaller than the mints is a low-density point, a center point with a number greater than or equal to the mints is recorded as a high-density point, and if one high-density point exists in a circle with another high-density point, the two high-density points are connected;
an outlier acquisition unit for continuously connecting the data points, connecting to the nearest high-density point as a boundary point if there is a low-density point in the high-density point circle, all the connected points forming one cluster, the low-density points not in any cluster being marked as outliers.
11. An intelligent fusion terminal, which is characterized by comprising:
a memory and a processor, the memory and the processor are in communication connection with each other, the memory stores computer instructions, and the processor executes the computer instructions, thereby executing the substation main device operation state detection method based on the multi-source time series data according to any one of claims 1 to 7.
12. A computer-readable storage medium, wherein computer instructions for causing a computer to execute the multi-source time series data-based substation main equipment operation state detection method according to any one of claims 1 to 7 are stored thereon.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310520060.5A CN116539994A (en) | 2023-05-09 | 2023-05-09 | Substation main equipment operation state detection method based on multi-source time sequence data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310520060.5A CN116539994A (en) | 2023-05-09 | 2023-05-09 | Substation main equipment operation state detection method based on multi-source time sequence data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116539994A true CN116539994A (en) | 2023-08-04 |
Family
ID=87453833
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310520060.5A Pending CN116539994A (en) | 2023-05-09 | 2023-05-09 | Substation main equipment operation state detection method based on multi-source time sequence data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116539994A (en) |
-
2023
- 2023-05-09 CN CN202310520060.5A patent/CN116539994A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111459778B (en) | Operation and maintenance system abnormal index detection model optimization method, device and storage medium | |
| US9996444B2 (en) | Automated methods and systems for calculating hard thresholds | |
| CN110717535B (en) | Automatic modeling method and system based on data analysis processing system | |
| CN110474808B (en) | Flow prediction method and device | |
| CN111459700A (en) | Method and apparatus for diagnosing device failure, diagnostic device, and storage medium | |
| CN114297036B (en) | Data processing method, device, electronic equipment and readable storage medium | |
| CN113837596B (en) | Fault determination method and device, electronic equipment and storage medium | |
| CN112116198A (en) | Data-driven process industrial state perception network key node screening method | |
| CN111179591A (en) | Road network traffic time sequence characteristic data quality diagnosis and restoration method | |
| JP7481902B2 (en) | Management computer, management program, and management method | |
| CN117078048A (en) | Digital twinning-based intelligent city resource management method and system | |
| CN114399321A (en) | Business system stability analysis method, device and equipment | |
| CN114648060A (en) | Fault signal standardization processing and classification method based on machine learning | |
| KR101960755B1 (en) | Method and apparatus of generating unacquired power data | |
| RU2632124C1 (en) | Method of predictive assessment of multi-stage process effectiveness | |
| CN115600102B (en) | Abnormal point detection method and device based on ship data, electronic equipment and medium | |
| CN114297453B (en) | Alarm prediction method and device, electronic equipment and storage medium | |
| CN116539994A (en) | Substation main equipment operation state detection method based on multi-source time sequence data | |
| CN115616408A (en) | Battery thermal management data processing method and system | |
| CN114860542A (en) | Trend prediction model optimization method, trend prediction model optimization device, electronic device, and medium | |
| CN113822441A (en) | Decision model training method and device, terminal equipment and storage medium | |
| CN113139332A (en) | Automatic model construction method, device and equipment | |
| CN112395167A (en) | Operation fault prediction method and device and electronic equipment | |
| CN116610484B (en) | Model training method, fault prediction method, system, equipment and medium | |
| CN112685933B (en) | Method for predicting residual service life of roller screw pair |
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 |