CN111723862A - Switch cabinet state evaluation method and device - Google Patents
Switch cabinet state evaluation method and device Download PDFInfo
- Publication number
- CN111723862A CN111723862A CN202010560757.1A CN202010560757A CN111723862A CN 111723862 A CN111723862 A CN 111723862A CN 202010560757 A CN202010560757 A CN 202010560757A CN 111723862 A CN111723862 A CN 111723862A
- Authority
- CN
- China
- Prior art keywords
- cluster
- switch cabinet
- data
- sample
- 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.)
- Granted
Links
- 238000011156 evaluation Methods 0.000 title claims description 30
- 238000001514 detection method Methods 0.000 claims abstract description 70
- 238000000034 method Methods 0.000 claims abstract description 28
- 238000007781 pre-processing Methods 0.000 claims description 7
- 238000005457 optimization Methods 0.000 claims description 5
- 238000012545 processing Methods 0.000 claims description 5
- 238000011425 standardization method Methods 0.000 claims description 4
- 238000012423 maintenance Methods 0.000 abstract description 21
- 230000000694 effects Effects 0.000 abstract description 10
- 230000007613 environmental effect Effects 0.000 description 7
- 238000009413 insulation Methods 0.000 description 6
- 238000012360 testing method Methods 0.000 description 5
- 239000013598 vector Substances 0.000 description 5
- 238000005259 measurement Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 230000032683 aging Effects 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000008676 import Effects 0.000 description 1
- 239000011810 insulating material Substances 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000008707 rearrangement Effects 0.000 description 1
- 238000006467 substitution reaction 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
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a method and a device for evaluating the state of a switch cabinet. The method comprises the steps of acquiring electrified detection data of the switch cabinet; carrying out state division on the charged detection data by adopting a DBSCAN clustering algorithm to determine the state of the switch cabinet; and optimizing parameters of the DBSCAN clustering algorithm through the contour coefficients. The state division is carried out on the charged detection data by adopting a DBSCAN clustering algorithm, so that the data can be classified more scientifically without depending on subjective weight. Moreover, the DBSCAN clustering algorithm can well divide non-spherical clusters and clusters with different sizes, the problem that the non-spherical clusters and the clusters with different sizes are difficult to process in a common mean clustering algorithm is solved, and more scientific and reasonable operation and maintenance suggestions can be provided for operation and maintenance personnel of the switch cabinet. And then, parameters of the DBSCAN clustering algorithm are optimized through the contour coefficients, so that the clustering effect of the DBSCAN clustering algorithm is good, and the rationality of operation and maintenance suggestions is further improved.
Description
Technical Field
The embodiment of the invention relates to the technical field of switch cabinet state evaluation, in particular to a switch cabinet state evaluation method and device.
Background
With the continuous development of power grid technology in China, a large number of switch cabinet devices enter the power grid operation, and the operation and maintenance of the switch cabinet are important links of the power grid operation. Usually, the inside long-term work of cubical switchboard is under the abominable condition of high temperature, high pressure high humidity, and the phenomenon of ageing, corruption appears easily in insulating material, and the loss will be punctured by the inside high voltage of cubical switchboard to a certain extent, leads to cubical switchboard insulating fault to appear easily. The overall structure of the switch cabinet is complex, and the live detection data is various. At present, according to the live detection data of the switch cabinet, there are many state evaluation methods, including a common state evaluation method such as a fuzzy comprehensive evaluation method, a normal cloud theory, a clustering algorithm and the like.
In a clustering evaluation algorithm of a switch cabinet, the state is graded according to Euclidean distance, but non-spherical clusters and clusters with different sizes are difficult to process. Meanwhile, in the state division evaluation of the clustering algorithm, the initial parameters of the clustering algorithm generally need to be set manually, and the final clustering effect is easily influenced to a certain extent.
Disclosure of Invention
The invention provides a switch cabinet state evaluation method and device, which are used for realizing more scientific classification of data and improving the clustering effect.
In a first aspect, an embodiment of the present invention provides a method for evaluating a status of a switch cabinet, including:
acquiring electrified detection data of a switch cabinet;
performing state division on the live detection data by adopting a DBSCAN clustering algorithm to determine the state of the switch cabinet;
and optimizing parameters of the DBSCAN clustering algorithm through the contour coefficients.
Optionally, performing state division on the live detection data by using a DBSCAN clustering algorithm to determine a state of the switch cabinet, including:
initializing a data sample and clustered clusters, wherein the data sample is the charged detection data, and the number of the clustered clusters is 0;
determining a core object according to the data sample, and allocating a new cluster according to the core object; wherein, the number of the core objects is more than or equal to MinPts in the neighborhood; the neighborhood of the core object is an open interval which takes the core object as a central point and has preset length as a radius, and MinPts is a preset integer which is greater than or equal to 1 and less than the number of the data samples;
determining clusters of the data samples according to the core object and a sample set in the neighborhood of the core object;
and determining the state of the switch cabinet according to the clusters, wherein each cluster corresponds to one state of the switch cabinet.
Optionally, determining the cluster of the data sample according to the core object and the sample set in the neighborhood thereof includes:
if the sample in the neighborhood of the core object is not allocated with a new cluster, allocating the sample to the cluster of the core object;
and if the sample in the neighborhood of the core object is another core object, combining the cluster of the core object and the cluster of the other core object into a new cluster.
Optionally, determining the state of the switchgear according to the cluster includes:
and determining the state of the switch cabinet according to the value of the cluster center point.
Optionally, when determining a core object according to the data sample and allocating a new cluster according to the core object, the method further includes:
and determining a noise point according to the data sample, wherein the noise point is a sample of which the number of the sample sets in the neighborhood is less than MinPts.
Optionally, optimizing parameters of the DBSCAN clustering algorithm by using contour coefficients includes:
calculating the contour coefficient of the DBSCAN clustering algorithm according to the clusters of the data samples;
adjusting the radius of the neighborhood and the value of MinPts according to the profile coefficient.
Optionally, calculating a contour coefficient of the DBSCAN clustering algorithm according to the cluster of the data samples includes:
calculating the average distance from the sample point in the cluster to other sample points in the same cluster, namely the intra-cluster dissimilarity of the sample points in the cluster;
calculating the average distance from the sample point in the cluster to the sample point in any other cluster, namely the dissimilarity degree between the sample point in the cluster and any other cluster;
calculating the contour coefficient of the sample point in the cluster according to the dissimilarity degree of the sample point in the cluster and any other cluster;
and calculating the average value of the contour coefficients of all the cluster sample points to serve as the contour coefficient of the DBSCAN clustering algorithm.
Optionally, after acquiring the live detection data of the switch cabinet, the method further includes:
preprocessing the charged detection data;
and carrying out data standardization on the preprocessed charged detection data by adopting a Z-score standardization method.
Optionally, the preprocessing the charged detection data includes:
and performing missing value interpolation processing on the charged detection data by adopting a Lagrange interpolation method.
In a second aspect, an embodiment of the present invention further provides a switch cabinet state evaluation apparatus, including:
the data acquisition module is used for acquiring the electrified detection data of the switch cabinet;
the state determining module is used for performing state division on the live detection data by adopting a DBSCAN clustering algorithm so as to determine the state of the switch cabinet;
and the parameter optimization module is used for optimizing the parameters of the DBSCAN clustering algorithm through the contour coefficients.
According to the technical scheme of the embodiment of the invention, the state division is carried out on the charged detection data by adopting the DBSCAN clustering algorithm, so that the data can be classified more scientifically without depending on subjective weight. Moreover, the DBSCAN clustering algorithm can well divide non-spherical clusters and clusters with different sizes, the problem that the non-spherical clusters and the clusters with different sizes are difficult to process in a common mean clustering algorithm is solved, and more scientific and reasonable operation and maintenance suggestions can be provided for operation and maintenance personnel of the switch cabinet. And then, parameters of the DBSCAN clustering algorithm are optimized through the contour coefficients, so that the clustering effect of the DBSCAN clustering algorithm is good, and the rationality of operation and maintenance suggestions is further improved.
Drawings
Fig. 1 is a flowchart of a switch cabinet state evaluation method according to an embodiment of the present invention;
fig. 2 is a flowchart of a switch cabinet state evaluation method according to a second embodiment of the present invention;
fig. 3 is a flowchart of a switch cabinet state evaluation method according to a third embodiment of the present invention;
fig. 4 is a schematic structural diagram of a switch cabinet state evaluation apparatus according to a fourth embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.
Example one
Fig. 1 is a flowchart of a switch cabinet state evaluation method according to an embodiment of the present invention, which is applicable to a situation of evaluating a health state of a switch cabinet, and the method can be executed by a switch cabinet state evaluation device, and specifically includes the following steps:
and S110, acquiring the electrified detection data of the switch cabinet.
Specifically, the fault of the switch cabinet is mainly an insulation fault, so that when the health state of the switch cabinet is evaluated, the insulation state of the switch cabinet can be mainly evaluated, and the acquired live detection data of the switch cabinet mainly include data affecting the insulation state of the switch cabinet. For example, the live detection data of the switch cabinet may include TEV detection data and ultrasonic detection data for reflecting the insulation state of the switch cabinet. In addition, the TEV detection data and the ultrasonic detection data may include data of measurement points at different positions of the switch cabinet, for example, six measurement points, i.e., upper, middle and lower, on the front and rear surfaces of the switch cabinet, and the maximum value of each measurement point in the TEV detection data and the ultrasonic detection data is used as the charged detection data.
The live detection data of the switch cabinet can also comprise environmental state temperature data, environmental state humidity data and historical input operation time data, the data have certain influence on the insulation state of the switch cabinet, and the data can be used for reflecting the insulation state of the switch cabinet.
And S120, carrying out state division on the electrified detection data by adopting a DBSCAN clustering algorithm to determine the state of the switch cabinet.
Specifically, the charged detection data is multi-source data. When the DBSCAN clustering algorithm is adopted to divide the state of the charged detection data, the area with high enough density of the charged detection data can be divided into clusters, clusters with any shapes can be formed, similar state quantities are quickly divided through the clusters, and the state of the switch cabinet is determined according to the divided similar state quantities. When the DBSCAN clustering algorithm is adopted to divide the charged detection data, compared with a fuzzy comprehensive evaluation method, the method can classify the data more scientifically without depending on subjective weight. Moreover, the DBSCAN clustering algorithm can well divide non-spherical clusters and clusters with different sizes, the problem that the non-spherical clusters and the clusters with different sizes are difficult to process in a common mean clustering algorithm is solved, and more scientific and reasonable operation and maintenance suggestions can be provided for operation and maintenance personnel of the switch cabinet.
S130, optimizing parameters of the DBSCAN clustering algorithm through the contour coefficients.
Specifically, the contour coefficient is used to measure the similarity between any data point and its cluster compared to other clusters, and ranges from-1 to 1, and a larger value indicates that the data point is more similar to the cluster. Namely, the contour coefficient can be used for evaluating the clustering effect of the DBSCAN clustering algorithm. When the DBSCAN clustering algorithm is adopted to carry out state division on the charged electromagnetic data, the parameters of the DBSCAN clustering algorithm are set randomly. For example, the values of the distance of the neighborhood and the number of core objects in the DBSCAN clustering algorithm are randomly set. By optimizing the parameters of the DBSCAN clustering algorithm, when the value of the contour coefficient is closer to 1, the clustering effect of the DBSCAN clustering algorithm is better, the optimization of the parameters of the DBSCAN clustering algorithm is realized, and the rationality of the operation and maintenance suggestion is further improved.
According to the technical scheme of the embodiment, the state division is carried out on the charged detection data by adopting the DBSCAN clustering algorithm, so that the data can be classified more scientifically without depending on subjective weight. Moreover, the DBSCAN clustering algorithm can well divide non-spherical clusters and clusters with different sizes, the problem that the non-spherical clusters and the clusters with different sizes are difficult to process in a common mean clustering algorithm is solved, and more scientific and reasonable operation and maintenance suggestions can be provided for operation and maintenance personnel of the switch cabinet. And then, parameters of the DBSCAN clustering algorithm are optimized through the contour coefficients, so that the clustering effect of the DBSCAN clustering algorithm is good, and the rationality of operation and maintenance suggestions is further improved.
Example two
Fig. 2 is a flowchart of a switch cabinet state evaluation method according to a second embodiment of the present invention, where on the basis of the second embodiment, the method includes:
and S210, acquiring the electrified detection data of the switch cabinet.
And S220, preprocessing the charged detection data.
Specifically, the charged detection data may include a variety of data, and the charged detection data is first constructed into a multi-dimensional feature data set. For example, the charging test data includes TEV test data, ultrasonic test data, environmental state temperature data, environmental state humidity data, and historical input operation time data, a new data set is created for the above 5 state feature quantities, the data sets of the 5 sets of feature quantities are respectively represented by k × 1 order column vectors, and the data set of the TEV test data is exemplarily U ═ U(1)U(2)… U(k)]TSimilarly, k-dimensional vectors of the other 4 sets of feature quantity data sets are established, and the data sets respectively used as ultrasonic detection data are Z ═ Z(1)Z(2)… Z(k)]TThe data set of the environmental state humidity data is H ═ H(1)H(2)… H(k)]TThe data set of the ambient temperature data is T ═ T(1)T(2)… T(k)]TAnd the data set of historical invested runtime data is S ═ S(1)S(2)… S(k)]T。
And (4) performing data cleaning on all data, and filling up vacant data by an interpolation method. Illustratively, the preprocessing the charged detection data includes:
and performing interpolation processing of missing values on the charged detection data by adopting a Lagrange interpolation method.
The method comprises the following specific steps: selecting 5 data before and after the missing value (skipping if the missing value is met in the data before and after the missing value), forming a group by the selected 10 data, and interpolating the missing data in the data set by adopting a Lagrange polynomial difference formula to obtain a new data set. For example, for TEV detection data, the data set U ═ U(1)U(2)… U(k)]TWhen interpolation is performed on the missing data in (1), a new data set U' is obtained as follows:
wherein x is the number missing from the switch cabinet, Uk' (x) is the interpolation result of the missing value, U(i)Is a non-missing value. Similarly, the data set Z of the ultrasonic detection data is sequentially subjected to the Lagrange interpolation method to obtain [ Z ═ Z [ ](1)Z(2)… Z(k)]TData set H ═ H for ambient state humidity data(1)H(2)… H(k)]TThe data set of ambient temperature data T ═ T(1)T(2)… T(k)]TAnd a data set of historical invested runtime data S ═ S(1)S(2)… S(k)]TAnd carrying out interpolation processing on the missing values to obtain preprocessed Z ', H', T 'and S', so that the data sets are k-dimensional vectors. By performing interpolation processing on missing data in the data set, the integrity of TEV detection data can be improved.
And S230, carrying out data standardization on the preprocessed charged detection data by adopting a Z-score standardization method.
Specifically, the preprocessed charged detection data are k-dimensional vectors, and then the preprocessed charged detection data are subjected to data standardization by adopting a Z-score standardization method, so that the preprocessed charged detection data are subjected to standard normal distribution. For example, the dataset U' of the preprocessed TEV detection data is normalized:
a data set U "of normalized TEV test data is obtained.
Wherein, U(k)″The normalized result is shown, μ is the mean of the sample data, and σ is the standard deviation of the sample data.
Similarly, the preprocessed data set Z 'of the ultrasonic detection data, the data set H' of the environmental state humidity data, the data set T 'of the environmental state temperature data and the data set S' of the historical investment running time data are normalized, and the normalized data sets Z ", H", T ", S" are obtained in sequence.
Then, each dataset U, Z, H, T, S is updated to a normalized dataset, creating a multi-dimensional feature dataset, represented by a k × 5 order matrix:
wherein the matrix [ X1X2… Xk]TEach 1X 5 order row vector X1、X2……XkAnd the monitoring state information quantity of each switch cabinet is represented. Wherein, Xj={Xj∈ D | j ═ 1,2, …, k } represents the state quantities of any switchgear.
S240, initializing data samples and clustered clusters, wherein the data samples are charged detection data, and the number of the clustered clusters is 0.
Specifically, during initialization, the import dataset D ═ (X)1,X2,…,Xm) As a data sample, a clustered cluster is initialized at the same time, so that the number of clusters is 0, and the current cluster is an empty set. The point T ═ D in the unprocessed data samples is initialized.
S250, determining a core object according to the data sample, and distributing a new cluster according to the core object; wherein, the number of the core objects is more than or equal to MinPts in the neighborhood; the neighborhood of the core object is an open interval which takes the core object as a central point and has preset length as a radius, and MinPts is a preset integer which is greater than or equal to 1 and less than the number of data samples.
Specifically, for any sample X in the data samplej∈ D, whose neighborhood is from sample X in data sample DjIs not greater than the set of samples. When arbitrary sample Xj∈ D, if the number of samples contained in its neighborhood is greater than or equal to MinPts, then sample XjIs a core object. The core object and its neighborhood have a sufficiently high density of samples so that the core object can be divided into a cluster.
It should be noted that the neighborhood and MinPts of any one of the data samples are randomly set, and the values thereof can be adjusted according to the accuracy requirement of the DBSCAN clustering algorithm.
And S260, determining a cluster of the data sample according to the core object and the sample set in the neighborhood of the core object.
Specifically, after determining the core object, the core object is divided into a cluster. And then sequentially accessing samples in the neighborhood thereof, and performing cluster division on the samples in the neighborhood. The method comprises the following steps:
if the sample in the neighborhood of the core object is not allocated with a new cluster, allocating the sample to the cluster of the core object;
if the sample in the neighborhood of the core object is another core object, the cluster of the core object and the cluster of the other core object are merged into a new cluster.
Specifically, when the samples in the neighborhood are not assigned to a new cluster, the samples in the neighborhood are not core objects, and the distance between the samples in the neighborhood and the core objects is in the neighborhood, the clusters formed by the core objects and the samples in the neighborhood have the characteristic of high density, so that the samples in the neighborhood can be assigned to the clusters of the core objects. When the sample in the neighborhood is another core object, the sample in the neighborhood is already allocated to a cluster when the core object is determined, and the cluster formed by the core object and the sample in the neighborhood has the characteristic of high density, so that the cluster of the core object and the cluster corresponding to the sample in the neighborhood can be merged into a new cluster, and the cluster division is reduced. In addition, when the sample in the neighborhood is another core object, the samples in the neighborhood of the other core object are sequentially accessed, and the process is the same as the above process, and is not described herein again. Until the samples in the neighborhood of the core object are all accessed, clusters corresponding to the core object are formed. When all the core objects are processed through the above process, a cluster C ═ C corresponding to the data sample can be formed1,C2,…,Ck}。
S270, determining the state of the switch cabinet according to the clusters, wherein each cluster corresponds to one state of the switch cabinet.
Specifically, a cluster is a set of similar state quantities in a data sample, and each cluster corresponds to one state of the switch cabinet. Determining a state of the switchgear cabinet from the cluster, comprising:
and determining the state of the switch cabinet according to the value of the cluster center point.
Wherein the center point of the cluster is the average value of the sample set in the cluster. And dividing corresponding health states of the switch cabinets in the clusters through the specific value of the center point of the clusters.
S280, optimizing parameters of the DBSCAN clustering algorithm through the contour coefficients.
According to the technical scheme, the condition that the switch cabinet records data loss can be effectively solved by preprocessing the electrified detection data, and the integrity of the data set is improved. And then, the state division is carried out on the charged detection data by adopting a DBSCAN clustering algorithm, so that the data can be classified more scientifically without depending on subjective weight. Moreover, the DBSCAN clustering algorithm can well divide non-spherical clusters and clusters with different sizes, the problem that the non-spherical clusters and the clusters with different sizes are difficult to process in a common mean clustering algorithm is solved, and more scientific and reasonable operation and maintenance suggestions can be provided for operation and maintenance personnel of the switch cabinet. And then, parameters of the DBSCAN clustering algorithm are optimized through the contour coefficients, so that the clustering effect of the DBSCAN clustering algorithm is good, and the rationality of operation and maintenance suggestions is further improved.
Optionally, with continuing reference to fig. 2, when determining a core object according to the data sample and allocating a new cluster according to the core object, the method further includes:
and S290, determining a noise point according to the data sample, wherein the noise point is a sample of which the number of the sample sets in the neighborhood is less than MinPts.
Specifically, when determining the core object, when the sample X isj∈ D, if the number of samples contained in the neighborhood is less than MinPts, the samples are noise points, therefore, when the DBSCAN clustering algorithm is adopted to carry out state division on the charged detection data, the noise in the data samples can be removed, the accuracy of state division of the DBSCAN clustering algorithm is improved, and the rationality of operation and maintenance suggestions is further improved.
EXAMPLE III
Fig. 3 is a flowchart of a switch cabinet state evaluation method according to a third embodiment of the present invention, where on the basis of the foregoing embodiments, the method includes:
and S310, acquiring the electrified detection data of the switch cabinet.
And S320, carrying out state division on the electrified detection data by adopting a DBSCAN clustering algorithm to determine the state of the switch cabinet.
S330, calculating the contour coefficient of the DBSCAN clustering algorithm according to the clusters of the data samples.
Specifically, calculating the profile coefficient of the DBSCAN clustering algorithm according to the clusters of the data samples includes:
and calculating the average distance from the sample point in the cluster to other sample points in the same cluster, namely the intra-cluster dissimilarity of the sample points in the cluster.
In particular, the sample point X within a clusterjThe average distance to other sample points in the same cluster can be denoted as dj. By sample point X in clusterjThe average distance to other sample points in the same cluster represents the sample point X in the clusterjInter-cluster dissimilarity.
And calculating the average distance from the sample point in the cluster to the sample point in any other cluster, namely the dissimilarity between the sample point in the cluster and any other cluster.
Specifically, a sample point X within a cluster is calculatedjTo other clusters CkIs measured over a time period of one sample pointijE is to beijReferred to as intra-cluster sample point XjAnd cluster CkDegree of dissimilarity of, define ej=min{ej1,ej2,...,ejmIs an in-cluster sample point XjInter-cluster dissimilarity.
And calculating the contour coefficient of the sample point in the cluster according to the intra-cluster dissimilarity of the sample point in the cluster and the dissimilarity of the sample point in the cluster and any other cluster.
In particular, according to the intra-cluster sample point XjWithin cluster dissimilarity of djDegree of dissimilarity e between clustersjDefining a sample point X within a clusterjThe profile factor of (a) is as follows:
wherein d isjIs an in-cluster sample point XjAverage distance to other samples in the same cluster, ejIs xjMinimum of average distance from all samples in other clusters.
And calculating the average value of the contour coefficients of all the sample points in the cluster as the contour coefficient of the data sample.
Specifically, the average of all sample points s (j) is calculatedAt this timeContour coefficients for DBSCAN clustering algorithm:
and S340, adjusting the radius of the neighborhood and the MinPts value according to the contour coefficient.
Specifically, the neighborhood radius and MinPts of the DBSCAN clustering algorithm are adjusted according to the value of the contour coefficient, so that the value of the contour coefficient tends to 1, and when the value of the contour coefficient tends to 1, the distance effect of the DBSCAN clustering algorithm is better, and the operation and maintenance suggestion given according to the DBSCAN clustering algorithm is more reasonable.
Example four
The fourth embodiment of the invention also provides a switch cabinet state evaluation device. Fig. 4 is a schematic structural diagram of a switch cabinet state evaluation apparatus according to a fourth embodiment of the present invention. As shown in fig. 4, the apparatus includes:
and the data acquisition module 10 is used for acquiring the electrified detection data of the switch cabinet.
And the state determining module 20 is configured to perform state division on the charged detection data by using a DBSCAN clustering algorithm to determine the state of the switch cabinet.
And the parameter optimization module 30 is used for optimizing parameters of the DBSCAN clustering algorithm through the contour coefficients.
According to the technical scheme of the embodiment, the state determination module adopts the DBSCAN clustering algorithm to perform state division on the charged detection data, so that the data can be classified more scientifically without depending on subjective weight. Moreover, the DBSCAN clustering algorithm can well divide non-spherical clusters and clusters with different sizes, the problem that the non-spherical clusters and the clusters with different sizes are difficult to process in a common mean clustering algorithm is solved, and more scientific and reasonable operation and maintenance suggestions can be provided for operation and maintenance personnel of the switch cabinet. And then the parameter optimization module optimizes the parameters of the DBSCAN clustering algorithm through the contour coefficients, so that the clustering effect of the DBSCAN clustering algorithm is good, and the rationality of operation and maintenance suggestions is further improved.
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims (10)
1. A switch cabinet state evaluation method is characterized by comprising the following steps:
acquiring electrified detection data of a switch cabinet;
performing state division on the live detection data by adopting a DBSCAN clustering algorithm to determine the state of the switch cabinet;
and optimizing parameters of the DBSCAN clustering algorithm through the contour coefficients.
2. The switch cabinet state evaluation method according to claim 1, wherein performing state division on the live detection data by using a DBSCAN clustering algorithm to determine the state of the switch cabinet comprises:
initializing a data sample and clustered clusters, wherein the data sample is the charged detection data, and the number of the clustered clusters is 0;
determining a core object according to the data sample, and allocating a new cluster according to the core object; wherein, the number of the core objects is more than or equal to MinPts in the neighborhood; the neighborhood of the core object is an open interval which takes the core object as a central point and has preset length as a radius, and MinPts is a preset integer which is greater than or equal to 1 and less than the number of the data samples;
determining clusters of the data samples according to the core object and a sample set in the neighborhood of the core object;
and determining the state of the switch cabinet according to the clusters, wherein each cluster corresponds to one state of the switch cabinet.
3. The switch cabinet state evaluation method of claim 2, wherein determining the cluster of data samples from the set of samples in the core object and its neighborhood comprises:
if the sample in the neighborhood of the core object is not allocated with a new cluster, allocating the sample to the cluster of the core object;
and if the sample in the neighborhood of the core object is another core object, combining the cluster of the core object and the cluster of the other core object into a new cluster.
4. The switch cabinet state evaluation method according to claim 2, wherein determining the state of the switch cabinet from the cluster comprises:
and determining the state of the switch cabinet according to the value of the cluster center point.
5. The switch cabinet state evaluation method according to claim 2, wherein when determining a core object from the data samples and assigning a new cluster from the core object, further comprising:
and determining a noise point according to the data sample, wherein the noise point is a sample of which the number of the sample sets in the neighborhood is less than MinPts.
6. The switch cabinet state evaluation method according to claim 2, wherein the optimizing the parameters of the DBSCAN clustering algorithm by the profile coefficients comprises:
calculating the contour coefficient of the DBSCAN clustering algorithm according to the clusters of the data samples;
adjusting the radius of the neighborhood and the value of MinPts according to the profile coefficient.
7. The switch cabinet state evaluation method of claim 6, wherein calculating the profile coefficients of the DBSCAN clustering algorithm from the clusters of data samples comprises:
calculating the average distance from the sample point in the cluster to other sample points in the same cluster, namely the intra-cluster dissimilarity of the sample points in the cluster;
calculating the average distance from the sample point in the cluster to the sample point in any other cluster, namely the dissimilarity degree between the sample point in the cluster and any other cluster;
calculating the contour coefficient of the sample point in the cluster according to the dissimilarity degree of the sample point in the cluster and any other cluster;
and calculating the average value of the contour coefficients of all the cluster sample points to serve as the contour coefficient of the DBSCAN clustering algorithm.
8. The method for evaluating the state of a switchgear cabinet according to claim 1, further comprising, after acquiring the live detection data of the switchgear cabinet:
preprocessing the charged detection data;
and carrying out data standardization on the preprocessed charged detection data by adopting a Z-score standardization method.
9. The switch cabinet state evaluation method according to claim 8, wherein preprocessing the live detection data comprises:
and performing missing value interpolation processing on the charged detection data by adopting a Lagrange interpolation method.
10. A switch cabinet condition assessment device, comprising:
the data acquisition module is used for acquiring the electrified detection data of the switch cabinet;
the state determining module is used for performing state division on the live detection data by adopting a DBSCAN clustering algorithm so as to determine the state of the switch cabinet;
and the parameter optimization module is used for optimizing the parameters of the DBSCAN clustering algorithm through the contour coefficients.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010560757.1A CN111723862B (en) | 2020-06-18 | 2020-06-18 | Switch cabinet state evaluation method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010560757.1A CN111723862B (en) | 2020-06-18 | 2020-06-18 | Switch cabinet state evaluation method and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111723862A true CN111723862A (en) | 2020-09-29 |
CN111723862B CN111723862B (en) | 2024-09-06 |
Family
ID=72567548
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010560757.1A Active CN111723862B (en) | 2020-06-18 | 2020-06-18 | Switch cabinet state evaluation method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111723862B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112459971A (en) * | 2020-11-30 | 2021-03-09 | 吉电(滁州)章广风力发电有限公司 | Abnormal vibration detection method for tower of wind generating set |
CN116304766A (en) * | 2023-05-25 | 2023-06-23 | 山东艾迈科思电气有限公司 | Multi-sensor-based quick assessment method for state of switch cabinet |
CN116381506A (en) * | 2023-04-11 | 2023-07-04 | 国网宁夏电力有限公司电力科学研究院 | Reconfigurable battery network system battery state sorting method based on data clustering |
CN118035768A (en) * | 2024-01-16 | 2024-05-14 | 上海栈略数据技术有限公司 | Group insurance policy clustering analysis method based on neural network encoder |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2016141978A1 (en) * | 2015-03-11 | 2016-09-15 | You Know Watt | Improved non-intrusive appliance load monitoring method and device |
CN107463751A (en) * | 2017-08-10 | 2017-12-12 | 山东师范大学 | A kind of crowd based on DBSCAN clustering algorithms by half is grouped evacuation emulation method and system |
CN107609709A (en) * | 2017-09-26 | 2018-01-19 | 上海爱优威软件开发有限公司 | Paths planning method and system based on scene classification |
CN110728322A (en) * | 2019-10-11 | 2020-01-24 | 深圳市前海随手数据服务有限公司 | Data classification method and related equipment |
CN110738245A (en) * | 2019-09-29 | 2020-01-31 | 上海大学 | automatic clustering algorithm selection system and method for scientific data analysis |
JP2020064286A (en) * | 2018-10-12 | 2020-04-23 | 株式会社東陽テクニカ | Abnormal sound detection system, device, method, and program |
CN111060153A (en) * | 2019-12-23 | 2020-04-24 | 北京中交兴路车联网科技有限公司 | Method and device for detecting cargo state of truck and storage medium |
CN111175626A (en) * | 2020-03-20 | 2020-05-19 | 广东电网有限责任公司 | Abnormal detection method for insulation state of switch cabinet |
-
2020
- 2020-06-18 CN CN202010560757.1A patent/CN111723862B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2016141978A1 (en) * | 2015-03-11 | 2016-09-15 | You Know Watt | Improved non-intrusive appliance load monitoring method and device |
CN107463751A (en) * | 2017-08-10 | 2017-12-12 | 山东师范大学 | A kind of crowd based on DBSCAN clustering algorithms by half is grouped evacuation emulation method and system |
CN107609709A (en) * | 2017-09-26 | 2018-01-19 | 上海爱优威软件开发有限公司 | Paths planning method and system based on scene classification |
JP2020064286A (en) * | 2018-10-12 | 2020-04-23 | 株式会社東陽テクニカ | Abnormal sound detection system, device, method, and program |
CN110738245A (en) * | 2019-09-29 | 2020-01-31 | 上海大学 | automatic clustering algorithm selection system and method for scientific data analysis |
CN110728322A (en) * | 2019-10-11 | 2020-01-24 | 深圳市前海随手数据服务有限公司 | Data classification method and related equipment |
CN111060153A (en) * | 2019-12-23 | 2020-04-24 | 北京中交兴路车联网科技有限公司 | Method and device for detecting cargo state of truck and storage medium |
CN111175626A (en) * | 2020-03-20 | 2020-05-19 | 广东电网有限责任公司 | Abnormal detection method for insulation state of switch cabinet |
Non-Patent Citations (1)
Title |
---|
意)吉安卡洛·扎克尼,(德)礼萨·卡里姆: "《TensorFlow深度学习 原书第2版》", 31 March 2020, 机械工业出版社, pages: 165 - 167 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112459971A (en) * | 2020-11-30 | 2021-03-09 | 吉电(滁州)章广风力发电有限公司 | Abnormal vibration detection method for tower of wind generating set |
CN116381506A (en) * | 2023-04-11 | 2023-07-04 | 国网宁夏电力有限公司电力科学研究院 | Reconfigurable battery network system battery state sorting method based on data clustering |
CN116304766A (en) * | 2023-05-25 | 2023-06-23 | 山东艾迈科思电气有限公司 | Multi-sensor-based quick assessment method for state of switch cabinet |
CN116304766B (en) * | 2023-05-25 | 2023-07-28 | 山东艾迈科思电气有限公司 | Multi-sensor-based quick assessment method for state of switch cabinet |
CN118035768A (en) * | 2024-01-16 | 2024-05-14 | 上海栈略数据技术有限公司 | Group insurance policy clustering analysis method based on neural network encoder |
Also Published As
Publication number | Publication date |
---|---|
CN111723862B (en) | 2024-09-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111723862A (en) | Switch cabinet state evaluation method and device | |
CN112911627B (en) | Wireless network performance detection method, device and storage medium | |
CN108919059A (en) | A kind of electric network failure diagnosis method, apparatus, equipment and readable storage medium storing program for executing | |
CN108491302A (en) | A method of detection spark clustered node states | |
CN111638988A (en) | Cloud host fault intelligent prediction method based on deep learning | |
CN111160404A (en) | Method and device for analyzing reasonable value of line loss marking pole of power distribution network | |
CN111709668A (en) | Power grid equipment parameter risk identification method and device based on data mining technology | |
CN111008726A (en) | Class image conversion method in power load prediction | |
CN113379116A (en) | Cluster and convolutional neural network-based line loss prediction method for transformer area | |
CN111198979A (en) | Method and system for cleaning big data for power transmission and transformation reliability evaluation | |
CN115797551A (en) | Laser point cloud data automatic modeling method based on two-step unsupervised clustering algorithm | |
CN111897888A (en) | Household variable relation identification method based on Spark framework and coacervation hierarchical clustering algorithm | |
CN116595465A (en) | High-dimensional sparse data outlier detection method and system based on self-encoder and data enhancement | |
CN107908807A (en) | A kind of System in Small Sample Situation Reliability Assessment Method based on bayesian theory | |
CN107729918B (en) | Classification method for cellular automaton emerging phenomenon based on cost-sensitive support vector machine | |
CN107274025B (en) | System and method for realizing intelligent identification and management of power consumption mode | |
CN113378889A (en) | Density clustering and binning method | |
CN117930012A (en) | Battery consistency assessment method and device, computer equipment and storage medium | |
CN115051363B (en) | Distribution network area user change relation identification method and device and computer storage medium | |
CN111459926A (en) | Park comprehensive energy anomaly data identification method | |
CN115017988A (en) | Competitive clustering method for state anomaly diagnosis | |
CN113589034A (en) | Electricity stealing detection method, device, equipment and medium for power distribution system | |
CN109657795B (en) | Hard disk failure prediction method based on attribute selection | |
CN113487080B (en) | Wind speed dynamic scene generation method, system and terminal based on wind speed classification | |
CN112884167B (en) | Multi-index anomaly detection method based on machine learning and application system thereof |
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 |