Method and device for monitoring running state of active power distribution network

By acquiring the potential sparseness and contribution optimization factors of multi-dimensional monitoring data in the active distribution network, the clustering process is optimized, and the inaccuracy of the operating state evaluation model caused by data imbalance is solved, and the accuracy of monitoring is improved.

CN118965052AInactive Publication Date: 2024-11-15DEZHOU POWER SUPPLY COMPANY OF STATE GRID SHANDONG ELECTRIC POWER

Patent Information

Application Number
CN202411019089.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-29
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

In the multi-dimensional monitoring data of the active distribution network, there are few abnormal samples, which leads to the center point of the cluster class tilting towards dense areas during clustering, resulting in inaccuracy of the operating state evaluation model.

Method used

By acquiring the potential sparseness of multi-dimensional monitoring of timing data and optimizing the membership of data points to the cluster center point using contribution optimization factors during cluster iteration, an operating state evaluation model of the active distribution network is constructed.

Benefits of technology

It improves the accuracy of operating status monitoring of active distribution networks, eliminates data imbalance problem, and enhances the performance of clustering results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118965052A_ABST
    Figure CN118965052A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of power distribution network monitoring, in particular to an active power distribution network operation state monitoring method and device, and the method comprises the steps: obtaining multi-dimensional monitoring time sequence data of an active power distribution network in a historical time period, correspondingly obtaining normalized time sequence data, and obtaining the potential sparse degree of each data point in the normalized time sequence data; constructing a multi-dimensional data space of the normalized time series data, and according to the potential sparse degree of each data point in the multi-dimensional data space, obtaining a contribution optimization factor of each data point in the multi-dimensional data space to each cluster center point during each clustering iteration; and optimizing the membership degree in each clustering iteration process by using the contribution optimization factor to obtain an operation state evaluation model of the active power distribution network, and performing real-time operation state monitoring on the real-time multi-dimensional monitoring data of the active power distribution network by using the operation state evaluation model. The accuracy of real-time operation state monitoring based on the operation state evaluation model is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of power distribution network monitoring, and in particular to a method and device for monitoring the operating status of an active power distribution network. Background Art

[0002] In the process of monitoring the operating status of the active distribution network, the multi-dimensional operating monitoring data of the active distribution network is clustered to obtain corresponding cluster clusters. According to the cluster clusters, the operating status of the active distribution network is divided into fuzzy states such as general, normal, severe and critical, so as to obtain an operating status evaluation model of the active distribution network. Subsequently, the operating status evaluation model of the active distribution network is used to monitor the real-time operating status of the active distribution network.

[0003] However, in the clustering process when obtaining the operation status assessment model of the active distribution network, there is an imbalance problem in the data space of the multidimensional operation monitoring data of the active distribution network that there are fewer abnormal samples than normal samples, that is, in the multidimensional monitoring data space of the active distribution network, most of the sample points are monitoring data points in normal status, and only a small number of sample points are monitoring data points in abnormal status, where the abnormal status is the two clustering results corresponding to the serious fault operation status and the critical fault operation status. This will cause the cluster center points in the clustering iteration process to tend to dense areas when clustering the operation status assessment model for the sample points in the multidimensional monitoring data space, resulting in inaccurate status assessment in the obtained operation status assessment model.

[0004] Therefore, how to improve the monitoring performance of the operating status assessment model of the active distribution network based on clustering acquisition to ensure the monitoring accuracy of the operating status of the active distribution network has become an urgent problem to be solved. Summary of the invention

[0005] In view of this, an embodiment of the present invention provides an active distribution network operating status monitoring method and device to solve the problem of how to improve the monitoring performance of an active distribution network operating status assessment model based on clustering acquisition to ensure the monitoring accuracy of the active distribution network operating status.

[0006] In a first aspect, an embodiment of the present invention provides a method for monitoring the operating status of an active power distribution network, the method comprising the following steps:

[0007] Acquire multidimensional monitoring time series data of the active distribution network in a historical period, perform data normalization processing on each dimension in the multidimensional monitoring time series data, and obtain corresponding normalized time series data, and obtain the potential sparsity degree of each data point in the normalized time series data according to the local outlier factor and local data point of each data point in each dimension in the normalized time series data;

[0008] Constructing a multidimensional data space of the normalized time series data, and in the process of performing fuzzy clustering on all data points in the multidimensional data space, obtaining a contribution optimization factor of each data point in the multidimensional data space to each cluster center point in each clustering iteration according to the potential sparsity of each data point in the multidimensional data space;

[0009] By utilizing the contribution optimization factor of each data point in the multidimensional data space to each cluster center point at each clustering iteration, the membership of each data point to each cluster center point during each clustering iteration is optimized to obtain an operation status evaluation model of the active distribution network, and the real-time operation status monitoring of the real-time multidimensional monitoring data of the active distribution network is performed using the operation status evaluation model.

[0010] Furthermore, the step of obtaining the potential sparsity degree of each data point in the normalized time series data according to the local outlier factor of each data point in each dimension and the local data point in the normalized time series data includes:

[0011] For any data point in the normalized time series data, a local window of a preset length is obtained in each dimension of the normalized time series data with the data point as the center, and a local outlier factor of the data point in each dimension is obtained according to the local window of the data point in each dimension;

[0012] For any dimension of the data point, obtain the data mean of all data points in the local window of the data point on the dimension, calculate the absolute value of the difference between the value of the data point on the dimension and the data mean, normalize the absolute value of the difference to obtain a corresponding normalized value, and use the product of the normalized value and the local outlier factor of the data point on the dimension as the eigenvalue of the data point on the dimension;

[0013] The characteristic values ​​of the data points in each dimension are accumulated, and the obtained accumulated value is used as the potential sparsity degree of the data points.

[0014] Furthermore, obtaining the contribution optimization factor of each data point in the multidimensional data space to each cluster center point in each clustering iteration according to the potential sparsity of each data point in the multidimensional data space includes:

[0015] In any clustering iteration process, all cluster center points in the clustering iteration process are determined, and for any data point in the multidimensional data space, the Euclidean distance between the data point and each of the cluster center points is obtained;

[0016] Obtain a preset number of nearest neighbor data points of the data point according to the Euclidean distances between the data points, calculate the sum of the Euclidean distances between the data point and all the nearest neighbor data points, and for any nearest neighbor data point of the data point, detect whether the preset number of nearest neighbor data points of the nearest neighbor data point contains the data point, and if it is detected that the preset number of nearest neighbor data points of the nearest neighbor data point contains the data point, use the nearest neighbor data point as the target point of the data point;

[0017] Counting the number of target points of the data point, calculating the number ratio between the number of nearest neighbor data points of the data point and the number of target points, normalizing the product of the number ratio and the sum of the Euclidean distances to obtain a corresponding normalized result, obtaining a multiplication result between the normalized result and the potential sparsity of the data point, and negatively mapping the multiplication result using an exponential function with a natural constant as the base to obtain a corresponding mapping value;

[0018] The normalized value of the product of the Euclidean distance between the data point and each of the cluster center points and the mapping value is used as the contribution optimization factor of the data point to each cluster center point in the clustering iteration process.

[0019] Furthermore, the contribution optimization factor of each data point in the multidimensional data space to each cluster center point in each cluster iteration is utilized to optimize the membership of each data point to each cluster center point in each cluster iteration process, so as to obtain the operation status evaluation model of the active distribution network, including:

[0020] The membership optimization formula is:

[0021]

[0022] Among them, μ ic represents the optimized membership of the i-th data point in the multidimensional data space to the c-th cluster center point during the clustering iteration process, ε ic represents the contribution optimization factor of the i-th data point in the multidimensional data space to the c-th cluster center point during the clustering iteration process, K represents the number of cluster centers, ε ij represents the contribution optimization factor of the i-th data point in the multidimensional data space to the j-th cluster center point during the clustering iteration process, ||x i -v c || 2 represents the Euclidean distance between the i-th data point in the multidimensional data space and the c-th cluster center point during the clustering iteration process, ||x i -v j || 2represents the Euclidean distance between the i-th data point in the multidimensional data space and the j-th cluster center point during the clustering iteration process, x i represents the value of the i-th data point in the multidimensional data space, v c Indicates the value of the center point of the cth cluster during the clustering iteration, v j represents the value of the jth cluster center point during the clustering iteration process, and m represents the preset fuzzy weighted index;

[0023] The optimized membership of each data point to each cluster center point in each cluster iteration process is obtained, and all data points in the multidimensional data space are divided into K clusters according to the optimized membership of each data point to each cluster center point in each cluster iteration process, one cluster is used as an operating state, and the K clusters constitute the operating state evaluation model of the active distribution network.

[0024] Furthermore, the use of the operating status assessment model to perform real-time operating status monitoring on the real-time multi-dimensional monitoring data of the active power distribution network includes:

[0025] Real-time multidimensional monitoring data of the active distribution network is obtained, and according to the Euclidean distance between the real-time multidimensional monitoring data and the cluster center of each cluster in the operation status evaluation model, the operation status corresponding to the cluster corresponding to the minimum Euclidean distance is used as the real-time operation status of the real-time multidimensional monitoring data.

[0026] In a second aspect, an embodiment of the present invention provides an active distribution network operating status monitoring device, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein when the processor executes the computer program, an active distribution network operating status monitoring method as described in the first aspect is implemented.

[0027] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0028] The present invention obtains multidimensional monitoring time series data of an active distribution network in a historical period, performs data normalization processing on each dimension in the multidimensional monitoring time series data, obtains corresponding normalized time series data, and obtains the potential sparsity degree of each data point in the normalized time series data according to the local outlier factor of each data point in each dimension and the local data point in the normalized time series data; constructs a multidimensional data space of the normalized time series data, and in the process of fuzzy clustering all data points in the multidimensional data space, obtains the contribution optimization factor of each data point in the multidimensional data space to each cluster center point in each clustering iteration according to the potential sparsity degree of each data point in the multidimensional data space; uses the contribution optimization factor of each data point in the multidimensional data space to each cluster center point in each clustering iteration to optimize the membership degree of each data point to each cluster center point in each clustering iteration process, obtains the operation status evaluation model of the active distribution network, and uses the operation status evaluation model to perform real-time operation status monitoring on the real-time multidimensional monitoring data of the active distribution network. Among them, the potential sparsity of each data point is obtained through the fluctuation state of the monitoring data in the multidimensional monitoring time series data of the active distribution network. The potential sparsity is a preliminary evaluation of the sparsity of the monitoring data, and the contribution optimization factor of each data point to each cluster center point in the clustering iteration process is obtained through distribution information in the multidimensional data space according to the potential sparsity of each data point, which is used to optimize the membership in the clustering iteration process. By eliminating the data imbalance problem in the multidimensional monitoring data space of the active distribution network, the performance of the operation status evaluation model obtained by the clustering result is better, and the accuracy of real-time operation status monitoring based on the operation status evaluation model is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0030] Figure 1 It is a method flow chart of a method for monitoring the operating status of an active distribution network provided in Embodiment 1 of the present invention. DETAILED DESCRIPTION

[0031] Embodiments of the present disclosure are described in detail below, and examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present disclosure, but should not be understood as limiting the present disclosure.

[0032] It should be noted that the terms "first", "second", etc. in the specification of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure.

[0033] In order to illustrate the technical solution of the present invention, specific embodiments are provided below for illustration.

[0034] See also Figure 1 , is a method flow chart of a method for monitoring the operating status of an active distribution network provided in Embodiment 1 of the present invention, such as Figure 1 As shown, the method may include:

[0035] Step S101, obtain multidimensional monitoring time series data of the active distribution network in a historical period, perform data normalization processing on each dimension in the multidimensional monitoring time series data, and obtain corresponding normalized time series data, and obtain the potential sparsity degree of each data point in the normalized time series data according to the local outlier factor of each data point in each dimension and the local data point in the normalized time series data.

[0036] In the process of monitoring the operating status of the active distribution network, the operating data of the active distribution network is first collected, that is, the data monitored during the operation. The monitoring data mainly targeted in the embodiment of the present invention include but are not limited to current, voltage, load rate, line current carrying capacity, line loss rate, power supply reliability rate, life coefficient, etc. One type of monitoring data is a dimension. For the monitoring data of each dimension mentioned above, after completing the collection of the monitoring time series data of the dimension, the monitoring time series data of multiple dimensions constitute a multi-dimensional monitoring time series data, and the multi-dimensional monitoring time series data is preprocessed, that is, the missing values ​​are filled with the front and back mean fitting, and the monitoring time series data of each dimension is normalized to eliminate the dimension dimension, and then the normalized monitoring time series data of all dimensions are composed of a normalized time series data.

[0037] In the traditional way, after obtaining the normalized time series data, a multidimensional data space of the normalized time series data is constructed, and a fuzzy clustering process is performed in the multidimensional data space to divide the data points in the multidimensional data space into different clusters. According to the data characteristics presented by the data points contained in each cluster, each cluster is defined as a different operating state, and one cluster corresponds to one operating state. It should be noted that the operating state of the active distribution network is usually divided into four operating states, namely normal, general, severe, and critical, corresponding to 4 clusters in the fuzzy clustering process.

[0038] However, in the fuzzy clustering process, there are more data points in normal and general states, which leads to the situation that the cluster centers tend to be closer to the data points in the dense area during the clustering iteration process, that is, the clustering result deviation problem caused by the uneven data distribution. Therefore, in an embodiment of the present invention, before fuzzy clustering is performed on the data points in the multidimensional data space, the potential sparsity of each data point is analyzed according to the fluctuation state of the monitoring data in the normalized time series data. The potential sparsity is a preliminary evaluation of the sparsity of the normalized time series data.

[0039] Among them, the potential sparsity of each data point is analyzed according to the fluctuation state of the monitoring data in the normalized time series data, including:

[0040] For any data point in the normalized time series data, a local window of a preset length is obtained in each dimension of the normalized time series data with the data point as the center, and a local outlier factor of the data point in each dimension is obtained according to the local window of the data point in each dimension;

[0041] For any dimension of the data point, obtain the data mean of all data points in the local window of the data point on the dimension, calculate the absolute value of the difference between the value of the data point on the dimension and the data mean, normalize the absolute value of the difference to obtain a corresponding normalized value, and use the product of the normalized value and the local outlier factor of the data point on the dimension as the eigenvalue of the data point on the dimension;

[0042] The characteristic values ​​of the data points in each dimension are accumulated, and the obtained accumulated value is used as the potential sparsity degree of the data points.

[0043] In one embodiment, taking the i-th data point in the normalized time series data as an example, first set the local window in the time series data, set the length of the local window L=20, and obtain the corresponding local window in each dimension in the normalized time series data with the i-th data point as the center. Then, for the q-th dimension, according to the local window of the i-th data point in the q-th dimension, obtain the COF local outlier factor of the i-th data point in the q-th dimension, where the COF local outlier factor belongs to the prior art and will not be repeated here. Finally, according to the COF local outlier factor of the i-th data point in each dimension and the local data point, obtain the potential sparsity of the i-th data point, and the calculation expression of the potential sparsity is:

[0044]

[0045] in, represents the potential sparsity of the i-th data point in the normalized time series data, N represents the total number of dimensions, and softmax q () indicates normalization of the qth dimension, g i,q Indicates the value corresponding to the i-th data point in the q-th dimension, g i,q,j represents the value corresponding to the jth data point in the local window of the i-th data point in the q-th dimension, || represents the absolute value symbol, COF i,q represents the COF local outlier factor of the ith data point in the qth dimension.

[0046] It should be noted that the potential sparsity degree is to evaluate the possibility of each data point being in a sparse area in the multidimensional data space through the local changes of the data points in the time series data. For the local window of the ith data point, the outliers presented by the data points in the local window in each dimension are used as the weights of each dimension to comprehensively evaluate the potential sparsity degree of the ith data point; and then, according to the COF local outlier factor of the ith data point in each dimension, the outliers presented in all dimensions are evaluated by weighted summation, that is, the weighted summation result is used as the potential sparsity degree of the ith data point.

[0047] At this point, the potential sparsity level of each data point in the normalized time series data can be obtained.

[0048] Step S102, construct a multidimensional data space of normalized time series data, and in the process of fuzzy clustering all data points in the multidimensional data space, obtain the contribution optimization factor of each data point in the multidimensional data space to each cluster center point in each clustering iteration according to the potential sparsity of each data point in the multidimensional data space.

[0049] After obtaining the potential sparsity of each data point in the normalized time series data, the membership in the clustering process can be optimized according to the potential sparsity of each data point in the normalized time series data, so as to ensure the accuracy of the acquisition of the operating status evaluation model of the active distribution network.

[0050] After obtaining the potential sparsity of each data point in the normalized time series data, further evaluation based on the potential sparsity can be performed through the neighboring data points of each data point in the multidimensional data space during the fuzzy clustering process. Specifically, first, a multidimensional data space of the normalized time series data is constructed, and each data point in the normalized time series data is mapped in the multidimensional data space. Then, for each data point, the contribution optimization factor of the data point to each cluster center point in each clustering iteration is obtained based on the potential sparsity through its M=20 neighboring data points in the multidimensional data space.

[0051] The method for obtaining the contribution optimization factor of each data point in the multidimensional data space to each cluster center point at each clustering iteration includes:

[0052] In any clustering iteration process, all cluster center points in the clustering iteration process are determined, and for any data point in the multidimensional data space, the Euclidean distance between the data point and each of the cluster center points is obtained;

[0053] Obtain a preset number of nearest neighbor data points of the data point according to the Euclidean distances between the data points, calculate the sum of the Euclidean distances between the data point and all the nearest neighbor data points, and for any nearest neighbor data point of the data point, detect whether the preset number of nearest neighbor data points of the nearest neighbor data point contains the data point, and if it is detected that the preset number of nearest neighbor data points of the nearest neighbor data point contains the data point, use the nearest neighbor data point as the target point of the data point;

[0054] Counting the number of target points of the data point, calculating the number ratio between the number of nearest neighbor data points of the data point and the number of target points, normalizing the product of the number ratio and the sum of the Euclidean distances to obtain a corresponding normalized result, obtaining a multiplication result between the normalized result and the potential sparsity of the data point, and negatively mapping the multiplication result using an exponential function with a natural constant as the base to obtain a corresponding mapping value;

[0055] The normalized value of the product of the Euclidean distance between the data point and each of the cluster center points and the mapping value is used as the contribution optimization factor of the data point to each cluster center point in the clustering iteration process.

[0056] In one embodiment, taking the i-th data point in the multidimensional data space as an example, for any clustering iteration process, all cluster center points in the clustering iteration process are determined, and according to the Euclidean distance between any two data points in the multidimensional data space, the M nearest neighbor data points of the i-th data point are obtained, and the Euclidean distance between the i-th data point and the c-th cluster center point is recorded as || g i -g c || 2 , and then according to the distance difference between the ith data point and its nearest neighbor data point, the potential sparsity of the ith data point and the Euclidean distance between the ith data point and the cth cluster center point, the contribution optimization factor of the ith data point to the cth cluster center point in the clustering iteration process is obtained. The calculation expression of the contribution optimization factor is:

[0057]

[0058] Among them, ε ic represents the contribution optimization factor of the ith data point to the cth cluster center point during the clustering iteration process, softmax() represents the normalized exponential function, exp() represents the exponential function with a natural constant as the base, Norm() represents the normalization function, N' M Represents the number of target points for the i-th data point, N M represents the total number of nearest neighbor data points of the i-th data point, ||g i -g j || 2 represents the Euclidean distance between the i-th data point and the j-th nearest neighbor data point, represents the potential sparsity of the i-th data point, ||g i -g c || 2 Represents the Euclidean distance between the i-th data point and the c-th cluster center point.

[0059] It should be noted that in the multidimensional data space, each data point corresponds to a potential sparsity degree. However, because the data points that appear as local outliers in the time series may have a large number of similar data points in the multidimensional data space, it is necessary to further evaluate the sparsity degree through the similarity of the local neighboring data points of each data point in the multidimensional data space, so as to more accurately judge the contribution degree required by each data point in the iterative process of fuzzy clustering, and select the initial cluster center based on this. Therefore, for the contribution optimization factor of the i-th data point to the c-th cluster center point in the clustering iteration process, the more the number of target points of the i-th data point, that is, The larger the value of , the more uniform the local distribution of the ith data point. The smaller the Euclidean distance between the ith data point and the nearest neighbor data point, the more likely the ith data point is to be in a dense area. Therefore, the potential sparsity of the ith data point is optimized through this measurement process, so as to obtain the final sparsity of the ith data point in the multidimensional data space. When the sparsity of the ith data point is higher, it is necessary to reduce the distance between the ith data point and the cluster center point, so as to achieve the effect of eliminating data imbalance. Finally, the exponential function is used for mapping to obtain the contribution optimization factor of the ith data point in the clustering process.

[0060] At this point, in the clustering iteration process, the contribution optimization factor of each data point in the multidimensional data space to each cluster center point can be obtained.

[0061] Step S103, using the contribution optimization factor of each data point in the multidimensional data space to each cluster center point in each clustering iteration, the membership of each data point to each cluster center point in each clustering iteration process is optimized to obtain an operation status evaluation model of the active distribution network, and the real-time operation status monitoring of the real-time multidimensional monitoring data of the active distribution network is performed using the operation status evaluation model.

[0062] After obtaining the contribution optimization factor of each data point in the multidimensional data space to each cluster center point at each clustering iteration, the membership evaluation between each data point and the cluster center point during the clustering iteration process can be optimized to reduce the contribution of data points in dense areas to cluster centers in sparse areas, thereby eliminating the cluster update bias caused by data imbalance during the clustering process.

[0063] In the embodiment of the present invention, the contribution optimization factor of each data point in the multidimensional data space to each cluster center point in each cluster iteration is used to optimize the membership of each data point to each cluster center point in each cluster iteration process, and obtain the operation status evaluation model of the active distribution network, including:

[0064] (1) The contribution optimization factor of each data point in the multidimensional data space to each cluster center point in each clustering iteration is used to optimize the membership of each data point to each cluster center point in each clustering iteration process.

[0065] Specifically, the membership optimization formula is:

[0066]

[0067] Among them, μ ic represents the optimized membership of the i-th data point in the multidimensional data space to the c-th cluster center point during the clustering iteration process, ε icrepresents the contribution optimization factor of the i-th data point in the multidimensional data space to the c-th cluster center point during the clustering iteration process, K represents the number of cluster centers, ε ij represents the contribution optimization factor of the i-th data point in the multidimensional data space to the j-th cluster center point during the clustering iteration process, ||x i -v c || 2 represents the Euclidean distance between the i-th data point in the multidimensional data space and the c-th cluster center point during the clustering iteration process, ||x i -v j || 2 represents the Euclidean distance between the i-th data point in the multidimensional data space and the j-th cluster center point during the clustering iteration process, x i represents the value of the i-th data point in the multidimensional data space, v c Indicates the value of the center point of the cth cluster during the clustering iteration, v j It represents the value of the j-th cluster center point during the clustering iteration process, and m represents the preset fuzzy weighted index.

[0068] It should be noted that The updating formula of the membership degree in the FCM clustering process belongs to the prior art. In the embodiment of the present invention, the Euclidean distance evaluation between each data point and the cluster center point is optimized by the contribution optimization factor, thereby eliminating the clustering result deviation caused by data imbalance in the clustering process.

[0069] (2) Obtaining the optimized membership of each data point to each cluster center point in each clustering iteration process, and dividing all data points in the multidimensional data space into K clusters according to the optimized membership of each data point to each cluster center point in each clustering iteration process, wherein one cluster is regarded as an operating state, and the K clusters constitute an operating state evaluation model of the active distribution network.

[0070] Specifically, based on the optimized membership of each data point to each cluster center point in each clustering iteration process, all data points in the multidimensional data space are divided into K clusters. This clustering process belongs to the FCM clustering process and belongs to the prior art, which will not be described here. Among them, one cluster is regarded as an operating state, and the K clusters constitute the operating state evaluation model of the active distribution network.

[0071] Furthermore, after obtaining the operation status evaluation model of the active distribution network, the operation status evaluation model can be used to perform real-time operation status monitoring on the real-time multi-dimensional monitoring data of the active distribution network. The specific monitoring process is:

[0072] Real-time multidimensional monitoring data of the active distribution network is obtained, and according to the Euclidean distance between the real-time multidimensional monitoring data and the cluster center of each cluster in the operation status evaluation model, the operation status corresponding to the cluster corresponding to the minimum Euclidean distance is used as the real-time operation status of the real-time multidimensional monitoring data.

[0073] In summary, the embodiment of the present invention obtains multidimensional monitoring time series data of an active distribution network within a historical period, performs data normalization processing on each dimension in the multidimensional monitoring time series data, and obtains corresponding normalized time series data. According to the local outlier factor and local data point of each data point in the normalized time series data in each dimension, the potential sparsity degree of each data point in the normalized time series data is respectively obtained; a multidimensional data space of the normalized time series data is constructed, and in the process of fuzzy clustering of all data points in the multidimensional data space, according to the potential sparsity degree of each data point in the multidimensional data space, a contribution optimization factor of each data point in the multidimensional data space to each cluster center point at each clustering iteration is obtained; using the contribution optimization factor of each data point in the multidimensional data space to each cluster center point at each clustering iteration, the membership degree of each data point to each cluster center point in each clustering iteration is optimized to obtain the operation status evaluation model of the active distribution network, and the operation status evaluation model is used to perform real-time operation status monitoring on the real-time multidimensional monitoring data of the active distribution network. Among them, the potential sparsity of each data point is obtained through the fluctuation state of the monitoring data in the multidimensional monitoring time series data of the active distribution network. The potential sparsity is a preliminary evaluation of the sparsity of the monitoring data, and the contribution optimization factor of each data point to each cluster center point in the clustering iteration process is obtained through distribution information in the multidimensional data space according to the potential sparsity of each data point, which is used to optimize the membership in the clustering iteration process. By eliminating the data imbalance problem in the multidimensional monitoring data space of the active distribution network, the performance of the operation status evaluation model obtained by the clustering result is better, and the accuracy of real-time operation status monitoring based on the operation status evaluation model is improved.

[0074] Based on the same inventive concept as the above method, an embodiment of the present invention also provides an active distribution network operation status monitoring device, including a memory, a processor, and a computer program stored in the memory and running on the processor, and when the processor executes the computer program, the steps of any one of the above-mentioned methods for monitoring the operation status of an active distribution network are implemented.

[0075] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A method for monitoring the operating status of an active distribution network, characterized in that: The method comprises: Acquire multidimensional monitoring time series data of the active distribution network in a historical period, perform data normalization processing on each dimension in the multidimensional monitoring time series data, and obtain corresponding normalized time series data, and obtain the potential sparsity degree of each data point in the normalized time series data according to the local outlier factor and local data point of each data point in each dimension in the normalized time series data; Constructing a multidimensional data space of the normalized time series data, and in the process of performing fuzzy clustering on all data points in the multidimensional data space, obtaining a contribution optimization factor of each data point in the multidimensional data space to each cluster center point in each clustering iteration according to the potential sparsity of each data point in the multidimensional data space; By utilizing the contribution optimization factor of each data point in the multidimensional data space to each cluster center point at each clustering iteration, the membership of each data point to each cluster center point during each clustering iteration is optimized to obtain an operation status evaluation model of the active distribution network, and the real-time operation status monitoring of the real-time multidimensional monitoring data of the active distribution network is performed using the operation status evaluation model.

2. The method for monitoring the operating status of an active power distribution network according to claim 1, characterized in that: The step of obtaining the potential sparsity degree of each data point in the normalized time series data according to the local outlier factor of each data point in each dimension and the local data point in the normalized time series data comprises: For any data point in the normalized time series data, a local window of a preset length is obtained in each dimension of the normalized time series data with the data point as the center, and a local outlier factor of the data point in each dimension is obtained according to the local window of the data point in each dimension; For any dimension of the data point, obtain the data mean of all data points in the local window of the data point on the dimension, calculate the absolute value of the difference between the value of the data point on the dimension and the data mean, normalize the absolute value of the difference to obtain a corresponding normalized value, and use the product of the normalized value and the local outlier factor of the data point on the dimension as the eigenvalue of the data point on the dimension; The characteristic values ​​of the data points in each dimension are accumulated, and the obtained accumulated value is used as the potential sparsity degree of the data points.

3. The method for monitoring the operating status of an active power distribution network according to claim 1, characterized in that: The step of obtaining, according to the potential sparsity degree of each data point in the multidimensional data space, a contribution optimization factor of each data point in the multidimensional data space to each cluster center point in each clustering iteration includes: In any clustering iteration process, all cluster center points in the clustering iteration process are determined, and for any data point in the multidimensional data space, the Euclidean distance between the data point and each of the cluster center points is obtained; Obtain a preset number of nearest neighbor data points of the data point according to the Euclidean distances between the data points, calculate the sum of the Euclidean distances between the data point and all the nearest neighbor data points, and for any nearest neighbor data point of the data point, detect whether the preset number of nearest neighbor data points of the nearest neighbor data point contains the data point, and if it is detected that the preset number of nearest neighbor data points of the nearest neighbor data point contains the data point, use the nearest neighbor data point as the target point of the data point; Counting the number of target points of the data point, calculating the number ratio between the number of nearest neighbor data points of the data point and the number of target points, normalizing the product of the number ratio and the sum of the Euclidean distances to obtain a corresponding normalized result, obtaining a multiplication result between the normalized result and the potential sparsity of the data point, and negatively mapping the multiplication result using an exponential function with a natural constant as the base to obtain a corresponding mapping value; The normalized value of the product of the Euclidean distance between the data point and each of the cluster center points and the mapping value is used as the contribution optimization factor of the data point to each cluster center point in the clustering iteration process.

4. The method for monitoring the operating status of an active power distribution network according to claim 1, characterized in that: The method uses the contribution optimization factor of each data point in the multidimensional data space to each cluster center point in each cluster iteration, optimizes the membership of each data point to each cluster center point in each cluster iteration process, and obtains the operating status evaluation model of the active distribution network, including: The membership optimization formula is: Among them, μ ic represents the optimized membership of the i-th data point in the multidimensional data space to the c-th cluster center point during the clustering iteration process, ε ic represents the contribution optimization factor of the i-th data point in the multidimensional data space to the c-th cluster center point during the clustering iteration process, K represents the number of cluster centers, ε ij represents the contribution optimization factor of the i-th data point in the multidimensional data space to the j-th cluster center point during the clustering iteration process, ||x i -v c || 2 represents the Euclidean distance between the i-th data point in the multidimensional data space and the c-th cluster center point during the clustering iteration process, ||x i -v j || 2 represents the Euclidean distance between the i-th data point in the multidimensional data space and the j-th cluster center point during the clustering iteration process, x i represents the value of the i-th data point in the multidimensional data space, v c Indicates the value of the center point of the cth cluster during the clustering iteration, v j represents the value of the jth cluster center point during the clustering iteration process, and m represents the preset fuzzy weighted index; The optimized membership of each data point to each cluster center point in each cluster iteration process is obtained, and all data points in the multidimensional data space are divided into K clusters according to the optimized membership of each data point to each cluster center point in each cluster iteration process, one cluster is used as an operating state, and the K clusters constitute the operating state evaluation model of the active distribution network.

5. The method for monitoring the operating status of an active power distribution network according to claim 4, characterized in that: The using the operation status evaluation model to perform real-time operation status monitoring on the real-time multi-dimensional monitoring data of the active power distribution network includes: Real-time multidimensional monitoring data of the active distribution network is obtained, and according to the Euclidean distance between the real-time multidimensional monitoring data and the cluster center of each cluster in the operation status evaluation model, the operation status corresponding to the cluster corresponding to the minimum Euclidean distance is used as the real-time operation status of the real-time multidimensional monitoring data.

6. An active power distribution network operation status monitoring device, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method for monitoring the operating status of an active power distribution network as described in any one of claims 1 to 5 are implemented.

Citation Information

Patent Citations

  • Electric reactor vibration monitoring system based on LOF-FCM fuzzy clustering algorithm

    CN114046873A

Cited By

  • Intelligent temperature control method for supercritical extraction of plant essential oil

    CN119620799A

  • An intelligent temperature control method for supercritical extraction of plant essential oils

    CN119620799B

  • Abnormal state assessment method used in solar cell operation process

    CN119813956A

  • Electric energy meter with tripping current recording and state analyzing functions

    CN121878596A

  • Electricity meter with tripping current recording and state analysis function

    CN121878596B