A method and system for identifying the same source of voltage sag

Abstract

The invention discloses a method and system for identifying the same source of voltage sags. The identification method comprises the following steps: obtaining sag monitoring data recorded by a power grid voltage sag monitoring device; based on the propagation law of voltage sag waveforms between multi-stage transformers, using a Hausdorff distance algorithm to measure the waveform similarity between multiple groups of sag monitoring data, and using an effective value absolute difference method to calculate the duration of the sag; deriving a transfer formula for a phase jump value through a transformer, using a linear interpolation method to accurately calculate the phase jump value in the sag process, and using the Euclidean distance to respectively calculate the similarity of the phase jump values ​​of multiple sag waveforms; obtaining characteristic indicators of voltage sag data, including waveform similarity, duration and phase jump three-dimensional features, forming voltage sag data points, and jointly forming a same source identification clustering matrix; counting the number of clustering result clusters based on the color of data points in a point cloud image, and outputting the number of sag same source data groups and the clustering results.

Description

Technical Field

[0001] The present invention relates to the technical field of power quality management, and in particular to a method and system for identifying the same source of voltage sag based on multivariate waveform features and improved DENCLUE clustering. Background Art

[0002] In recent years, a large number of automated industrial equipment and user-side precision electronic equipment have been put into use. These devices are very sensitive to voltage changes. When a voltage sag occurs and the voltage suddenly drops, sensitive equipment will be affected and may even be paralyzed, causing huge economic losses to users and industrial production. In order to propose effective and targeted governance methods and reasonably evaluate the impact and harm of sags, power grid companies install a large number of power quality monitoring devices to carry out large-scale, high-density real-time monitoring. However, as the monitoring scope continues to expand, the amount of monitoring data also continues to increase. When the voltage sag propagates in the power grid, it will cause multiple monitoring devices to repeatedly record the same fault event, resulting in redundancy of monitoring data, which has an adverse effect on the evaluation and governance of the sag. Therefore, it is necessary to realize the same-source monitoring and identification of voltage sag data, eliminate the interference of redundant information, determine the source of voltage sag, and grasp the true level and state of voltage sag.

[0003] At present, the research on the identification of the same source of voltage sag is still in its infancy. Existing research only describes voltage sag events through two single characteristic quantities, namely, the amplitude and duration of the sag, which can only reflect some characteristics of the sag event and make it difficult to comprehensively analyze the sag waveform, resulting in the loss of sag information and reducing the accuracy of the same source identification. In addition, the sag data monitored by the power grid are relatively concentrated, and the distance similarity between multiple events and the benchmark event is relatively close, which will affect the accuracy of the existing same source identification method.

[0004] In summary, the existing voltage sag homologous source identification methods have the defects of single characteristic index and poor data resolution, which leads to insufficient identification accuracy. In order to improve the accuracy of voltage sag homologous source identification, it is necessary to introduce new voltage sag feature quantities. Summary of the invention

[0005] In response to the problems and needs raised above, this solution proposes a method and system for identifying the same source of voltage sags. Due to the adoption of the following technical features, the above technical objectives can be achieved and many other technical effects can be brought about.

[0006] The present invention proposes a method for identifying the same source of voltage sag, comprising the following steps:

[0007] S10: Acquire the voltage sag monitoring data recorded by the grid voltage sag monitoring device;

[0008] S20: Based on the propagation law of voltage sag waveforms between multi-stage transformers, the Hausdorff distance algorithm is used to measure the waveform similarity between multiple groups of sag monitoring data, and the effective value absolute difference method is used to calculate the sag duration;

[0009] S30: deriving a transfer formula of the phase jump value through the transformer, using a linear interpolation method to accurately calculate the phase jump value during the sag process, and using the Euclidean distance to calculate the similarity of the phase jump values ​​of multiple sag waveforms;

[0010] S40: Based on the above steps S10 to S30, characteristic indicators of voltage sag data are obtained, including waveform similarity, duration and phase jump three-dimensional characteristics, to form voltage sag data points, which together constitute a homologous identification clustering matrix;

[0011] S50: Count the number of clusters of clustering results based on the colors of the data points in the point cloud image, and output the number of temporarily homologous data groups and the clustering results.

[0012] In addition, the method for identifying the same source of voltage sag according to the present invention may also have the following technical features:

[0013] In an example of the present invention, step S10 includes:

[0014] S11: Select n voltage sag data monitored within a period of time, select the sag data with the earliest triggering time as the reference data, and the remaining n-1 data as non-reference data;

[0015] S12: comparing the voltage sag amplitudes of each phase, and selecting the phase with the lowest voltage sag amplitude as the voltage sag phase;

[0016] S13: retain the sag phase in the sag monitoring data and remove other data to obtain the sag monitoring data.

[0017] In one example of the present invention, in step S20, using the Hausdorff distance algorithm to measure the waveform similarity between multiple groups of sag monitoring data includes the following steps:

[0018] S21: Based on the propagation characteristics of voltage sag between multi-stage transformers, the voltage sag phase voltage transfer matrix T1~T8 under the action of different types and numbers of transformers is derived, and the effective values ​​of the three-phase voltage of non-reference data in the sag data are multiplied by T1~T8 respectively to obtain the set of all transformed sag waveforms:

[0019] [Q i ]=[S i1 S i2 …S i8 ]

[0020] In the formula, Qi S is the temporary drop data i The waveform set after the transformer transfer transformation; S ij is the transformer matrix T j Multiply the temporary drop data S i The sag waveform obtained later, j is 1, 2…8;

[0021] S22: The waveform similarity matrix is ​​obtained by using the Hausdorff distance between the reference data and the non-reference data:

[0022] D i =[d(S i1 ,S z )d(S i2 ,S z )…d(S i8 ,S z )] 1×8

[0023] In the formula, S z It is the reference temporary drop recording wave data;

[0024] S23: Let the similarity ratio between the reference data and its own waveform be 1, and obtain the waveform similarity between the reference data for homology identification and n-1 non-reference data as:

[0025] (D) n×8 =[D1D2…D n ] T

[0026] Where D n The waveform similarity matrix is ​​obtained by using the Hausdorff distance between the benchmark sag recording data and the non-benchmark data.

[0027] In an example of the present invention, in step S20, calculating the sag duration using the effective value absolute difference method includes the following steps:

[0028] S24: Use the effective value absolute difference method to calculate the start time and the sag time of the voltage sag, and calculate the duration of the voltage sag as:

[0029]

[0030] In the formula, k e is the sampling point at the end of the sag, k s is the sampling point at the start of the sag; k n is the total number of sampling points of the sag waveform; t is the duration of the sag, in seconds;

[0031] S25: Calculate the start and end time of n voltage sag data to obtain the duration characteristic matrix of n voltage sag data:

[0032] L n×1 =[t1t2…t n ] T

[0033] In the formula, t n is the duration of the nth recorded data.

[0034] In an example of the present invention, in step S30, the method of using linear interpolation to accurately calculate the phase jump value during the sag process includes the following steps:

[0035] S31: Calculate the approximate zero-crossing point during the voltage sag process, and determine whether there is voltage distortion or oscillation at the starting position of the voltage sag, and if so, exclude the oscillation zero-crossing point;

[0036] S32: Detect and select sampling points adjacent to the approximate zero-crossing point and with opposite voltage signs, and use linear interpolation to find each zero-crossing point for calculation;

[0037] S33: Calculate the phase jump value of the zero-crossing point of the sag signal. The calculation formula is as follows:

[0038]

[0039] Where, b = 2, 3, 4, ..., m; It refers to the phase jump value of the bth zero-crossing point of the sag signal; n b is the sampling point number of the zero-crossing point; N is the total number of sampling points; m refers to the total number of zero-crossing points of the sag signal;

[0040] S34: The phase jump angle α of the temporary drop without passing through the transformer is

[0041]

[0042] Where: R S is the system resistance, R F is the system reactance, X S is the fault line resistance, X F is the fault line reactance;

[0043] S35: Based on the expression of the temporary phase voltage drop without being transmitted through the transformer, multiply it with the phase voltage equivalent transfer matrix T1~T8 to obtain the effective value of the three-phase voltage transmitted through the transformer, derive the phase jump transmission law in the three cases of single-phase short circuit, two-phase short circuit and two-phase grounding fault, and obtain the expression of phase jump transmission between multi-stage transformers.

[0044] In one example of the present invention, in step S40, it includes: using the clustering data points C1 to C n , together constitute the homology recognition clustering matrix Zn×4 :

[0045]

[0046] Where: C is the cluster data point vector of the measured data; D, L, and β represent the first, second, and third eigenvalues ​​of the temporary drop data, respectively.

[0047] In an example of the present invention, in the step S50, the DENCLUE clustering parameters are further optimized to adapt to the actual voltage sag homology identification work requirements, specifically including the following steps:

[0048] S51: The density entropy formula in kernel density estimation is defined as follows. When H takes the minimum value, the corresponding σ is the optimal value of the parameter σ in the temporary drop data set:

[0049]

[0050] Where p f (x μ ) is the data point x μ The density of , w is the number of data points in the data set;

[0051] S52: Define variable merging threshold ξ μυ To determine whether the centrally defined clusters can be merged, the expression is as follows:

[0052]

[0053] Where: and is the kernel density estimate of the density attractor corresponding to the two center-defined clusters μ and υ, and r is the dimension of the data space.

[0054] Another object of the present invention is to provide a voltage sag homology identification system, comprising:

[0055] A data monitoring module configured to obtain voltage sag monitoring data recorded by a power grid voltage sag monitoring device;

[0056] A first calculation module is configured to measure the waveform similarity between multiple groups of sag monitoring data based on the propagation law of the voltage sag waveform between the multi-stage transformers using the Hausdorff distance algorithm, and calculate the sag duration using the effective value absolute difference method;

[0057] The second calculation module is configured to derive a transfer formula of a phase jump value through a transformer, accurately calculate the phase jump value during a sag using a linear interpolation method, and calculate the similarity of phase jump values ​​of multiple sag waveforms using a Euclidean distance;

[0058] A matrix construction module is configured to form voltage sag data points based on voltage sag data characteristic indicators obtained by the data monitoring module, the first calculation module and the second calculation module, including waveform similarity, duration and phase jump three-dimensional characteristics, to jointly form a homologous identification clustering matrix;

[0059] The data output module is configured to count the number of clusters of clustering results based on the color of data points in the point cloud image, and output the number of temporarily dropped homologous data groups and the clustering results.

[0060] In one example of the present invention, the data monitoring module includes:

[0061] A data selection unit is configured to select n voltage sag data monitored within a period of time, select a voltage sag data with the earliest triggering time as the reference data, and the remaining n-1 data as the non-reference data;

[0062] A data comparison unit is configured to compare the sag amplitudes of the voltages of the phases, and select the phase with the lowest sag amplitude as the sag phase;

[0063] The data acquisition unit is configured to retain the temporary sag phase in the temporary sag monitoring data and remove other data to obtain the temporary sag monitoring data.

[0064] In one example of the present invention, the first calculation module includes:

[0065] The sag waveform calculation unit is configured to derive the voltage sag phase voltage transfer matrix T1-T8 under the action of different types and numbers of transformers based on the propagation characteristics of the voltage sag between the multi-stage transformers, and multiply the three-phase voltage effective values ​​of the non-reference data in the sag data by T1-T8 to obtain all transformed sag waveform sets:

[0066] [Q i ]=[S i1 S i2 …S i8 ]

[0067] In the formula, Q i S is the temporary drop data i The waveform set after the transformer transfer transformation; S ij is the transformer matrix T j Multiply the temporary drop data S i The sag waveform obtained later, j is 1, 2…8;

[0068] The first similarity calculation unit is configured to obtain a waveform similarity matrix from the Hausdorff distance between the reference data and the non-reference data:

[0069] D i =[d(S i1,S z )d(S i2 ,S z )…d(S i8 ,S z )] 1×8

[0070] In the formula, S z It is the reference temporary drop recording wave data;

[0071] The second similarity calculation unit is configured to set the waveform similarity ratio between the reference data and its own waveform to be 1, and obtain the waveform similarity between the reference data for homology identification and n-1 non-reference data as follows:

[0072] (D) n×8 =[D1D2…D n ] T

[0073] Where D n The waveform similarity matrix is ​​obtained by using the Hausdorff distance between the benchmark sag recording data and the non-benchmark data.

[0074] The best embodiment for carrying out the present invention will be described in more detail below with reference to the accompanying drawings so that the features and advantages of the present invention can be easily understood. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings of the embodiments of the present invention, wherein the drawings are only used to illustrate some embodiments of the present invention, but not to limit all embodiments of the present invention thereto.

[0076] Figure 1 It is a schematic diagram of the overall process of the method for identifying the same source of voltage sag according to an embodiment of the present invention;

[0077] Figure 2 The figure is a flow chart of phase jump value calculation according to an embodiment of the present invention. DETAILED DESCRIPTION

[0078] In order to make the purpose, technical solution and advantages of the technical solution of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described in conjunction with the drawings of specific embodiments of the present invention. The same figure marks in the drawings represent the same parts. It should be noted that the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the described embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0079] Unless otherwise defined, the technical terms or scientific terms used herein shall have the usual meanings understood by persons of ordinary skill in the field to which the present invention belongs. The words "first", "second" and similar words used in the patent application specification and claims of the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. Similarly, words such as "one" or "one" do not necessarily indicate a quantity limitation. Words such as "include" or "comprise" mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Words such as "connect" or "connected" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0080] According to a first aspect of the present invention, a method for identifying a voltage sag from the same source is provided. Figure 1 As shown, the following steps are included:

[0081] S10: Acquire the voltage sag monitoring data recorded by the grid voltage sag monitoring device;

[0082] S20: Based on the propagation law of voltage sag waveforms between multi-stage transformers, the Hausdorff distance algorithm is used to measure the waveform similarity between multiple groups of sag monitoring data, and the effective value absolute difference method is used to calculate the sag duration;

[0083] S30: deriving a transfer formula of the phase jump value through the transformer, using a linear interpolation method to accurately calculate the phase jump value during the sag process, and using the Euclidean distance to calculate the similarity of the phase jump values ​​of multiple sag waveforms;

[0084] S40: Based on the above steps S10 to S30, characteristic indicators of voltage sag data are obtained, including waveform similarity, duration and phase jump three-dimensional characteristics, to form voltage sag data points, which together constitute a homologous identification clustering matrix;

[0085] S50: Count the number of clusters of clustering results based on the colors of the data points in the point cloud image, and output the number of temporarily homologous data groups and the clustering results.

[0086] Firstly, this method constructs a three-dimensional feature matrix and uses characteristic indicators such as the similarity, duration and phase jump of the voltage sag waveform to describe and represent the voltage sag event. This method can more comprehensively represent the characteristic information of the voltage sag event, which is beneficial to subsequent data processing and analysis.

[0087] Secondly, by deriving the transmission law of voltage sag characteristic indicators through multi-stage transformers, this method can more accurately estimate the impact of voltage sag events on various nodes in the power grid, which helps power grid operators to better understand the transmission and diffusion of voltage sag events in the system and take targeted measures to adjust and repair them.

[0088] In addition, this method also proposes a waveform similarity calculation method based on Hausdorff distance and a phase jump similarity measurement method based on Euclidean distance, which can quantitatively evaluate the similarity between different voltage sag events, help to quickly identify and classify voltage sag events, and enable power grid operators to quickly and accurately identify important events when faced with a large amount of voltage sag data, and take corresponding treatment measures.

[0089] Finally, this method uses the improved DENCLUE clustering method to perform cluster analysis on the measurement results, which can classify similar voltage sag events together, thereby further mining the hidden laws and correlations in the data, and providing valuable reference for power grid fault detection, fault warning and fault diagnosis.

[0090] To sum up, the role and significance of this method lies in providing a more accurate and efficient voltage sag characteristic analysis method, which can overcome the problems existing in traditional methods and improve the accuracy and efficiency of power grid voltage sag monitoring data processing, which has important practical significance for power grid operation and maintenance, and helps to improve the reliability and stability of power grid operation.

[0091] In an example of the present invention, step S10 includes:

[0092] S11: Select n voltage sag data monitored within a period of time, select the voltage sag data with the earliest triggering time as the reference data, and the remaining n-1 data as non-reference data;

[0093] S12: comparing the voltage sag amplitudes of each phase, and selecting the phase with the lowest voltage sag amplitude as the voltage sag phase;

[0094] S13: retain the sag phase in the sag monitoring data and remove other data to obtain the sag monitoring data.

[0095] In one example of the present invention, in step S20, using the Hausdorff distance algorithm to measure the waveform similarity between multiple groups of sag monitoring data includes the following steps:

[0096] S21: Based on the propagation characteristics of voltage sag between multi-stage transformers, the voltage sag phase voltage transfer matrix T1~T8 under the action of different types and numbers of transformers is derived, and the effective values ​​of the three-phase voltage of non-reference data in the sag data are multiplied by T1~T8 respectively to obtain the set of all transformed sag waveforms:

[0097] [Q i ]=[S i1 S i2 …S i8 ]

[0098] In the formula, Q i S is the temporary drop data i The waveform set after the transformer transfer transformation; S ij is the transformer matrix T j Multiply the temporary drop data S i The sag waveform obtained later, j is 1, 2…8;

[0099] S22: The waveform similarity matrix is ​​obtained by using the Hausdorff distance between the reference data and the non-reference data:

[0100] D i =[d(S i1 ,S z ) d(S i2 ,S z )…d(S i8 ,S z )] 1×8

[0101] In the formula, S z It is the reference temporary drop recording wave data;

[0102] S23: Let the similarity ratio between the reference data and its own waveform be 1, and obtain the waveform similarity between the reference data for homology identification and n-1 non-reference data as:

[0103] (D) n×8 =[D1D2…D n ] T

[0104] Where D n The waveform similarity matrix is ​​obtained by using the Hausdorff distance between the benchmark sag recording data and the non-benchmark data.

[0105] Specifically, based on the propagation characteristics of voltage sag between multi-stage transformers, the voltage sag phase voltage transfer matrix T1~T8 under the action of different types and numbers of transformers is derived. The specific expression is shown in Table 1:

[0106] Table 1 Phase voltage equivalent transfer matrix

[0107]

[0108] The time of the first sampling point on the leftmost side of the cycle before the start of the temporary sag waveform is recorded as 0, and the time axis is translated rightward to the last sampling point of the cycle after the temporary sag ends to obtain the temporary sag data S1~S1 at the same scale coordinate. n , multiply the three-phase voltage effective values ​​of the reference data in the sag data by T1~T8 respectively to obtain the set of all transformed sag waveforms:

[0109] [Q i ]=[S i1 S i2 …S i8 ]

[0110] Where: Q i S is the temporary drop data i The waveform set after the transformer transfer transformation; S i8 The transformer matrix T8 is multiplied by the temporary drop data S i The resulting sag waveform.

[0111] Select the phase sag data with the lowest effective value. For non-reference sag S i ={S i1 ,…,S i8} and the reference sag recording data S z ={S z1 ,…,S z8}, the Hausdorff distance between them is:

[0112] H(S i ,S z )=max[h(S i ,S z ),h(S z ,S i )]

[0113]

[0114] Where: |·| is the data S i With S z The distance norm of .

[0115] The waveform similarity matrix is ​​obtained by the Hausdorff distance between the benchmark data and the non-benchmark data:

[0116] D i =[d(S i1 ,S z )d(S i2 ,S z )…d(S i8 ,S z)] 1×8

[0117] Let the similarity ratio between the reference data and its own waveform be 1, and the waveform similarity between the reference data for homology identification and (n-1) non-reference data is:

[0118] (D) n×8 =[D1D2…D n ] T

[0119] In an example of the present invention, in step S20, calculating the sag duration using the effective value absolute difference method includes the following steps:

[0120] S24: Use the effective value absolute difference method to calculate the start time and the sag time of the voltage sag, and calculate the duration of the voltage sag as:

[0121]

[0122] In the formula, k e is the sampling point at the end of the sag, k s is the sampling point at the start of the sag; k n is the total number of sampling points of the sag waveform; t is the duration of the sag, in seconds;

[0123] S25: Calculate the start and end time of n voltage sag data to obtain the duration characteristic matrix of n voltage sag data:

[0124] L n×1 =[t1t2…t n ] T

[0125] In the formula, t n is the duration of the nth recorded data.

[0126] In one example of the present invention, in step S30, Figure 2 As shown, the method of using linear interpolation to accurately calculate the phase jump value during the sag process includes the following steps:

[0127] S31: Calculate the approximate zero-crossing point during the voltage sag process, and determine whether there is voltage distortion or oscillation at the starting position of the voltage sag, and if so, exclude the oscillation zero-crossing point;

[0128] S32: Detect and select sampling points adjacent to the approximate zero-crossing point and with opposite voltage signs, and use linear interpolation to find each zero-crossing point for calculation;

[0129] S33: Calculate the phase jump value of the zero-crossing point of the sag signal. The calculation formula is as follows:

[0130]

[0131] Where, b = 2, 3, 4, ..., m; It refers to the phase jump value of the bth zero-crossing point of the sag signal; n b is the sampling point number of the zero-crossing point; N is the total number of sampling points; m refers to the total number of zero-crossing points of the sag signal;

[0132] S34: The phase jump angle α of the temporary drop without passing through the transformer is

[0133]

[0134] Where: R S is the system resistance, R F is the system reactance, X S is the fault line resistance, X F is the fault line reactance;

[0135] S35: Based on the expression of the temporarily dropped phase voltage without being transmitted through the transformer, it is multiplied by the phase voltage equivalent transfer matrix T1~T8 to obtain the effective value of the three-phase voltage transmitted through the transformer, and the phase jump transmission law under the three conditions of single-phase short circuit, two-phase short circuit and two-phase grounding fault is derived. The expression of the phase jump transmission between multi-stage transformers is obtained as shown in Table 2:

[0136] Table 2 The changing rules of voltage sag phase jump passing through a single or multiple transformers

[0137]

[0138]

[0139] The three features D, L and S of voltage sag data are obtained by the above method. The three features D, L and S of n sag measured data constitute the data points for density clustering:

[0140] C i =[i D i L i S i ]

[0141] Where: C i is the clustered data point vector of the ith measured data; i is the number of the voltage sag measured data; D i , L i , S i They represent the 1st, 2nd, and 3rd eigenvalues ​​of the i-th data respectively.

[0142] In one example of the present invention, in step S40, it includes: using the clustering data points C1 to C n, together constitute the homology recognition clustering matrix Z n×4 :

[0143]

[0144] Where: C is the cluster data point vector of the measured data; D, L, and β represent the first, second, and third eigenvalues ​​of the temporary drop data, respectively.

[0145] In an example of the present invention, in the step S50, the DENCLUE clustering parameters are further optimized to adapt to the actual voltage sag homology identification work requirements, specifically including the following steps:

[0146] S51: The density entropy formula in kernel density estimation is defined as follows. When H takes the minimum value, the corresponding σ is the optimal value of the parameter σ in the temporary drop data set:

[0147]

[0148] Where p f (x μ ) is the data point x μ The density of , w is the number of data points in the data set;

[0149] When the data set is constant, the density entropy is only related to σ. The kernel density estimate of each data point is converted according to the above formula. When H takes the minimum value, the corresponding σ is the optimal value of the parameter σ in the temporary drop data set.

[0150] It should be noted that in view of the large difference in data point density and the difficulty in selecting a suitable step size parameter in the actual voltage sag homology identification work, an adaptive parameter δ solution method is proposed. The specific implementation steps are as follows:

[0151] Select a data point x in the data set, set x(0) = x, δ x (0) = 0.5 × d min (x), k = 1; when k ≥ 1, calculate the gradient of x(k-1) ▽f(x(k-1)); get x(k) by the following formula:

[0152]

[0153] Find the data point x that is closest to x(k) μ , δ x (k) = 0.5 × d min (x i ); if f(x(k-1))>f(x(k)), then the density of data point x attracts point x * Labeled as x(k-1), that is, x * =x(k-1); otherwise k=k+1.

[0154] S52: Define variable merging threshold ξ μυ To determine whether the centrally defined clusters can be merged, and is the kernel density estimate of the density attraction points corresponding to the two center-defined clusters μ and υ. If there is a path between the two center-defined clusters, and the kernel density estimates on the path are all greater than ξ μυ , then the two center-defined clusters can be merged, ξ μυ The value of is:

[0155]

[0156] Where: and is the kernel density estimate of the density attractor corresponding to the two center-defined clusters μ and υ, and r is the dimension of the data space.

[0157] When implementing it, Figure 1 As shown, use the DENCLUE clustering algorithm, set the clustering parameters, and enter Z n×4 , identify the same source of voltage sag, and output the number of data groups with the same source of voltage sag and the clustering results.

[0158] The homologous identification method proposes to construct a three-dimensional feature matrix based on voltage sag waveform similarity, duration and phase jump characteristic indicators, derives the transmission law of voltage sag characteristic indicators through multi-stage transformers, proposes a waveform similarity calculation method based on Hausdorff distance and a phase jump similarity measurement method based on Euclidean distance, and finally uses an improved DENCLUE clustering method to perform cluster analysis on the measurement results, thereby overcoming the problem of over-estimation of power grid voltage sag level caused by redundant voltage sag recording data, significantly reducing the calculation amount and complexity of voltage sag monitoring data processing. The present invention overcomes the problem of too single sag characteristic indicator in the existing voltage sag homologous identification method and improves the accuracy of voltage sag homologous identification.

[0159] A voltage sag homologous identification system according to a second aspect of the present invention comprises:

[0160] A data monitoring module configured to obtain voltage sag monitoring data recorded by a power grid voltage sag monitoring device;

[0161] A first calculation module is configured to measure the waveform similarity between multiple groups of sag monitoring data based on the propagation law of the voltage sag waveform between the multi-stage transformers using the Hausdorff distance algorithm, and calculate the sag duration using the effective value absolute difference method;

[0162] The second calculation module is configured to derive a transfer formula of a phase jump value through a transformer, accurately calculate the phase jump value during a sag using a linear interpolation method, and calculate the similarity of phase jump values ​​of multiple sag waveforms using a Euclidean distance;

[0163] A matrix construction module is configured to form voltage sag data points based on voltage sag data characteristic indicators obtained by the data monitoring module, the first calculation module and the second calculation module, including waveform similarity, duration and phase jump three-dimensional characteristics, to jointly form a homologous identification clustering matrix;

[0164] The data output module is configured to count the number of clusters of clustering results based on the color of data points in the point cloud image, and output the number of temporarily dropped homologous data groups and the clustering results.

[0165] First, the system constructs a three-dimensional feature matrix and uses characteristic indicators such as the similarity, duration and phase jump of the voltage sag waveform to describe and represent the voltage sag event. It can more comprehensively represent the characteristic information of the voltage sag event, which is conducive to subsequent data processing and analysis.

[0166] Secondly, by deriving the transmission rules of voltage sag characteristic indicators through multi-stage transformers, the system can more accurately estimate the impact of voltage sag events on various nodes in the power grid, which helps power grid operators better understand the transmission and spread of voltage sag events in the system and take targeted measures to adjust and repair them.

[0167] In addition, the system also proposes a waveform similarity calculation method based on Hausdorff distance and a phase jump similarity measurement method based on Euclidean distance, which can quantitatively evaluate the similarity between different voltage sag events, help to quickly identify and classify voltage sag events, and enable power grid operators to quickly and accurately identify important events when faced with a large amount of voltage sag data, and take corresponding treatment measures.

[0168] Finally, the system uses the improved DENCLUE clustering method to perform cluster analysis on the measurement results, which can classify similar voltage sag events together, thereby further mining the hidden rules and correlations in the data, and providing valuable reference for power grid fault detection, fault warning and fault diagnosis.

[0169] In summary, the role and significance of this system is to provide a more accurate and efficient voltage sag characteristic analysis method, which can overcome the problems existing in traditional methods and improve the accuracy and efficiency of grid voltage sag monitoring data processing. This has important practical significance for grid operation and maintenance, and helps to improve the reliability and stability of grid operation.

[0170] In one example of the present invention, the data monitoring module includes:

[0171] A data selection unit is configured to select n voltage sag data monitored within a period of time, select a voltage sag data with the earliest triggering time as the reference data, and the remaining n-1 data as the non-reference data;

[0172] A data comparison unit is configured to compare the sag amplitudes of the voltages of the phases, and select the phase with the lowest sag amplitude as the sag phase;

[0173] The data acquisition unit is configured to retain the temporary sag phase in the temporary sag monitoring data and remove other data to obtain the temporary sag monitoring data.

[0174] In one example of the present invention, the first calculation module includes:

[0175] The sag waveform calculation unit is configured to derive the voltage sag phase voltage transfer matrix T1-T8 under the action of different types and numbers of transformers based on the propagation characteristics of the voltage sag between the multi-stage transformers, and multiply the three-phase voltage effective values ​​of the non-reference data in the sag data by T1-T8 to obtain all transformed sag waveform sets:

[0176] [Q i ]=[S i1 S i2 …S i8 ]

[0177] In the formula, Q i S is the temporary drop data i The waveform set after the transformer transfer transformation; S ij is the transformer matrix T j Multiply the temporary drop data S i The sag waveform obtained later, j is 1, 2…8;

[0178] The first similarity calculation unit is configured to obtain a waveform similarity matrix based on the Hausdorff distance between the reference data and the non-reference data:

[0179] D i =[d(S i1 ,S z )d(S i2 ,S z )…d(S i8 ,S z )] 1×8

[0180] In the formula, S z It is the reference temporary drop recording wave data;

[0181] The second similarity calculation unit is configured to set the waveform similarity ratio between the reference data and its own waveform to be 1, and obtain the waveform similarity between the reference data for homology identification and n-1 non-reference data as follows:

[0182] (D) n×8 =[D1D2…D n ] T

[0183] Where D n The waveform similarity matrix is ​​obtained by using the Hausdorff distance between the benchmark sag recording data and the non-benchmark data.

[0184] The homologous identification system proposes to construct a three-dimensional feature matrix based on voltage sag waveform similarity, duration and phase jump characteristic indicators, derives the transmission law of voltage sag characteristic indicators through multi-stage transformers, proposes a waveform similarity calculation method based on Hausdorff distance and a phase jump similarity measurement method based on Euclidean distance, and finally uses an improved DENCLUE clustering method to perform cluster analysis on the measurement results, thereby overcoming the problem of over-estimation of power grid voltage sag level caused by redundant voltage sag recording data, significantly reducing the calculation amount and complexity of voltage sag monitoring data processing. The present invention overcomes the problem of too single sag characteristic indicator in the existing voltage sag homologous identification method, and improves the accuracy of voltage sag homologous identification.

[0185] The exemplary implementation of the voltage sag homologous identification method and system proposed in the present invention is described in detail above with reference to the preferred embodiments. However, it can be understood by those skilled in the art that, without departing from the concept of the present invention, various modifications and variations can be made to the above-mentioned specific embodiments, and various technical features and structures proposed in the present invention can be combined in various ways without exceeding the protection scope of the present invention, which is determined by the attached claims.

Claims

1. A method for identifying the same source of voltage sag, characterized in that: The steps include: S10: Acquire the voltage sag monitoring data recorded by the grid voltage sag monitoring device; S20: Based on the propagation law of voltage sag waveforms between multi-stage transformers, the Hausdorff distance algorithm is used to measure the waveform similarity between multiple groups of sag monitoring data, and the effective value absolute difference method is used to calculate the sag duration; S30: deriving a transfer formula of the phase jump value through the transformer, using a linear interpolation method to accurately calculate the phase jump value during the sag process, and using the Euclidean distance to calculate the similarity of the phase jump values ​​of multiple sag waveforms; S40: Based on the above steps S10 to S30, characteristic indicators of voltage sag data are obtained, including waveform similarity, duration and phase jump three-dimensional characteristics, to form voltage sag data points, which together constitute a homologous identification clustering matrix; S50: Counting the number of clusters of clustering results based on the colors of the data points in the point cloud image, and outputting the number of temporarily decreasing homologous data groups and the clustering results; Wherein, the step S10 includes: S11: Select n voltage sag data monitored within a period of time, select the voltage sag data with the earliest triggering time as the reference data, and the remaining n-1 data as non-reference data; S12: comparing the voltage sag amplitudes of each phase, and selecting the phase with the lowest voltage sag amplitude as the voltage sag phase; S13: retaining the sag phase in the sag monitoring data and removing other data to obtain the sag monitoring data; Wherein, in step S20, using the Hausdorff distance algorithm to measure the waveform similarity between multiple groups of sag monitoring data includes the following steps: S21: Based on the propagation characteristics of voltage sag between multi-stage transformers, the voltage sag phase voltage transfer matrix T1~T8 under the action of different types and numbers of transformers is derived, and the effective values ​​of the three-phase voltage of non-reference data in the sag data are multiplied by T1~T8 respectively to obtain the set of all transformed sag waveforms: [Q i ]=[S i1 S i2 …S i8 ] In the formula, Q i S is the temporary drop data i The waveform set after the transformer transfer transformation; S ij is the transformer matrix T j Multiply the temporary drop data S i The sag waveform obtained later, j is 1, 2…8; S22: The waveform similarity matrix is ​​obtained by using the Hausdorff distance between the reference data and the non-reference data: D i =[d(S i1 ,S z )d(S i2 ,S z )…d(S i8 ,S z )] 1×8 In the formula, S z It is the reference temporary drop recording wave data; S23: Let the similarity ratio between the reference data and its own waveform be 1, and obtain the waveform similarity between the reference data for homology identification and n-1 non-reference data as: (D) n×8 =[D1D2…D n ] T Where D n The waveform similarity matrix is ​​obtained by using the Hausdorff distance between the benchmark sag recording data and the non-benchmark data.

2. The method for identifying the same source of voltage sag according to claim 1, characterized in that: In step S20, calculating the sag duration using the effective value absolute difference method includes the following steps: S24: Use the effective value absolute difference method to calculate the start time and the sag time of the voltage sag, and calculate the duration of the voltage sag as: In the formula, k e is the sampling point at the end of the sag, k s is the sampling point at the start of the sag; k n is the total number of sampling points of the sag waveform; t is the duration of the sag, in seconds; S25: Calculate the start and end time of n voltage sag data to obtain the duration characteristic matrix of n voltage sag data: L n×1 =[t1t2…t n ] T Where, t n is the duration of the nth recorded data.

3. The method for identifying the same source of voltage sag according to claim 1, characterized in that: In the step S30, the method of using linear interpolation to accurately calculate the phase jump value during the sag process includes the following steps: S31: Calculate the approximate zero-crossing point during the voltage sag process, and determine whether there is voltage distortion or oscillation at the starting position of the voltage sag, and if so, exclude the oscillation zero-crossing point; S32: Detect and select sampling points adjacent to the approximate zero-crossing point and with opposite voltage signs, and use linear interpolation to find each zero-crossing point for calculation; S33: Calculate the phase jump value of the zero-crossing point of the sag signal. The calculation formula is as follows: Where, b = 2, 3, 4, ..., m; It refers to the phase jump value of the bth zero-crossing point of the sag signal; n b is the sampling point number of the zero-crossing point; N is the total number of sampling points; m refers to the total number of zero-crossing points of the sag signal; S34: The phase jump angle α of the temporary drop without passing through the transformer is Where: R S is the system resistance, R F is the system reactance, X S is the fault line resistance, X F is the fault line reactance; S35: Based on the expression of the temporary phase voltage drop without being transmitted through the transformer, multiply it with the phase voltage equivalent transfer matrix T1~T8 to obtain the effective value of the three-phase voltage transmitted through the transformer, derive the phase jump transmission law in the three cases of single-phase short circuit, two-phase short circuit and two-phase grounding fault, and obtain the expression of phase jump transmission between multi-stage transformers.

4. The method for identifying the same source of voltage sag according to claim 1, characterized in that: In the step S40, the method includes: using the clustered data points C1 to C n , together constitute the homology recognition clustering matrix Z n×4 : Where: C is the cluster data point vector of the measured data; D, L, and β represent the first, second, and third eigenvalues ​​of the temporary drop data, respectively.

5. The method for identifying the same source of voltage sag according to claim 1, characterized in that: In the step S50, the DENCLUE clustering parameters are optimized to adapt to the actual voltage sag homology identification work needs, which specifically includes the following steps: S51: The density entropy formula in kernel density estimation is defined as follows. When H takes the minimum value, the corresponding σ is the optimal value of the parameter σ in the temporary drop data set: Where p f (x μ ) is the data point x μ The density of , w is the number of data points in the data set; S52: Define variable merging threshold ξ μυ To determine whether the centrally defined clusters can be merged, the expression is as follows: Where: and is the kernel density estimate of the density attractor corresponding to the two center-defined clusters μ and υ, and r is the dimension of the data space.

6. A voltage sag homologous identification system, characterized in that: include: A data monitoring module configured to obtain voltage sag monitoring data recorded by a power grid voltage sag monitoring device; A first calculation module is configured to measure the waveform similarity between multiple groups of sag monitoring data based on the propagation law of the voltage sag waveform between the multi-stage transformers using the Hausdorff distance algorithm, and calculate the sag duration using the effective value absolute difference method; The second calculation module is configured to derive a transfer formula of a phase jump value through a transformer, accurately calculate the phase jump value during a sag using a linear interpolation method, and calculate the similarity of phase jump values ​​of multiple sag waveforms using a Euclidean distance; A matrix construction module is configured to form voltage sag data points based on voltage sag data characteristic indicators obtained by the data monitoring module, the first calculation module and the second calculation module, including waveform similarity, duration and phase jump three-dimensional characteristics, to jointly form a homologous identification clustering matrix; A data output module is configured to count the number of clusters of clustering results based on the color of data points in the point cloud image, and output the number of temporarily dropped homologous data groups and clustering results; The data monitoring module comprises: A data selection unit is configured to select n voltage sag data monitored within a period of time, select a voltage sag data with the earliest triggering time as the reference data, and the remaining n-1 data as the non-reference data; A data comparison unit is configured to compare the sag amplitudes of the voltages of the phases, and select the phase with the lowest sag amplitude as the sag phase; a data acquisition unit configured to retain the sag phase in the sag monitoring data and remove other data to obtain the sag monitoring data; The first calculation module includes: The sag waveform calculation unit is configured to derive the voltage sag phase voltage transfer matrix T1-T8 under the action of different types and numbers of transformers based on the propagation characteristics of the voltage sag between the multi-stage transformers, and multiply the three-phase voltage effective values ​​of the non-reference data in the sag data by T1-T8 to obtain all transformed sag waveform sets: [Q i ]=[S i1 S i2 …S i8 ] In the formula, Q i S is the temporary drop data i The waveform set after the transformer transfer transformation; S ij is the transformer matrix T j Multiply the temporary drop data S i The sag waveform obtained later, j is 1, 2…8; The first similarity calculation unit is configured to obtain a waveform similarity matrix from the Hausdorff distance between the reference data and the non-reference data: D i =[d(S i1 ,S z ) d(S i2 ,S z )…d(S i8 ,S z )] 1×8 In the formula, S z It is the reference temporary drop recording wave data; The second similarity calculation unit is configured to set the waveform similarity ratio between the reference data and its own waveform to be 1, and obtain the waveform similarity between the reference data for homology identification and n-1 non-reference data as follows: (D) n×8 =[D1D2…D n ] T Where D n The waveform similarity matrix is ​​obtained by using the Hausdorff distance between the benchmark sag recording data and the non-benchmark data.

Images