Power supply quality coupling analysis method for source load disturbance

Abstract

The application relates to the technical field of measuring electric variables, and discloses a power supply quality coupling analysis method for source-load disturbance, which comprises the following steps: collecting electric data and environmental data of nodes, the nodes including transformer substation low-voltage side nodes, photovoltaic nodes and load nodes, the electric data including current, voltage and power; marking target transformer substation low-voltage side nodes, target photovoltaic nodes and target load nodes, respectively determining the correlation coefficients of the target transformer substation low-voltage side nodes and the target load nodes and the correlation coefficients of the target transformer substation low-voltage side nodes and the target photovoltaic nodes, and obtaining abnormal characteristic indexes of the nodes; screening optimal particles, obtaining clustering clusters of abnormal characteristic value sequences, and respectively determining the influence weights of the target load nodes and the target photovoltaic nodes; and combining the abnormal characteristic indexes to obtain power supply quality coupling analysis results for source-load disturbance. The application aims to ensure stable power supply quality when new source-load disturbance occurs.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of measuring electrical variables, and particularly relates to a power supply quality coupling analysis method for source-load disturbance. BACKGROUND

[0002] Nowadays, the traditional centralized power generation mode is gradually changing to distributed power generation. However, in the case of high proportion of new energy access, the output of distributed power generation is intermittent and uncertain, the source-load coupling interaction in the distribution network becomes more complex, and the power quality disturbance problem is increasingly prominent. The transient power quality disturbance in the power quality disturbance has the characteristics of short duration, high randomness, diversity and strong complexity, and detection and analysis are more difficult.

[0003] The traditional power quality monitoring means mainly aims at steady-state disturbance and analyzes based on a single disturbance source, and has limited ability to capture transient disturbances, which is difficult to effectively deal with the complexity brought by new source-load disturbance. The neglect of the interaction between multiple distributed power sources and the comprehensive influence of distributed power sources on the distribution network often leads to lag and inaccuracy of power supply control measures, thereby affecting the overall stability and safety of the distribution network. SUMMARY

[0004] The present application provides a power supply quality coupling analysis method for source-load disturbance to solve the problem of unstable power supply quality caused by neglecting the interaction between distributed power sources during new source-load disturbance. The technical solution adopted is as follows:

[0005] One embodiment of the present application provides a power supply quality coupling analysis method for source-load disturbance, which comprises the following steps:

[0006] Collecting electrical and environmental data at different collection times of nodes in the distribution network, the nodes including substation low-voltage side nodes, photovoltaic nodes and load nodes, the electrical data including current, voltage and power;

[0007] Any one of the substation low-voltage side nodes, photovoltaic nodes and load nodes is respectively denoted as a target substation low-voltage side node, a target photovoltaic node and a target load node, the correlation coefficient of the target substation low-voltage side node and the target load node is determined, the similar nodes of each node are determined respectively, the abnormal feature index and the abnormal feature value of the node at the same collection time are determined according to all electrical data of the node at the same collection time, the abnormal feature value sequence of the node is determined, and the correlation coefficient of the target substation low-voltage side node and the target photovoltaic node is determined according to the similarity between all electrical data of the target photovoltaic node and the target substation low-voltage side node, and the similarity between the abnormal feature values of the target substation low-voltage side node and the similar nodes of the target photovoltaic node;

[0008] According to the electrical data, environmental data and abnormal characteristic value, the optimal particle in the particle composed of the node and similar nodes of the node is screened, the abnormal characteristic value sequence of all nodes is clustered, the clustering cluster is obtained, the influence weight of the target load node is determined according to the correlation coefficient of all load nodes in the same clustering cluster as the target load node, and the influence weight of the target photovoltaic node is determined according to the difference between the environmental data of all nodes in the optimal particle and the target photovoltaic node, the difference between the power, and the correlation coefficient of the target photovoltaic node;

[0009] According to the abnormal characteristic index of all nodes and all influence weights, the power supply quality coupling analysis result facing source load disturbance is obtained.

[0010] Further, the correlation coefficient determination method of the target substation low-voltage side node and the target load node is:

[0011] Arranging all electrical data of the same kind of the target substation low-voltage side node according to the collection time of the electrical data, the electrical data sequence of the same kind of the target substation low-voltage side node is obtained.

[0012] Arranging all electrical data of the same kind of the target load node according to the collection time of the electrical data, the electrical data sequence of the same kind of the target load node is obtained.

[0013] The absolute value of the similarity between the electrical data sequence of the same kind of the target substation low-voltage side node and the target load node is recorded as the correlation degree of the electrical data of the same kind of the target substation low-voltage side node and the target load node, and the average value of the correlation degrees of all electrical data of the same kind of the target substation low-voltage side node and the target load node is recorded as the correlation coefficient of the target substation low-voltage side node and the target load node.

[0014] Further, the determination method of the similar node of the node is:

[0015] Any one of all nodes is recorded as a target node, and any one node different from the target node is recorded as a target comparison node.

[0016] Arranging all environmental data of the same kind of the target node according to the collection time of the environmental data, the environmental data sequence of the same kind of the target node is obtained, and arranging all environmental data of the same kind of the target comparison node according to the collection time of the environmental data, the environmental data sequence of the same kind of the target comparison node is obtained.

[0017] The dtw distance between the target node and the target comparison node is recorded as a one-class direct difference of the same kind of environment class data of the target node and the target comparison node, and the average of the differences of all the same kinds of environment class data of the target node and the target comparison node is recorded as a two-class direct difference of the target node and the target comparison node.

[0018] The similar node of the target node is determined according to the lower quartile of all the two-class direct differences corresponding to the target node.

[0019] Further, the specific method for determining the abnormal feature index and the abnormal feature value of the node at the same collection time according to all the electrical class data of the node at the same collection time, and the abnormal feature value sequence of the node comprises:

[0020] The electrical class data of the target node at any collection time is recorded as a target collection time.

[0021] The short-time flicker value, the total harmonic distortion rate and the three-phase imbalance degree of the target collection time are calculated respectively, and the results of the normalization processing of the short-time flicker value, the total harmonic distortion rate and the three-phase imbalance degree of the target collection time are all recorded as the abnormal feature index of the target collection time.

[0022] The average of all the abnormal feature indexes of the target node at the target collection time is recorded as the abnormal feature value of the target node at the target collection time.

[0023] The abnormal feature value sequence of the target node is obtained by arranging the abnormal feature values of the target node according to the collection time corresponding to the abnormal feature values.

[0024] Further, the specific method for determining the correlation coefficient between the target transformer substation low-voltage side node and the target photovoltaic node according to the similarity between all the electrical class data of the target photovoltaic node and the target transformer substation low-voltage side node, and the similarity between the abnormal feature values of the target transformer substation low-voltage side node and the target photovoltaic node and the similar node of the target photovoltaic node respectively comprises:

[0025] All the same kinds of electrical class data of the target photovoltaic node are arranged according to the collection time of the electrical class data to obtain the sequence of the same kinds of electrical class data of the target photovoltaic node; the absolute value of the similarity between the sequence of the same kinds of electrical class data of the target transformer substation low-voltage side node and the target photovoltaic node is recorded as the first correlation degree of the same kinds of electrical class data of the target transformer substation low-voltage side node and the target photovoltaic node; and the average of the first correlation degrees of all the same kinds of electrical class data of the target transformer substation low-voltage side node and the target photovoltaic node is recorded as the second correlation degree of the target transformer substation low-voltage side node and the target photovoltaic node.

[0026] According to the similarity between the abnormal characteristic value of the target transformer substation low-voltage side node and the target photovoltaic node and the similar node of the target photovoltaic node, the abnormal characteristic value similarity and the similar node abnormal characteristic value similarity between the target transformer substation low-voltage side node and the target photovoltaic node are determined respectively;

[0027] According to the second correlation degree, the abnormal characteristic value similarity, and the similar node abnormal characteristic value similarity between the target transformer substation low-voltage side node and the target photovoltaic node, the correlation coefficient between the target transformer substation low-voltage side node and the target photovoltaic node is determined, and the calculation formula is:

[0028]

[0029] In the formula, The correlation coefficient between the target transformer substation low-voltage side node and the target photovoltaic node is represented; The similar node abnormal characteristic value similarity between the target transformer substation low-voltage side node and the target photovoltaic node is represented; The abnormal characteristic value similarity between the target transformer substation low-voltage side node and the target photovoltaic node is represented; The second correlation degree between the target transformer substation low-voltage side node and the target photovoltaic node is represented.

[0030] Further, the method of determining the abnormal characteristic value similarity and the similar node abnormal characteristic value similarity between the target transformer substation low-voltage side node and the target photovoltaic node according to the similarity between the abnormal characteristic value of the target transformer substation low-voltage side node and the target photovoltaic node and the similar node of the target photovoltaic node includes the following specific methods:

[0031] The absolute value of the similarity between the similar node of the target photovoltaic node and the abnormal characteristic value sequence of the target transformer substation low-voltage side node is recorded as the abnormal characteristic value similarity between the target transformer substation low-voltage side node and the target photovoltaic node;

[0032] The average value of the absolute value of the similarity between the abnormal characteristic value sequence of all similar nodes and the abnormal characteristic value sequence of the target transformer substation low-voltage side node is recorded as the similar node abnormal characteristic value similarity between the target transformer substation low-voltage side node and the target photovoltaic node.

[0033] Further, the method of screening the optimal particle from the particles composed of nodes and similar nodes of the nodes according to the electrical data, environmental data, and abnormal characteristic values includes the following specific methods:

[0034] The node and the similar node of the node are taken as particles, the electrical class data and the environmental class data of the node are taken as the corresponding data of the node, the accumulation and minimum of the abnormal characteristic values of all collection moments are taken as the target function, and an optimization algorithm is used to obtain the optimal particle.

[0035] Further, the method for determining the influence weight of the target load node is:

[0036] The mean value of the correlation coefficients of all load nodes in the same cluster as the target load node is denoted as the correlation coefficient benchmark of the target load node, and the normalized value of the ratio of the mean value of all correlation coefficients of the target load node to the correlation coefficient benchmark is denoted as the influence weight of the target load node.

[0037] Further, the method for determining the influence weight of the target photovoltaic node according to the differences between the environmental class data of all nodes in the optimal particle and the target photovoltaic node, the differences between the powers, and the correlation coefficient of the target photovoltaic node comprises the following specific method:

[0038] The optimal power difference between the target photovoltaic node and any one node in the optimal particle is determined according to the differences between the environmental class data of the optimal particle and the target photovoltaic node and the differences between the powers.

[0039] The optimal environmental difference corresponding to the target photovoltaic node is taken as the independent variable, the optimal power difference corresponding to the target photovoltaic node is taken as the dependent variable, the difference fitting curve of the target photovoltaic node is obtained, the value of the optimal environmental difference of the node that is not a similar node in the optimal particle and the target photovoltaic node is taken as the value of the independent variable, the value of the dependent variable corresponding to the difference fitting curve is calculated, and the value of the dependent variable is denoted as the difference fitting value of the target photovoltaic node.

[0040] The optimal power difference between the node that is not a similar node in the optimal particle and the target photovoltaic node is denoted as the difference real value of the target photovoltaic node.

[0041] The ratio of the difference real value to the difference fitting value of the target photovoltaic node is denoted as the first ratio of the target photovoltaic node, and the normalized value of the product of the mean value of all correlation coefficients of the target photovoltaic node and the first ratio is denoted as the influence weight of the target photovoltaic node.

[0042] Further, the method for obtaining the power supply quality coupling analysis result facing the source load disturbance according to the abnormal characteristic indexes of all nodes and all influence weights comprises the following specific method:

[0043] According to the abnormal characteristic indexes of the substation low-voltage side node, the photovoltaic node and the load node, an abnormal characteristic index matrix is established; according to the influence weight of the substation low-voltage side node, the photovoltaic node and the load node, an influence weight matrix is established; the product of the abnormal characteristic index matrix and the influence weight matrix is denoted as a power supply quality matrix of the distribution network;

[0044] The principal component analysis algorithm is adopted to obtain a load matrix of the power supply quality matrix of the distribution network, and all the numerical values contained in the load matrix are used to screen a disturbance node, and the disturbance node is a power supply quality coupling analysis result facing source and load disturbance.

[0045] The present application has the following beneficial effects:

[0046] Since the power supply of the distributed photovoltaic power generation presents completely different trends in distribution characteristics and randomness, the stability of power generation will obviously fluctuate, and the power supply quality of the distribution network will also be affected by the coupling of various nodes, therefore, the present application evaluates the correlation between the load node and the substation low-voltage side node according to the similarity between the electrical data of the substation low-voltage side node and the load node, determines the correlation coefficient of the target substation low-voltage side node and the target load node, and further, since the photovoltaic power generation is more affected by the environment, the correlation coefficient of the substation low-voltage side node and the photovoltaic node is obtained based on the similarity evaluation between the electrical data, and further combined with the similarity between the abnormal characteristic values of the substation low-voltage side node and the photovoltaic node and the similar nodes of the photovoltaic node; the mutual influence between different nodes may lead to more serious coupling in the distribution network, affecting the regulation and control of the power supply quality, since the fluctuation of power quality is reflected through the change of power, the present application further analyzes the influence of each node on the power quality of the distribution network according to the power and environmental data, compares the photovoltaic node with the optimal particle, and determines the influence weight of the target photovoltaic node, since the load node is less affected by the environment, the influence weight of the target load node is directly determined according to the correlation coefficients of all the load nodes in the same cluster as the target load node; finally, the power supply quality coupling analysis result facing source and load disturbance is obtained according to the abnormal characteristic indexes of all the nodes and all the influence weights, the problem of unstable power supply quality caused by ignoring the interaction between distributed power sources during new source and load disturbance is solved, and the power supply quality during new source and load disturbance is improved. BRIEF DESCRIPTION OF DRAWINGS

[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a schematic flowchart of a power quality coupling analysis method for source-load disturbances provided in an embodiment of the present invention.

[0049] Figure 2 This is a flowchart illustrating the process of obtaining the correlation coefficient between the low-voltage side node and the load node in a substation, as provided in one embodiment of the present invention. Detailed Implementation

[0050] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] Please see Figure 1 The diagram illustrates a flowchart of a power quality coupling analysis method for source-load disturbances provided by an embodiment of the present invention. The method includes the following steps:

[0052] Step S001: Collect electrical and environmental data at different collection times within a preset time period at the nodes of the distribution network. The nodes include low-voltage side nodes of substations, photovoltaic nodes, and load nodes. The electrical data includes current, voltage, and power.

[0053] Different types of nodes are set up in the distribution network, including three types: low-voltage side nodes of substations, photovoltaic nodes, and load nodes.

[0054] There is one and only one low-voltage side node in the substation. The low-voltage side node is located at the low-voltage distribution network of the substation and is used to monitor the overall power quality of the low-voltage distribution network. The photovoltaic node is located at the distributed photovoltaic power source and is used to monitor the power quality of the distributed photovoltaic power source output. It helps to assess the impact of new source-load disturbances on the distribution network. Load nodes are set at the bus, feeder outlet, and the location where large-capacity industrial users access the distribution network. The load nodes are used to monitor the load changes of the distribution network.

[0055] At each node, smart meters are used to collect different types of electrical data, including current, voltage, and power.

[0056] Preferably, in one embodiment of the present application, when collecting electrical data, the sampling frequency of each type of electrical data in this embodiment is 1 KHz, and electrical data within 30 minutes before the current sampling time is collected. In actual application, as other implementation manners, the implementer can determine the sampling frequency and sampling time according to actual conditions, and the present application does not make special limitations.

[0057] In the photovoltaic system of the power distribution network, the irradiance, temperature, humidity and air pressure are collected by using a pyranometer, a temperature sensor, a humidity sensor and an air pressure sensor respectively. The irradiance, temperature, humidity and air pressure are all recorded as environmental data.

[0058] Preferably, in one embodiment of the present application, when collecting environmental data, the sampling frequency of each type of environmental data in this embodiment is 1 minute, and environmental data within 30 minutes before the current sampling time is collected. In actual application, as other implementation manners, the implementer can determine the sampling frequency and sampling time according to actual conditions, and the present application does not make special limitations.

[0059] It should be noted that, in order to facilitate calculation, all electrical data and environmental data involved in the operation in this embodiment are subjected to data preprocessing, thereby canceling the influence of dimension. This embodiment adopts the Z-Score standard normalization method to perform dimensionless processing on electrical data and environmental data of the same type. In actual application, the implementer can use other methods such as the existing technology of maximum and minimum value normalization method to perform dimensionless processing, which is not limited herein.

[0060] At this point, the electrical data and environmental data within the preset time period before the current sampling time are obtained.

[0061] Step S002: Label any low-voltage side node, photovoltaic node, and load node of a substation as the target substation low-voltage side node, target photovoltaic node, and target load node, respectively. Determine the correlation coefficient between the target substation low-voltage side node and the target load node based on the similarity of their electrical data. Determine similar nodes for each node based on the differences in their environmental data. Determine the abnormal characteristic indicators and abnormal characteristic values ​​of each node at the same acquisition time based on all electrical data collected at the same acquisition time, as well as the sequence of abnormal characteristic values ​​for each node. Determine the correlation coefficient between the target substation low-voltage side node and the target photovoltaic node based on the similarity of all electrical data between the target photovoltaic node and the target substation low-voltage side node, and the similarity of the abnormal characteristic values ​​between the target substation low-voltage side node and the target photovoltaic node and its similar nodes.

[0062] When a high proportion of distributed photovoltaic (PV) power sources are connected to the distribution network, compared to traditional power generation models, the power supply of distributed PV power exhibits completely different characteristics in terms of distribution and randomness. This leads to significant fluctuations in power generation stability, and the power quality of the distribution network is also affected by the coupling effects of various nodes. Therefore, there is a certain correlation between changes in electrical data at different nodes and changes in electrical data of the distribution network. Thus, it is necessary to perform decoupling processing to determine the impact of each node on the power quality of the distribution network, thereby ensuring the power quality of the distribution network and guaranteeing its safe and stable operation.

[0063] Any low-voltage side node of a substation is designated as the target low-voltage side node; any photovoltaic node is designated as the target photovoltaic node; and any load node is designated as the target load node.

[0064] The correlation coefficient between the low-voltage side node of the target substation and the target load node is determined based on the similarity between all electrical data of the same type between the low-voltage side node of the target substation and the target load node.

[0065] Arranging all electrical class data of the same kind of the target substation low-voltage side node according to the collection time of the electrical class data to obtain the electrical class data sequence of the same kind of the target substation low-voltage side node; arranging all electrical class data of the same kind of the target load node according to the collection time of the electrical class data to obtain the electrical class data sequence of the same kind of the target load node; taking the absolute value of the similarity between the electrical class data sequence of the same kind of the target substation low-voltage side node and the target load node as the correlation degree of the electrical class data of the same kind of the target substation low-voltage side node and the target load node, and taking the average of the correlation degrees of all electrical class data of the same kind of the target substation low-voltage side node and the target load node as the correlation coefficient of the target substation low-voltage side node and the target load node.

[0066] Preferably, the similarity between the electrical class data sequence of the same kind of the target substation low-voltage side node and the target load node can be the Pearson correlation coefficient. As another embodiment, other methods such as cosine similarity can be used to obtain the similarity between two sequences for the purpose of measuring the similarity between two sequences, and the present application does not make special limitations.

[0067] Thus, the correlation coefficient of the target substation low-voltage side node and the target load node is obtained.

[0068] The correlation coefficient of any substation low-voltage side node and any load node can be obtained by the same method, and the flow chart of obtaining the correlation coefficient of the substation low-voltage side node and the load node is shown in FIG. Figure 2

[0069] The greater the similarity between the electrical class data of the same kind of the load node and the substation low-voltage side node, the closer the power generation of the load node and the substation low-voltage side node, and the more consistent the influence of the load node on the power supply quality of the distribution network. At this time, the correlation coefficient of the substation low-voltage side node and the load node is greater, and the influence of the load node on the power quality of the distribution network is greater.

[0070] Any one of all nodes is taken as a target node, any one node different from the target node is taken as a target comparison node, and the target node and the target comparison node are compared.

[0071] ​Arranging all the same kind of environmental class data of the target node according to the collection time of the environmental class data, obtaining the sequence of the same kind of environmental class data of the target node; arranging all the same kind of environmental class data of the target comparison node according to the collection time of the environmental class data, obtaining the sequence of the same kind of environmental class data of the target comparison node; recording the dtw distance between the sequences of the same kind of environmental class data of the target node and the target comparison node as the one-class direct difference of the same kind of environmental class data of the target node and the target comparison node, recording the mean of the differences of all the same kind of environmental class data of the target node and the target comparison node as the two-class direct difference of the target node and the target comparison node.

[0072] The two-class direct difference of the target node and any different node can be obtained in the same way.

[0073] The lower quartile of all the two-class direct differences corresponding to the target node is calculated, and all the nodes corresponding to the two-class direct differences less than the lower quartile are recorded as the similar nodes of the target node.

[0074] The similar nodes of all the nodes can be obtained in the same way.

[0075] Among them, the calculation of the dtw distance between the sequences and the calculation of the lower quartile are both well-known technologies and will not be repeated.

[0076] The electrical class data of the target node at any collection time is recorded as the target collection time.

[0077] The short-time flicker value Pst, the total harmonic distortion rate and the three-phase imbalance degree of the target collection time are calculated respectively, the short-time flicker value, the total harmonic distortion rate and the three-phase imbalance degree of the target collection time are normalized respectively, and the results after normalization are recorded as the abnormal feature indicators of the target collection time.

[0078] Among them, the calculation of the short-time flicker value, the total harmonic distortion rate and the three-phase imbalance degree of the target collection time is all well-known technology and will not be repeated, specifically, the short-time flicker value is used to measure the voltage fluctuation, the total harmonic distortion rate is used to measure the harmonic pollution of the current, and the three-phase imbalance degree is used to measure the three-phase balance of the power supply quality.

[0079] It should be noted that the Z-Score standard normalization method is used for normalization in this embodiment, and other methods such as the maximum and minimum value normalization method, the sigmoid function and other methods of prior art can be used for normalization in actual application, which is not limited here.

[0080] The mean value of all abnormal characteristic indexes of the target node at the target acquisition time is recorded as the abnormal characteristic value of the target node at the target acquisition time.

[0081] In the same way, the similar nodes of any one of all nodes and the abnormal characteristic indexes and abnormal characteristic values of any one of all nodes at any acquisition time can be obtained, wherein the abnormal characteristic indexes include the results of normalization of the short-time flicker value, the total harmonic distortion rate, and the three-phase unbalance degree, respectively.

[0082] The abnormal characteristic values of the target node are arranged according to the acquisition times corresponding to the abnormal characteristic values to obtain the abnormal characteristic value sequence of the target node.

[0083] The second correlation degree between the target photovoltaic node and the target substation low-voltage side node is determined according to the similarity between all electrical class data of the same kind between the target photovoltaic node and the target substation low-voltage side node.

[0084] The electrical class data sequence of the same kind of the target photovoltaic node is obtained by arranging all electrical class data of the same kind of the target photovoltaic node according to the acquisition times of the electrical class data. The absolute value of the similarity between the electrical class data sequence of the same kind of the target photovoltaic node and the target substation low-voltage side node is recorded as the first correlation degree of the electrical class data of the same kind of the target photovoltaic node and the target substation low-voltage side node. The mean value of the first correlation degrees of all electrical class data of the same kind of the target photovoltaic node and the target substation low-voltage side node is recorded as the second correlation degree of the target photovoltaic node and the target substation low-voltage side node.

[0085] Preferably, the similarity between the electrical class data sequence of the same kind of the target photovoltaic node and the target substation low-voltage side node can be the Pearson correlation coefficient. As other embodiments, for the purpose of measuring the similarity of two sequences, the implementer can use other methods such as cosine similarity in the prior art to obtain the similarity of two sequences, and the present application does not make special limitations.

[0086] The abnormal characteristic value similarity and the similar node abnormal characteristic value similarity between the target substation low-voltage side node and the target photovoltaic node are respectively determined according to the similarity between the abnormal characteristic values of the target substation low-voltage side node and the target photovoltaic node and the similar node of the target photovoltaic node.

[0087] The absolute value of the similarity between the similar node of the target photovoltaic node and the abnormal characteristic value sequence of the target substation low-voltage side node is recorded as the abnormal characteristic value similarity between the target substation low-voltage side node and the target photovoltaic node.

[0088] The mean value of the absolute value of the similarity between the abnormal eigenvalue sequence of all similar nodes and the abnormal eigenvalue sequence of the target substation low-voltage side node is denoted as the similarity of the abnormal eigenvalue of the similar node between the target substation low-voltage side node and the target photovoltaic node.

[0089] According to the second correlation degree, the abnormal eigenvalue similarity, and the similarity of the abnormal eigenvalue of the similar node between the target substation low-voltage side node and the target photovoltaic node, the correlation coefficient between the target substation low-voltage side node and the target photovoltaic node is determined. Specifically, the calculation formula of the correlation coefficient between the target substation low-voltage side node and the target photovoltaic node is as follows:

[0090]

[0091] In the formula, the correlation coefficient between the target substation low-voltage side node and the target photovoltaic node is denoted as the correlation coefficient between the target substation low-voltage side node and the target photovoltaic node. The abnormal eigenvalue similarity between the target substation low-voltage side node and the target photovoltaic node is denoted as the abnormal eigenvalue similarity between the target substation low-voltage side node and the target photovoltaic node. The similarity of the abnormal eigenvalue of the similar node between the target substation low-voltage side node and the target photovoltaic node is denoted as the similarity of the abnormal eigenvalue of the similar node between the target substation low-voltage side node and the target photovoltaic node. The second correlation degree between the target substation low-voltage side node and the target photovoltaic node is denoted as the second correlation degree between the target substation low-voltage side node and the target photovoltaic node. The abnormal eigenvalue similarity between the target substation low-voltage side node and the target photovoltaic node is denoted as the abnormal eigenvalue similarity between the target substation low-voltage side node and the target photovoltaic node. The similarity of the abnormal eigenvalue of the similar node between the target substation low-voltage side node and the target photovoltaic node is denoted as the similarity of the abnormal eigenvalue of the similar node between the target substation low-voltage side node and the target photovoltaic node. The second correlation degree between the target substation low-voltage side node and the target photovoltaic node is denoted as the second correlation degree between the target substation low-voltage side node and the target photovoltaic node.

[0092] At this point, the correlation coefficient between the target substation low-voltage side node and the target photovoltaic node is obtained.

[0093] The correlation coefficient between any substation low-voltage side node and any photovoltaic node can be obtained by the same method. At this point, the correlation coefficient between each substation low-voltage side node and all nodes that are not substation low-voltage side nodes is obtained, that is, the correlation coefficient between each substation low-voltage side node and each photovoltaic node is obtained, and the correlation coefficient between each substation low-voltage side node and each load node is obtained.

[0094] It can be understood that the closer the similarity between the same type of electrical data of the photovoltaic node and the substation low-voltage side node, and the greater the similarity between the abnormal eigenvalues of the substation low-voltage side node and the photovoltaic node and the similar nodes of the photovoltaic node, the closer the power generation of the photovoltaic node and the substation low-voltage side node, and the more consistent the influence of the load node on the power supply quality of the power distribution network. At this time, the correlation coefficient between the photovoltaic node and the substation low-voltage side node is greater, and the influence of the photovoltaic node on the power quality of the power distribution network is greater.

[0095] ​​​​​​So far, the correlation coefficient of each substation low-voltage side node and each photovoltaic node is obtained, and the correlation coefficient of each substation low-voltage side node and each load node is obtained.

[0096] In step S003, the optimal particle in the particle composed of the node and the similar node of the node is screened according to the electrical data, the environmental data and the abnormal feature value, the abnormal feature value sequence of all nodes is clustered to obtain a clustering cluster, the influence weight of the target load node is determined according to the correlation coefficient of all load nodes in the same clustering cluster as the target load node, and the influence weight of the target photovoltaic node is determined according to the difference between the environmental data of all nodes in the optimal particle and the target photovoltaic node, the difference between the power, and the correlation coefficient of the target photovoltaic node.

[0097] There is mutual influence between different nodes, which may cause more serious coupling in the power distribution network and affect the regulation and control of power supply quality. Since the fluctuation of power quality is reflected through the change of power, it is necessary to further analyze the influence of each node on the power quality of the power distribution network according to the power.

[0098] The node and the similar node of the node are taken as particles, the electrical data and the environmental data of the node are taken as the corresponding data of the node, the accumulation and the minimum of the abnormal feature values at all collection moments are taken as a target function, and a particle swarm optimization algorithm is used to obtain an optimal particle.

[0099] The particle swarm optimization algorithm for obtaining the optimal particle is a known technology and will not be described herein. In this embodiment, the maximum iteration number of the particle swarm optimization algorithm is set to 100, the inertia weight is set to 0.5, and the two initial learning factors are both set to 1.5.

[0100] The abnormal feature value sequence of all nodes is clustered to obtain a clustering cluster.

[0101] In this embodiment, the DBSCAN clustering algorithm is used to cluster the abnormal feature value sequence of all nodes. The use of the DBSCAN clustering algorithm to obtain a clustering cluster is a known technology and will not be described herein. In this embodiment, the maximum neighborhood radius of the DBSCAN clustering algorithm is determined by a k-distance diagram. The value of the minimum neighbor number minPts of the DBSCAN clustering algorithm is 15, the maximum neighborhood radius of the DBSCAN clustering algorithm is determined by a k-distance diagram, and the use of the DBSCAN clustering algorithm to obtain a clustering cluster is a known technology and will not be described herein.

[0102] The influence weight of the target load node is determined according to the correlation coefficient of all load nodes in the same clustering cluster as the target load node.

[0103] ​The mean of the correlation coefficients of all the load nodes in the same cluster as the target load node is denoted as the correlation coefficient benchmark of the target load node, and the normalized value of the ratio of the mean of all the correlation coefficients of the target load node to the correlation coefficient benchmark is denoted as the influence weight of the target load node.

[0104] The influence weight of any load node can be obtained in the same way.

[0105] The optimal power difference between the target photovoltaic node and any node in the optimal particle is determined according to the difference between the optimal particle and the environmental class data of the target photovoltaic node, and the difference between the power.

[0106] The absolute value of the difference between the corresponding environmental class data of any node in the optimal particle and the target photovoltaic node is denoted as the optimal environmental difference between the target photovoltaic node and the node in the optimal particle, and the absolute value of the difference between the power of any node in the optimal particle and the target photovoltaic node is denoted as the optimal power difference between the target photovoltaic node and the node in the optimal particle.

[0107] It can be understood that since each node and the similar node of the node are taken as a particle, the optimal particle contains multiple nodes, and for any node in the optimal particle, there is a corresponding optimal environmental difference and a corresponding optimal power difference with the target photovoltaic node.

[0108] The influence weight of the target photovoltaic node is obtained according to the optimal environmental difference and the optimal power difference corresponding to the target photovoltaic node, and all the correlation coefficients of the target photovoltaic node.

[0109] The optimal environmental difference corresponding to the target photovoltaic node is taken as the independent variable, the optimal power difference corresponding to the target photovoltaic node is taken as the dependent variable, a curve fitting is performed using a polynomial fitting technique to obtain a difference fitting curve of the target photovoltaic node. The optimal environmental difference of the node in the optimal particle that is not a similar node and the target photovoltaic node is taken as the value of the independent variable, and is substituted into the difference fitting curve of the target photovoltaic node, and the corresponding dependent variable value of the difference fitting curve is denoted as the difference fitting value of the target photovoltaic node. The optimal power difference of the node in the optimal particle that is not a similar node and the target photovoltaic node is denoted as the difference real value of the target photovoltaic node.

[0110] The ratio of the difference real value to the difference fitting value of the target photovoltaic node is denoted as the first ratio of the target photovoltaic node, and the normalized value of the product of the mean of all the correlation coefficients of the target photovoltaic node and the first ratio is denoted as the influence weight of the target photovoltaic node.

[0111] The influence weight of any photovoltaic node can be obtained in the same way.

[0112] It can be understood that the input power quality of the photovoltaic node in the power distribution network has a high correlation with the operating environment, so that the difference fitting curve obtained by the node in the optimal particle and the similar node can more accurately measure the difference between the photovoltaic node and the optimal particle, and the influence weight of any photovoltaic node is obtained through the difference between the difference fitting value and the difference true value and the correlation coefficient, to judge the severity of the power quality problem of the photovoltaic node and the influence degree of the input power of the photovoltaic node on the power supply quality of the power distribution network. When the influence weight of the photovoltaic node is larger, the influence degree of the input power of the photovoltaic node on the power supply quality of the power distribution network is larger.

[0113] At this point, the influence weights of all photovoltaic nodes and load nodes are obtained.

[0114] Step S004, according to the abnormal characteristic indexes of all nodes and all influence weights, obtaining the power supply quality coupling analysis result facing source and load disturbance.

[0115] The abnormal characteristic indexes of the nodes on the low-voltage side of the transformer substation, all photovoltaic nodes and all load nodes are filled in the rows of the abnormal characteristic index matrix in the order from top to bottom, and the abnormal characteristic index matrix is obtained. The influence weights of the nodes on the low-voltage side of the transformer substation, all photovoltaic nodes and all load nodes are arranged into a row to obtain the influence weight matrix. The product of the abnormal characteristic index matrix and the influence weight matrix is denoted as the power distribution network power supply quality matrix.

[0116] Wherein, each row of the abnormal characteristic index matrix corresponds to a node, and in each row of the abnormal characteristic index matrix, the abnormal characteristic indexes are arranged in the order of short-time flicker value, total harmonic distortion rate and three-phase unbalance degree; the order of the nodes corresponding to the first row to the last row of the abnormal characteristic index matrix is the same as the order of the nodes corresponding to the first column to the last column of the influence weight matrix.

[0117] The power distribution network power supply quality matrix is processed by using the principal component analysis algorithm to obtain the loading matrix of the power distribution network power supply quality matrix. All values contained in the loading matrix are divided using the maximum inter-class variance method to obtain a division threshold, and the nodes corresponding to the values greater than the division threshold among all values contained in the loading matrix are denoted as disturbance nodes.

[0118] The loading matrix can determine the weight of each node on different principal components, that is, the contribution degree of the node to the principal component, and the greater the contribution degree of the node to the principal component, the greater the disturbance of the node to the power supply quality of the power distribution network; the disturbance node is the node that has a disturbance influence on the power supply quality of the power distribution network, and the disturbance node is the power supply quality coupling analysis result facing source and load disturbance.

[0119] Wherein, the principal component analysis algorithm is adopted to obtain the load matrix, and the maximum inter-class variance method is used to obtain the division threshold, which are all known technologies and will not be described herein.

[0120] Up to now, the power supply quality coupling analysis oriented to source load disturbance is completed.

[0121] The above merely describes the preferred embodiments of the present application and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the principles of the present application shall be included in the protection scope of the present application.

Claims

1. A power quality coupling analysis method for source-load disturbances, characterized in that, The method includes the following steps: Electrical and environmental data collected at different times at nodes in the distribution network, including low-voltage side nodes of substations, photovoltaic nodes, and load nodes, and electrical data including current, voltage, and power; Let any low-voltage side node, photovoltaic node, and load node of a substation be denoted as the target substation low-voltage side node, target photovoltaic node, and target load node, respectively. Determine the correlation coefficient between the target substation low-voltage side node and the target load node. Determine the similar nodes for each node. Based on all electrical data of the node at the same acquisition time, determine the abnormal characteristic indicators and abnormal characteristic values ​​of the node at the same acquisition time, as well as the abnormal characteristic value sequence of the node. Based on the similarity between all electrical data of the target photovoltaic node and the target substation low-voltage side node, and the similarity between the abnormal characteristic values ​​of the target substation low-voltage side node and the target photovoltaic node and the similar nodes of the target photovoltaic node, respectively, determine the correlation coefficient between the target substation low-voltage side node and the target photovoltaic node. Based on electrical data, environmental data, and abnormal feature values, the optimal particle is selected from the particles composed of nodes and similar nodes. The abnormal feature value sequences of all nodes are clustered to obtain clusters. The influence weight of the target load node is determined based on the correlation coefficient of all load nodes in the same cluster as the target load node. The influence weight of the target photovoltaic node is determined based on the differences in environmental data and power between all nodes in the optimal particle and the target photovoltaic node, as well as the correlation coefficient of the target photovoltaic node. Based on the abnormal characteristic indicators of all nodes and all influence weights, the power quality coupling analysis results oriented towards source-load disturbance are obtained.

2. The power quality coupling analysis method for source-load disturbances according to claim 1, characterized in that, The method for determining the correlation coefficient between the low-voltage side node of the target substation and the target load node is as follows: Arrange all electrical data of the same type at the low-voltage side node of the target substation according to the time of data collection, and obtain the electrical data sequence of the same type at the low-voltage side node of the target substation. Arrange all electrical data of the same type for the target load node according to the time of data collection to obtain the electrical data sequence of the same type for the target load node; The absolute value of the similarity between the electrical data sequences of the same type between the low-voltage side node of the target substation and the target load node is denoted as the correlation degree of the electrical data of the same type between the low-voltage side node of the target substation and the target load node. The mean value of the correlation degree of all electrical data of the same type between the low-voltage side node of the target substation and the target load node is denoted as the correlation coefficient between the low-voltage side node of the target substation and the target load node.

3. The power quality coupling analysis method for source-load disturbances according to claim 1, characterized in that, The method for determining the similar nodes of the node is as follows: Let any one of the nodes be the target node, and let any node that is different from the target node be the target comparison node. Arrange all environmental data of the same type at the target node according to the collection time of the environmental data to obtain the environmental data sequence of the same type at the target node; arrange all environmental data of the same type at the target comparison node according to the collection time of the environmental data to obtain the environmental data sequence of the same type at the target comparison node. The dtw distance between the target node and the target comparison node of the same type of environmental data sequence is denoted as the first type of direct difference between the target node and the target comparison node of the same type of environmental data. The mean of the differences between the target node and the target comparison node of all environmental data of the same type is denoted as the second type of direct difference between the target node and the target comparison node. The similar nodes of the target node are determined based on the lower quartiles of all direct differences of the second type corresponding to the target node.

4. The power quality coupling analysis method for source-load disturbances according to claim 1, characterized in that, The specific method for determining the abnormal characteristic indicators and abnormal characteristic values ​​of a node at the same acquisition time, as well as the sequence of abnormal characteristic values ​​of the node, based on all electrical data of the node at the same acquisition time, includes the following: The electrical data of the target node at any acquisition time is recorded as the target acquisition time. The short-time flicker, total harmonic distortion rate, and three-phase imbalance at the target electrical acquisition time are calculated respectively. The normalized results of the short-time flicker, total harmonic distortion rate, and three-phase imbalance at the target acquisition time are all recorded as the abnormal characteristic indexes at the target acquisition time. The average value of all abnormal feature indicators of the target node at the target acquisition time is recorded as the abnormal feature value of the target node at the target acquisition time. The abnormal feature values ​​of the target node are arranged according to the acquisition time corresponding to the abnormal feature values ​​to obtain the abnormal feature value sequence of the target node.

5. The power quality coupling analysis method for source-load disturbances according to claim 1, characterized in that, The method for determining the correlation coefficient between the target photovoltaic node and the target photovoltaic node based on the similarity between all electrical data of the target photovoltaic node and the low-voltage side node of the target substation, and the similarity between the low-voltage side node of the target substation and the abnormal feature values ​​of the target photovoltaic node and similar nodes of the target photovoltaic node, includes the following specific methods: All electrical data of the same type for the target photovoltaic node are arranged according to the time of data collection to obtain the electrical data sequence of the same type for the target photovoltaic node; the absolute value of the similarity between the electrical data sequence of the same type between the low-voltage side node of the target substation and the target photovoltaic node is recorded as the first correlation degree between the electrical data of the same type between the low-voltage side node of the target substation and the target photovoltaic node; the mean of the first correlation degree of all electrical data of the same type between the low-voltage side node of the target substation and the target photovoltaic node is recorded as the second correlation degree between the low-voltage side node of the target substation and the target photovoltaic node. Based on the similarity between the abnormal feature values ​​of the low-voltage side nodes of the target substation and the target photovoltaic node and the similar nodes of the target photovoltaic node, the similarity between the abnormal feature values ​​of the low-voltage side nodes of the target substation and the target photovoltaic node and the similarity between the abnormal feature values ​​of the similar nodes are determined respectively. Based on the second correlation degree, the similarity of abnormal eigenvalues, and the similarity of abnormal eigenvalues ​​of similar nodes between the low-voltage side nodes of the target substation and the target photovoltaic node, the correlation coefficient between the low-voltage side nodes of the target substation and the target photovoltaic node is determined. The calculation formula is as follows: ； In the formula, Indicates the low-voltage side node of the target substation With the target photovoltaic node The correlation coefficient; Indicates the low-voltage side node of the target substation With the target photovoltaic node Similarity of anomalous feature values ​​of similar nodes; Indicates the low-voltage side node of the target substation With the target photovoltaic node Similarity of abnormal feature values; Indicates the low-voltage side node of the target substation With the target photovoltaic node The second degree of correlation.

6. The power quality coupling analysis method for source-load disturbances according to claim 5, characterized in that, The method for determining the similarity of abnormal feature values ​​between the low-voltage side nodes of the target substation and the target photovoltaic node, and the similarity of abnormal feature values ​​between similar nodes of the target photovoltaic node, is as follows: The absolute value of the similarity between the similar nodes of the target photovoltaic node and the abnormal feature value sequence of the low-voltage side node of the target substation is denoted as the abnormal feature value similarity between the low-voltage side node of the target substation and the target photovoltaic node. The mean of the absolute values ​​of the similarity between the abnormal feature value sequences of all similar nodes and the abnormal feature value sequences of the low-voltage side nodes of the target substation is denoted as the similarity of the abnormal feature values ​​of the low-voltage side nodes of the target substation and the target photovoltaic nodes.

7. The power quality coupling analysis method for source-load disturbances according to claim 1, characterized in that, The specific method for selecting the optimal particle from the particles composed of nodes and similar nodes based on electrical data, environmental data, and abnormal feature values ​​includes: Nodes and their similar nodes are treated as particles, and the electrical and environmental data of a node are treated as the corresponding data of the node. The sum of the abnormal feature values ​​at all acquisition times is taken as the objective function, and an optimization algorithm is used to obtain the optimal particle.

8. The power quality coupling analysis method for source-load disturbances according to claim 1, characterized in that, The method for determining the influence weight of the target load node is as follows: The mean of the correlation coefficients of all load nodes in the same cluster as the target load node is denoted as the baseline correlation coefficient of the target load node; the normalized value of the ratio of the mean of all correlation coefficients of the target load node to the baseline correlation coefficient is denoted as the influence weight of the target load node.

9. The power quality coupling analysis method for source-load disturbances according to claim 1, characterized in that, The method for determining the influence weight of the target photovoltaic node based on the differences in environmental data and power between all nodes in the optimal particle and the target photovoltaic node, as well as the correlation coefficient of the target photovoltaic node, includes the following specific methods: Based on the differences in environmental data between the optimal particle and the target photovoltaic node, as well as the differences in power, the optimal power difference between the target photovoltaic node and any node among the optimal particles is determined. Using the optimal environmental difference corresponding to the target photovoltaic node as the independent variable and the optimal power difference corresponding to the target photovoltaic node as the dependent variable, the difference fitting curve of the target photovoltaic node is obtained; the optimal environmental difference between the nodes that are not similar nodes in the optimal particles and the target photovoltaic node is taken as the value of the independent variable, the value of the dependent variable corresponding to the difference fitting curve is calculated, and the value of the dependent variable is recorded as the difference fitting value of the target photovoltaic node. The difference between the optimal power of non-similar nodes in the optimal particle and the target photovoltaic node is denoted as the true difference value of the target photovoltaic node. The ratio of the actual difference value to the fitted difference value of the target photovoltaic node is denoted as the first ratio of the target photovoltaic node. The normalized value of the product of the mean of all correlation coefficients of the target photovoltaic node and the first ratio is denoted as the influence weight of the target photovoltaic node.

10. The power quality coupling analysis method for source-load disturbances according to claim 1, characterized in that, The specific method for obtaining power quality coupling analysis results oriented towards source-load disturbance based on the abnormal characteristic indicators of all nodes and all influence weights includes: An abnormal characteristic index matrix is ​​established based on the abnormal characteristic indicators of the low-voltage side nodes, photovoltaic nodes, and load nodes of the substation; an influence weight matrix is ​​established based on the influence weights of the low-voltage side nodes, photovoltaic nodes, and load nodes of the substation; the product of the abnormal characteristic index matrix and the influence weight matrix is ​​denoted as the power supply quality matrix of the distribution network. The load matrix of the power supply quality matrix of the distribution network is obtained by using the principal component analysis algorithm. Disturbance nodes are screened based on all the values ​​contained in the load matrix. The disturbance nodes are the power supply quality coupling analysis results oriented towards source-load disturbance.

Images