Fault detection method for new energy access considering three-phase imbalance in distribution system

Abstract

The present application relates to the field of electric power grid technology, and specifically to a method for detecting faults in the access of new energy sources that takes into account three-phase imbalance in a distribution system. The method comprises: obtaining electrical parameters of distribution network nodes and environmental parameters of distributed power access nodes to form a sample; obtaining a random measure of three-phase imbalance changes based on the distribution of voltage imbalance characteristics in the sample; analyzing similarities in trend changes of voltage and current in each phase to obtain an average similarity measure; obtaining a grid oscillation coefficient; analyzing the distribution characteristics of environmental parameters and electrical parameters to obtain the degree of influence of light intensity, instantaneous wind speed, and continuous wind speed, and obtaining an abnormal fluctuation coefficient of the distributed power access node; and performing fault detection on the sample in combination with the grid oscillation coefficient. The present application aims to improve the accuracy of fault detection and ensure the safe and stable operation of the distribution system.

Description

Technical Field

[0001] The present application relates to the field of electric power grid technology, and in particular to a method for detecting faults in the access of new energy sources taking into account three-phase imbalance in a power distribution system. Background Art

[0002] With the growing global demand for clean energy, renewable energy sources (such as solar and wind) are increasingly being used in power systems. Traditional distribution systems primarily rely on stable and reliable fossil fuel power generation, which offers relatively stable and controllable power supply characteristics. However, renewable energy generation exhibits significant intermittent and fluctuating characteristics, significantly impacted by natural environmental factors such as sunlight intensity and wind speed. This leads to unstable output power and even significant fluctuations in power supply conditions over short periods of time. This uncertainty not only impacts grid stability but also complicates fault detection and protection. Furthermore, the large number and dispersion of renewable energy access points complicates the grid structure, further complicating fault detection.

[0003] Existing fault detection methods for renewable energy integration typically identify faults based on changes in electrical parameters over a period of time. This method demonstrates high accuracy in fault detection for traditional power generation environments, but struggles with the rapid fluctuations in renewable energy power supply. The intermittent and volatile nature of renewable energy generation can cause significant changes in the effective values and phase relationships of voltage and current within a short period of time. This is particularly true for three-phase imbalance. Due to sudden changes in environmental parameters, the output power of photovoltaic or wind power systems can fluctuate dramatically, resulting in brief three-phase imbalance in the grid. Traditional methods can misinterpret these normal fluctuations as serious faults, triggering unnecessary protective actions and reducing grid reliability. Summary of the Invention

[0004] In view of the above, it is necessary to provide a new energy access fault detection method considering the three-phase imbalance of the distribution system to solve the above problems.

[0005] One embodiment of the present application provides a method for detecting a new energy access fault that takes into account three-phase imbalance in a power distribution system. The method includes:

[0006] The electrical parameters of the distribution network nodes and the environmental parameters of the distributed power access nodes collected within a preset time period are combined into a sample; the types of the electrical parameters and environmental parameters are at least 3, where the electrical parameters include voltage and current; and the environmental parameters include wind speed and light intensity;

[0007] Based on the distribution of voltage imbalance characteristics within all acquisition cycles in the sample, a random measure of three-phase imbalance change is obtained by combining the run-length test algorithm. The similarity of the trend changes of voltage and current in each phase of each distribution network node in each acquisition cycle is analyzed to obtain the average similarity measure. Combined with the change characteristics of other electrical parameters, the grid oscillation coefficient is obtained.

[0008] Extract the mutation points of wind speed and light intensity in the preset time period respectively; divide the first time interval between adjacent mutation points corresponding to the light intensity, and combine the distribution characteristics of the electrical parameters to obtain the degree of influence of the light intensity; analyze the distribution characteristics of the electrical parameters in the second time interval between adjacent mutation points corresponding to the wind speed to obtain the instantaneous wind speed influence degree and the continuous wind speed influence degree, and combine the light intensity influence degree and the distribution characteristics of other environmental parameters to obtain the abnormal fluctuation coefficient of the distributed power access node;

[0009] Based on the numerical value of the abnormal fluctuation coefficient and in combination with the power grid oscillation coefficient, fault detection is performed on the sample.

[0010] Preferably, the random measurement of the three-phase imbalance change is obtained as follows:

[0011] Calculate the voltage imbalance of each distribution network node in each acquisition cycle, and take the average of the voltage imbalances of all distribution network nodes in each acquisition cycle as the three-phase voltage imbalance of each acquisition cycle;

[0012] When the three-phase voltage imbalance is greater than a preset threshold, the imbalance value of the corresponding acquisition period is marked as a first preset value, otherwise it is marked as a second preset value; wherein the first preset value is not equal to the second preset value;

[0013] The sequence of imbalance values from all acquisition cycles is recorded as an imbalance value sequence. A runs test algorithm is used to obtain a probability value and a Z-statistic. When the probability value is less than the significance level of the runs test, 0 is used as a measure of the randomness of the three-phase imbalance change; otherwise, the Z-statistic is used as a measure of the randomness of the three-phase imbalance change.

[0014] Preferably, the average similarity measure is obtained as follows:

[0015] Obtain the trend sequence corresponding to the voltage of each distribution network node in each acquisition period; calculate the average trend sequence of the trend sequence of all nodes in each phase voltage in each acquisition period; obtain the mean of the similarity coefficients between the average trend sequences of all two-phase voltage combinations in each acquisition period, recorded as the first mean; record the mean of the first mean values of all acquisition periods as the voltage similarity measure;

[0016] Accordingly, the current similarity measure is obtained based on the trend sequence corresponding to the current of each node in all acquisition cycles;

[0017] The average of the voltage similarity metric and the current similarity metric is recorded as the average similarity metric.

[0018] Preferably, the grid oscillation coefficient is obtained as follows:

[0019] The average value of the three-phase voltage imbalance in all acquisition cycles is used as the three-phase imbalance measure of the distribution network, denoted as N;

[0020] Calculate the average change rate of the power factor of each distribution network node, and record the average of the average change rates of all distribution network nodes as the average change rate of the power factor of the distribution network, which is recorded as Q;

[0021] The grid oscillation coefficient is recorded as A, and its formula is: Where, is the average similarity measure; It is a measure of the randomness of the three-phase unbalance variation; is the tuning coefficient.

[0022] Preferably, the process of obtaining the degree of influence of light intensity is:

[0023] A threshold segmentation algorithm is applied to all first time intervals corresponding to the light intensity to obtain a first segmentation threshold, and a time period corresponding to a first time interval greater than the first segmentation threshold is used as a photovoltaic effective time;

[0024] The product of the discreteness of the output power of each distributed power access node in each photovoltaic effective time and the average change rate of the output power at all adjacent moments is obtained, and the mean of the product of each distributed power access node in all photovoltaic effective times is forward fused with the mean of the first time interval corresponding to all photovoltaic effective times to obtain the degree of influence of light intensity.

[0025] Preferably, the process of obtaining the instantaneous wind speed impact degree and the continuous wind speed impact degree is specifically as follows:

[0026] For each distributed power access node, calculate the time interval between adjacent mutation points obtained according to the wind speed to obtain a second time interval;

[0027] A threshold segmentation algorithm is applied to all second time intervals corresponding to the wind speed to obtain a second segmentation threshold, and a time period corresponding to a second time interval greater than the second segmentation threshold is used as a continuous wind speed time; and a time period corresponding to the remaining second time intervals is used as an instantaneous wind speed time;

[0028] Based on the number of instantaneous wind speed periods corresponding to each distributed power access node and the degree of change in output power within each instantaneous wind speed period, the degree of influence of the instantaneous wind speed on the distributed power access node is obtained; wherein output power is an electrical parameter obtained;

[0029] According to the number of continuous wind speed periods corresponding to each distributed generation access node and the distribution characteristics of the output power within each continuous wind speed period, the influence degree of the continuous wind speed of the distributed generation access node is obtained.

[0030] Preferably, the specific formula for the instantaneous wind speed influence degree of the distributed power access node is: ;in, Indicates the distributed power access node The degree of influence of instantaneous wind speed; is the number of instantaneous wind speed times; They are The variance and average rate of change of output power within a certain instantaneous wind speed period; It is a distributed power access node The average wind speed in all instantaneous wind speed time periods; Represents a natural constant.

[0031] Preferably, the formula for the influence degree of the continuous wind speed of the distributed power supply access node is: ;in, Indicates the distributed power access node The degree of impact of sustained wind speed; is the number of hours of sustained wind speed, They are The variance and mean of the output power during the continuous wind speed period, It is a distributed power access node The mean value of all sustained wind speed times corresponding to the second time interval.

[0032] Preferably, the abnormal fluctuation coefficient of the distributed power access node is obtained as follows:

[0033] Calculate the mean variance of all environmental parameters except light intensity and wind speed at the distributed generation access node;

[0034] When the distributed power access node belongs to the photovoltaic access node set, the sum of the light intensity impact degree and the variance mean is used as the abnormal fluctuation coefficient of the distributed power access node;

[0035] When the distributed power access node data peak point accesses the node set, the cumulative sum of the instantaneous wind speed impact degree, the continuous wind speed impact degree, and the variance mean is used as the abnormal fluctuation coefficient of the distributed power access node.

[0036] Preferably, the fault detection of the sample based on the numerical value of the abnormal fluctuation coefficient in combination with the grid oscillation coefficient is specifically performed as follows:

[0037] The mean of the abnormal fluctuation coefficients of all distributed power access nodes in each sample is recorded as the sample abnormal fluctuation coefficient of each sample;

[0038] The vector consisting of the power grid oscillation coefficient and the sample abnormal fluctuation coefficient is used as the feature vector of each sample;

[0039] Based on the value range of the abnormal fluctuation coefficient of the sample in all samples, all characteristic vectors are equally divided; each characteristic vector is clustered, and the cluster with the smallest mean value of the power grid oscillation coefficient is taken as the optimal cluster;

[0040] Perform curve fitting based on all elements in all optimal clusters to obtain a fitting curve;

[0041] Obtain the difference between the function value of the grid oscillation coefficient of each sample and the sample abnormal fluctuation coefficient on the fitting curve as the fault detection value of each sample;

[0042] The fault detection value of each sample detection data is clustered with the fault detection value of the sample detection data of known faults to determine the fault of each sample.

[0043] This application has at least the following beneficial effects:

[0044] This application first analyzes the changes in electrical parameters of distribution network nodes, and based on the distribution of voltage imbalance characteristics in all acquisition cycles in the sample, combined with the run-length test algorithm, obtains a random measure of three-phase imbalance changes, which helps to improve the detection accuracy and reliability of grid imbalance faults in dynamic grid environments; analyzes the similarity of the trend changes of voltage and current in each phase of each distribution network node in each acquisition cycle, obtains an average similarity measure, evaluates the grid operation status, and combines the change characteristics of other electrical parameters to obtain the grid oscillation coefficient, which can fully reflect the power quality fluctuations in the distribution network and help to enhance the accuracy of three-phase imbalance fault identification; then analyzes the changes in environmental parameters of distributed power access nodes, and extracts the mutation points of wind speed and light intensity in preset time periods; divides the time intervals between adjacent mutation points corresponding to light intensity, and combines the distribution characteristics of electrical parameters to obtain The degree of influence of light intensity is combined with the distribution characteristics of electrical parameters to evaluate the impact of light intensity on the power grid, and identify the degree of influence of sudden changes in light intensity on the changes in power grid load; further analyze the distribution characteristics of electrical parameters in the time interval between the mutation points corresponding to wind speed to obtain the degree of influence of instantaneous wind speed and the degree of influence of continuous wind speed, which helps to better understand the impact of wind speed on power grid load fluctuations; combined with the degree of influence of light intensity and the distribution characteristics of other environmental parameters, the abnormal fluctuation coefficient of distributed power access nodes is obtained, which can enhance the accuracy of identifying abnormal fluctuations in new energy power generation according to different new energy types; finally, using these indicators, fault detection can be dynamically adjusted according to the abnormal situation of new energy power generation fluctuations. This fault detection method adapted to the access of new energy can help reduce power quality problems caused by fluctuations in new energy power generation and ensure the stability and reliability of power supply. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A flow chart of a new energy access fault detection method considering three-phase imbalance in the power distribution system provided in this application;

[0046] Figure 2 This is a flowchart for obtaining the fault detection value provided by this application. DETAILED DESCRIPTION

[0047] In the description of the embodiments of this application, words such as "exemplary," "or," and "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of this application should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary," "or," and "for example" is intended to present the relevant concepts in a concrete manner.

[0048] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art in the art of this application. The terms used in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application.

[0049] It should also be noted that the terms "first" and "second" in this application and its accompanying drawings are used to distinguish similar objects, rather than to describe a specific order or precedence. The methods disclosed in the embodiments of this application or the methods shown in the flowcharts include one or more steps for implementing the methods. Without departing from the scope of protection of this application, the order of execution of multiple steps can be interchanged with each other, and some steps can also be deleted.

[0050] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0051] This application proposes a new energy access fault detection method considering the three-phase imbalance of the distribution system, which is applied to the technical field of power grid technology. Figure 1 , the method comprises the following steps:

[0052] S1: The electrical parameters of the distribution network nodes and the environmental parameters of the distributed power access nodes collected within a preset time period are combined into a sample; the electrical parameters include at least voltage and current; the environmental parameters include at least wind speed and light intensity.

[0053] At the distribution network nodes, voltage and current data are collected through voltage and current transformers respectively; the output power of the distributed power access node is collected through a power meter; and the power factor data is collected through a power factor meter. In this embodiment, the collection frequency of all electrical parameters is set to 100Hz, with 1 minute as a collection cycle. The implementer can adjust the collection frequency of electrical parameters according to actual conditions.

[0054] In this application, the location of the distribution network nodes includes but is not limited to the high-voltage side of the transformer, the low-voltage side of the transformer, the busbar, and the feeder. The electrical parameters of the distribution network nodes can reflect the changes in the power quality in the distribution network, but because the electrical parameters of different distribution network nodes are quite different, such as the high-voltage side and low-voltage side of the transformer, the data mean of the same type of distribution network nodes in one acquisition cycle is used as the electrical parameter of this type of node in one acquisition cycle; the same type of distribution network nodes are specifically distribution network nodes in the same location area, for example, all nodes at the high-voltage side of the transformer are one type of node, and all nodes at the busbar position are one type of node. The various electrical parameters of the distribution network nodes and each type of electrical parameter are normalized. In this embodiment, the maximum and minimum value normalization method is adopted.

[0055] Furthermore, the environmental parameters of the distributed power access nodes in the new energy power generation system are collected. This embodiment takes into account the access of two types of new energy sources: wind power generation and photovoltaic power generation. Among them, wind power generation is mainly affected by wind speed, temperature and humidity, and photovoltaic power generation is mainly affected by light intensity, temperature and humidity. Therefore, this application collects wind speed data through the anemometer in the wind power generation system; collects light intensity data through the light intensity sensor in the photovoltaic system; and collects temperature data and humidity data of each new energy power generation system through temperature sensors and humidity sensors.

[0056] In this embodiment, wind speed and light intensity are collected once per second, while temperature and humidity are collected once per minute. The various environmental parameters collected by each new energy power generation system are organized into corresponding data sequences in chronological order, namely wind speed, light intensity, temperature, and humidity. Each environmental parameter sequence is then normalized, using the maximum and minimum value normalization method in this embodiment.

[0057] All electrical parameters and environmental parameters collected within a preset time period are combined into a sample. In this embodiment, the preset time period is set to 1 hour, and the implementer can adjust it according to actual conditions.

[0058] S2: Based on the preset acquisition cycle, the distribution of voltage imbalance characteristics within all acquisition cycles in the sample is combined with the run-length test algorithm to obtain a random measure of the three-phase imbalance change. The similarity of the trend changes of voltage and current in each phase of each distribution network node in each acquisition cycle is analyzed to obtain the average similarity measure. Combined with the change characteristics of the other electrical parameters, the grid oscillation coefficient is obtained.

[0059] The integration of a large number of distributed renewable energy generation systems into the distribution network will cause certain disturbances to the power quality of the distribution network. These disturbances can be mainly divided into two categories: one is the generation of faults, such as three-phase imbalance faults; the other is the disturbance of the distribution network caused by power generation fluctuations due to the random characteristics of renewable energy generation. Traditional methods determine whether there is a three-phase imbalance fault by calculating the three-phase imbalance in the power system. This method can provide a relatively accurate judgment when the three-phase imbalance fault is obvious or more serious. However, in the early stages of a three-phase imbalance fault or when it is affected by large fluctuations in renewable energy generation, the accuracy of traditional methods is often low and they cannot effectively identify these relatively weak fluctuations. There is an urgent need to improve their detection sensitivity.

[0060] When three-phase imbalance exists in the system, it causes asymmetry in the voltage and current of different phases. This asymmetry manifests as three-phase imbalance, and the changing trends and characteristics of the electrical parameters between the phases are also different, and the duration is relatively long. Secondly, as the load on each phase and the access of distributed energy resources change, the corresponding electrical parameters of each phase will also change, making the changes in three-phase imbalance have a strong random characteristic. At the same time, the imbalance will cause the reactive power of the power system to increase, resulting in a significant decrease in the power factor. However, the change in the power factor caused by the change in renewable energy generation is relatively small.

[0061] Specifically, the voltage imbalance of each distribution network node in each acquisition cycle is calculated separately, and the average voltage imbalance of all distribution network nodes in the same acquisition cycle is the three-phase voltage imbalance of the acquisition cycle. When the three-phase voltage imbalance degree of a collection cycle exceeds the threshold When the imbalance value of the corresponding collection period is marked as the first preset value, otherwise it is marked as the second preset value; in this embodiment, the first preset value is 1 and the second preset value is 0, which can be adjusted by the implementer; the imbalance values of all collection periods are arranged in the order of the collection periods to form an imbalance value sequence; then the imbalance value sequence and the significance level are compared. As input, the runs test is used to obtain the probability value q and the Z-statistic. ; Where S represents the randomness measure of three-phase unbalance change; They are the probability value and Z-statistic output by the runs test, is the significance level.

[0062] In this embodiment, the threshold The value range of , in this example, the value is 2%, and the significance level is The recommended value range is [0.01, 0.05]. In this embodiment, 0.05 is taken, and the implementer can adjust it according to the actual situation. In addition, the calculation of voltage imbalance and run test are existing well-known technologies, and this application will not elaborate on them.

[0063] The voltage and current data of each distribution network node in each acquisition cycle are used as input, and the seasonal trend decomposition using Loess (STL) time series decomposition algorithm is used to output the corresponding trend sequence. Then, using voltage as an example, the average trend sequence of each phase in the same acquisition cycle is calculated. The elements in the average trend sequence are the means of the elements at the same index position in the trend sequence of each distribution network node on the same phase line. The phase line to which each node belongs is obtained through the distribution management system. In addition, STL time series decomposition is a well-known technology and will not be described in detail in this application.

[0064] Obtain the mean of the similarity coefficients between the average trend sequences of all two-phase voltage combinations in each acquisition cycle, and record it as the first mean. Record the mean of the first mean values across all acquisition cycles as the voltage similarity metric. Correspondingly, obtain the mean of the similarity coefficients between the average trend sequences of all two-phase current combinations in each acquisition cycle, and record it as the second mean. Record the mean of the second mean values across all acquisition cycles as the current similarity metric. It should be understood that six similarity coefficients can be obtained for the same acquisition cycle, namely, three for voltage and three for current. In this embodiment, the similarity coefficients can be obtained using the Pearson correlation coefficient; in other embodiments, they can be obtained using the Spearman correlation coefficient. Record the mean of the voltage similarity metric and the current similarity metric as the average similarity metric.

[0065] Based on the above analysis, the grid oscillation coefficient is calculated to measure the oscillation fluctuation of power quality in the distribution network.

[0066] : is the grid oscillation coefficient, is the average similarity measure; It is a measure of the randomness of the three-phase unbalance variation; It is a measure of the three-phase imbalance of the distribution network and is determined by the mean of the three-phase voltage imbalance over all acquisition cycles; is the average rate of change of the power factor in the distribution network, which is determined by the mean of the average rate of change of the power factor of all nodes; It is a parameter adjustment coefficient to avoid the denominator being 0, and its value is 1.

[0067] It is understandable that when the distribution network is connected to new energy and there is a three-phase imbalance fault, the three-phase imbalance is large, and the correlation between the change trends of the electrical parameters of each phase due to the three-phase imbalance is small, that is, the smaller the average similarity measure, the more random the three-phase imbalance change is, and the power factor drops significantly, that is, the rate of change of the drop is large, so the corresponding grid oscillation coefficient is larger.

[0068] S3: Extract the mutation points of wind speed and light intensity in the preset time period respectively; divide the first time interval between adjacent mutation points corresponding to the light intensity, and obtain the degree of influence of light intensity in combination with the distribution characteristics of electrical parameters; analyze the distribution characteristics of electrical parameters in the second time interval between adjacent mutation points corresponding to wind speed, obtain the degree of influence of instantaneous wind speed and the degree of influence of continuous wind speed, and obtain the abnormal fluctuation coefficient of the distributed power access node in combination with the degree of influence of light intensity and the distribution characteristics of other environmental parameters.

[0069] The grid oscillation coefficient can, to a certain extent, measure whether there is a three-phase unbalanced fault in the distribution network when the new energy is connected. However, due to the complex operating environment of the distribution network, the fluctuation of the new energy power supply will also cause disturbances to the distribution network, and the randomness of this disturbance is relatively strong, which is difficult to accurately predict. As a result, the grid oscillation coefficient can only be clearly distinguished when the fault is obvious in the distribution network. When the three-phase unbalanced fault is relatively minor, various characteristics are not obvious enough. In addition, due to a certain degree of disturbance caused by the new energy power supply, there is a certain error in the grid oscillation coefficient. Therefore, it is necessary to further optimize the grid oscillation coefficient in combination with the changes in new energy power generation.

[0070] The new energy sources in this embodiment include photovoltaic power generation and wind power generation. The main factors affecting the quality of photovoltaic power generation are changes in light intensity and temperature and humidity, while the main factors affecting the quality of wind power generation are changes in wind speed and temperature and humidity. Under normal circumstances, changes in light intensity and temperature and humidity are relatively gentle and thus do not significantly affect the quality of photovoltaic power generation. When there is a sudden change in light intensity, it will cause a decrease in output power. However, due to the relatively slow response speed of the photovoltaic system, such a sudden change usually does not immediately cause significant output power fluctuations. Therefore, only when the sudden change in light intensity is persistent will it cause significant fluctuations in output power. Secondly, in a high temperature and high humidity environment, a water film may form on the surface of the photovoltaic cell, affecting the transmittance of light and thus reducing the output power. At the same time, the corresponding resistivity in a high temperature and high humidity environment is also relatively large, which will further cause fluctuations in power generation quality. Therefore, temperature and humidity fluctuations have a positive correlation with output power fluctuations.

[0071] Under normal circumstances, wind speed changes are relatively gradual and usually do not cause significant fluctuations in wind power generation systems. When wind speed changes abnormally, it will cause fluctuations in wind power output. Wind speed changes are divided into instantaneous changes and continuous changes. Instantaneous wind speed changes can cause a sudden increase or decrease in the force on the wind turbine blades, which has a significant impact on the system and causes large fluctuations in output power. Such fluctuations may trigger the protection mechanism in the distribution network, which may be interpreted as a serious system failure. Sustained high wind speeds can cause the output power to remain at a high level for a long time, which may be misdiagnosed as a serious power quality problem or equipment failure, resulting in unnecessary protection action. Secondly, rapid changes in temperature and humidity can also cause changes in blade material or friction, which in turn affects the balance and performance of the blades, causing significant fluctuations in wind power output, which can be misdiagnosed as an equipment failure or serious power quality problem, resulting in unnecessary protection action.

[0072] However, in these cases, the fluctuations in the output power of the renewable energy power generation system are not caused by faults, but by normal environmental influences. Therefore, it is necessary to distinguish between the impact of environmental fluctuations and the impact of real faults based on environmental factors.

[0073] The light intensity or wind speed sequence collected by each distributed power access node is used as input. A mutation point detection algorithm is used to output the mutation points. This algorithm is a well-known technique and will not be described in detail here. At each distributed power access node, the absolute value of the difference in collection time between adjacent mutation points is calculated based on the acquired mutation points, and this is recorded as the collection time difference. All acquisition time differences acquired by each distributed power access node are used as input. The Otsu threshold segmentation method is used to determine the segmentation threshold for that node. This algorithm is a well-known technique and will not be described in detail here.

[0074] Distributed power generation access nodes are divided into photovoltaic access nodes and wind power access nodes. For a photovoltaic access node, the time period of all acquisition time differences greater than the split threshold is used as the photovoltaic effective time. For a wind power access node, the time period of the acquisition time differences greater than the split threshold is used as the continuous wind speed time, and the time period less than the split threshold is used as the instantaneous wind speed time. The acquisition time difference corresponding to the photovoltaic access node is recorded as the first time interval, and the acquisition time difference corresponding to the wind power access node is recorded as the second time interval.

[0075] Based on the above analysis, the abnormal fluctuation coefficient is calculated to measure the abnormal fluctuation of the environment, so as to determine the power supply fluctuation of each new energy power generation system and connect the distributed power supply to the node. The abnormal fluctuation coefficient is recorded as , its formula form is: ; It is a distributed power access node Abnormal fluctuation coefficient; It is a distributed power access node Temperature and humidity fluctuations; They are the photovoltaic access node set and the wind power access node set; is the degree of influence of light intensity on the distributed generation access node; Distributed power access nodes The degree of influence of instantaneous wind speed and the degree of influence of continuous wind speed. In this embodiment, It is a distributed power access node The mean of the variance of the temperature series and humidity series.

[0076] The method for obtaining the degree of influence of light intensity is as follows: obtain the product of the dispersion of the output power of each node in each photovoltaic effective time and the average change rate of the output power at all adjacent moments, and perform forward fusion on the mean of the product of each node in all photovoltaic effective times and the mean of the first time interval corresponding to all photovoltaic effective times to obtain the degree of influence of light intensity. In this embodiment, the average change rate is specifically the average change rate of the output power of the next moment relative to the previous moment in each photovoltaic effective time; the average of the product of each node in all photovoltaic effective times is recorded as , the mean value of each node in the first time interval corresponding to all photovoltaic effective times is recorded as , then the formula for the influence of light intensity is: ; Among them, exp() is an exponential function with a natural constant as the base.

[0077] Furthermore, distributed power access nodes The impact of instantaneous wind speed The formula is: ; is the number of instantaneous wind speed times, They are The variance and average rate of change of output power within the instantaneous wind speed period, It is a distributed power access node The average wind speed in all instantaneous wind speed time periods, Represents a natural constant.

[0078] Distributed power access node The impact of sustained wind speed The formula is: ; is the number of hours of sustained wind speed, They are The variance and mean of the output power during the continuous wind speed period, It is a distributed power access node The variance is used to measure the fluctuation of output power, and the mean is used to measure whether it is in a continuous high power operation state.

[0079] It is understandable that when distributed power is connected to the node For photovoltaic access nodes, the longer the duration of sustained sudden changes in light intensity and the greater the fluctuation in output power during the corresponding time period, the more severe the impact of photovoltaic power generation on the power quality of the distribution network. Furthermore, the greater the temperature and humidity fluctuations, the greater the fluctuation in output power. In this case, the abnormal fluctuation coefficient is large. In this case, the output power fluctuation is not caused by a true system fault, but rather by the impact of environmental changes on the power generation system, not a fault. Conversely, when environmental fluctuations are small or have a smaller impact on output power, the corresponding abnormal fluctuation coefficient is also small, and the abnormal fluctuation is more likely to be caused by a true system fault.

[0080] Similarly, when distributed power is connected to the node It is a wind power access node. If the instantaneous wind speed or continuous wind speed and the temperature and humidity change, which leads to the change of the blade work, and thus affects the output power of the wind turbine motor, the abnormal fluctuation coefficient will be larger, and the corresponding abnormal fluctuation is more likely to be caused by environmental fluctuations; on the contrary, when the abnormal fluctuation is more likely to be a real system failure, the abnormal fluctuation coefficient is smaller.

[0081] S4: Based on the numerical value of the abnormal fluctuation coefficient and in combination with the power grid oscillation coefficient, fault detection is performed on the sample.

[0082] Collect multiple samples of known faults and obtain the grid oscillation coefficient and the average abnormal fluctuation coefficient of all distributed power access nodes of each sample. These two indicators constitute the characteristic vector of the sample. The number of samples in this embodiment is 100. The value intervals of the abnormal fluctuation coefficients of all samples are divided into Segments, all feature vectors of each segment are taken as input, clustering algorithm is used to output the corresponding cluster clusters. In this embodiment The value is 5, and the clustering algorithm used is the DBSCAN clustering algorithm, which is a well-known technology and will not be described in detail. Among all the clusters corresponding to each segment, the cluster with the smallest mean value of the power grid oscillation coefficient is taken as the optimal cluster, and the optimal clusters in all segments are obtained.

[0083] For the elements in the optimal cluster, the average abnormal fluctuation coefficient is used as the horizontal coordinate and the grid oscillation coefficient is used as the vertical coordinate. Polynomial fitting is performed on all elements to output the optimal fitting curve. This fitting curve reflects the relatively minimum electrical parameter fluctuation of the renewable energy power generation system under different environmental fluctuations, and to a certain extent measures the disturbance of renewable energy to the power quality of the distribution network.

[0084] When subsequently performing fault detection on a sample, the average abnormal fluctuation coefficient of the sample is first input into the optimal fitting curve. The difference between the actual calculated grid oscillation coefficient and the function value of the fitting curve is used as the fault detection value for the sample. This fault detection value is used to detect faults when new energy is connected to the distribution system. Specifically, this method can be used to obtain fault detection values for each sample under different circumstances, cluster all fault detection values, and determine the fault condition of the sample to be detected.

[0085] Among them, the flow chart for obtaining the fault detection value is as follows: Figure 2 shown.

[0086] The present application provides a new energy access fault detection method considering the three-phase imbalance of the distribution system, the method comprising: firstly analyzing the changes in electrical parameters of the distribution network nodes, and based on the distribution of the imbalance characteristics of the voltage in all acquisition cycles in the sample, combining with the run-length test algorithm, obtaining a random measure of the three-phase imbalance change, which helps to improve the detection accuracy and reliability of the grid imbalance fault in the dynamic grid environment; analyzing the similarity of the trend changes of the voltage and current of each phase of each distribution network node in each acquisition cycle, obtaining an average similarity measure, evaluating the grid operation state, and combining the change characteristics of the remaining electrical parameters to obtain the grid oscillation coefficient, which can fully reflect the power quality fluctuation in the distribution network, and helps to enhance the accuracy of the three-phase imbalance fault identification; then analyzing the changes in the environmental parameters of the distributed power access nodes, respectively extracting the mutation points of the wind speed and light intensity in the preset time period; and performing the time interval between adjacent mutation points corresponding to the light intensity. Segmentation, combined with the distribution characteristics of electrical parameters, the degree of influence of light intensity is obtained, and the distribution characteristics of electrical parameters are combined to evaluate the impact of light intensity on the power grid, and the degree of influence of light intensity mutation on the power grid load change is identified; further analysis of the distribution characteristics of electrical parameters in the time interval between the mutation points corresponding to wind speed is carried out to obtain the degree of influence of instantaneous wind speed and the degree of influence of continuous wind speed, which helps to better understand the impact of wind speed on power grid load fluctuations; combined with the degree of influence of light intensity and the distribution characteristics of other environmental parameters, the abnormal fluctuation coefficient of distributed power access node is obtained, which can enhance the accuracy of identifying abnormal fluctuations in new energy power generation according to different new energy types; finally, using these indicators, fault detection can be dynamically adjusted according to the abnormal situation of new energy power generation fluctuations. This fault detection method adapted to the access of new energy can help reduce the power quality problems caused by fluctuations in new energy power generation and ensure the stability and reliability of power supply.

[0087] The flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the systems, methods and computer program products according to the embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of the code, and the part of the module, program segment or code contains one or more executable instructions for realizing the specified logical function. In some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, which can depend on the functions involved. In the description corresponding to the flowcharts and block diagrams in the accompanying drawings, the operations or steps corresponding to different boxes can also occur in an order different from that disclosed in the description, and sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, which can depend on the functions involved. Each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented by a dedicated hardware-based system that performs the specified function or action, or may be implemented by a combination of dedicated hardware and computer instructions.

[0088] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A new energy access fault detection method considering three-phase imbalance in the power distribution system is characterized by: The method comprises the following steps: The electrical parameters of the distribution network nodes and the environmental parameters of the distributed power access nodes collected within a preset time period are combined into a sample; the types of the electrical parameters and environmental parameters are at least 3, where the electrical parameters include voltage and current; and the environmental parameters include wind speed and light intensity; A collection cycle is preset, and based on the distribution of voltage imbalance characteristics in all collection cycles in the sample, the imbalance value sequence of all collection cycles is obtained. The run test algorithm is used to obtain the probability value and the Z-statistic. When the probability value is less than the significance level of the run test, 0 is used as the randomness measure of the three-phase imbalance change; otherwise, the Z-statistic is used as the randomness measure of the three-phase imbalance change; the similarity of the trend changes of voltage and current in each phase of each distribution network node in each collection cycle is analyzed to obtain the average similarity measure, and the mean of the three-phase voltage imbalance of all collection cycles is used as the three-phase imbalance measure of the distribution network, denoted as N; the average change rate of the power factor of each distribution network node is calculated, and the average of the average change rates of all distribution network nodes is recorded as the average change rate of the power factor of the distribution network, denoted as Q; the grid oscillation coefficient is recorded as A, and its formula form is: ; Where P is the average similarity measure; Z is the randomness measure of the three-phase imbalance change; is the parameter adjustment coefficient; Extract the mutation points of wind speed and light intensity in the preset time period respectively; segment the first time interval between adjacent mutation points corresponding to the light intensity, and obtain the degree of influence of the light intensity by combining the distribution characteristics of the electrical parameters; use the threshold segmentation algorithm for all second time intervals corresponding to the wind speed to obtain the second segmentation threshold, and use the time period corresponding to the second time interval greater than the second segmentation threshold as the continuous wind speed time; use the time period corresponding to the remaining second time interval as the instantaneous wind speed time, analyze the distribution characteristics of the electrical parameters in different wind speed times, and obtain the instantaneous wind speed influence degree and the continuous wind speed influence degree. The specific formula for the instantaneous wind speed influence degree of the distributed power supply access node is: ;in, Indicates the impact of the instantaneous wind speed at the distributed generation access node j; is the number of instantaneous wind speed times; 、 They are The variance and average rate of change of output power within a certain instantaneous wind speed period; It is a distributed power access node The average wind speed in all instantaneous wind speed time periods; e represents a natural constant; the mean variance of all environmental parameters other than light intensity and wind speed of the distributed power access node is calculated; when the distributed power access node belongs to the photovoltaic access node set, the sum of the light intensity influence degree and the mean variance is used as the abnormal fluctuation coefficient of the distributed power access node; when the distributed power access node data peak point access node set, the cumulative sum of the instantaneous wind speed influence degree, the continuous wind speed influence degree, and the mean variance is used as the abnormal fluctuation coefficient of the distributed power access node; Based on the numerical value of the abnormal fluctuation coefficient and in combination with the power grid oscillation coefficient, fault detection is performed on the sample.

2. The method for detecting a new energy access fault considering three-phase imbalance in a power distribution system according to claim 1, characterized in that: The imbalance value sequence of all acquisition cycles is obtained as follows: Calculate the voltage imbalance of each distribution network node in each acquisition cycle, and take the average of the voltage imbalances of all distribution network nodes in each acquisition cycle as the three-phase voltage imbalance of each acquisition cycle; When the three-phase voltage imbalance is greater than a preset threshold, the imbalance value of the corresponding acquisition period is marked as a first preset value, otherwise it is marked as a second preset value; wherein the first preset value is not equal to the second preset value; The sequence consisting of the imbalance values of all acquisition cycles is recorded as the imbalance value sequence.

3. The method for detecting a new energy access fault considering three-phase imbalance in a power distribution system according to claim 1, wherein: The average similarity measure is obtained as follows: Obtain the trend sequence corresponding to the voltage of each distribution network node in each acquisition period; calculate the average trend sequence of the trend sequence of all nodes in each phase voltage in each acquisition period; obtain the mean of the similarity coefficients between the average trend sequences of all two-phase voltage combinations in each acquisition period, recorded as the first mean; Recording the mean of the first mean values of all acquisition periods as a voltage similarity metric; Accordingly, the current similarity measure is obtained based on the trend sequence corresponding to the current of each node in all acquisition cycles; The average of the voltage similarity metric and the current similarity metric is recorded as the average similarity metric.

4. The method for detecting a new energy access fault considering three-phase imbalance in a power distribution system according to claim 1, wherein: The process of obtaining the influence degree of light intensity is as follows: A threshold segmentation algorithm is applied to all first time intervals corresponding to the light intensity to obtain a first segmentation threshold, and a time period corresponding to a first time interval greater than the first segmentation threshold is used as a photovoltaic effective time; The product of the discreteness of the output power of each distributed power access node in each photovoltaic effective time and the average change rate of the output power at all adjacent moments is obtained, and the mean of the product of each distributed power access node in all photovoltaic effective times is forward fused with the mean of the first time interval corresponding to all photovoltaic effective times to obtain the degree of influence of light intensity.

5. The method for detecting new energy access faults considering three-phase imbalance in a power distribution system according to claim 1, characterized in that: The specific process of obtaining the degree of influence of the sustained wind speed is as follows: According to the number of continuous wind speed periods corresponding to each distributed generation access node and the distribution characteristics of the output power within each continuous wind speed period, the influence degree of the continuous wind speed of the distributed generation access node is obtained.

6. The method for detecting a new energy access fault considering three-phase imbalance in a power distribution system according to claim 5, characterized in that: The formula for the influence degree of the sustained wind speed of the distributed power access node is: ;in, Indicates the impact of the continuous wind speed at the distributed generation access node j; is the number of hours of sustained wind speed, 、 They are The variance and mean of the output power during the continuous wind speed period, It is a distributed power access node The mean value of all sustained wind speed times corresponding to the second time interval.

7. The method for detecting a new energy access fault considering three-phase imbalance in a power distribution system according to claim 1, wherein: The fault detection of the sample based on the numerical value of the abnormal fluctuation coefficient and the grid oscillation coefficient is specifically as follows: The mean of the abnormal fluctuation coefficients of all distributed power access nodes in each sample is recorded as the sample abnormal fluctuation coefficient of each sample; The vector consisting of the power grid oscillation coefficient and the sample abnormal fluctuation coefficient is used as the feature vector of each sample; Based on the value range of the abnormal fluctuation coefficient of the sample in all samples, all characteristic vectors are equally divided; each characteristic vector is clustered, and the cluster with the smallest mean value of the power grid oscillation coefficient is taken as the optimal cluster; Perform curve fitting based on all elements in all optimal clusters to obtain a fitting curve; Obtain the difference between the function value of the grid oscillation coefficient of each sample and the sample abnormal fluctuation coefficient on the fitting curve as the fault detection value of each sample; The fault detection value of each sample detection data is clustered with the fault detection value of the sample detection data of known faults to determine the fault of each sample.

Images