A fault diagnosis system and method suitable for microgrid

Abstract

The present application discloses a fault diagnosis system and method suitable for a microgrid, which relates to the technical field of microgrid fault diagnosis, and solves the technical problem that the fault of the microgrid may be caused by equipment failure or transmission line, and only the electrical parameters on the output side of each device in the microgrid are considered to diagnose the microgrid fault, resulting in low diagnostic accuracy; a number of detection areas are obtained by obtaining the device coordinates of each microgrid device and performing cluster analysis; a fault risk coefficient of the detection area is generated according to the device type and the number of historical faults; detection nodes in the detection area are set according to the fault risk coefficient; change data of the detection node is generated according to preprocessed data; a detection abnormality coefficient of the detection node is generated according to the change data; an abnormal area and an alarm signal are generated according to the detection abnormality coefficient, the number of detection nodes and the location of the detection nodes are dynamically set, the abnormal area is accurately found and an alarm is issued, and the detection accuracy of the abnormal area is improved.

Description

Technical Field

[0001] The present application relates to the field of microgrid fault diagnosis, and specifically to a fault diagnosis system and method suitable for microgrids. Background Art

[0002] With the rapid development of distributed energy and renewable energy, microgrids are becoming increasingly important as an important form of achieving efficient energy utilization and distributed management. A microgrid is an autonomous power system that can achieve self-control, protection and management. It can be connected to an external power grid or operate in an isolated grid. However, the complexity and diversity of microgrids make fault diagnosis a very challenging task. Traditional fault diagnosis methods often have problems such as slow response speed and low accuracy, which makes it difficult to meet the real-time and accurate diagnosis needs of microgrids.

[0003] The prior art (invention patent with announcement number CN112072656B) discloses a microgrid fault handling method and centralized monitoring system, which realizes online automatic identification of fault points or equipment that really need to be dispatched offline for maintenance. The microgrid includes: multiple devices B1, and devices A1, A2, ..., and Am, which are cascaded to the power grid in sequence, m≥0; each device B1 has a communication connection with the centralized monitoring system; the input end of device A1 is connected to multiple parallel output devices B1. The method includes: judging whether each device B1 has an abnormal output side electrical parameter; if at least one device B1 has an abnormal output side electrical parameter, judging whether the proportion of devices B1 with abnormal output side electrical parameters exceeds a preset value; if it exceeds, reporting that at least one of devices A1, A2, ..., Am, and the power grid has a fault, and each device B1 has no fault; if it does not exceed, reporting that the device B1 with abnormal output side electrical parameters has a fault.

[0004] The above-mentioned microgrid fault monitoring system obtains the output side electrical parameters of each device to determine whether the output side electrical parameters of each device are abnormal; and then determines whether each device is abnormal; because the failure of the microgrid may be caused by equipment failure or transmission line failure, only the output side electrical parameters of each device in the microgrid are considered to diagnose the microgrid fault; resulting in low accuracy of microgrid fault diagnosis, so the method of microgrid fault diagnosis still needs to be further improved. Summary of the invention

[0005] The present application aims to solve at least one of the technical problems existing in the prior art; to this end, the present application proposes a fault diagnosis system and method suitable for a microgrid, which is used to solve the technical problem that the fault of the microgrid may be caused by equipment failure or transmission line failure, and only the electrical parameters on the output side of each device in the microgrid are considered to diagnose the microgrid fault, resulting in low accuracy of microgrid fault diagnosis.

[0006] To achieve the above-mentioned purpose, the first aspect of the present application provides a fault diagnosis system suitable for a microgrid, comprising: a node setting module, a data acquisition module, a data analysis module, an early warning module and a database;

[0007] The node setting module: obtains the device coordinates of each microgrid device; performs cluster analysis on the device coordinates to obtain a number of detection areas; obtains the device type corresponding to the microgrid device in each detection area and the corresponding historical fault number through the database connected thereto; generates the fault risk coefficient of each detection area according to each of the device types and the historical fault number; and sets the detection nodes in each detection area according to the fault risk coefficient;

[0008] The data acquisition module is used to obtain the detection data of each detection node in real time and pre-process the detection data to obtain pre-processed detection data;

[0009] The data analysis module: obtains pre-processed detection data of each detection node; generates change data corresponding to each detection node according to the pre-processed data; generates detection anomaly coefficients of each detection node according to the change data; generates abnormal areas and alarm signals according to the detection anomaly coefficients;

[0010] The early warning module is used to issue an early warning according to the alarm signal.

[0011] The present application obtains the device coordinates of each microgrid device, performs cluster analysis on the device coordinates to obtain several detection areas; obtains the device type corresponding to the microgrid device in each detection area and its corresponding historical fault number; generates a fault risk coefficient for each detection area according to each device type and the historical fault number; sets the detection nodes in each detection area according to the fault risk coefficient; obtains the detection data of each detection node; and preprocesses it to obtain preprocessed detection data; obtains the preprocessed detection data of each detection node; generates change data corresponding to each detection node according to the preprocessed data; generates a detection anomaly coefficient for each detection node according to the change data; generates an abnormal area and an alarm signal according to the detection anomaly coefficient, dynamically sets the number of detection nodes and the location of detection nodes, identifies detection node anomalies through changes in detection values ​​of detection nodes, accurately finds abnormal areas and issues alarms, and improves the detection accuracy of abnormal areas.

[0012] Furthermore, cluster analysis is performed on the device coordinates to obtain several detection areas, including:

[0013] Step 1: Obtain the coordinate data of each device, and perform data standardization operations on each coordinate data to obtain standard coordinate data;

[0014] Step 2: Select K-Means algorithm for cluster analysis; select K points as initial cluster centers, K is set based on experience;

[0015] Step 3: Calculate the distance from each standard coordinate data to all cluster centers and assign it to the cluster where the nearest cluster center is located; for each cluster, recalculate the mean of all points in it;

[0016] Step 4: Repeat step 3 until the cluster center no longer changes significantly or the preset number of iterations is reached;

[0017] Step 5: The microgrid device corresponding to the final cluster and the physical link between each microgrid device are used as a detection area.

[0018] Furthermore, the failure risk coefficient of each detection area is generated according to each device type and the number of historical failures, including:

[0019] The maximum value of the number of failures of several scrapped microgrid devices of the same type and the same manufacturer as the device is obtained through the database and marked as ZG; the maximum value of the number of failures of several scrapped microgrid devices of the same type and the same manufacturer as the device is obtained and marked as CG;

[0020] The number of historical faults is marked as DGi; i is the number of the microgrid equipment in the detection area;

[0021] The risk coefficient FXi corresponding to the microgrid equipment is calculated by the formula FXi=α1×exp[-(ZG-DGi)]+α2×exp[-(CG-DGi)]; wherein α1 and α2 are weight coefficients;

[0022] The fault risk coefficient GX of the corresponding detection area is calculated by the formula GX=∑(FXi); wherein i=1, 2, ..., N; and N is the total number of microgrid devices in the corresponding detection area.

[0023] Furthermore, the setting of detection nodes in each detection area according to the fault risk coefficient includes:

[0024] Obtain the fault risk coefficient GX of each detection area and the total number N of microgrid devices corresponding to the detection area;

[0025] By formula The number of detection nodes JS in the corresponding detection area is calculated; k is the proportional coefficient; DX is the risk coefficient threshold of the corresponding detection area; is the rounding symbol;

[0026] Obtain the microgrid devices in each detection area and the connection relationship between each microgrid device; and establish a microgrid topology map corresponding to each detection area;

[0027] Obtain the physical link of the device node corresponding to each microgrid device in the microgrid topology diagram; set the detection node position according to the physical link corresponding to each microgrid device and the number of detection nodes corresponding to the corresponding detection area.

[0028] This application calculates the minimum number of standard detection nodes that should be set in the detection area by obtaining the fault risk coefficient of the detection area and the total number of microgrid equipment in the detection area. While meeting the detection requirements of each detection area, the number of detection points is set as small as possible, thereby reducing resource waste.

[0029] Furthermore, the microgrid topology diagram is constructed by the following steps:

[0030] Obtain the total number and coordinate information of each microgrid device in each detection area, as well as the physical links between each microgrid device;

[0031] Each microgrid device in the detection area is regarded as a device node, and the physical link between each microgrid device is regarded as the physical link of each device node; a microgrid topology map corresponding to each detection area is constructed.

[0032] Further, the setting of the detection node position according to the physical link corresponding to each microgrid device and the number of detection nodes corresponding to the corresponding detection area includes:

[0033] Obtain the number of detection nodes JS in each detection area and the risk factor FXi of each microgrid device in the corresponding detection area; and the link attributes of the physical link between each microgrid device;

[0034] Generate a distribution coefficient FP corresponding to each physical link according to the risk factor of each microgrid device and the link attribute of the physical link;

[0035] The allocation coefficients corresponding to each physical link in the detection area are normalized to obtain the normalized allocation coefficients; Calculate the number of link detection nodes LJ corresponding to the corresponding physical link; calculate the number of link detection nodes of each physical link in the corresponding detection area in turn;

[0036] According to the number of link detection nodes corresponding to each physical link, the number of link detection nodes and the preset detection node positions are evenly set on each corresponding physical link; the risk coefficients of the microgrid devices connected on both sides of the physical link are obtained and marked as FXa and FXb respectively;

[0037] Determine whether |FXa-FXb| is greater than the set difference threshold;

[0038] If yes, then obtain the distance JDL between each preset detection node position and the microgrid devices connected on both sides of the physical link with a large risk coefficient; calculate the offset distance TZJ of the corresponding preset detection node position by the formula TZJ=exp[-DFJ / JDL×max(FXa, FXb)]×DFJ, and obtain the offset distance of each preset detection node in turn; and adjust each preset detection node position toward the microgrid device with a large risk coefficient according to the offset distance to obtain the detection node position;

[0039] If not, obtain the risk coefficient corresponding to the microgrid device with the shorter distance between each preset detection node position and the microgrid devices connected on both sides of the physical link, marked as FXx; its corresponding distance is marked as JDLx; calculate the offset distance TZJ of each preset detection node position by the formula TZJ=exp[-DFJ / JDLx×FXx]×DFJ; obtain the offset distance of each preset detection node in turn; and adjust the position of each preset detection node toward the microgrid device with the shorter distance according to the offset distance to obtain the detection node position; wherein DFJ is the distance between each preset detection node.

[0040] The present application obtains the number of detection nodes in the detection area and the risk coefficient of each microgrid device in the corresponding detection area, as well as the link properties of the physical links between the microgrid devices; generates an allocation coefficient for each physical link according to the risk coefficient of each microgrid device and the link properties of the physical link; dynamically allocates a corresponding number of nodes to each physical link according to the allocation coefficient; and then dynamically sets the detection node position in the physical link according to the link properties of each physical link and the risk coefficient of the microgrid device connected to the physical link, thereby improving the effectiveness of the detection node setting.

[0041] Furthermore, generating a distribution coefficient corresponding to each physical link according to the risk coefficient of each microgrid device and the link attribute of the physical link includes:

[0042] The link length in the link attribute corresponding to each physical link is obtained and marked as LC; and the sum of the risk factors of the microgrid devices on both sides of the corresponding physical link is marked as HFX;

[0043] The allocation coefficient FP of the corresponding physical link is calculated by the formula FP=a×arctan(β1×LC / DLC+β2×HFX); wherein a is the proportional coefficient; β1 and β2 are weight coefficients; and DLC is the unit link length.

[0044] Furthermore, the preprocessing includes data cleaning and normalization operations.

[0045] Furthermore, the generating of change data corresponding to each detection node according to the preprocessed data includes:

[0046] Obtain several detection values ​​corresponding to each detection node in the detection area in the preprocessed data; obtain several detection values ​​corresponding to each detection node in the detection area in the corresponding last preprocessed data; make a difference between several identical detection values ​​of the detection nodes in the detection area to obtain the change values ​​of several detection values ​​of the corresponding detection nodes; integrate the change values ​​of several detection values ​​of each detection node into change data.

[0047] Furthermore, the generating of the detection anomaly coefficient of each detection node in the detection area according to the change data includes:

[0048] Extract the change value BHj corresponding to several detection values ​​in the change data corresponding to each detection node; calculate the detection abnormality coefficient YC of the corresponding detection node by the formula YC=∑[βj×exp(-DWj / (BHj+WTj))]; wherein DWj is the unit change coefficient corresponding to the detection value numbered j; WTj is the adjustment coefficient corresponding to the detection value numbered j; βj is the proportional coefficient corresponding to the detection value numbered j; j is the number corresponding to the detection value, and j=1, 2,…, Q; Q is the total number of corresponding change values.

[0049] Further, the generating of abnormal area and alarm signal according to the detected abnormal coefficient includes:

[0050] Obtain the detection anomaly coefficient of each detection node;

[0051] When the detection abnormality coefficient corresponding to the detection node is greater than the abnormality threshold; the detection node is marked as an abnormal node and an alarm signal is generated; otherwise, the detection node is marked as a normal node;

[0052] The detection anomaly coefficients of all abnormal nodes are obtained; each detection anomaly coefficient is input into the anomaly radius generation model to obtain the anomaly distance corresponding to each abnormal node; the circular area with the abnormal node as the center and the abnormal distance as the radius is the abnormal area corresponding to the abnormal node; the abnormal radius generation model is obtained through artificial intelligence model training.

[0053] Another aspect of the present invention provides a fault diagnosis method applicable to a microgrid, comprising:

[0054] S0: Obtain the device coordinates of each microgrid device; perform cluster analysis on the device coordinates to obtain several detection areas; obtain the device type corresponding to the microgrid device in each detection area and the corresponding historical fault number; generate the fault risk coefficient of each detection area according to each of the device types and historical fault numbers; set the detection nodes in each detection area according to the fault risk coefficient;

[0055] S1: Acquire detection data of each detection node; and preprocess it to obtain preprocessed detection data;

[0056] S2: Generate change data corresponding to each detection node according to the preprocessed data; generate detection anomaly coefficients of each detection node according to the change data; generate abnormal areas and alarm signals according to the detection anomaly coefficients;

[0057] S3: Issue early warning based on alarm signals.

[0058] Compared with the prior art, the beneficial effects of this application are:

[0059] 1. The present application obtains the device coordinates of each microgrid device, performs cluster analysis on the device coordinates to obtain several detection areas; obtains the device type corresponding to the microgrid device in each detection area and its corresponding historical fault number; generates a fault risk coefficient for each detection area according to each device type and the historical fault number; sets the detection nodes in each detection area according to the fault risk coefficient; obtains the detection data of each detection node; and preprocesses it to obtain preprocessed detection data; obtains the preprocessed detection data of each detection node; generates change data corresponding to each detection node according to the preprocessed data; generates a detection anomaly coefficient for each detection node according to the change data; generates an abnormal area and an alarm signal according to the detection anomaly coefficient, dynamically sets the number of detection nodes and the location of detection nodes, identifies detection node anomalies through changes in detection values ​​of detection nodes, accurately finds abnormal areas and issues alarms, and improves the detection accuracy of abnormal areas.

[0060] 2. This application calculates the minimum number of standard detection nodes that should be set in the detection area by obtaining the fault risk coefficient of the detection area and the total number of microgrid equipment in the detection area. It can dynamically allocate the corresponding number of nodes to different detection areas, and set the number of detection points as small as possible while meeting the detection requirements of each detection area, thereby reducing resource waste.

[0061] 3. The present application obtains the number of detection nodes in the detection area and the risk coefficient of each microgrid device in the corresponding detection area, as well as the link properties of the physical links between each microgrid device; generates an allocation coefficient for each physical link according to the risk coefficient of each microgrid device and the link properties of the physical link; dynamically allocates a corresponding number of nodes to each physical link according to the allocation coefficient; and then dynamically sets the detection node position in the physical link according to the link properties of each physical link and the risk coefficient of the microgrid device connected to the physical link, thereby improving the effectiveness of the detection node setting. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0063] Figure 1 A schematic diagram of the principle of a fault diagnosis system applicable to a microgrid in the present application;

[0064] Figure 2 A flowchart of a fault diagnosis method applicable to a microgrid in the present application;

[0065] Figure 3 The microgrid topology diagram of this application;

[0066] Figure 4 A flowchart for setting node positions for this application;

[0067] Figure 5 Generate example maps for the anomaly regions in this application. DETAILED DESCRIPTION

[0068] The technical solution of the present application will be described clearly and completely below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0069] See also Figure 1-Figure 2 , the first aspect of the present application provides a fault diagnosis system suitable for a microgrid, including: a node setting module, a data acquisition module, a data analysis module, an early warning module and a database;

[0070] Node setting module: obtain the device coordinates of each microgrid device; perform cluster analysis on the device coordinates to obtain several detection areas; and obtain the device type corresponding to the microgrid device in each detection area and the corresponding historical fault number through the database connected to it; generate the fault risk coefficient of each detection area according to each device type and the historical fault number; set the detection nodes in each detection area according to the fault risk coefficient; the detection area refers to the clustering area of ​​the microgrid equipment obtained by analyzing the device location of the microgrid equipment through cluster analysis; the equipment type includes distributed power supply and energy storage device, etc.; the fault risk coefficient refers to the possibility of failure, and the greater the fault risk coefficient, the more likely it is to fail; the detection node is a node set to detect data in the operation of the microgrid;

[0071] Data acquisition module: acquires the detection data of each detection node in real time; and preprocesses it to obtain preprocessed detection data; the detection data includes several detection values ​​corresponding to each detection node, including current detection values ​​and voltage detection values, etc.;

[0072] Data analysis module: obtain pre-processed detection data of each detection node; generate change data corresponding to each detection node according to the pre-processed data; generate detection anomaly coefficients of each detection node according to the change data; generate abnormal areas and alarm signals according to the detection anomaly coefficients; change data refers to the change amount corresponding to each detection value in each detection data; detection anomaly coefficient refers to the possibility of abnormality in the detection data;

[0073] Early warning module: issue early warning based on alarm signals.

[0074] This embodiment obtains the device coordinates of each microgrid device, performs cluster analysis on the device coordinates to obtain several detection areas; obtains the device type corresponding to the microgrid device in each detection area and its corresponding historical fault number; generates a fault risk coefficient for each detection area according to each device type and the historical fault number; sets the detection nodes in each detection area according to the fault risk coefficient; obtains the detection data of each detection node; and preprocesses it to obtain preprocessed detection data; obtains the preprocessed detection data of each detection node; generates change data corresponding to each detection node according to the preprocessed data; generates a detection anomaly coefficient of each detection node according to the change data; generates an abnormal area and an alarm signal according to the detection anomaly coefficient, dynamically sets the number of detection nodes and the location of detection nodes, identifies detection node anomalies through changes in detection values ​​of detection nodes, accurately finds abnormal areas and issues alarms, and improves the detection accuracy of abnormal areas.

[0075] Cluster analysis of device coordinates yields several detection areas, including:

[0076] Step 1: Obtain the coordinate data of each device, and perform data standardization on each coordinate data to obtain standard coordinate data; the coordinate data includes longitude data and latitude data; the longitude data range is -180° to 180°, and the latitude data range is -90° to 90°, and the minimum-maximum scaling method is selected for data standardization;

[0077] Step 2: Select K-Means algorithm for cluster analysis; select K points as initial cluster centers, K is set based on experience;

[0078] Step 3: Calculate the distance from each standard coordinate data to all cluster centers and assign it to the cluster where the nearest cluster center is located; for each cluster, recalculate the mean of all points in it;

[0079] Step 4: Repeat step 3 until the cluster center no longer changes significantly or the preset number of iterations is reached;

[0080] Step 5: The microgrid device corresponding to the final cluster and the physical link between each microgrid device are used as a detection area.

[0081] The failure risk coefficient of each detection area is generated according to each device type and the number of historical failures, including: obtaining the maximum number of failures of several scrapped microgrid devices with the same device type and the same manufacturer through the database and marking it as ZG; obtaining the maximum number of failures of several scrapped microgrid devices with the same device type but different manufacturers and marking it as CG; the number of failures that occur when the current device reaches scrapping is less than the maximum number of failures of scrapped microgrid devices of the same manufacturer as the device, so the maximum number of failures of the scrapped microgrid device is selected as the factor affecting the risk coefficient of the microgrid device.

[0082] The number of historical faults is marked as DGi; the number of historical faults is the number of faults that occurred on the current device since it left the factory; i is the number of the microgrid device in the detection area;

[0083] The risk coefficient FXi corresponding to the microgrid equipment is calculated by the formula FXi=α1×exp[-(ZG-DGi)]+α2×exp[-(CG-DGi)]; where α1 and α2 are weight coefficients, and α1>α2, α1 and α2 are set according to experience. The failure risk of the current equipment is closely related to its same manufacturer, so α1>α2; as the number of historical failures increases, the risk coefficient corresponding to the current microgrid equipment becomes larger;

[0084] This embodiment obtains the probability of abnormality of each microgrid device as a risk coefficient through the above formula; when the number of failures of the microgrid is closer to the maximum number of failures, the corresponding risk coefficient is higher; it can be understood that the evaluation criteria for the number of failures of microgrid devices in use and scrapped microgrid devices are the same; scrapped microgrid equipment is a microgrid device that is used normally until the service life of the corresponding equipment reaches the set number of years and is scrapped.

[0085] The fault risk coefficient GX of the corresponding detection area is calculated by the formula GX=∑(FXi); wherein i=1, 2, …, N; N is the total number of microgrid devices in the corresponding detection area; as the number of microgrid devices in the detection area increases and the risk coefficient of the microgrid devices increases, the fault risk coefficient of the detection area will also increase accordingly.

[0086] Setting detection nodes in each detection area according to the fault risk coefficient includes: obtaining the fault risk coefficient GX of each detection area and the total number N of microgrid devices corresponding to the detection area;

[0087] By formula The number of detection nodes JS corresponding to the detection area is calculated; where k is the proportional coefficient, which is set based on experience; is the rounding symbol; as the fault risk coefficient of the detection area increases, the number of detection nodes in the corresponding detection area will also increase; DX is the risk coefficient threshold of the corresponding detection area; the risk coefficient threshold DX is set to a fixed value based on experience; it can also be calculated by the formula DX=∑{α1×exp[-(ZG-DGi)]+α2×exp[-(CG-DGi)]} to obtain a non-fixed value set for different detection areas, where DGi=0, i=1, 2, …, N; where α1 and α2 are weight coefficients, and α1>α2, α1 and α2 are set based on experience; in calculating the detection nodes of the detection area The method of rounding up when calculating the number of points is to obtain an integer number of detection nodes; the risk coefficient threshold is expressed as the risk coefficient when the current microgrid device has no faults. If the maximum value ZG of the number of failures of several scrapped microgrid devices with the same device type and the same manufacturer and the maximum value CG of the number of failures of several scrapped microgrid devices with the same device type and different manufacturers are larger, it means that the corresponding microgrid device has a greater probability of abnormality during use, and more detection nodes are required in the later stage; when the risk coefficient is similar, the more detection nodes should be set, and the corresponding risk coefficient threshold should decrease accordingly.

[0088] Obtain the microgrid devices in each detection area and the connection relationship between each microgrid device; and establish a microgrid topology map corresponding to each detection area;

[0089] Obtain the physical link of the device node corresponding to each microgrid device in the microgrid topology diagram; set the detection node position according to the physical link corresponding to each microgrid device and the number of detection nodes corresponding to the corresponding detection area.

[0090] This embodiment calculates the minimum standard number of detection nodes that should be set in the detection area by obtaining the fault risk coefficient of the detection area and the total number of microgrid devices in the detection area. The corresponding number of nodes can be dynamically allocated to different detection areas. Different numbers of detection nodes are equipped according to the risk coefficients of different areas. This ensures that the detection data corresponding to the detection nodes in each detection area is sufficient, while minimizing the waste of resources caused by setting the number of nodes.

[0091] See also Figure 3 ,The microgrid topology diagram is constructed through the following steps:

[0092] Obtain the total number and coordinate information of each microgrid device in each detection area, as well as the physical links between each microgrid device;

[0093] Each microgrid device in the detection area is regarded as a device node, and the physical link between each microgrid device is regarded as the physical link of each device node; a microgrid topology map corresponding to each detection area is constructed.

[0094] See also Figure 4 , the detection node positions are set according to the physical links corresponding to each microgrid device and the number of detection nodes corresponding to the corresponding detection area, including:

[0095] Obtain the number of detection nodes JS in each detection area and the risk factor FXi of each microgrid device in the corresponding detection area; and the link attributes of the physical link between each microgrid device, the link attributes including the link length, etc.;

[0096] Generate a distribution coefficient FP corresponding to each physical link according to the risk coefficient of each microgrid device and the link attribute of the physical link; the distribution coefficient refers to the standard for allocating the number of detection nodes to each physical link. The larger the distribution coefficient, the more detection nodes are allocated;

[0097] The allocation coefficients corresponding to each physical link in the detection area are normalized to obtain the normalized allocation coefficients; Calculate the number of link detection nodes LJ corresponding to the corresponding physical link; calculate the number of link detection nodes of each physical link in the corresponding detection area in turn; in the formula, the larger the allocation coefficient is and the more detection nodes are in the detection area, the more the number of detection nodes of the corresponding physical link will increase, and the rounding up method is used for calculation in order to obtain the optimal number of detection nodes; the number of detection nodes in the detection area in this embodiment is a minimum standard value, and in practice the sum of the number of detection nodes allocated to each physical link may be slightly more than the number of detection nodes in the detection area;

[0098] According to the number of link detection nodes corresponding to each physical link, the number of link detection nodes and the preset detection node positions are evenly set on each corresponding physical link; the risk coefficients of the microgrid devices connected on both sides of the physical link are obtained and marked as FXa and FXb respectively;

[0099] Determine whether |FXa-FXb| is greater than the set difference threshold; the difference threshold is set based on experience;

[0100] If yes, then obtain the distance JDL between each preset detection node position and the microgrid devices connected on both sides of the physical link with a large risk coefficient; calculate the offset distance TZJ of the corresponding preset detection node position through the formula TZJ=exp[-DFJ / JDL×max(FXa, FXb)]×DFJ, and obtain the offset distance of each preset detection node in turn; and adjust each preset detection node position toward the microgrid device with a large risk coefficient according to the offset distance to obtain the detection node position; as FXa or FXb increases, the offset distance of the detection node position farther from the microgrid device with a large risk coefficient will increase accordingly;

[0101] If not, obtain the risk coefficient corresponding to the microgrid device with the shorter distance between each preset detection node position and the microgrid devices connected on both sides of the physical link, marked as FXx; its corresponding distance is marked as JDLx; calculate the offset distance TZJ of each preset detection node position through the formula TZJ=exp[-DFJ / JDLx×FXx]×DFJ; obtain the offset distance of each preset detection node in turn; and adjust the position of each preset detection node toward the microgrid device with the shorter distance according to the offset distance to obtain the detection node position; wherein DFJ is the distance between each preset detection node; as JDLx increases, the offset distance of the detection node will increase accordingly.

[0102] This embodiment obtains the number of detection nodes in the detection area and the risk coefficient of each microgrid device in the corresponding detection area, as well as the link properties of the physical links between the microgrid devices; generates an allocation coefficient for each physical link according to the risk coefficient of each microgrid device and the link properties of the physical link; dynamically allocates a corresponding number of nodes to each physical link according to the allocation coefficient; and then dynamically sets the detection node position in the physical link according to the link properties of each physical link and the risk coefficient of the microgrid device connected to the physical link, thereby improving the effectiveness of the detection node setting.

[0103] Generate a distribution coefficient corresponding to each physical link according to the risk coefficient of each microgrid device and the link attribute of the physical link, including: obtaining a link length in the link attribute corresponding to each physical link marked as LC; and a sum of the risk coefficients of the microgrid devices on both sides of the corresponding physical link marked as HFX;

[0104] The allocation coefficient FP of the corresponding physical link is calculated by the formula FP=a×arctan(β1×LC / DLC+β2×HFX); where a is the proportional coefficient; β1 and β2 are weight coefficients; they are set according to experience; DLC is the unit link length; as the link length of the physical link becomes longer and the sum of the risk coefficients of the microgrid devices connected to the physical link becomes larger, the allocation coefficient of the corresponding physical link will become larger.

[0105] In this embodiment, a set of data is obtained, in which the link length LC is 2m, the sum of the risk coefficients of the microgrid devices at both ends of the link is 1.5, the unit link length DLC is set to 1m, β1 and β2 are set to 0.3 and 0.7, and a is set to 0.9; the allocation coefficient FP of the physical link is calculated by the formula, FP=0.9×arctan(0.3×(2m / 1m)+0.7×1.5)=0.9×1.03=0.927.

[0106] Preprocessing includes data cleaning and normalization operations, including:

[0107] Data cleaning: directly remove missing values ​​from the acquired test data to obtain normal data;

[0108] Normalization operation: Normalize the normal data using the maximum and minimum normalization method to obtain the preprocessed detection data.

[0109] Generate change data corresponding to each detection node based on the preprocessed data, including:

[0110] Obtain several detection values ​​corresponding to each detection node in the detection area in the preprocessing data; the detection values ​​include current detection values ​​and voltage detection values, etc.; obtain several detection values ​​corresponding to each detection node in the detection area in the corresponding last preprocessing data; make a difference between several identical detection values ​​of the detection nodes in the detection area to obtain the change values ​​of several detection values ​​of the corresponding detection nodes; integrate the change values ​​of several detection values ​​of each detection node into change data. When the detected microgrid is an AC microgrid, the time interval between this preprocessing data and the corresponding last preprocessing data is set according to experience; it should be set to an integer multiple of the least common multiple of the corresponding cycles of each detection value; when the microgrid is a DC microgrid, the time interval can be set according to experience; when the microgrid is a DC-AC hybrid grid, the time interval is set according to the type of power transmission of the line corresponding to the detection point. If it is AC, it is set to an integer multiple of the least common multiple of the corresponding cycles of each detection value; for DC, it can be set according to experience.

[0111] Generate the detection anomaly coefficient of each detection node in the detection area based on the change data, including:

[0112] Extract the change values ​​BHj corresponding to several detection values ​​in the change data corresponding to each detection node; calculate the detection anomaly coefficient YC of the corresponding detection node by the formula YC=∑[βj×exp(-DWj / (BHj+WTj))]; wherein DWj is the unit change coefficient corresponding to the detection value numbered j; WTj is the adjustment coefficient corresponding to the detection value numbered j; βj is the proportional coefficient corresponding to the detection value numbered j; j is the number corresponding to the detection value, and j=1, 2,…, Q; Q is the total number of corresponding change values; as the change value corresponding to the detection value increases, the detection anomaly coefficient of the detection node will also increase accordingly;

[0113] The proportionality coefficient βj is obtained by the probability of abnormality of the corresponding detection value in each historical fault data; specifically, the difference ZCjk between each detection value and the normal value in several historical fault data, and the number of times each difference is not zero are obtained; and the number of times corresponding to each detection value is normalized to obtain the normalized number of times, and marked as GCj; the proportionality coefficient βj corresponding to the detection value numbered j is calculated by the formula βj=GCj×∑[arctan(|ZCjk|)]; wherein k is the number of the historical fault data, k=1, 2,…, K; K is the total number of historical fault data; as the absolute value of the difference between each detection value and the normal value increases, its corresponding proportionality coefficient will also increase accordingly.

[0114] Generate abnormal areas and alarm signals based on the detected abnormality coefficients, including:

[0115] Obtain the detection anomaly coefficient of each detection node;

[0116] When the detection abnormality coefficient corresponding to the detection node is greater than the abnormality threshold; the detection node is marked as an abnormal node and an alarm signal is generated; otherwise, the detection node is marked as a normal node;

[0117] The detection anomaly coefficients of all abnormal nodes are obtained; each detection anomaly coefficient is input into the anomaly radius generation model to obtain the anomaly distance corresponding to each abnormal node; the circular area with the abnormal node as the center and the abnormal distance as the radius is the abnormal area corresponding to the abnormal node; the anomaly radius generation model is obtained through artificial intelligence model training.

[0118] The alarm signal in this embodiment may also be generated by the abnormal area; when the number of times the overlapping parts of the abnormal area overlap is greater, the level of the alarm signal of the corresponding overlapping area is higher.

[0119] See also Figure 5In this embodiment, five detection nodes are set. The abnormal nodes include abnormal node 1, abnormal node 2 and abnormal node 3. The abnormal node 1 corresponds to abnormal area 1, the abnormal node 2 corresponds to abnormal area 2, and the abnormal node 3 corresponds to abnormal area 3. The normal nodes include normal node 1 and normal node 2. The abnormal area 3 overlaps 3 times and has the highest corresponding alarm level.

[0120] The anomaly radius generation model is obtained through artificial intelligence model training, including:

[0121] The abnormal coefficients of several detection nodes and the line distance between the fault location and the corresponding detection node are obtained through the database; the line distance is 1.15 times the length of the shortest connection line from the fault location to the detection node, and the number of times the shortest connection line length can be set according to the experience of relevant experts; the line distance is used as the abnormal radius;

[0122] The anomaly coefficient and its corresponding anomaly radius are integrated into training data and test data; the training data is used to train the artificial intelligence model; the trained artificial intelligence model is tested using the test data, and finally an anomaly radius generation model with the anomaly coefficient as input and the anomaly radius as output is obtained; the artificial intelligence model includes a BP neural network model or an RBF neural network model.

[0123] Another aspect of the present application provides a fault diagnosis method applicable to a microgrid, including:

[0124] S0: Obtain the device coordinates of each microgrid device; perform cluster analysis on the device coordinates to obtain several detection areas; obtain the device type corresponding to the microgrid device in each detection area and its corresponding historical fault count; generate the fault risk coefficient of each detection area according to each device type and historical fault count; set the detection nodes in each detection area according to the fault risk coefficient;

[0125] S1: Acquire detection data of each detection node; and preprocess it to obtain preprocessed detection data;

[0126] S2: Generate change data corresponding to each detection node according to the preprocessed data; generate detection anomaly coefficients of each detection node according to the change data; generate abnormal areas and alarm signals according to the detection anomaly coefficients;

[0127] S3: Issue early warning based on alarm signals.

[0128] Part of the data in the above formula is calculated by removing the dimension and taking its numerical value. The formula is a formula closest to the actual situation obtained by software simulation of a large amount of collected data; the preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained through simulation of a large amount of data.

[0129] The working principle of the present application is as follows: by obtaining the device coordinates of each microgrid device, clustering the device coordinates to obtain a number of detection areas; obtaining the device type corresponding to the microgrid device in each detection area and its corresponding historical fault number; generating a fault risk coefficient for each detection area according to each device type and the historical fault number; setting detection nodes in each detection area according to the fault risk coefficient; obtaining the detection data of each detection node; and preprocessing it to obtain preprocessed detection data; obtaining the preprocessed detection data of each detection node; generating change data corresponding to each detection node according to the preprocessed data; generating a detection anomaly coefficient for each detection node according to the change data; generating an abnormal area and an alarm signal according to the detection anomaly coefficient, dynamically setting the number of detection nodes and the location of the detection nodes, identifying the abnormality of the detection node through the change of the detection value of the detection node, accurately finding the abnormal area and issuing an alarm, thereby improving the detection accuracy of the abnormal area, avoiding the problem that the failure of the microgrid may be caused by equipment failure or transmission line failure, and only considering the output side electrical parameters of each device in the microgrid to diagnose the microgrid fault, resulting in low diagnostic accuracy.

[0130] The above embodiments are only used to illustrate the technical method of the present application and are not intended to limit it. Although the present application has been described in detail with reference to the preferred embodiments, a person of ordinary skill in the art should understand that the technical method of the present application may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present application.

Claims

1. A fault diagnosis system suitable for a microgrid, characterized in that: include: Node setting module, data collection module, data analysis module, early warning module and database; The node setting module: obtains the device coordinates of each microgrid device; performs cluster analysis on the device coordinates to obtain a number of detection areas; obtains the device type and the corresponding historical fault number of the microgrid device in each detection area through a database connected thereto; generates a fault risk coefficient for each detection area according to each of the device types and the historical fault number; Set detection nodes in each detection area according to the fault risk factor; The failure risk coefficient of each detection area is generated according to each device type and the number of historical failures, including: The maximum value of the number of failures of several scrapped microgrid devices of the same type and the same manufacturer as the device is obtained through the database and marked as ZG; the maximum value of the number of failures of several scrapped microgrid devices of the same type and the same manufacturer as the device is obtained and marked as CG; The number of historical faults is marked as DGi; i is the number of the microgrid equipment in the detection area; The risk coefficient FXi corresponding to the microgrid equipment is calculated by the formula FXi=α1×exp[-(ZG-DGi)]+α2×exp[-(CG-DGi)]; wherein α1 and α2 are weight coefficients; The fault risk coefficient GX of the corresponding detection area is calculated by the formula GX=∑(FXi); wherein i=1, 2, ..., N; N is the total number of microgrid devices in the corresponding detection area; Set detection nodes in each detection area according to the fault risk factor, including: Obtain the fault risk coefficient GX of each detection area and the total number N of microgrid devices corresponding to the detection area; By formula The number of detection nodes JS in the corresponding detection area is calculated; k is the proportional coefficient; DX is the risk coefficient threshold of the corresponding detection area; is the rounding symbol; Obtain the microgrid devices in each detection area and the connection relationship between each microgrid device; and establish a microgrid topology map corresponding to each detection area; Obtain the physical links of the device nodes corresponding to each microgrid device in the microgrid topology diagram; set the detection node positions according to the physical links corresponding to each microgrid device and the number of detection nodes corresponding to the corresponding detection area; The detection node positions are set according to the physical links corresponding to each microgrid device and the number of detection nodes corresponding to the corresponding detection area, including: Obtain the number of detection nodes JS in each detection area and the risk factor FXi of each microgrid device in the corresponding detection area; and the link attributes of the physical link between each microgrid device; Generate a distribution coefficient FP corresponding to each physical link according to the risk factor of each microgrid device and the link attribute of the physical link; The allocation coefficients corresponding to each physical link in the detection area are normalized to obtain the normalized allocation coefficients; Calculate the number of link detection nodes LJ corresponding to the corresponding physical link; calculate the number of link detection nodes of each physical link in the corresponding detection area in turn; According to the number of link detection nodes corresponding to each physical link, the number of link detection nodes and the preset detection node positions are evenly set on each corresponding physical link; the risk coefficients of the microgrid devices connected on both sides of the physical link are obtained and marked as FXa and FXb respectively; Determine whether |FXa-FXb| is greater than the set difference threshold; If yes, then obtain the distance JDL between each preset detection node position and the microgrid devices connected on both sides of the physical link with a large risk coefficient; calculate the offset distance TZJ of the corresponding preset detection node position by the formula TZJ=exp[-DFJ / JDL×max(FXa, FXb)]×DFJ, and obtain the offset distance of each preset detection node in turn; and adjust each preset detection node position toward the microgrid device with a large risk coefficient according to the offset distance to obtain the detection node position; If not, the risk coefficient corresponding to the microgrid device with the shortest distance between each preset detection node position and the microgrid devices connected on both sides of the physical link is obtained, and marked as FXx; its corresponding distance is marked as JDLx; the offset distance TZJ of each preset detection node position is calculated by the formula TZJ=exp[-DFJ / JDLx×FXx]×DFJ; the offset distance of each preset detection node is obtained in turn; and the position of each preset detection node is adjusted to the microgrid device with the shortest distance according to the offset distance to obtain the detection node position; wherein DFJ is the distance between each preset detection node; The data acquisition module is used to obtain the detection data of each detection node in real time and pre-process the detection data to obtain pre-processed detection data; The data analysis module: obtains pre-processed detection data of each detection node; generates change data corresponding to each detection node according to the pre-processed data; generates detection anomaly coefficients of each detection node according to the change data; generates abnormal areas and alarm signals according to the detection anomaly coefficients; The early warning module is used to issue an early warning according to the alarm signal.

2. A fault diagnosis system suitable for a microgrid according to claim 1, characterized in that: Cluster analysis is performed on the device coordinates to obtain several detection areas, including: Step 1: Obtain the coordinate data of each device, and perform data standardization operations on each coordinate data to obtain standard coordinate data; Step 2: Select K-Means algorithm for cluster analysis; select K points as initial cluster centers, K is set based on experience; Step 3: Calculate the distance from each standard coordinate data to all cluster centers and assign it to the cluster where the nearest cluster center is located; for each cluster, recalculate the mean of all points in it; Step 4: Repeat step 3 until the cluster center no longer changes significantly or the preset number of iterations is reached; Step 5: The microgrid device corresponding to the final cluster and the physical link between each microgrid device are used as a detection area.

3. A fault diagnosis system suitable for a microgrid according to claim 1, characterized in that: The step of generating a distribution coefficient corresponding to each physical link according to the risk coefficient of each microgrid device and the link attribute of the physical link includes: The link length in the link attribute corresponding to each physical link is obtained and marked as LC; and the sum of the risk factors of the microgrid devices on both sides of the corresponding physical link is marked as HFX; The allocation coefficient FP of the corresponding physical link is calculated by the formula FP=a×arctan(β1×LC / DLC+β2×HFX); wherein a is the proportional coefficient; β1 and β2 are weight coefficients; and DLC is the unit link length.

4. A fault diagnosis system suitable for a microgrid according to claim 1, characterized in that: The step of generating change data corresponding to each detection node according to the preprocessed data includes: Obtain several detection values ​​corresponding to each detection node in the detection area in the preprocessed data; obtain several detection values ​​corresponding to each detection node in the detection area in the corresponding last preprocessed data; make a difference between several identical detection values ​​of the detection nodes in the detection area to obtain the change values ​​of several detection values ​​of the corresponding detection nodes; integrate the change values ​​of several detection values ​​of each detection node into change data.

5. A fault diagnosis system suitable for a microgrid according to claim 1, characterized in that: The step of generating the detection anomaly coefficient of each detection node in the detection area according to the change data includes: Extract the change value BHj corresponding to several detection values ​​in the change data corresponding to each detection node; calculate the detection abnormality coefficient YC of the corresponding detection node by the formula YC=∑[βj×exp(-DWj / (BHj+WTj))]; wherein DWj is the unit change coefficient corresponding to the detection value numbered j; WTj is the adjustment coefficient corresponding to the detection value numbered j; βj is the proportional coefficient corresponding to the detection value numbered j; j is the number corresponding to the detection value, and j=1, 2,…, Q; Q is the total number of corresponding change values.

6. A fault diagnosis system suitable for a microgrid according to claim 1, characterized in that: The generating of abnormal area and alarm signal according to the detected abnormal coefficient comprises: Obtain the detection anomaly coefficient of each detection node; When the detection abnormality coefficient corresponding to the detection node is greater than the abnormality threshold; the detection node is marked as an abnormal node and an alarm signal is generated; otherwise, the detection node is marked as a normal node; The detection anomaly coefficients of all abnormal nodes are obtained; each detection anomaly coefficient is input into the anomaly radius generation model to obtain the anomaly distance corresponding to each abnormal node; the circular area with the abnormal node as the center and the abnormal distance as the radius is the abnormal area corresponding to the abnormal node; the abnormal radius generation model is obtained through artificial intelligence model training.

7. A fault diagnosis method applicable to a microgrid, applied to the operation of a fault diagnosis system applicable to a microgrid as claimed in any one of claims 1 to 6, characterized in that: The following steps are involved: S0: Obtain the device coordinates of each microgrid device; perform cluster analysis on the device coordinates to obtain several detection areas; obtain the device type corresponding to the microgrid device in each detection area and the corresponding number of historical failures; generate the failure risk coefficient of each detection area according to each of the device types and historical failure times; Set detection nodes in each detection area according to the fault risk factor; S1: Acquire detection data of each detection node; and preprocess it to obtain preprocessed detection data; S2: Generate change data corresponding to each detection node according to the preprocessed data; generate detection anomaly coefficients of each detection node according to the change data; generate abnormal areas and alarm signals according to the detection anomaly coefficients; S3: Issue early warning based on alarm signals.

Images