A method for identifying the fault section and fault branch of a wind farm collector line

Abstract

The present invention belongs to the technical field of line fault identification, and specifically discloses a method for identifying fault sections and fault branches of a wind farm collector line. By segmenting the wind farm collector line and using the fault records of each line segment to screen out the fault-prone sections, and then analyzing the fault evolution trend of the historical fault data of these fault-prone sections. On this basis, in the operation monitoring of the current fault-prone sections, even if no abnormality is found, the occurrence time of potential faults can be predicted by using the fault evolution trend, so as to realize the forward-looking identification of hidden faults in the wind farm collector line, significantly improving the timeliness of fault identification. At the same time, the operation monitoring frequency is adjusted according to the predicted fault occurrence time of the current normal line sections, reflecting the preventive measures before the faults occur, ensuring high-frequency monitoring during key time periods, and thus being able to reduce the probability of missing monitoring during the period when the faults have not occurred to a certain extent.

Description

Technical Field

[0001] The present invention belongs to the technical field of line fault identification, and specifically discloses a method for identifying fault sections and fault branches of a wind farm collector line. Background Art

[0002] The collector line of a wind farm is an important part that collects the electric power generated by each wind turbine and transmits it to a booster station or a substation. Since a wind power generation system is usually located in a remote area and the scale of the wind farm is large, the collector line is long and complex, and is prone to failures. In order to ensure the reliable operation of the wind farm, it is necessary to effectively identify the fault sections and fault branches in the collector line.

[0003] Generally, the faults in the wind farm collector line can be divided into two categories: significant faults and hidden faults. Significant faults can be directly identified through physical observation or simple electrical measurement. Such faults usually cause obvious damage or interruption to the system when they occur, such as open circuit and short circuit. Hidden faults are not easily directly observable in the initial stage and gradually deteriorate over time, eventually leading to a decline in system performance or failure. Common hidden faults include insulation aging, overload, and poor contact. Compared with significant faults, hidden faults are more likely to occur in the harsh environment of the wind farm collector line and have a longer identification period. Therefore, more strict monitoring methods need to be adopted.

[0004] At present, the identification of hidden faults in the wind farm collector line mainly relies on real-time monitoring technology. However, the real-time monitoring technology cannot predict the fault occurrence time for the current situation where no fault has occurred, resulting in a lack of predictability in fault identification and the inability to take preventive measures in advance, such as adjusting the monitoring frequency. In addition, during the real-time monitoring process, the monitoring data may be lost or omitted due to reasons such as equipment failure, network interruption, and data transmission error, which will reduce the accuracy and timeliness of fault identification and increase the risk of potential faults not being discovered in time. Summary of the Invention

[0005] For this reason, an object of an embodiment of the present application is to provide a method for identifying fault sections and fault branches of a wind farm collector line for hidden faults in the wind farm collector line. By using historical fault data to increase the prediction of the hidden fault occurrence time, the optimization of hidden fault identification is realized, and the problems mentioned in the background art are effectively solved.

[0006] The object of the present invention can be achieved by the following technical solutions: A method for identifying fault sections and fault branches of a wind farm collector line, comprising the following steps: constructing a power grid topology model according to the layout diagram of the wind farm collector line, capturing key nodes in the line on this basis, and then dividing the line segments according to the connection relationship between adjacent nodes.

[0007] Retrieve the fault records of each line segment, identify the segments prone to faults from them, and at the same time analyze the fault characteristics of the segments prone to faults, specifically the types of faults prone to occur and the fault evolution trends.

[0008] Monitor the operation of the segments prone to faults according to the types of faults prone to occur in the segments prone to faults, and based on the monitoring data, identify whether there are faults, extract the segments prone to faults with faults from them and mark them as the current fault segments, and at the same time locate the position where the current fault segments are located.

[0009] Based on the current monitoring data of the segments prone to faults, screen out the segments prone to faults that do not have faults currently and mark them as the current normal segments, and use the fault evolution trends and the current monitoring data of the current normal segments to predict the fault occurrence time.

[0010] Adjust the operation monitoring frequency according to the predicted fault occurrence time of the current normal segments.

[0011] Combining all the above technical solutions, the positive effects of the present invention are as follows: 1. By segmenting the collector lines of the wind farm and using the fault records of each line segment to screen out the segments prone to faults, and then analyzing the fault evolution trends of the historical fault data of these segments prone to faults, on this basis, in the operation monitoring of the current segments prone to faults, even if no abnormalities are found, the fault evolution trends can be used to predict the fault occurrence time of potential faults, so as to realize the forward-looking identification of hidden faults in the collector lines of the wind farm and significantly improve the timeliness of fault identification.

[0012] 2. By adjusting the operation monitoring frequency based on the predicted fault occurrence time of the current normal line segments, it reflects the preventive measures before the faults occur, ensures high-frequency monitoring during key time periods, and thus can reduce the probability of missing monitoring during the period when the faults do not occur to a certain extent and enhance the timeliness of fault identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] The present invention is further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation to the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained according to the following drawings.

[0014] Figure 1 It is a flowchart of the method implementation steps of the present invention.

[0015] Figure 2 It is a schematic diagram of locating the position where the current fault segment is located as the fault section in the present invention.

[0016] Figure 3 It is a schematic diagram of locating the position where the current fault segment is located as the fault branch in the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] Next, in combination with the accompanying drawings in the embodiments of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0018] See Figure 1 As shown, the present invention proposes a method for identifying fault sections and fault branches of a wind farm collector line, including the following steps: constructing a power grid topology model based on the layout diagram of the wind farm collector line, capturing the nodes in the line on this basis, and then dividing the line segments according to the connection relationship between adjacent nodes.

[0019] It should be noted that the topology diagram of the wind farm collector line is a graphical representation used to show the connection relationship and layout among various devices such as wind turbines, transformers, switchyards, and cable lines within the wind farm.

[0020] Furthermore, it should be noted that the nodes in the line can include various types of support structures such as utility poles and suspension points. The main functions of these nodes are to support and fix the conductors to ensure the safety and reliability of power transmission. By dividing the line segments based on adjacent nodes, physical isolation can be provided between the nodes, reducing electromagnetic interference. On long lines, electromagnetic interference may propagate along the entire line. In segmented monitoring, the electromagnetic interference in each segment is restricted to a local range, reducing the impact on other segments. Therefore, by dividing the wind farm collector line using adjacent nodes, it is more convenient to monitor the operating status of the line.

[0021] Retrieve the fault records of each line segment, identify the fault-prone segments from them, and at the same time analyze the fault characteristics of the fault-prone segments, specifically the fault types and fault evolution trends.

[0022] Applied to the above solution, the process of identifying the fault-prone segments is as follows: Select a specified historical period and count the proportion of the fault records corresponding to each line segment during this period.

[0023] It should be clear that selecting a specified historical period can ensure that the analyzed fault data is from the recent period, and these data can better reflect the current operating status and fault trends of the system. At the same time, selecting an appropriate historical period can ensure that there are sufficient fault record samples. A too short time period may result in insufficient sample quantity, and the results are easily affected by accidental factors. An appropriate historical period can reduce contingency and improve the reliability of the analysis results. Exemplarily, the fault records in the past 12 months can be selected.

[0024] Extract the occurrence time of each fault from the fault records of the same line segment, calculate the time interval between adjacent fault records, and thus obtain the average fault interval duration of each line segment.

[0025] Substitute the average fault interval duration of each line segment and the proportion of fault records into the expression Calculate the fault proneness index of each line segment , where 、 respectively represent the proportion of fault records and the average fault interval duration of the th line segment, represents the line segment number, , represents the duration of the specified historical period, represents the natural constant.

[0026] It can be seen from the above formula that the larger the proportion of fault records of the line segment, the shorter the average fault interval duration, and the larger the fault proneness index.

[0027] Compare the fault proneness index of each line segment with the preset warning value. Exemplarily, the warning value is 0.7. Screen out the line segments whose fault proneness index reaches or exceeds the warning value, and mark them as fault-prone segments.

[0028] The purpose of identifying the fault-prone segments by means of the historical fault records of the line segments in the present invention is to be able to further use the fault characteristics of the fault-prone segments for targeted fault handling. For non-fault-prone segments, due to the lack of fault tendency performance, regular routine inspections and maintenance are carried out on non-fault-prone segments to ensure the good basic operating state of these segments, concentrate more maintenance resources and energy on fault-prone segments, and the input of maintenance resources for non-fault-prone segments can be appropriately reduced to improve resource utilization efficiency.

[0029] Further applied to the above scheme, the fault types prone to occur in the fault-prone segments are as follows: Extract the fault types of each fault from the fault records of the fault-prone line segments, and then summarize the occurrence frequencies of the same fault types.

[0030] Calculate the standard deviation of the occurrence frequencies of each fault type corresponding to the fault-prone section, and compare it with the preset allowable standard deviation. The allowable standard deviation is preset, and its purpose is to determine whether the difference in the occurrence frequencies of fault types is within an acceptable range. If the standard deviation is greater than the allowable standard deviation, it indicates that the difference in the occurrence frequencies of fault types is relatively large. At this time, the fault type with the highest occurrence frequency is taken as the prone fault type of the fault-prone section. This is because the fault type with the highest occurrence frequency best represents the main fault mode on this line; conversely, it indicates that the frequencies of fault types are relatively uniform, which means that each fault type has a certain occurrence frequency, but there is no obvious dominant fault type. Then calculate the occurrence proportion of each fault type and compare it with the median occurrence proportion, where the median occurrence proportion is 0.5, and take the fault type with an occurrence proportion higher than the median occurrence proportion as the prone fault type.

[0031] It should be understood that when the occurrence frequencies of fault types are relatively uniform, selecting the fault type with an occurrence proportion higher than the median occurrence proportion can ensure that the selected fault type has a high representativeness among all fault types.

[0032] Furthermore, applied to the above scheme, the fault evolution trend of the prone fault type corresponding to the fault-prone section is as follows: Extract the fault records that meet the prone fault type from the fault records retrieved from the fault-prone section based on the prone fault type, and record them as valid fault records.

[0033] Retrieve the operation monitoring records of the fault-prone section within a specified historical period, and extract the monitoring time from the operation monitoring records, and then perform an association mapping with the valid fault records to construct a set of operation monitoring records associated with the valid fault records.

[0034] It should be noted that when identifying that a line section has an operation fault, it is achieved by real-time monitoring of its operation status. When the operation status monitoring data shows an abnormality, it will be determined that the line section has an operation fault. Therefore, a series of operation monitoring records correspond to the front end of each fault record.

[0035] In the specific implementation of the above scheme, when performing the association mapping of the operation monitoring records for the valid fault records, by comparing the monitoring time of the retrieved operation monitoring records with the fault occurrence time of the valid fault records, the operation monitoring records with a monitoring time before the fault occurrence time are screened out as the operation monitoring records associated with the valid fault records.

[0036] As an example of the above implementation, assume that at 11:00 on March 1, 2020, the system detected an abnormal insulation resistance in line segment A, marked as an insulation aging fault. At this time, a fault record was generated, recording information such as the time of fault occurrence, fault type, and fault location. For example, the fault record is as follows: Fault ID: 001, Fault time: 2020-03-01 11:00:00, Fault type: Insulation aging, Fault location: Line segment A. A series of operation monitoring records of this line segment before the fault occurrence time were retrieved and compared with the fault occurrence time of this valid fault record. The operation monitoring records with monitoring time before the fault occurrence time were extracted from them as the operation monitoring records associated with the valid fault record.

[0037] Extract the monitoring data from the operation monitoring records, and construct a coordinate system with the monitoring time as the horizontal axis and the monitoring data as the vertical axis. For each valid fault record in the fault-prone section, extract the monitoring time and monitoring data from its associated operation monitoring record set, and generate a fault evolution trend curve corresponding to each valid fault record within the constructed coordinate system.

[0038] It should be noted that the monitoring data extracted from the operation monitoring records above are the required monitoring parameters. When the fault type is insulation aging, the monitoring data extracted from the operation monitoring records is the insulation resistance.

[0039] Monitor the fault-prone sections according to the fault-prone types of the fault-prone sections, and based on the monitoring data, identify whether there is a fault, and extract the fault-prone sections with faults and mark them as the current fault sections, and at the same time locate the location of the current fault sections.

[0040] Preferably, the operation monitoring of the fault-prone sections according to the fault-prone types of the fault-prone sections is implemented as follows: Summarize the number of fault-prone types analyzed from the fault-prone sections. If there is only one fault-prone type, monitor the fault-prone sections based on the required monitoring parameters corresponding to this fault-prone type to obtain real-time monitoring data. If there is more than one fault-prone type, arrange the various fault-prone types in descending order of their occurrence frequencies in the fault records to obtain the monitoring order of the fault-prone types, and accordingly monitor the required monitoring parameters corresponding to the various fault-prone types of the fault-prone sections in sequence.

[0041] In the example of the above preferred operation, the data parameters to be monitored for different fault types are different. For example, when the fault type is poor contact, the required monitoring parameter is the contact resistance. Another example is that when the fault type is insulation aging, the required monitoring parameter is the insulation resistance. Another example is that when the fault type is overload, the required monitoring parameters are the conductor temperature and current.

[0042] It should be noted that when there is more than one type of fault prone to occur, if the demand monitoring parameters for all types of faults prone to occur are monitored simultaneously, this means that the system collects and processes the monitoring data of multiple fault types at the same time, which may lead to resource tension and increase the complexity of the system and the difficulty of data processing. By monitoring the demand monitoring parameters of the types of faults prone to occur in sequence according to the frequency of occurrence of the fault types, this means that the system monitors the fault types one by one in a certain order, giving priority to monitoring the fault types with a high frequency of occurrence, which can utilize resources more efficiently, give priority to monitoring high-risk fault types, and can respond to potential problems faster. In addition, monitoring one by one reduces the pressure of processing a large amount of data at the same time and simplifies the data processing process.

[0043] It should be further noted that if resources are sufficient, a simultaneous monitoring strategy can be considered, because this can ensure that all fault types can be monitored in a timely manner and provide more comprehensive fault information.

[0044] Further preferably, the operation of identifying whether there is a fault is as follows: compare the monitoring data of the fault prone section corresponding to the type of fault prone to occur with the normal operation data. If the monitoring data of a certain type of fault prone to occur does not conform to the normal operation data, it is identified that there is a fault, and this type of fault prone to occur is used as the current fault type.

[0045] In the example of the above operation, when the type of fault prone to occur is insulation aging, compare the insulation resistance of the fault prone section corresponding to insulation aging with the normal insulation resistance. The normal insulation resistance corresponding to the fault prone section can be obtained by referring to the product manual provided by the manufacturer of the cable or equipment used in the line section, which usually contains the recommended normal insulation resistance value.

[0046] Even further preferably, the location of the current fault section is located as follows: locate the two end nodes corresponding to the current fault section from the power grid topology model. If the two end nodes are located on the main line, see Figure 2 as shown, it is identified that the current fault section belongs to the fault section. If the two end nodes are located on the branch line, it is identified that the current fault section belongs to the fault branch, see Figure 3 as shown.

[0047] It should be added that when monitoring the operation of the fault prone section, the focus is on monitoring the demand monitoring parameters of the types of faults prone to occur, which does not mean that other parameters are not monitored. By focusing on monitoring the demand monitoring parameters of the types of faults prone to occur, the monitoring resources can be concentrated on the key parameters most likely to cause faults, which can utilize the limited monitoring resources more efficiently. At the same time, by monitoring the key parameters, the precursors of faults can be detected earlier, and preventive measures can be taken in time to avoid the expansion of faults.

[0048] Based on the current monitoring data of the prone-to-failure sections, select the prone-to-failure sections that currently have no faults and mark them as the current normal sections, and use the fault evolution trend of the current normal sections and the current monitoring data to predict the fault occurrence time.

[0049] Adjust the operation monitoring frequency according to the predicted fault occurrence time of the current normal sections.

[0050] In the above, the specific operation of predicting the fault occurrence time is as follows: Extract the fault evolution trend curves of the corresponding prone-to-failure types of the current normal sections in each valid fault record, and then capture the current monitoring data of the corresponding prone-to-failure types of the current normal sections from the fault evolution trend curves. The point where the current monitoring data is located on the curve is used as the current state point.

[0051] Respectively obtain the time on the horizontal axis of the current state point in the fault evolution trend curves of each valid fault record of the prone-to-failure type as the historical time of the current state in each valid fault record, and obtain the month in which the historical time is located.

[0052] In the example of the above operation, assume that the time on the horizontal axis of the current state point is 14:00 on October 4, 2019, then the month in which the historical time is located is October.

[0053] Compare the month in which the historical time of the current state point is located in each valid fault record with the month in which the current time is located, and select the valid fault records in which the month in which the historical time is located is the same as the month in which the current time is located as the available fault records.

[0054] Obtain the duration on the horizontal axis between the current state point captured by the fault evolution trend curve of the corresponding prone-to-failure type of the current normal section in the available fault records and the end point on the curve as the fault hidden duration.

[0055] Add the current time to the fault hidden duration to obtain the predicted fault occurrence time.

[0056] It should be understood that since the occurrence process of the hidden fault is highly related to the environmental conditions (such as temperature and humidity) of the line. Taking the insulation aging fault as an example, in a high-temperature and high-humidity environment, the insulation aging will deteriorate rapidly. Therefore, in the case where the current insulation resistances are all 200 MΩ and all decay to 100 MΩ, the high-temperature and high-humidity environment will shorten the decay duration of the insulation resistance. Therefore, when predicting the fault occurrence time based on the current monitoring data, the influence of the environmental conditions of the line location on the prediction must be considered.

[0057] Based on this consideration, the month in which the monitoring time of the current monitoring data is located is incorporated into the prediction model because there is a certain correlation between the environmental conditions of a region and the seasonal months. The environmental conditions in the same month in different years are mostly similar. By selecting, based on the month in which the monitoring time of the current monitoring data is located, the valid fault records with the same month of historical time as the current month from the fault evolution trend curves of each valid fault record as the available fault records, it can be ensured to the greatest extent that the prediction of the fault occurrence time is carried out under the same environmental conditions, thereby improving the accuracy and reliability of the prediction results.

[0058] In the innovation of the above solution, the process of predicting the fault occurrence time by using the fault evolution trend of the current normal segment and the current monitoring data also includes the following: when comparing the month of historical time of the current state in each valid fault record with the current month and there is no valid fault record with the same month of historical time as the current month, locate the region spanned by the current normal segment, and thus obtain the meteorological type to which the region spanned by the current normal segment belongs.

[0059] Obtain the seasonal distribution months corresponding to the region spanned by the current normal segment based on the meteorological type to which the region spanned by the current normal segment belongs.

[0060] Exemplarily, the climate types include tropical rainforest climate, tropical savannah climate, tropical monsoon climate, subtropical monsoon climate, etc. Since the climate characteristics corresponding to different climate types are different, the resulting climate distributions are different, which makes the seasonal distribution months corresponding to different climate types different. For example, the spring distribution months corresponding to the subtropical monsoon climate are from March to May, the summer distribution months are from June to August, the autumn distribution months are from September to November, and the winter distribution months are from December to February of the following year.

[0061] Convert the month of historical time of the current state in each valid fault record to the corresponding season, and then conduct a comparison of the same season, and select the valid fault records with the same season of historical time as the current season as the available fault records.

[0062] In the further implementation of the above solution, adjust the operation monitoring frequency according to the predicted fault occurrence time of the current normal segment as follows: compare the fault hiding duration of the fault type prone to occur corresponding to the current normal segment with the set reference fault hiding duration to calculate the fault urgency of the fault type prone to occur , and the specific calculation formula is , where in the formula represents the fault hiding duration, represents the reference fault hiding duration.

[0063] It should be added that the purpose of setting the reference fault hiding duration is to eliminate the dimension of the fault hiding duration in the calculation of the fault urgency. The setting of the reference fault hiding duration should meet the condition that when the fault hiding duration reaches the reference fault hiding duration, the fault is presented as not urgent, and it can be monitored according to the current monitoring frequency. When the fault hiding duration is less than the reference fault hiding duration, the fault is presented as urgent, and the shorter the fault hiding duration, the greater the fault urgency. The specific reference fault hiding duration can be consulted with experienced electrical engineers. Exemplarily, the reference fault hiding duration is 10 days.

[0064] Obtain the current monitoring frequency of the demand monitoring parameters corresponding to the easily-occurring fault types , and substitute it together with the fault urgency into the expression to obtain the adjusted monitoring frequency of the demand monitoring parameters corresponding to the easily-occurring fault types .

[0065] It should be noted that the above-mentioned monitoring frequency is the monitoring frequency within a unit time period. For example, when the unit time period is one month, when the fault urgency is 0, it can be monitored according to the current monitoring frequency. When the fault urgency is greater, the required monitoring frequency is greater than the current monitoring frequency, that is, the monitoring frequency within the unit time period is more.

[0066] Monitor the demand monitoring parameters of the easily-occurring fault types according to the adjusted monitoring frequency.

[0067] The present invention adjusts the operation monitoring frequency based on the predicted fault occurrence time of the current normal line segment, which reflects the preventive measures before the fault occurs, ensures high-frequency monitoring during the critical time period, and thus can reduce the probability of missing monitoring during the period when the fault has not occurred to a certain extent, enhancing the timeliness of fault identification.

[0068] The above content is only an example and explanation of the structure of the present invention. Those skilled in the art of the present technology can make various modifications or supplements to the described specific embodiments or use similar methods to replace them, as long as they do not deviate from the structure of the invention or exceed the scope defined by the present invention, they should all fall within the protection scope of the present invention.

Claims

1. A method for identifying faulty sections and faulty branches of a wind farm collector line, characterized in that: The following steps are involved: The grid topology model is constructed based on the layout diagram of the wind farm collection line, and on this basis, the nodes in the line are captured, and then the line segments are divided according to the connection relationship between adjacent nodes; Retrieve the fault records of each line section, identify the fault-prone sections, and analyze the fault characteristics of the fault-prone sections. The fault characteristics of the fault-prone sections are the fault types and fault evolution trends. Perform operation monitoring on the fault-prone segments according to the fault-prone types of the fault-prone segments, identify whether there are faults based on the monitoring data, extract the fault-prone segments with faults and mark them as current fault segments, and locate the current fault segments; Based on the current monitoring data of the fault-prone segments, the fault-prone segments that do not currently have faults are screened out and marked as the current normal segments. The fault evolution trend of the current normal segments and the current monitoring data are used to predict the time of fault occurrence. Adjust the operation monitoring frequency according to the predicted fault occurrence time of the current normal segment; The process of identifying the fault-prone segment is as follows: Select a specified historical period and count the proportion of fault records corresponding to each line segment within this period; Extract the occurrence time of each fault from the fault records of the same line section, calculate the time interval between adjacent fault records, and thus obtain the average fault interval time of each line section; Substitute the average fault interval time of each line section and the fault record ratio into the expression Calculate the fault susceptibility index of each line section , where , Respectively represent The proportion of fault records and the average fault interval time of the line section, Indicates the line segment number. , Indicates the duration of the specified historical period. represents a natural constant; Compare the fault-prone index of each line section with the preset warning value, select the line section whose fault-prone index reaches or exceeds the warning value, and mark it as a fault-prone section; The operation monitoring frequency of the normal segment is adjusted according to the predicted fault occurrence time as follows: The fault hiding time of the current normal segment corresponding to the prone fault type The reference fault hiding time is set Compare and calculate the fault urgency of the prone fault types , ; Get the current monitoring frequency of the required monitoring parameters corresponding to the prone fault type , and substitute it into the expression with the fault urgency Get the adjustment monitoring frequency of the required monitoring parameters corresponding to the prone fault type ; Perform operational monitoring of the required monitoring parameters for the fault-prone types according to the adjusted monitoring frequency.

2. A method for identifying faulty sections and faulty branches of a wind farm collector line according to claim 1, characterized in that: The types of faults that are prone to occur refer to the following analysis process: Extract the fault type of each fault from the fault records of the prone line section, and then summarize the occurrence frequency of the same fault type; The standard deviation of the occurrence frequency of each fault type corresponding to the fault-prone segment is calculated and compared with the preset allowable standard deviation. If the standard deviation is greater than the allowable standard deviation, the fault type with the highest occurrence frequency is taken as the prone fault type of the fault-prone segment; otherwise, the occurrence ratio of each fault type is calculated and compared with the median occurrence ratio, and the fault type with an occurrence ratio higher than the median occurrence ratio is taken as the prone fault type.

3. A method for identifying faulty sections and faulty branches of a wind farm collector line according to claim 1, characterized in that: The fault evolution trend is shown in the following analysis process: Based on the prone fault type, the fault records retrieved from the prone fault segment are extracted to match the prone fault type and recorded as valid fault records; Retrieve the operation monitoring records of the fault-prone section within the specified historical period, extract the monitoring time from the operation monitoring records, and then associate and map with the valid fault records to build a set of operation monitoring records associated with the valid fault records; Monitoring data are extracted from the operation monitoring records, and a coordinate system is constructed with monitoring time as the horizontal axis and monitoring data as the vertical axis. For each valid fault record in the fault-prone segment, the monitoring time and monitoring data are extracted from its associated set of operation monitoring records, and a fault evolution trend curve corresponding to each valid fault record is generated in the constructed coordinate system.

4. A method for identifying faulty sections and faulty branches of a wind farm collector line according to claim 1, characterized in that: The operation monitoring of the fault-prone section according to the fault-prone section's fault type is implemented as follows: Summarize the number of fault-prone types obtained from the analysis of the fault-prone segments. If there is only one fault-prone type, perform operation monitoring on the fault-prone segment based on the demand monitoring parameters corresponding to the fault-prone type to obtain real-time monitoring data. If there is more than one fault-prone type, arrange each fault-prone type from high to low according to the frequency of occurrence in the fault record to obtain the monitoring order of the fault-prone types, and monitor the demand monitoring parameters corresponding to each fault-prone type in the fault-prone segment in this order.

5. A method for identifying faulty sections and faulty branches of a wind farm collector line according to claim 1, characterized in that: The identification of whether there is a fault is performed as follows: The monitoring data of the prone fault type corresponding to the prone fault segment is compared with the normal operation data. If the monitoring data of a prone fault type does not match the normal operation data, the presence of a fault is identified and the prone fault type is used as the current fault type.

6. A method for identifying faulty sections and faulty branches of a wind farm collector line according to claim 1, characterized in that: The location of the current fault segment is as follows: The two end nodes corresponding to the current fault section are located from the power grid topology model. If the two end nodes are located on the main line, the current fault section is identified as a fault section. If the two end nodes are located on the branch line, the current fault section is identified as a fault branch.

7. A method for identifying faulty sections and faulty branches of a wind farm collector line according to claim 3, characterized in that: The prediction of the fault occurrence time using the fault evolution trend of the current normal segment and the current monitoring data is performed as follows: Extract the fault evolution trend curve of the prone fault type corresponding to the current normal segment in each valid fault record, and then capture the current monitoring data of the prone fault type corresponding to the current normal segment from the fault evolution trend curve to the point where the current monitoring data is located on the curve as the current state point; The time of the current state point on the horizontal axis of the fault evolution trend curve of each valid fault record of the prone fault type is obtained respectively as the historical time of the current state in each valid fault record, and the month in which the historical time is located is obtained; Compare the historical month of the current state point in each valid fault record with the current month, and select the valid fault record whose historical month is consistent with the current month as the available fault record; Obtain the time between the current state point captured by the fault evolution trend curve of the available fault record corresponding to the prone fault type in the current normal segment and the end point on the curve on the horizontal axis as the fault hiding time; Add the current time and the fault hiding time to get the predicted fault occurrence time.

8. A method for identifying faulty sections and faulty branches of a wind farm collector line according to claim 7, characterized in that: The prediction of the fault occurrence time by using the fault evolution trend of the current normal segment and the current monitoring data also includes the following process: When comparing the historical month of the current state in each valid fault record with the current month, if there is no valid fault record whose historical month is consistent with the current month, the area spanned by the current normal section is located, thereby obtaining the meteorological type of the area spanned by the current normal section; Based on the meteorological type of the area that the current normal segment crosses, obtain the seasonal distribution month corresponding to the area that the current normal segment crosses; The historical month of the current status in each valid fault record and the current month are converted to the season of the historical time, and then a seasonal consistency comparison is performed, and the valid fault records whose historical season is consistent with the current season are selected as available fault records.

Images