Intelligent detection method and system for transformer

Abstract

The invention relates to the technical field of transformer detection, in particular to an intelligent detection method and system for a transformer, and the system comprises a detection state dynamic management module which carries out the dynamic detection of a detection state according to an association detection chain between detection units in an abnormal state of the transformer. And dynamically adjusting the operation state of each detection unit in the detection period. When the operation state of the transformer is detected in real time through various sensors, the energy consumption of the sensors in the detection process is reduced by adopting an intermittent detection mode; by combining the historical detection data of the transformer to obtain an associated detection chain between the detection units in the abnormal state of the transformer and the detection data of the corresponding detection units, the dynamic adjustment of the operation state of the detection units is realized, and the influence of the intermittent control mode of the detection units on the detection precision of the abnormal operation state of the transformer is reduced; effective management and control of transformer detection are realized, and high practical value and economic benefit are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of transformer detection, and particularly to an intelligent detection method and system for a transformer. Background Technique

[0002] A transformer is an important device commonly used in the power system, and its performance and operating status directly affect the stable operation and safety of the power system. Therefore, the research and development of an intelligent detection system for transformers are of great significance.

[0003] The existing transformer detection systems mainly adopt the traditional manual inspection and regular maintenance methods, and there are the following problems: First, the efficiency of manual inspection is low, and the operating status of the transformer cannot be monitored in real time; second, it is impossible to conduct a comprehensive and in-depth detection and analysis of the transformer, and it is difficult to discover potential fault hazards in a timely manner; third, it is impossible to realize the real-time monitoring and remote management of the transformer operation data, which limits the timely grasp and handling of the transformer operating status.

[0004] With the rapid development of sensor technology, people can detect the operating status of transformers in real time through a variety of sensors. However, the continuous operation of a variety of sensors consumes a relatively large amount of energy; if the intermittent detection method is adopted, although it can reduce the energy consumption to a certain extent, there is a large interference with the detection accuracy of the abnormal operating status of the transformer. Therefore, how to control the sensors to achieve the intelligent detection of transformers is an urgent problem to be solved in the current industry. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent detection method and system for a transformer to solve the problems raised in the above background technique.

[0006] To solve the above technical problems, the present invention provides the following technical solution: An intelligent detection method for a transformer, the method includes the following steps: S100. During the detection period, each detection unit is used to detect the operating status of the transformer respectively, and the detection data corresponding to each detection unit at different time points in the corresponding detection period are obtained, and the original detection data is generated; S200. Obtain the number of times of abnormal operating status of the transformer in the historical data, the corresponding abnormal time interval each time the operating status of the transformer is abnormal, and the detection data in the original detection data corresponding to each detection unit each time the operating status of the transformer is abnormal, and construct the abnormal detection analysis data; S300. According to the correlation characteristics between the detection data corresponding to each detection unit in the constructed abnormal detection analysis data, obtain the correlation detection chain between the detection units under the abnormal state of the transformer; S400. Dynamically adjust the operating states of each detection unit within the detection period according to the association detection chain between the detection units under abnormal transformer conditions.

[0007] Further, the detection period is the preset duration in the database; When detecting the operating state of the transformer, the operating state of the transformer is detected by multiple detection units. Each detection unit corresponds to a detection parameter, and the detection parameter is acquired through sensors set on the transformer. Different detection units correspond to different detection parameters; The detection unit is used to detect the operating state of the transformer, and the operating state of the transformer includes an abnormal operating state and a normal operating state.

[0008] Further, when generating the original detection data in S100, each detection period corresponds to an original detection data set. Each original detection data set includes original detection data groups respectively corresponding to multiple detection units, and each detection unit corresponds to an original detection data group; The original detection data group includes the detection data corresponding to each time point of the corresponding detection unit within the corresponding detection period. When the operating state of the corresponding detection unit within the detection period is in the sleep state, it is determined that the detection data corresponding to the corresponding time point is empty; the operating state of the detection unit includes the sleep state and the working state.

[0009] Further, the method for constructing the abnormal detection analysis data in S200 includes the following steps: S201. Obtain the number of times the transformer operating state is abnormal in the historical data, the corresponding abnormal time interval each time the transformer operating state is abnormal, and the detection data in the original detection data corresponding to each detection unit each time the transformer operation is abnormal; Denote the abnormal time interval corresponding to the nth abnormal operating state of the transformer in the historical data as Qn, and denote the set composed of the detection data in the original detection data corresponding to the ith detection unit when the transformer operating state is abnormal for the nth time in the historical data as H （n，i） ; Any element in the H （n，i） corresponds to a time belonging to Qn; S202. Extract the elements in H （n，i） that do not belong to the normal value range of the detection data corresponding to the ith detection unit in the database. The normal value range of the detection data corresponding to the detection unit is preset in the database; Denote the set composed of the extracted elements from H （n，i） as YH （n，i） ; Denote the minimum value of the time corresponding to the elements extracted from H （n，i） as TminH （n，i） Denote the...（n，i） The maximum value of the time corresponding to the elements extracted from （n，i） is denoted as TmaxH （n，i） All the elements in （n，i） whose corresponding time belongs to [TminH （n，i） , TmaxH （n，i） are arranged in ascending order of the corresponding time to obtain the anomaly detection segment in S203. Obtain all the detection data within the anomaly detection segment corresponding to （n，i） . For any two elements that both belong to YH （n，i） , the maximum value of the number of all detection data between them and there are no elements in YH （n，i） between them is taken as the screening and determination threshold for the anomaly detection data segment corresponding to the i-th detection unit when the transformer is in the n-th abnormal operation state in the historical data; S204. Obtain the set of all detection data within the unit time before the minimum time point in Qn from the original detection data corresponding to the i-th detection unit when the transformer is in the n-th abnormal operation state in the historical data, and denote it as DH （n，i） ; the unit time is a constant preset in the database; S205. Dynamically select a reference point in DH （n，i） until the final reference point corresponding to the i-th detection unit when the transformer is in the n-th abnormal operation state is obtained; When dynamically selecting a reference point, the minimum time point in Qn is used as the initial reference point. The reference point with the smallest corresponding time point among the obtained reference points is denoted as the reference point to be analyzed. The minimum time point of the detection data that does not belong to the normal value range of the detection data corresponding to the i-th detection unit in the database among all the detection data between the first time point and the reference point to be analyzed is used as a new reference point, and the operation of dynamically selecting a reference point is continued; the first time point is the time point before the reference point to be analyzed and the time interval from the reference point to be analyzed is equal to the screening and determination threshold of the anomaly detection data segment corresponding to the i-th detection unit when the transformer is in the n-th abnormal operation state in the historical data; If all the detection data between the first time point and the reference point to be analyzed belong to the normal value range of the detection data corresponding to the i-th detection unit in the database, the operation of dynamically selecting a reference point is not continued, and the reference point with the smallest corresponding time point among the obtained reference points is used as the final reference point corresponding to the i-th detection unit when the transformer is in the n-th abnormal operation state.

[0010] S206. In the original detection data corresponding to the i-th detection unit when the transformer is in the n-th abnormal operation state in the historical data, from the final reference point corresponding to the i-th detection unit when the transformer is in the n-th abnormal operation state to TmaxH （n，i）All the detection data between them are arranged in ascending order according to the corresponding time to obtain the abnormal detection data segment corresponding to the i-th detection unit when the transformer is in the n-th abnormal operation state in the historical data; The set composed of the abnormal detection data segments corresponding to each detection unit when the transformer is in the n-th abnormal operation state in the historical data is used as the abnormal detection analysis data corresponding to the n-th abnormal operation state of the transformer.

[0011] In the process of constructing the abnormal detection analysis data of the present invention, parameters such as the number of times the transformer is in an abnormal operation state in the historical data, the corresponding abnormal time interval each time the transformer is in an abnormal operation state, and the detection data in the original detection data corresponding to each detection unit each time the transformer operates abnormally are considered; the obtained screening and determination threshold in the present invention is dynamically changed and changes with the change of the corresponding abnormal detection segment, and the obtained screening and determination threshold will directly affect the subsequent screening results of the reference point and the final reference point, and further affect the finally obtained abnormal detection analysis data. （n，i） And it changes with the change of the corresponding abnormal detection segment, and the obtained screening and determination threshold will directly affect the subsequent screening results of the reference point and the final reference point, and further affect the finally obtained abnormal detection analysis data.

[0012] Further, the method for obtaining the associated detection chain between the detection units in the abnormal state of the transformer in S300 includes the following steps: S301. Obtain the abnormal detection analysis data corresponding to each time the transformer is in an abnormal operation state in the historical data; S302. Obtain the abnormal detection data segments corresponding to each detection unit at different times when the transformer is in an abnormal operation state; S303. Construct different associated analysis pairs, where the associated analysis pair includes a first analysis object and a second analysis object, the first analysis object is one or more detection units, and the second analysis object is one detection unit; S304. Combining the obtained results of S301 and S302, calculate the associated characteristic value between each associated analysis pair, and screen the associated analysis pairs with the associated characteristic value greater than the preset value to construct associated detection chain nodes, and form the associated detection chain between the detection units in the abnormal state of the transformer; the associated detection chain includes one or more associated detection chain nodes, and each associated detection chain node represents the corresponding relationship between the first analysis object and the second analysis object in the associated analysis pair; When calculating the associated characteristic value between each associated analysis pair, denote the associated characteristic value between the j-th associated analysis pair as Gj, , where, TC （n，j，m）It represents the difference between the minimum time point corresponding to the abnormal detection data segment of the second analysis object in the j-th correlation analysis pair and the minimum time point corresponding to the abnormal detection data segment of the m-th detection unit in the first analysis object in the j-th correlation analysis pair when the transformer is in the n-th abnormal operating state; If, when the transformer is in the n-th abnormal operating state, the abnormal detection data segments corresponding to the second analysis object in the j-th correlation analysis pair and the m-th detection unit in the first analysis object do not exist simultaneously, then it is determined that TC （n，j，m） = 0; PS （j，m） It represents the quotient of the number of values greater than 0 corresponding to each TC when n takes different values and the first abnormal statistical count Cjm; when Cjm = 0, then PS （n，j，m） = 0; （j，m） = 0; The first abnormal statistical count Cjm represents the number of times the transformer is in an abnormal operating state when the abnormal detection data segments corresponding to the second analysis object in the j-th correlation analysis pair and the m-th detection unit in the first analysis object both exist; n1 represents the number of times the transformer is in an abnormal operating state in the historical data; m1 represents the number of detection units in the first analysis object in the j-th correlation analysis pair; When constructing the correlation detection chain node, extract and screen out each correlation analysis pair whose correlation feature value is greater than the preset value and the corresponding second analysis object is the same, and use the corresponding relationship between the union of the sets corresponding to the first analysis object in each correlation analysis pair and the corresponding second analysis object as a correlation detection chain node.

[0013] When the present invention makes an abnormal judgment on the detection data obtained by the detection unit, if the detection data is empty, it is determined that the corresponding detection data belongs to the normal value range of the corresponding detection data; in the process of constructing the correlation detection chain node of the present invention, the relationship fusion between the obtained correlation analysis pairs is realized, and the relationship between the first analysis object and the second analysis object corresponding to one or more correlation analysis pairs is fused into a group of relationships (the relationship between the union of the sets corresponding to the first analysis object and the corresponding second analysis object), providing data support for the dynamic adjustment of the operating states of each detection unit in the subsequent steps.

[0014] Further, when the S400 dynamically adjusts the operating states of each detection unit within the detection period, When, within the previous unit time based on the current time, there is no abnormal situation in the detection data obtained by each detection unit, then the working states of each detection unit are intermittently controlled, and each detection unit is successively maintained to work for the first unit time and rest for the second unit time, and both the first unit time and the second unit time are constants preset in the database; When the detection data obtained by each detection unit in the previous unit time based on the current time has abnormal conditions, the associated detection chain between the detection units under the abnormal state of the transformer is obtained, and the set of detection units corresponding to the monitored abnormal detection data is compared with each associated detection chain node in the obtained associated detection chain. If the obtained set is a subset of the corresponding set of the first analysis object in the associated detection chain node, the elements in the obtained set and the detection units corresponding to the second analysis object in the corresponding associated detection chain node are controlled to exit the intermittent control state, while the corresponding detection units are kept in a continuous working state; otherwise, the detection units corresponding to the elements in the obtained set are controlled to exit the intermittent control state, while the corresponding detection units are kept in a continuous working state, and the intermittent control state is still maintained for the detection units corresponding to the second analysis object in the corresponding associated detection chain node.

[0015] An intelligent detection system for a transformer, the system comprising the following modules: A raw data acquisition module, wherein the raw data acquisition module detects the operating status of the transformer through each detection unit in a detection cycle, obtains detection data corresponding to each detection unit at different time points in the corresponding detection cycle, and generates raw detection data; An abnormal data extraction module, wherein the abnormal data extraction module obtains the number of abnormal transformer operating states in historical data, the abnormal time interval corresponding to each abnormal transformer operating state, and the detection data in the original detection data corresponding to each detection unit each time the transformer operating state is abnormal, and constructs abnormal detection analysis data; A detection unit abnormality detection association module, wherein the detection unit abnormality detection association module obtains an association detection chain between detection units under abnormal conditions of the transformer according to association features between detection data corresponding to each detection unit in the constructed abnormality detection analysis data; A detection status dynamic management module is provided, wherein the detection status dynamic management module dynamically adjusts the operating status of each detection unit within a detection cycle according to the associated detection chain between the detection units under the abnormal state of the transformer.

[0016] Furthermore, the detection unit abnormality detection association module includes an association feature analysis unit and an association detection chain construction unit. The correlation feature analysis unit analyzes the correlation features between the detection data corresponding to each detection unit in the constructed abnormal detection analysis data; The associated detection chain construction unit combines the results obtained by the associated feature analysis unit to obtain an associated detection chain between the detection units under the abnormal state of the transformer.

[0017] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: When the present invention detects the operating state of a transformer in real time through multiple sensors, an intermittent detection method is adopted to reduce the energy consumption of the sensors during the detection process; and the associated detection chain between detection units and the detection data of the corresponding detection units under abnormal states of the transformer obtained by combining the historical detection data of the transformer are used to dynamically adjust the operating state of the detection units, reduce the influence of the intermittent control method of the detection units on the detection accuracy of the abnormal operating state of the transformer, and achieve effective control over the detection of the transformer, having high practical value and economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention, and do not constitute a limitation to the present invention. In the drawings: Figure 1 is a schematic structural diagram of an intelligent detection system for a transformer according to the present invention; Figure 2 is a schematic flow diagram of an intelligent detection method for a transformer according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of 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.

[0020] Please refer to Figure 1 , the present invention provides a technical solution: an intelligent detection system for a transformer, the system includes the following modules: An original data acquisition module, which, within a detection period, detects the operating state of the transformer through each detection unit respectively, obtains the detection data corresponding to each detection unit at different time points within the corresponding detection period, and generates original detection data; An abnormal data extraction module, which obtains the number of times of abnormal operating states of the transformer in historical data, the corresponding abnormal time intervals for each abnormal operating state of the transformer, and the detection data within the original detection data corresponding to each detection unit for each abnormal operating state of the transformer, and constructs abnormal detection analysis data; A detection unit abnormal detection association module, which obtains the associated detection chain between detection units under the abnormal state of the transformer according to the association characteristics between the detection data corresponding to each detection unit in the constructed abnormal detection analysis data; The detection status dynamic management module dynamically adjusts the operating status of each detection unit within the detection period according to the associated detection chain among the detection units under the abnormal state of the transformer.

[0021] The detection unit abnormal detection association module includes an association feature analysis unit and an association detection chain construction unit. The association feature analysis unit is based on the association features between the detection data corresponding to each detection unit in the constructed abnormal detection analysis data. The association detection chain construction unit combines the results obtained by the association feature analysis unit to obtain the association detection chain among the detection units under the abnormal state of the transformer.

[0022] As Figure 2 shown, an intelligent detection method for a transformer, the method includes the following steps: S100. Within the detection period, each detection unit detects the operating status of the transformer respectively, obtains the detection data corresponding to each detection unit at different time points in the corresponding detection period, and generates the original detection data. The detection period is the preset duration in the database. When detecting the operating status of the transformer, the operating status of the transformer is detected by multiple detection units. Each detection unit corresponds to a detection parameter, and the detection parameter is acquired through a sensor set on the transformer. Different detection units correspond to different detection parameters. The detection unit is used to detect the operating status of the transformer, and the operating status of the transformer includes an abnormal operating status and a normal operating status.

[0023] When generating the original detection data in S100, each detection period corresponds to an original detection data set. Each original detection data set includes the original detection data groups corresponding to multiple detection units respectively, and each detection unit corresponds to an original detection data group. The original detection data group includes the detection data corresponding to each time point of the corresponding detection unit in the corresponding detection period. When the operating status of the corresponding detection unit in the detection period is in the sleep state, it is determined that the detection data corresponding to the corresponding time point is empty; the operating status of the detection unit includes the sleep state and the working state.

[0024] S200. Obtain the number of times the operating status of the transformer is abnormal in the historical data, the corresponding abnormal time interval each time the operating status of the transformer is abnormal, and the detection data in the original detection data corresponding to each detection unit each time the operating status of the transformer is abnormal, and construct the abnormal detection analysis data. The method for constructing the abnormal detection analysis data in S200 includes the following steps: S201. Obtain the number of times of abnormal operation status of the transformer in the historical data, the corresponding abnormal time interval each time the transformer operation status is abnormal, and the detection data in the original detection data corresponding to each detection unit each time the transformer operation is abnormal; Record the abnormal time interval corresponding to the nth abnormal operation status of the transformer in the historical data as Qn, and record the set composed of the detection data in the original detection data corresponding to the ith detection unit when the transformer has the nth abnormal operation status in the historical data as H （n，i） ; Any element in the said H （n，i） has a corresponding time belonging to Qn; S202. Extract the elements in H （n，i） that do not belong to the normal value range of the detection data corresponding to the ith detection unit in the database, and the normal value range of the detection data corresponding to the detection unit is preset in the database; Record the set composed of the elements extracted from H （n，i） as YH （n，i） ; Record the minimum value of the time corresponding to the elements extracted from H （n，i） as TminH （n，i） , record the maximum value of the time corresponding to the elements extracted from H （n，i） as TmaxH （n，i） , and arrange all the elements in H （n，i） whose corresponding time belongs to [TminH （n，i） , TmaxH （n，i） in ascending order of the corresponding time to obtain the abnormal detection segment in H （n，i） ; S203. Obtain, within the abnormal detection segment corresponding to H （n，i） , the maximum value of the number of all detection data between any two elements that both belong to YH （n，i） and there are no elements in YH （n，i） between the corresponding two elements, and take the obtained maximum value as the screening and determination threshold of the abnormal detection data segment corresponding to the ith detection unit when the transformer has the nth abnormal operation status in the historical data; S204. Obtain the set composed of all detection data within the unit time before the minimum time point in Qn in the original detection data corresponding to the ith detection unit when the transformer has the nth abnormal operation status in the historical data, and record it as DH （n，i） ; The unit time is a constant preset in the database; S205. Dynamically select the reference points in DH （n，i） until the final reference point corresponding to the ith detection unit when the transformer has the nth abnormal operation status is obtained; When dynamically selecting a reference point, the minimum time point in Qn is used as the initial reference point. The reference point with the minimum corresponding time point among the obtained reference points is denoted as the reference point to be analyzed. The minimum time point of the detection data that does not belong to the normal value range of the detection data corresponding to the i-th detection unit in the database among all the detection data between the first time point and the reference point to be analyzed is used as a new reference point, and the operation of dynamically selecting a reference point is continued; the first time point is the time point before the reference point to be analyzed and the interval duration from the reference point to be analyzed is equal to the screening and determination threshold of the abnormal detection data segment corresponding to the i-th detection unit during the n-th abnormal operation state of the transformer in the historical data. If all the detection data between the first time point and the reference point to be analyzed belong to the normal value range of the detection data corresponding to the i-th detection unit in the database, the operation of dynamically selecting a reference point is not continued, and the reference point with the minimum corresponding time point among the obtained reference points is used as the final reference point corresponding to the i-th detection unit during the n-th abnormal operation state of the transformer.

[0025] In this embodiment, if the minimum time point in Qn is Tr1, the screening and determination threshold corresponding to Qn is RY, and if the unit time is TDW; Then, among the detection data corresponding to the time within [Tr1 - TDW, Tr1], all the abnormal detection data that do not belong to the normal value range of the corresponding detection data and the corresponding detection time belongs to [Tr1 - RY, Tr1] are obtained. If there are two pieces of abnormal detection data that meet the above conditions, the detection times corresponding to these two pieces of abnormal detection data are Tr2 and Tr3 respectively, and the time point corresponding to Tr2 is smaller than the time point corresponding to Tr3, then Tr2 is used as the new reference point. If the corresponding time of a piece of abnormal detection data that does not belong to the normal value range of the corresponding detection data belongs to [Tr1 - TDW, Tr1] and does not belong to [Tr1 - RY, Tr1], or the corresponding time of a piece of abnormal detection data that does not belong to the normal value range of the corresponding detection data belongs to [Tr1 - TDW, Tr1] and belongs to [Tr1 - RY, Tr1], then this piece of abnormal detection data cannot be used as the new reference point in this determination stage.

[0026] S206. Among the original detection data corresponding to the i-th detection unit during the n-th abnormal operation state of the transformer in the historical data, all the detection data between the final reference point corresponding to the i-th detection unit during the n-th abnormal operation state of the transformer and TmaxH （n，i） are arranged in ascending order according to the corresponding time to obtain the abnormal detection data segment corresponding to the i-th detection unit during the n-th abnormal operation state of the transformer in the historical data. Take the set composed of the abnormal detection data segments corresponding to each detection unit when the transformer is in the nth abnormal operation state in the historical data as the abnormal detection analysis data corresponding to the nth abnormal operation state of the transformer.

[0027] S300. Obtain the associated detection chain between the detection units under the abnormal state of the transformer according to the association characteristics between the detection data corresponding to each detection unit in the constructed abnormal detection analysis data. The method for obtaining the associated detection chain between the detection units under the abnormal state of the transformer in S300 includes the following steps: S301. Obtain the abnormal detection analysis data corresponding to each abnormal operation state of the transformer in the historical data. S302. Obtain the abnormal detection data segments corresponding to each detection unit at different times when the transformer is in an abnormal operation state. S303. Construct different association analysis pairs, where the association analysis pair includes a first analysis object and a second analysis object. The first analysis object is one or more detection units, and the second analysis object is a detection unit. S304. Combine the acquisition results of S301 and S302, calculate the association characteristic values between each association analysis pair, and screen the association analysis pairs with association characteristic values greater than the preset value to construct the associated detection chain nodes, and form the associated detection chain between the detection units under the abnormal state of the transformer. The associated detection chain includes one or more associated detection chain nodes, and each associated detection chain node represents the corresponding relationship between the first analysis object and the second analysis object in the association analysis pair. When calculating the association characteristic values between each association analysis pair, denote the association characteristic value between the jth association analysis pair as Gj. , where TC （n，j，m） represents the difference between the minimum time point corresponding to the abnormal detection data segment of the second analysis object in the jth association analysis pair and the minimum time point corresponding to the abnormal detection data segment of the mth detection unit in the first analysis object of the jth association analysis pair when the transformer is in the nth abnormal operation state. If the abnormal detection data segments corresponding to the second analysis object and the mth detection unit in the first analysis object in the jth association analysis pair do not exist simultaneously when the transformer is in the nth abnormal operation state, then it is determined that TC （n，j，m） = 0; PS （j，m） represents the quotient of the number of values of each TC （n，j，m） corresponding to different values of n greater than 0 and the first abnormal statistical number Cjm. When Cjm = 0, then PS （j，m） = 0; The first abnormal statistical number Cjm represents the number of abnormal transformer operating status when both the abnormal detection data segments corresponding to the second analysis object in the jth association analysis pair and the mth detection unit in the first analysis object exist; n1 represents the number of abnormal transformer operating conditions in historical data; m1 represents the number of detection units in the first analysis object in the jth association analysis pair; When constructing an association detection chain node, extract and filter each association analysis pair whose association characteristic value is greater than a preset value and whose corresponding second analysis object is the same, and use the correspondence between the union of the corresponding sets of the first analysis objects in each association analysis pair and the corresponding second analysis object as an association detection chain node.

[0028] S400 , dynamically adjusting the operating state of each detection unit within a detection cycle according to the associated detection chain between the detection units in the abnormal state of the transformer.

[0029] When the S400 dynamically adjusts the operating status of each detection unit within the detection cycle, When there is no abnormality in the detection data obtained by each detection unit in the previous unit time based on the current time, the working state of each detection unit is intermittently controlled, and each detection unit is kept working for a first unit time and resting for a second unit time in turn, and the first unit time and the second unit time are both constants preset in the database; When the detection data obtained by each detection unit in the previous unit time based on the current time has abnormal conditions, the associated detection chain between the detection units under the abnormal state of the transformer is obtained, and the set of detection units corresponding to the monitored abnormal detection data is compared with each associated detection chain node in the obtained associated detection chain. If the obtained set is a subset of the corresponding set of the first analysis object in the associated detection chain node, the elements in the obtained set and the detection units corresponding to the second analysis object in the corresponding associated detection chain node are controlled to exit the intermittent control state, while the corresponding detection units are kept in a continuous working state; otherwise, the detection units corresponding to the elements in the obtained set are controlled to exit the intermittent control state, while the corresponding detection units are kept in a continuous working state, and the intermittent control state is still maintained for the detection units corresponding to the second analysis object in the corresponding associated detection chain node.

[0030] It should be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device.

[0031] Finally, it should be noted that the above are only preferred embodiments of the present invention and are not used to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A transformer intelligent detection method, characterized in that: The method comprises the following steps: S100, in a detection cycle, the operating state of the transformer is detected by each detection unit respectively, and detection data corresponding to each detection unit at different time points in the corresponding detection cycle are obtained to generate original detection data; S200, obtaining the number of abnormal transformer operating states in historical data, the abnormal time interval corresponding to each abnormal transformer operating state, and the detection data in the original detection data corresponding to each detection unit each time the transformer operating state is abnormal, and constructing abnormal detection analysis data; S300, obtaining an associated detection chain between detection units under an abnormal state of the transformer according to the associated features between detection data corresponding to each detection unit in the constructed abnormal detection analysis data; S400 , dynamically adjusting the operating state of each detection unit within a detection cycle according to the associated detection chain between the detection units in the abnormal state of the transformer.

2. The intelligent detection method for a transformer according to claim 1, characterized in that: The detection period is the duration preset in the database; When detecting the operating state of the transformer, the operating state of the transformer is detected by multiple detection units, each detection unit corresponds to a detection parameter, and the detection parameter is acquired by a sensor set on the transformer, and different detection units correspond to different detection parameters; The detection unit is used to detect the operating state of the transformer, and the operating state of the transformer includes an abnormal operating state and a normal operating state.

3. The intelligent detection method for a transformer according to claim 1, characterized in that: When the original detection data is generated in S100, each detection cycle corresponds to an original detection data set, each original detection data set includes original detection data groups corresponding to multiple detection units, and each detection unit corresponds to an original detection data group; The original detection data group includes the detection data corresponding to the corresponding detection unit at each time point within the corresponding detection cycle. When the operating state of the corresponding detection unit within the detection cycle is a dormant state, the detection data corresponding to the corresponding time point is determined to be empty; the operating state of the detection unit includes a dormant state and a working state.

4. The intelligent detection method for a transformer according to claim 1, characterized in that: The method for constructing abnormality detection analysis data in S200 includes the following steps: S201, obtaining the number of times the transformer operating state is abnormal in the historical data, the abnormal time interval corresponding to each time the transformer operating state is abnormal, and the detection data in the original detection data corresponding to each detection unit each time the transformer operates abnormally; The abnormal time interval corresponding to the abnormal operation state of the transformer for the nth time in the historical data is recorded as Qn, and the set of detection data in the original detection data corresponding to the i-th detection unit when the transformer is abnormal in the nth operation state in the historical data is recorded as H （n，i） ; The H （n，i） The time corresponding to any element in belongs to Qn; S202, extract H （n，i） The elements in the database that do not belong to the normal value interval of the detection data corresponding to the i-th detection unit, where the normal value interval of the detection data corresponding to the detection unit is preset in the database; Will start from H （n，i） The set of elements extracted from is denoted as YH （n，i） ; will be from H （n，i） The minimum time corresponding to the elements extracted is recorded as TminH （n，i） , will be from H （n，i） The maximum value of the time corresponding to the element extracted is recorded as TmaxH （n，i） , H （n，i） The corresponding time belongs to [TminH （n，i） , TmaxH （n，i） ] are arranged in ascending order according to the corresponding time, and H （n，i） Anomaly detection snippet in ; S203, obtain H （n，i） The corresponding anomaly detection segments all belong to YH （n，i） The maximum value of all the detection data between any two elements of the corresponding two elements does not contain YH （n，i） The internal element is used as the maximum value obtained as the screening judgment threshold of the abnormal detection data segment corresponding to the i-th detection unit when the transformer is abnormal in the n-th operation state in the historical data; S204 obtains the set of all detection data in the original detection data corresponding to the ith detection unit when the transformer is abnormal in the nth operation state in the historical data, and before the minimum time point in Qn, which is recorded as DH （n，i） ; The unit time is a constant preset in the database; S205, Dynamic Selection of DH （n，i） until the final reference point corresponding to the i-th detection unit when the transformer is in abnormal operation state for the nth time is obtained; When dynamically selecting reference points, the minimum time point in Qn is used as the initial reference point, the reference point with the minimum corresponding time point among the obtained reference points is recorded as the reference point to be analyzed, and the minimum time point corresponding to the detection data that does not belong to the normal value interval of the detection data corresponding to the i-th detection unit in the database among all the detection data between the first time point and the reference point to be analyzed is used as a new reference point, and the operation of dynamically selecting reference points is continued; the first time point is the time point before the reference point to be analyzed and the interval time with the reference point to be analyzed is equal to the screening judgment threshold of the abnormal detection data segment corresponding to the i-th detection unit when the transformer is abnormal for the nth time in the operating state in the historical data; If all the detection data between the first time point and the reference point to be analyzed belong to the normal value interval of the detection data corresponding to the i-th detection unit in the database, the operation of dynamically selecting the reference point will not be continued, and the reference point with the smallest corresponding time point among the obtained reference points will be used as the final reference point corresponding to the i-th detection unit when the transformer is in an abnormal operating state for the nth time; S206, the final reference point corresponding to the i-th detection unit when the transformer is in an abnormal operating state for the nth time in the historical data corresponds to the original detection data, and the i-th detection unit when the transformer is in an abnormal operating state for the nth time corresponds to TmaxH （n，i） All the detection data between are arranged in ascending order according to the corresponding time, and the abnormal detection data fragment corresponding to the i-th detection unit when the transformer is abnormal in the n-th operation state in the historical data is obtained; A set of abnormal detection data fragments corresponding to each detection unit when the transformer is in an abnormal operating state for the nth time in the historical data is used as the abnormal detection analysis data corresponding to the transformer in the abnormal operating state for the nth time.

5. The intelligent detection method of a transformer according to claim 4, characterized in that: The method for obtaining the associated detection chain between the detection units in the abnormal state of the transformer in S300 includes the following steps: S301, obtaining abnormal detection and analysis data corresponding to each abnormal operation state of the transformer in historical data; S302, obtaining abnormal detection data fragments corresponding to each detection unit at different times when the transformer operating state is abnormal; S303, constructing different association analysis pairs, wherein the association analysis pairs include a first analysis object and a second analysis object, wherein the first analysis object is one or more detection units, and the second analysis object is one detection unit; S304, combining the results obtained in S301 and S302, calculating the correlation characteristic value between each correlation analysis pair, and screening the correlation analysis pairs whose correlation characteristic value is greater than the preset value to construct the correlation detection chain node, and forming the correlation detection chain between the detection units under the abnormal state of the transformer; the correlation detection chain includes one or more correlation detection chain nodes, each of which represents the corresponding relationship between the first analysis object and the second analysis object in the correlation analysis pair; When calculating the correlation eigenvalue between each correlation analysis pair, the correlation eigenvalue between the jth correlation analysis pair is recorded as Gj. ， Among them, TC （n，j，m） represents the difference between the minimum time point corresponding to the abnormal detection data segment corresponding to the second analysis object in the jth association analysis pair and the minimum time point corresponding to the abnormal detection data segment corresponding to the mth detection unit in the first analysis object in the jth association analysis pair when the transformer is in an abnormal operating state for the nth time; If the transformer is in abnormal operation state for the nth time, the abnormal detection data fragments corresponding to the second analysis object in the jth association analysis pair and the mth detection unit in the first analysis object do not exist at the same time, then TC is determined. （n，j，m） =0; PS （j，m） When n is different, each TC （n，j，m） The quotient of the number of values ​​greater than 0 and the first abnormal statistical number Cjm; when Cjm=0, PS （j，m） =0; The first abnormal statistical number Cjm represents the number of abnormal transformer operating status when both the abnormal detection data segments corresponding to the second analysis object in the jth association analysis pair and the mth detection unit in the first analysis object exist; n1 represents the number of abnormal transformer operating conditions in historical data; m1 represents the number of detection units in the first analysis object in the jth association analysis pair; When constructing an association detection chain node, extract and filter each association analysis pair whose association characteristic value is greater than a preset value and whose corresponding second analysis object is the same, and use the correspondence between the union of the corresponding sets of the first analysis objects in each association analysis pair and the corresponding second analysis object as an association detection chain node.

6. The intelligent detection method of a transformer according to claim 5, characterized in that: When the S400 dynamically adjusts the operating status of each detection unit within the detection cycle, When there is no abnormality in the detection data obtained by each detection unit in the previous unit time based on the current time, the working state of each detection unit is intermittently controlled, and each detection unit is kept working for a first unit time and resting for a second unit time in turn, and the first unit time and the second unit time are both constants preset in the database; When the detection data obtained by each detection unit in the previous unit time based on the current time has abnormal conditions, the associated detection chain between the detection units under the abnormal state of the transformer is obtained, and the set of detection units corresponding to the monitored abnormal detection data is compared with each associated detection chain node in the obtained associated detection chain. If the obtained set is a subset of the corresponding set of the first analysis object in the associated detection chain node, the elements in the obtained set and the detection units corresponding to the second analysis object in the corresponding associated detection chain node are controlled to exit the intermittent control state, while the corresponding detection units are kept in a continuous working state; otherwise, the detection units corresponding to the elements in the obtained set are controlled to exit the intermittent control state, while the corresponding detection units are kept in a continuous working state, and the intermittent control state is still maintained for the detection units corresponding to the second analysis object in the corresponding associated detection chain node.

7. An intelligent detection system for a transformer, characterized in that: The system includes the following modules: A raw data acquisition module, wherein the raw data acquisition module detects the operating status of the transformer through each detection unit in a detection cycle, obtains detection data corresponding to each detection unit at different time points in the corresponding detection cycle, and generates raw detection data; An abnormal data extraction module, wherein the abnormal data extraction module obtains the number of abnormal transformer operating states in historical data, the abnormal time interval corresponding to each abnormal transformer operating state, and the detection data in the original detection data corresponding to each detection unit each time the transformer operating state is abnormal, and constructs abnormal detection analysis data; A detection unit abnormality detection association module, wherein the detection unit abnormality detection association module obtains an association detection chain between detection units under abnormal conditions of the transformer according to association features between detection data corresponding to each detection unit in the constructed abnormality detection analysis data; A detection status dynamic management module is provided, wherein the detection status dynamic management module dynamically adjusts the operating status of each detection unit within a detection cycle according to the associated detection chain between the detection units under the abnormal state of the transformer.

8. The intelligent detection system for transformer according to claim 7, characterized in that: The detection unit abnormality detection association module includes an association feature analysis unit and an association detection chain construction unit. The correlation feature analysis unit analyzes the correlation features between the detection data corresponding to each detection unit in the constructed abnormal detection analysis data; The associated detection chain construction unit combines the results obtained by the associated feature analysis unit to obtain an associated detection chain between the detection units under the abnormal state of the transformer.

Images