Power equipment monitoring method and system based on machine learning

Abstract

The invention discloses a power equipment monitoring method and system based on machine learning, and relates to the technical field of machine learning, and the method comprises the steps: obtaining the historical operation data of power equipment, and building a historical operation data set of each power equipment; acquiring a time point of historical operation data of the power equipment, and acquiring environment data corresponding to the time point; based on the operation environment data and the historical operation data, obtaining a fault cause sequence of the power equipment; obtaining monitoring causes of the power equipment based on the fault cause sequence of the power equipment; acquiring historical data of the monitoring incentives of the power equipment in a time period before the fault occurs, acquiring actually measured values of the monitoring incentives of the power equipment in a time period with the same length, and obtaining a monitoring result based on the historical data and the actually measured values; and based on the power connection relationship between the power devices, obtaining a fault cause sequence of the fault-associated power devices, and obtaining a monitoring result of the fault-associated power devices. According to the invention, the monitoring precision and range of the power equipment are realized.

Description

Technical Field

[0001] This application belongs to the technical field of machine learning. Specifically, it belongs to a method and system for monitoring power equipment based on machine learning. Background Art

[0002] Regarding the detection of the operating status of power equipment, various methods have been developed currently. However, from the common monitoring process, discrete numerical values during the operation of power equipment are obtained, and then the obtained data is compared to determine whether there is a fault in the power equipment. It can be determined that this method is mainly a post-analysis method for faults, and obviously it is not a method that can detect faults in advance. Moreover, due to directly using discrete data for one-by-one comparison of data, compared with the continuous data comparison method, the monitoring accuracy is significantly too low. In addition, the current method is a method for simultaneously detecting and analyzing all generated operation data. This method requires obtaining a large amount of operation data and processing it, resulting in a large operating burden on the monitoring system. In addition, in the current monitoring, usually only single monitoring of power equipment is carried out, and the fault transmission caused by the connection relationship between power equipment is not considered. As a result, in fault monitoring, it is impossible to carry out advance monitoring of the transmission results of power equipment faults, and it is impossible to determine in advance the monitoring plan for power equipment associated with faults, resulting in potential faults in the power system being difficult to detect in time.

[0003] Therefore, how to formulate in advance a monitoring plan for the operating status of power equipment, reduce the monitored data and ensure the monitoring accuracy, and realize timely monitoring of power equipment with fault correlation relationships is a technical problem that those skilled in the art urgently need to solve. Summary of the Invention

[0004] The technical objective of this application is to ensure that potential faults of power equipment in the power system can be detected in time by adopting fewer monitoring targets of power equipment, ensuring the monitoring accuracy, and simultaneously realizing timely monitoring of power equipment that may have fault correlation relationships. Specifically, it includes: First aspect: A method for monitoring power equipment based on machine learning, the monitoring method includes: Obtain the historical operation data of power equipment and establish a historical operation data group for each power equipment; Obtain the time points of the historical operation data of power equipment and obtain the environmental data corresponding to the time points; Based on the operating environment data and the historical operation data, obtain the sorting of fault incentives of power equipment; Based on the sorting of the fault causes of the power equipment, obtain the monitoring causes of the power equipment; Obtain the historical data of the monitoring causes of the power equipment in the time period before the fault occurs, and obtain the measured values of the monitoring causes of the power equipment in the same length time period. Based on the historical data and the measured values, obtain the monitoring results; Based on the power connection relationship between power equipment, obtain the sorting of the fault causes of the fault-associated power equipment, and obtain the monitoring results of the fault-associated power equipment.

[0005] Optionally, the obtaining of the historical operation data of the power equipment and establishing the historical operation data group for each power equipment includes: Obtain the unique identification code of each power equipment in the power system, and establish an equipment data group based on the unique identification code; Add the historical operation data of the power equipment into the equipment data group to obtain the historical operation data group of the power equipment; Further includes: Obtain the acquisition time node of the historical operation data of the power equipment to obtain the data acquisition time point; In the historical operation data group, establish the corresponding relationship between the data acquisition time point and the historical operation data of the power equipment.

[0006] Optionally, the obtaining of the time point of the historical operation data of the power equipment and obtaining the corresponding environmental data at the time point includes: Based on the unique identification code of the power equipment, obtain the corresponding environmental data of the power equipment; Based on the data acquisition time point, obtain the corresponding environmental data of the power equipment corresponding to the data acquisition time point, and establish a corresponding relationship; Supplement the corresponding environmental data of the power equipment corresponding to the data acquisition time point into the historical operation data group corresponding to the power equipment to obtain the historical data group of the power equipment.

[0007] Optionally, the obtaining of the sorting of the fault causes of the power equipment based on the operating environment data and the historical operation data includes: Based on the power data acquisition time point corresponding to the power equipment fault data, obtain the historical operation data of the power equipment in the previous time period to obtain the operation data before the fault; Based on the operation data before the fault, obtain the change rate of the historical operation data; Based on the change rate, obtain the regression coefficient corresponding to each change rate. The equation for determining the regression coefficient corresponding to the change rate is: ; Wherein, Represents the regression coefficient corresponding to the change rate, Represents the change rate, m Represents the category index of historical operation data, n Represents the total amount of categories of historical operation data; Sort the regression coefficients corresponding to the change rate to obtain the sorting of fault inducements of the power equipment; Further includes: Based on the sorting of fault inducements of the power equipment, obtain the types of fault inducements, and obtain the sorting of the power equipment's own fault inducements and the sorting of operation environment inducements.

[0008] Optionally, the obtaining of the monitoring inducement of the power equipment based on the sorting of the fault inducements of the power equipment includes: Based on the sorting of the power equipment's own fault inducements, compare the regression coefficients corresponding to the own fault inducements with the preset regression coefficients, and obtain the fault inducements not lower than the preset regression coefficients to obtain high-impact fault inducements; Based on the sorting of the environment inducements, compare the regression coefficients corresponding to the environment inducements with the preset environment regression coefficients, and obtain the regression coefficients of the environment inducements corresponding to not lower than the preset environment regression coefficients to obtain high-impact environment inducements; Based on the high-impact environment inducements, obtain the power equipment failure probability caused by the high-impact environment inducements. The power equipment failure probability determination equation is: ; Wherein, Represents the failure probability caused by the high-impact environment inducement; Represents the intercept term, which is the logit when all independent variables are 0; Represents the regression coefficient of the high-impact environment inducement; Represents the specific parameter of the high-impact environment inducement; i Represents the index of the high-impact environment inducement; Compare the power equipment failure probability caused by the corresponding environment data with the preset power equipment failure probability. If the power equipment failure probability caused by the corresponding environment is not lower than the preset power equipment failure probability, determine that the high-impact environment inducement is the cause of the power equipment failure; Determine the high-impact fault inducement, or, the high-impact fault inducement and the high-impact environment inducement as the monitoring inducement of the power equipment.

[0009] Optionally, the obtaining of the historical data of the monitoring inducement of the power equipment in the time period before the fault occurs, and obtaining the measured values of the monitoring inducement of the power equipment in the same length time period, and obtaining the monitoring result based on the historical data and the measured values includes: Based on the pre-fault operating data and the monitoring incentives of the power equipment, obtain the historical monitoring incentive values of the power equipment; Based on the monitoring incentives of the power equipment, obtain the measured values of the monitoring incentives of the power equipment; Compare the historical monitoring incentive values of the power equipment with the measured values of the monitoring incentives of the power equipment to obtain an incentive value comparison result; Based on the incentive value comparison result and the regression coefficient corresponding to the monitoring incentives of the power equipment, obtain the monitoring result of the power equipment.

[0010] Optionally, the obtaining of the fault incentive ranking of the fault-related power equipment and the monitoring result of the fault-related power equipment based on the power connection relationship between power equipment includes: Based on the power connection relationship between power equipment, obtain the historical operation data groups of the power equipment connected to the faulty power equipment to obtain an associated data group; Obtain the power equipment fault data of the associated data group and the occurrence time node of the power equipment fault data of the associated data group; If the deviation between the occurrence time node of the power equipment fault data of the associated data group and the power data acquisition time point corresponding to the power equipment fault data is not higher than the preset deviation, determine that there is a fault association relationship between the power equipment to obtain the fault-related power equipment; Obtain the monitoring incentives of the fault-related power equipment, and based on the measured data of the monitoring incentives of the fault-related power equipment and the historical monitoring incentive data of the fault-related power equipment, obtain the monitoring result of the fault-related power equipment.

[0011] Second aspect: A machine learning-based power equipment monitoring system for implementing a power equipment monitoring method, including: a data acquisition module, a data preprocessing module, an association analysis module, a data analysis module, a power equipment monitoring module, and a database; The data acquisition module is connected to the power equipment monitoring module, the data preprocessing module, the association analysis module, and the database, and is used to obtain the measured values of the monitoring incentives of the power equipment, the historical operation data of the power equipment, and send the measured values of the monitoring incentives of the power equipment to the data preprocessing module, the association analysis module, and the database; The association analysis module is further connected to the data analysis module, and is used to analyze the power connection relationship between power equipment and obtain the association relationship between the associated data group, the high-impact fault incentives, and the high-impact environmental incentives; The data analysis module obtains the monitoring result of the power equipment based on the monitoring incentives of the power equipment.

[0012] Optionally, it further includes: The data preprocessing module is also connected to the database, and is used for preprocessing the measured data of the monitoring incentives of the power equipment obtained by the power equipment monitoring module, and then sending it to the database for storage; The data preprocessing module is also connected to the correlation analysis module, and is used for preprocessing the measured data of the monitoring incentives of the power equipment obtained by the power equipment monitoring module, and then sending it to the correlation analysis module for analyzing the correlation between high-impact fault incentives and high-impact environmental incentives.

[0013] Optionally, it further includes: The data acquisition module also obtains the power connection relationship of the power equipment from the database to obtain the fault correlation relationship between power equipment.

[0014] The beneficial effects of this application include: 1. It realizes the reduction of the number of monitoring objects. In the technical solution of this application, for the operation faults of power equipment, both the faults of the power equipment itself and the faults caused by environmental factors are analyzed, and then the correlation between these two types of factors and the occurrence of power equipment faults is determined. The fault influencing factors with higher influence degree are obtained from this, and then these influencing factors are used as the monitoring objects in the monitoring process of power equipment. In this way, it is not necessary to obtain and process all the operation data, and only a part of the monitoring data needs to be obtained to achieve high-precision monitoring of the operation state of power equipment.

[0015] 2. It realizes the determination of the fault correlation relationship of power equipment. In the technical solution of this application, the connection relationship between power equipment is analyzed, especially the fault manifestations of power equipment with connection relationships are analyzed, and then the fault incentives with such connection relationships are determined and also used in the monitoring, so as to realize the fault correlation analysis of power equipment, and thus ensure that based on the connection relationship of power equipment, all possible faults in the current power system can be monitored.

[0016] 3. It realizes the early determination of the monitoring objects for the operation faults of power equipment. In the technical solution of this application, based on the historical operation data of power equipment, the fault monitoring objects are analyzed in advance. Then, in the actual monitoring, based on the early determination of such fault monitoring objects, the parameters of these determined monitoring objects can be obtained during the actual monitoring. By the early determination of the operation fault monitoring objects, the fault determination efficiency in the monitoring operation of power equipment is improved. In addition, in the specific monitoring, instead of using discrete data for processing, the data within a time period is used for monitoring, which improves the monitoring accuracy. Description of the Drawings

[0017] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the accompanying drawings required in the embodiments of the present application or the prior art. Obviously, the following descriptions are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. The drawings are used to provide a further understanding of the present disclosure and constitute a part of the specification, and are used together with the following specific embodiments to explain the present disclosure, but do not constitute a limitation to the present disclosure. In the drawings: Figure 1 It is a flowchart of a power equipment monitoring method based on machine learning provided by an embodiment of the present application; Figure 2 It is a schematic diagram of a power equipment monitoring system based on machine learning provided by an embodiment of the present application. Specific Embodiments

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, rather than all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application. In addition, in the embodiments of the present application, "first", "second", etc. are used to distinguish similar objects and do not necessarily need to describe a specific order or sequence.

[0019] In the monitoring of power equipment, the currently adopted method is mainly based on setting preset values for corresponding monitoring parameters of the power equipment, and then analyzing whether there is a phenomenon exceeding the preset value during the actual operation of the power equipment. If so, it is considered that there is a high probability of a fault occurring. This method essentially belongs to a post-monitoring analysis method, that is: it is necessary to determine that the current power equipment has a fault when the actual operation data corresponds to the fault manifestation. This method no longer meets the current requirements for early identification and monitoring of faults in the operation of power equipment. In addition, the connection relationship between power equipment is not considered, especially the problem of fault transmission caused by the connection between power equipment, which has not been processed, resulting in only being able to monitor the operation status of each single power equipment, rather than obtaining all the operation performances that may be caused by the connection relationship of power equipment, which makes it difficult to timely discover some potential defects or faults in power equipment.

[0020] In the technical solution of this application, for the monitoring of power equipment, the fault causes are determined based on historical operation data, and during this process, the faults caused by the power equipment itself and the fault factors caused by the environment are determined simultaneously. Then, the monitoring causes are determined, the monitoring parameters are obtained based on the types of monitoring causes, and whether the power equipment will fail is determined in advance based on the performance of the monitoring parameters, so as to realize the early monitoring of faults. And based on the connection relationship of the power equipment, the monitoring scheme for other power equipment is determined, and the corresponding monitoring parameters are obtained to determine the current operation state of the power equipment.

[0021] In the technical solution of this application, on the one hand, this application discloses a power equipment monitoring method based on machine learning, as Figure 1 shown, which is a flowchart of a power equipment monitoring method based on machine learning provided by an embodiment of this application. Specifically: S110. Obtain the historical operation data of the power equipment and establish a historical operation data group for each power equipment; S120. Obtain the time points of the historical operation data of the power equipment and obtain the environmental data corresponding to the time points; S130. Based on the operation environment data and the historical operation data, obtain the fault cause ranking of the power equipment; S140. Based on the fault cause ranking of the power equipment, obtain the monitoring cause of the power equipment; S150. Obtain the historical data of the monitoring cause of the power equipment in the time period before the fault occurs, and obtain the measured values of the monitoring cause of the power equipment in the same length of time period. Based on the historical data and the measured values, obtain the monitoring result; S160. Based on the power connection relationship between power equipment, obtain the fault cause ranking of the fault-related power equipment and obtain the monitoring result of the fault-related power equipment.

[0022] The purpose of the above steps is that in the monitoring of power equipment, the key fault performance data can be monitored, and at the same time, based on the connection relationship between power equipment, the monitoring causes of the fault-related power equipment are determined. Thus, in the simultaneous monitoring of power equipment and fault-related power equipment, the operation state of the power equipment and the possible fault performance can be predicted based on the monitoring results to achieve pre-judgment. And because the connection relationship of the power equipment is considered, it can ensure the monitoring and identification of potential faults that may exist in the power system.

[0023] Next, the above steps will be specifically described. Specifically: As described in step S110, the purpose of this step is to establish a historical operation data group for each power device in the power system, so as to obtain the historical operation data of each power device during subsequent operation monitoring, laying a foundation for subsequent monitoring incentive acquisition work. Specifically: S111. Obtain the unique identification code of each power device in the power system, and establish a device data group based on the unique identification code.

[0024] The purpose of this step is that for the established historical operation data group, it is obvious that it is necessary to know the power device corresponding to the data in the data group. Considering that in the power system settings, a unique identification code will be set for each power device to determine the information of the power device, so by using this existing resource, the identification of the historical operation data of the power device can be realized.

[0025] Among them, for all set power devices, obtain the unique identification code corresponding to the power device from the already stored information library; In some embodiments, when the unique identification code of the power device has not been entered into the information library, it is newly created based on the use, type, and RFID chip information of the power device.

[0026] Among them, after determining the unique identification code, establish a blank data group, and associate the data group with the corresponding unique identification code, then the blank data group is the device data group.

[0027] In some embodiments, the established device data group includes two parts. One is the unique identification code of the device, and the other is the data to be filled, such as the form {r23t05|< >,< >,< >,···,< >}, where "r23t05" is the unique identification code of the device, and the "< >" part is the data to be filled area.

[0028] S112. Add the historical operation data of the power device into the device data group to obtain the historical operation data group of the power device.

[0029] The purpose of this step is that the technical solution of this application is that the monitoring process of the power device needs to be based on the historical operation data of the power device to determine the fault manifestations therein, especially to obtain the data performance during a period of time before the fault occurs. Therefore, in the process, it is necessary to be able to establish a dedicated historical operation data group for each power device for subsequent analysis.

[0030] Among them, for the historical operation data of the power device that has been stored, directly obtain the historical operation data based on the unique identification code of the power device, and then fill the obtained historical operation data directly into the established device data group.

[0031] In some embodiments, all the operation data generated during the operation of the power equipment is preprocessed, and then such data is all applied as historical operation data and filled into the equipment data group.

[0032] In some embodiments, when it is found that there are data losses or data interferences in the historical operation data of the power equipment, the obtained data needs to be processed and then stored. For example, if it is found that there are data losses, the data change rates in a period of time before and after the lost data are analyzed, and then data is supplemented based on the data change rate, and the supplemented data is filled into the equipment data group.

[0033] S113. It further includes: Obtain the acquisition time node of the historical operation data of the power equipment to obtain the data acquisition time point; In the historical operation data group, establish the corresponding relationship between the data acquisition time point and the historical operation data of the power equipment.

[0034] The purpose of this step is that in the monitoring of power equipment in this application, in order to better determine the fault manifestation and monitoring accuracy, it is determined based on the monitoring values in the time series. At the same time, considering the relevance of the operation environment change and time node of the power equipment in terms of data storage, therefore, in order to facilitate the subsequent determination of the monitoring target and the acquisition of operation environment data, it is also necessary to obtain the time node corresponding to the historical operation data for subsequent parameter setting.

[0035] Among them, the historical monitoring incentive value of the power equipment refers to obtaining the operation data and operation environment data of the same type of power equipment that are the same as the monitoring incentive of the power equipment from the historical operation data group of the power equipment.

[0036] Among them, for all historical operation data, determine the acquisition time of such data, directly determine the obtained acquisition time as the data acquisition time point, and mark it in the obtained historical operation data group. For example, the marked historical operation data group is expressed as: {r23t05|< >(t1),< >(t2),< >(t3),···,< >(t4)}, where the information in the small parentheses is the time node information.

[0037] In some embodiments, the data acquisition time point does not need to be marked in the historical operation data group, and the corresponding relationship between the time node and other parameters can be directly established.

[0038] As described in step S120, the purpose of this step is that for the operating faults of power equipment, in addition to being caused by its own characteristics, the environment in which the power equipment operates during the operation process may also cause the operating faults of the power equipment or increase the probability of the occurrence of faults. In order to ensure the effective monitoring of the operating state of the power equipment, in the analysis of the monitoring data of this application, both the characteristics of the power equipment itself and environmental factors are considered. Therefore, in the specific processing, it is necessary to integrate and analyze the specific associations of these two factors, and then obtain the analysis results. Specifically: S121. Based on the unique identification code of the power equipment, obtain the corresponding environmental data of the power equipment.

[0039] The purpose of this step is that in the current power system, monitoring devices for the operating environment of power equipment have been generally built, and the environmental data is also recorded, and the environmental data and the power equipment are associated. Therefore, in the specific processing, based on the unique identification code of the power, the corresponding environmental data of the power equipment can be directly determined from the database to simplify the process of obtaining environmental data.

[0040] Among them, the process of determining the unique identification code of the power equipment can be directly obtained based on the currently stored unique identification code and the corresponding environmental data.

[0041] Among them, for the obtained environmental data, the time information is also synchronously recorded, and these two types of data are obtained simultaneously.

[0042] In some embodiments, for the obtained data, preprocessing is also required according to the specific situation of the data. For example, if it is found that the environmental data corresponding to a certain time node is missing, then based on the data in a period of time before and after this time node, it is processed. For example: in a certain time node, it is found that there are absences of environmental temperature and humidity, and the change rates of environmental temperature and humidity in the previous period of time are basically 0, and the average values of environmental temperature and humidity are 23°C and 56% respectively, then these two values are supplemented to the missing time point.

[0043] S122. Based on the data acquisition time point, obtain the corresponding environmental data of the power equipment corresponding to the data acquisition time point and establish a corresponding relationship.

[0044] The purpose of this step is that in the analysis of the operating faults and other monitoring requirements caused by the operating environment and the self - operating state of the power equipment, in order to achieve the correlation degree between the environmental data in the whole system and the characteristics of the power equipment itself, it is obvious that strong correlation needs to be ensured between these two types of data. For the time node used in this process, it is obvious that it can be used for the correlation analysis process.

[0045] Among them, based on the historical operation data set of the power equipment, the acquisition time nodes of each historical operation data can be determined. Based on the determined data acquisition time points, the environmental data corresponding to such data acquisition time points can be directly obtained.

[0046] Among them, after determining the environmental data corresponding to the data acquisition time points, it is essentially considered that the logical association between these two types of data has been established, and then it can be directly used.

[0047] S123. Supplement the corresponding environmental data of the power equipment corresponding to the data acquisition time point into the historical operation data set corresponding to the power equipment to obtain the historical data set of the power equipment.

[0048] The purpose of this step is to associate these two types of data for the obtained historical operation data and environmental data of the power equipment. At the same time, considering the overall technical solution of this application, in addition to establishing the monitoring relationship for a single power equipment, it is also necessary to perform corresponding monitoring on other power equipment with a connection relationship based on the connection relationship between power equipment, so as to form an effective monitoring network. In order to clearly determine the co-determination of the equipment's own characteristics and operating environment data of other power equipment with an associated relationship, it is necessary to incorporate the obtained environmental data into the historical operation data set. Then, in the subsequent monitoring of power equipment with a connection relationship, the environmental data in this data set can be directly used.

[0049] Among them, for the corresponding environmental data of the power equipment, obtain the data acquisition time point of the corresponding environmental data. Then, based on this corresponding relationship, input the corresponding environmental data into the established historical operation data set to obtain the historical data set of the power equipment. Among them, the historical data set of the power equipment at least includes the historical operation data and operating environment data of the power equipment. For example: the established historical data set of the power equipment is {r23t05|< >(t1)[e1],< >(t2)[e2],<>(t3)[e3],···,< >(t n )[e n}, where the data in the square brackets is the corresponding environmental data.

[0050] In some embodiments, the historical data set of the power equipment established may not include time data, but in subsequent monitoring, based on the data included in the historical data set of the power equipment, the corresponding time nodes are called again to establish the corresponding relationship.

[0051] As described in step S130, the purpose of this step is that in the technical solution of the present application, a beneficial effect is to reduce the target data to be monitored. Based on this technical concept, further analysis is required for the operation environment data and historical operation data of power equipment. Since both of these types of parameters are the fault incentives of power equipment, the priorities of the incentives are sorted to determine the data monitoring intensity for different incentives. For the results obtained in this way, they need to be obtained and analyzed based on historical data. Specifically: S131. Based on the power data acquisition time point corresponding to the power equipment fault data, obtain the historical operation data of the power equipment in the previous time period to obtain the operation data before the fault.

[0052] The purpose of this step is that in the technical solution of the present application, the monitoring of power equipment is not based on a single data to obtain the operation status, but is based on a time period for comparison. Therefore, for historical operation data and operation environment data, a time series is set to obtain them, and then the monitoring data is compared to achieve higher monitoring accuracy. Therefore, in this step, by obtaining the operation data of the power equipment in the previous time period, it can lay a foundation for the subsequent comparison of the monitoring data.

[0053] Among them, for the established historical operation data group, among the historical operation data of all power equipment, the current power equipment fault data can be obtained based on the analysis of the data. Therefore, directly obtain the fault data in the historical operation data group, and obtain the occurrence time point of the fault data. Use this time point as the reference data to obtain the data therein.

[0054] Among them, after obtaining the power data acquisition time point corresponding to the relevant fault data, obtain the previous time period of this time point. In the specific setting of this previous time period, it can be directly obtained based on the specific performance of the data. For example: obtain the data based on a 2h, 4h, or 6h time period as the acquisition basis.

[0055] In some embodiments, after obtaining the power data acquisition time point corresponding to the power equipment fault data, the specific time length of the previous time period is separately set based on the characteristics and types of the power equipment itself to obtain specific detection results.

[0056] S132. Based on the operation data before the fault, obtain the change rate of the historical operation data.

[0057] The purpose of this step is that for historical operation data, there will inevitably be a mutation in operation parameters between normal operation data and fault manifestation data. This mutation will be manifested in the form of the data change rate. Similarly, for the same or similar power equipment operation faults, there will also be similar data manifestations. Therefore, in data analysis, it is necessary to determine the change rate of historical operation data for further analysis of the current fault manifestation.

[0058] Among them, after filtering out the historical operation data, the historical operation data corresponding to the fault time point is regarded as independent operation data, and the mean value of other operation data is obtained, and the deviation degree between this mean value and the fault data is calculated.

[0059] In some embodiments, for all the obtained historical operation data, calculate the change rate between such data and generate a change rate curve.

[0060] In some embodiments, set a data curve for the obtained historical operation data, and then obtain the change rate between adjacent historical operation data based on this curve.

[0061] S133. Based on the change rate, obtain the regression coefficient corresponding to each change rate. The determination equation of the regression coefficient corresponding to the change rate is: ; Among them, represents the regression coefficient corresponding to the change rate, represents the change rate, m represents the category index of historical operation data, n represents the total amount of categories of historical operation data.

[0062] The purpose of this step is that for the obtained historical operation data, it obviously includes the operation data and operation environment data of power equipment. Both of these data may affect the operation state of power equipment. In order to obtain a specific analysis of the impact of different types of data on power equipment, it is necessary to analyze the correlation degree of relevant parameters among them in order to obtain the impact of different historical operation data on the failure rate of power equipment.

[0063] Among them, since the change rate of historical data has been obtained, and different indicators have different impacts on the failure rate of power equipment, it is necessary to process the degree of impact based on this parameter. Specifically, it is obtained by establishing a regression equation. This regression equation can be expressed as: ; Among them, represents the regression coefficient corresponding to the change rate, represents the change rate, mIndicates the category index of historical operation data, n Indicates the total amount of categories of historical operation data. The 1 on the left side of the equation indicates that under the current operating conditions of the power equipment, fault manifestations have occurred.

[0064] Among them, the reason for the establishment of this equation within the technical solution disclosed in this application is that based on the analysis of the causes and manifestations of faults occurring in the power system, the current result is that during the operation of power equipment, if a fault occurs or is about to occur, for the main causes of this fault, greater changes will occur within this time period. For example, when the transformer load is too large and causes a fuse to blow, it is found that the load will increase significantly within a short period of time until it exceeds the load capacity of the transformer. Therefore, in this application, based on this phenomenon, a regression equation is determined, and then each influencing parameter therein is determined based on this regression equation.

[0065] Among them, for the determination of the regression coefficients in this equation, it can be filled based on the already obtained historical operation data, and then overall calculations can be performed on different historical operation data based on the values in this equation.

[0066] Among them, for all the obtained regression coefficients, the larger the regression coefficient, the greater the degree of the fault caused by the historical operation data corresponding to this regression coefficient to the power equipment.

[0067] S134. Sort the regression coefficients corresponding to the change rate to obtain the sorting of the fault inducements of the power equipment.

[0068] The purpose of this step is that after sorting based on the regression coefficients corresponding to the change rate, a corresponding sequence of the impacts on the power equipment faults can be obtained. Of course, considering the regression coefficients of the historical operation data, which include environmental data and the own parameters of the power equipment, and when the power equipment fails, it is not necessarily caused by the joint action of these two types of factors. It may be that the fault is caused only by the own factors of the power equipment, or simply caused by the operating environment. Therefore, in the analysis of fault inducements, these two types of data need to be obtained separately in order to determine the correlation degree of these two types of data in the subsequent analysis.

[0069] Among them, the self-fault inducements described here refer to the factors that can cause faults based on the own characteristics of the power equipment under the corresponding conditions of the own attributes, types, and models of the power equipment, such as the load parameters of a certain type of transformer, the parameters of the transformer oil protection device inside the transformer, etc.

[0070] Among them, the environmental inducements described here refer to the factors that cause faults due to the operating environment of the power equipment during its operation, such as heavy precipitation, strong winds, and high temperatures, etc.

[0071] Among them, for the obtained environmental incentive factors, it is necessary to sort the regression coefficients corresponding to all the change rates. And for the obtained regression coefficients, it is obvious that they correspond to the change rates. At the same time, the change rates also correspond to the corresponding historical operation data. Therefore, establishing the correspondence between the regression coefficients and the historical operation data is essentially equivalent to establishing the correspondence between the regression coefficients and the incentives that cause power equipment failures.

[0072] Among them, for the correspondence between the regression coefficients and the failure incentives, based on the sorting of the regression coefficients, the sorting of the failure incentives is obtained.

[0073] Among them, for the historical operation data, the type of incentive can be determined. Correspondingly, for the power equipment failure incentives caused by the state of the power equipment itself, they are determined as its own failure incentives, and for the sorting of the failure incentives caused by environmental factors, they are determined as the environmental incentive sorting.

[0074] Among them, for the sorting of the regression coefficients, several larger data types can be screened out for acquisition and determination, so as to obtain the corresponding incentive sorting. For example: for the obtained regression coefficient sorting is [0.92, 0.85, 0.82, 0.64, 0.64, 0.31, 0.25, 0.10, 0.06], considering the obvious change in the parameter size in the specific calculation of the regression coefficients, in order to facilitate the acquisition of monitoring data, the two values [0.10, 0.06] are removed, or 0.31 and 0.25 are also removed. For the obtained regression coefficient sequence, the failure incentives corresponding to the regression coefficients are sorted, and the obtained result is the failure incentive sorting.

[0075] In some embodiments, the specific content of all failure incentives is also described and can be marked in the obtained regression coefficient sorting.

[0076] S135. It further includes: Based on the sorting of the failure incentives of the power equipment, obtain the type of the failure incentives, and obtain the sorting of the power equipment's own failure incentives and the operation environment incentives.

[0077] The purpose of this step is that considering that there are two types of failure incentives for the power equipment that have been obtained, including the power equipment's own failure incentives and the operation environment incentives, and the specific influence degrees of these two incentives are different. Therefore, in actual processing, both of these two factors need to be processed. And for the power equipment failures caused by different failure incentives, it is necessary to analyze the specific monitoring of different influencing factors to perform more accurate monitoring. Therefore, in the acquisition of historical operation data, the analysis of specific incentives can be monitored correspondingly based on the monitoring requirements.

[0078] Among them, for the sorted failure causes of the power equipment that have been obtained, they are classified according to the specific types of the causes.

[0079] Among them, for the classified failure causes, they are sorted according to the magnitudes of the corresponding regression coefficients. For example, for the regression coefficient sorting result in step S134, the failure causes therein are analyzed separately. Among them, the regression coefficients corresponding to the self-failure causes of the power equipment are [0.85, 0.82, 0.64, 0.31, 0.25, 0.10], and the regression coefficients corresponding to the operating environment causes are [0.92, 0.64, 0.06]. When sorting, the result can be directly obtained based on the sorting of the failure causes of the power equipment.

[0080] Among them, after obtaining all the failures, specific labels also need to be made for different failure types. For example, the regression coefficients of the self-failure causes are expressed as [0.85 (insufficient capacity), 0.82 (overload), 0.64 (circuit chaos), 0.31 (too many obstacles), 0.25 (line breakage), 0.10 (increased downstream power consumption)], and the regression coefficients of the environmental causes are expressed as: [0.92 (too high environmental temperature), 0.64 (too high air humidity), 0.06 (short-term precipitation)]. Based on this recording method, the common representation of all failure cause types and the corresponding regression coefficients can be realized.

[0081] As described in step S140, the purpose of this step is that in the monitoring of the power system, for all the causes that trigger failures, different causes have different impacts on the failures of the power equipment. For example, when a certain failure occurs, although the environmental cause will also have a certain promoting effect on the failure of the power equipment, for the same failure, it is found that the failure of the power equipment will also occur in the normal operating environment of the power equipment, which indicates that this failure type is mainly caused by the self-failure causes of the power equipment, and the impact caused by the environmental cause is very small. That is to say, for the same power equipment, the formation reasons of its failure types are different, and it is not necessarily caused by the self-failure causes and the operating environment causes of the equipment at the same time. Therefore, the correlation degree between these two types of causes needs to be analyzed. Specifically: S141. Based on the sorting of the self-failure causes of the power equipment, compare the regression coefficients corresponding to the self-failure causes with the preset regression coefficients, and obtain the failure causes not lower than the preset regression coefficients to obtain high-impact failure causes.

[0082] The purpose of this step is to rank the causes of the power equipment's own faults. In fact, not all of these fault causes are valuable for subsequent monitoring. For example, for a certain relay fault, obviously, although a loose connection of the relay can also cause a fault, the regression coefficient of the fault is extremely low. This means that if subsequent monitoring is carried out, the amount of data generated during the monitoring process will be huge, and the influence of this type of data is very small. Monitoring it will only increase the resource consumption during the data processing process and has no impact on the processing accuracy of the results. Therefore, it is necessary to determine the high-impact fault causes.

[0083] Among them, after determining the ranking of the causes of the power equipment's own faults, obtain the regression coefficients corresponding to the fault causes.

[0084] Among them, compare the obtained regression coefficients with the preset regression coefficients. Only the causes of the power equipment's own faults that are not lower than the preset regression coefficients can be determined as high-impact fault causes.

[0085] Among them, in the determination of the preset review coefficient, it is set based on the monitoring accuracy standards of technical personnel for power equipment.

[0086] In some embodiments, calculate the average value of the regression coefficients corresponding to all the obtained causes of the power equipment's own faults, and this average value result is the preset regression coefficient.

[0087] S142. Based on the ranking of the environmental causes, compare the regression coefficients corresponding to the environmental causes with the preset environmental regression coefficients, and obtain the regression coefficients corresponding to the environmental causes that are not lower than the preset environmental regression coefficients, so as to obtain high-impact environmental causes.

[0088] The purpose of this step is that for the obtained environmental causes, considering that different environmental causes have different effects on promoting the faults of power equipment. For example, for the power equipment in the distribution box, the influence of the wind force factor is obviously very low. If the wind force parameters are also monitored in subsequent monitoring, there is no improvement effect in terms of monitoring requirements and accuracy, and it will also increase the resource consumption during the data processing process, which is not helpful for the implementation of the technical solution. Therefore, in the technical solution of this application, the environmental causes that need to be monitored are determined in advance.

[0089] Among them, the ranking of the environmental causes is determined based on the regression coefficients corresponding to the environmental causes. In this case, accurately determine the ranking of the environmental causes.

[0090] Among them, for the preset environmental regression coefficient, this parameter can be set based on the monitoring accuracy requirements for power equipment.

[0091] In some embodiments, for a preset environmental coefficient, the mean value of the regression coefficients corresponding to all environmental incentives can be calculated, and the obtained mean value result can be used as the preset regression coefficient for application.

[0092] Among them, only when it is determined that the regression coefficient corresponding to the environmental incentive is not lower than the preset environmental coefficient, can this environmental incentive be directly set as a high-impact environmental incentive.

[0093] S143. Based on the high-impact environmental incentives, obtain the power equipment failure probability caused by the high-impact environmental incentives. The power equipment failure probability determination equation is: ; Among them, represents the failure probability caused by high-impact environmental incentives; represents the intercept term, which is the logit when all independent variables are 0; represents the regression coefficient of high-impact environmental incentives; represents the specific parameter of high-impact environmental incentives; i represents the index of high-impact environmental incentives.

[0094] The purpose of this step is that among the operating failures of power equipment, not all failures are caused by both its own failure incentives and operating environmental incentives. Only when it is determined that there is a strong correlation between the two, can it be considered that the failure of the power equipment is caused by both its own failure incentives and operating environmental incentives. Therefore, in this step, the power equipment failure probability is established, and all the parameters therein are operating environment data to analyze the operating environmental incentives.

[0095] Among them, for the above equation, parameter settings are made for all the operating environmental incentives, so that based on this equation, when a failure occurs in the current power equipment, the failure probability caused only by the operating environmental incentives can be determined.

[0096] Among them, for the operating environmental incentives in the above equation, the specific parameters of the high-impact environmental incentives that have been obtained are selected to obtain the results.

[0097] Among them, the specific parameters of the high-impact environmental incentives are obtained based on historical operating environment parameters.

[0098] Among them, for the regression coefficient of the high-impact environmental incentives, it is obtained based on the regression coefficient determination method mentioned above, which will not be elaborated here.

[0099] S144. Compare the power equipment failure probability caused by the corresponding environmental data with the preset power equipment failure probability. If the power equipment failure probability caused by the corresponding environment is not lower than the preset power equipment failure probability, determine that the high-impact environmental factor is the cause of the power equipment failure.

[0100] The purpose of this step is to conduct a specific analysis of the environmental factors in the power equipment failure probability results that have been obtained. From the overall results, only when the probability of power equipment failure caused by environmental factors reaches a certain value can it be considered that the power equipment failure can be caused by environmental factors alone.

[0101] Among them, the preset probability of power equipment failure needs to be set based on the monitoring accuracy of the power equipment.

[0102] In some embodiments, the preset probability of power equipment failure may be directly set using the probability of corresponding failure of current power equipment.

[0103] The preset probability of power equipment failure refers to the probability of a corresponding failure of the power equipment based solely on the influence of the operating environment, without considering the failure caused by the power equipment's own factors.

[0104] S145. Determine the high-impact fault cause, or the high-impact fault cause and the high-impact environmental cause as monitoring causes for the electric power equipment.

[0105] The purpose of this step is that, although two high-impact inducements have been obtained in the monitoring of power equipment, namely: high-impact fault inducements and high-impact environmental inducements, based on the purpose and beneficial effects of step S144, it can be known that the high-impact environmental inducements do not necessarily lead to operational failures of power equipment. At the same time, considering that for the inspection and monitoring of power equipment, the parameters that need to be monitored are uniformly configured, which can substantially reduce the complexity of the system. Therefore, it is necessary to determine the monitoring inducements of the power equipment.

[0106] Among them, the monitoring inducement of power equipment refers to all inducements that can cause power failures during the operation of the power equipment, and the inducement does not distinguish between failures caused by the power equipment's own factors and failures caused by the operating environment, but uniformly regards them as inducements that need to be monitored.

[0107] Among them, considering that when a power device fails, the operating parameters of the power device itself will inevitably change greatly. Therefore, in order to better identify the faults of the power device, for the high-impact fault incentives generated, regardless of whether the fault of the power device is caused by the operating environment of the power device, it is necessary to obtain and monitor the high-impact fault incentives to determine whether a fault will occur. At this time, there is no need to use the high-impact environmental incentives as the monitoring object, and only the high-impact fault incentives need to be used as the monitoring incentives for the power device.

[0108] Among them, when it is determined that the high-impact environmental incentive is the cause of the power device fault, it means that during the operation of the power device, the fault of the power device will be caused by the combined action of the high-impact fault incentive and the high-impact environmental incentive of the power device. Therefore, in this case, it is necessary to jointly set the high-impact fault incentive and the high-impact environmental incentive as the monitoring incentives for the power device.

[0109] In some embodiments, regardless of whether the high-impact environmental incentive is the cause of the power device fault, two types of data, namely the high-impact fault incentive and the high-impact environmental incentive, are obtained and used as the monitoring incentives for the power device.

[0110] As described in step S150, the purpose of this step is that for the monitoring incentives of the power device, in fact, in all the above steps, the analysis results obtained are descriptions in words or corresponding numbers of various faults, rather than numerical values. However, in the specific monitoring process, it is obvious that the monitoring results need to be determined based on numerical values. Therefore, corresponding data comparison methods need to be used to obtain the monitoring results. Specifically: S151. Based on the pre-fault operating data and the monitoring incentives of the power device, obtain the historical monitoring incentive values of the power device.

[0111] The purpose of this step is that in the monitoring of the power device, it is necessary to set the reference values for monitoring and comparison. Therefore, the purpose of this step is to set the standard values used in the monitoring process.

[0112] Among them, for the pre-fault operating data, explanations have been made in the above steps, that is: the operating data of the power device within a period of time before the fault occurs. Considering that when the power device is about to fail, the changes in various historical operating data are more obvious, so the historical operating data within this period can be used as the standard numerical values for monitoring.

[0113] Among them, for the monitoring incentives of the power device of the present application, there will be various different parameters, that is: different incentives, and each different incentive corresponds to different pre-fault operating data. Therefore, in the process of processing, it is necessary to obtain the historical monitoring incentive values of the power device based on the monitoring incentives of the power device.

[0114] Among them, for the historical monitoring incentive value, from the corresponding historical operation data group, the specific type of the monitoring incentive of the corresponding power equipment is used to obtain the historical operation parameters, and the corresponding relationship between the monitoring incentive of the power equipment and the historical monitoring incentive data is established, so as to determine the specific historical monitoring incentive value.

[0115] S152. Based on the monitoring incentive of the power equipment, obtain the measured value of the monitoring incentive of the power equipment.

[0116] The purpose of this step is to achieve the monitoring purpose of the power equipment. Obviously, the actual operation data of the power equipment also needs to be obtained. However, considering that there may be fluctuations in the monitoring data during the operation of the power equipment, and in order to improve the processing accuracy of the monitoring results, during the processing, it is necessary to compare the measured values obtained within a time period to obtain the monitoring results.

[0117] Among them, according to the monitoring incentive type of the power equipment, the operation values of the relevant types of monitoring incentives are obtained.

[0118] Among them, for the time period before the occurrence of the fault that has been obtained, its length is known. When obtaining the measured value of the monitoring incentive of the power equipment, it is necessary to obtain the measured value of the monitoring incentive of the power equipment according to the length of the time period before the occurrence of the fault.

[0119] Among them, taking the time period before the occurrence of the fault that has been obtained as the benchmark for obtaining the measured value, based on the number of sampling time points within this time period, the measured value of the monitoring incentive of the power equipment is updated. For example: The data acquisition time points and the corresponding time data groups within the time period before the occurrence of the fault are expressed as I 1 (11:00), I 2 (11:05), I 3 (11:10), ···, I n (11:00 + 5n)], which means that the current data of the power equipment is obtained, and sampling is performed every 5 minutes. The total sampling time is 5n minutes, which is the same as the length of the time period before the occurrence of the fault. In the next sampling cycle, the data group changes to I 2 (11:05), I 3 (11:10), I 4 (11:15), ···, I n+1(11:00 + 5n + 5), where, for I n+1 is the value corresponding to a new sampling time point, and the sampling time is adjusted accordingly, but the total sampling time length remains unchanged.

[0120] Among them, for the time length between data acquisition time points, it also needs to be set according to the historical data acquisition time points corresponding to the time period before the fault occurred. For example, if the interval between historical data acquisition time points is 2 minutes, then in the acquisition of measured data, the time interval between adjacent time points also needs to be set to 2 minutes.

[0121] In some embodiments, when it is found that there are data missing points among them, based on the data corresponding to the adjacent time points of the time missing point, the data of the missing point is supplemented based on the slope of the previous time point and the next time point.

[0122] Among them, for the monitoring incentives of power equipment, all the monitoring incentives included therein need to obtain their historical monitoring incentive values and measured values. For example, the historical monitoring incentive values obtained are expressed as: ; Among them, the data corresponding to each column in this matrix is the monitoring incentive category of different power equipment. The first column is current, the second column is voltage, the third column is temperature, and the last column is humidity, etc. The value of each row represents the historical monitoring incentive value corresponding to the data acquisition time node. The measured value is expressed as: ; Among them, the data corresponding to each column in this matrix is the measured value of the monitoring incentive category of different power equipment. The first column is the measured value of current, the second column is the measured value of voltage, the third column is the measured value of temperature, and the last column is the measured value of humidity, etc. The value of each row represents the measured value corresponding to the data acquisition time node.

[0123] S153. Compare the historical monitoring incentive value of the power equipment with the measured value of the monitoring incentive of the power equipment to obtain a comparison result of the incentive values.

[0124] The purpose of this step is that in the monitoring of power equipment, it is necessary to obtain the comparison result between these two types of data based on the historical data and the obtained comparison result of the measured values, so as to obtain the monitoring result of the power equipment to determine whether the currently obtained measured value means that the current power equipment will have an operation failure.

[0125] Among them, the measured values that have been obtained are compared with the historical monitoring incentive monitoring values to determine the deviation amount between these two types of data. Specifically, all the values among them are compared horizontally. The so-called horizontal comparison means obtaining the values corresponding to the same acquisition time node and the monitoring incentives of the same device, and obtaining the deviation degree between these two values. For example, for the comparison between and , its determination equation is: ; Using this method to calculate the deviation degree of all the obtained values, when all the obtained deviation degrees are within the set preset deviation degree range, it is considered that the currently obtained measured value and the historical value are extremely similar. Considering that the final result of the historical value is that the power equipment has a fault, the current measured value also indicates that the power equipment will have the same fault.

[0126] In some embodiments, it is also possible to determine the ratio of the change rates of adjacent values within the measured value and historical value system, and compare the obtained calculation result with the preset change rate ratio. When all the obtained change rate ratios are within the set preset change rate ratio, it is considered that the currently obtained measured value and the historical value are extremely similar. Considering that the final result of the historical value is that the power equipment has a fault, the current measured value also indicates that the power equipment will have the same fault.

[0127] S154. Based on the comparison result of the incentive values and the regression coefficient corresponding to the monitoring incentive of the power equipment, obtain the monitoring result of the power equipment.

[0128] The purpose of this step is that for the monitoring result of the power equipment, based on the historical operation data, the regression coefficients corresponding to the monitoring incentives of different power equipment have been obtained. Different regression coefficients mean that the monitoring incentives of the relevant power equipment have different possibilities for the finally formed fault. In order to better obtain the determination of the probability of the fault occurrence, the result needs to be obtained based on the regression coefficient.

[0129] Among them, based on the determination of the comparison result of the incentive values, analyze the comparison result of the monitoring incentive of the corresponding power equipment. When it is found that the obtained comparison result is extremely similar to the historical value, it is considered that the current measured value needs to be applied to the calculation of the fault monitoring result.

[0130] Among them, when it is found that the comparison result of the monitoring incentive of a certain power equipment does not mean that the results of the historical data and the measured data are similar enough, the measured value cannot be applied to the calculation of the fault monitoring result.

[0131] Among them, when the application method of the measured value is determined, the occurrence of a fault is determined, and the determination equation is: ; Among them, represents the probability of fault occurrence, j represents the index of the measured value of the fault cause of the power equipment, k represents the total index of the measured values of the fault causes of the power equipment, p represents the index of the fault cause type of the power equipment, q represents the total index of the fault cause types of the power equipment.

[0132] Among them, for the obtained result of the probability of fault occurrence, it is necessary to determine the probability level of fault occurrence of the power equipment according to the specific value of the probability of fault occurrence. For example: when the probability is in the range of 0.1 - 0.5, it is considered that a fault is not likely to occur; when it is in the range of 0.5 - 0.7, it is considered that a fault may occur; when it is in the range of 0.7 - 0.8, it is considered that a fault is likely to occur; when it is in the range of 0.8 - 0.9, it is considered that a fault is extremely likely to occur; when it is in the range of 0.9 - 1.0, it is considered that a fault will definitely occur, etc.

[0133] As described in step S160, the purpose of this step is that during the operation of the power system, there may be a phenomenon of fault transfer between power equipment, and even a skip - connection transfer phenomenon may occur for some of these faults. In order to be able to monitor other power equipment after a fault occurs, it is also necessary to set up a specific monitoring plan for other power equipment based on the power connection relationship between power equipment. Specifically: S161. Based on the power connection relationship between power equipment, obtain the historical operation data group of the power equipment connected to the faulty power equipment to obtain an associated data group.

[0134] The purpose of this step is that during the operation of power equipment, there is obviously a power connection relationship between each power equipment, which may result in a fault transfer relationship between power equipment. In order to obtain the fault transfer relationship between power equipment, by obtaining the associated data group, the transfer relationship of faults can be analyzed.

[0135] Among them, to obtain the power connection relationship between power equipment, the power connection relationship includes direct connection, skip - connection, coupling, cascade, and secondary connection, etc.

[0136] Among them, based on the power equipment connection relationship, obtain the corresponding historical operation data group and establish an association relationship between the historical operation data groups.

[0137] In some embodiments, for the historical operation data that has been stored, analyze the occurrence of synchronization failures therein, or the failures that occur in other power equipment within a set time, and further determine the fault transmission relationship.

[0138] In some embodiments, based on the power equipment that has failed, determine the fault type, and then based on the specific construction mode of the power system, analyze the fault transmission situation between this fault type and other power equipment to obtain the fault transmission relationship.

[0139] S162. Obtain the power equipment fault data of the associated data group, and obtain the occurrence time node of the power equipment fault data of the associated data group.

[0140] The purpose of this step is that when determining the fault data in the associated data group, essentially the same determination method as that of the power equipment monitoring method can be adopted. Therefore, it is also necessary to obtain other data according to the occurrence time node of the power equipment fault value to determine the monitoring scheme.

[0141] Among them, for the occurrence time node of the power equipment fault data, the specific determination method is the same as that of step S122, and will not be elaborated here.

[0142] S163. If the deviation amount between the occurrence time node of the power equipment fault data of the associated data group and the power data acquisition time point corresponding to the power equipment fault data is not higher than the preset deviation amount, determine that there is a fault association relationship between the power equipment, and obtain the fault-associated power equipment.

[0143] The purpose of this step is that in the analysis of the associated data of power equipment, in actual performance, the faults between different power equipment occur synchronously or with a time delay. Therefore, based on the time parameter and the power connection relationship, the association relationship of different faults can be determined.

[0144] Among them, based on the fault manifestation of the power equipment, determine the occurrence time node of the fault in the associated data group. For example, if it is found that the occurrence time nodes of a certain two types of data are the same and there is a power connection relationship, then there is a fault connection relationship between these two power equipment.

[0145] Among them, for the set preset deviation amount, it can be based on the fault type of the power equipment and the time difference between the occurrence times of the faults of other power equipment with an association relationship, and then use this time difference value as the preset deviation amount.

[0146] S164. Obtain the self-fault causes of the fault-associated power equipment, or sort the self-fault causes and environmental causes of the fault-associated power equipment to obtain the fault cause ranking of the fault-associated power equipment.

[0147] The purpose of this step is that for faulty associated power equipment, sorting the fault causes of various types of faulty associated power equipment can be used to obtain the specific fault causes of the faulty associated power equipment. Based on this fault cause, the fault cause of the faulty associated power equipment can be determined.

[0148] Among them, the method adopted in this step is the same as that in step S140, which will not be elaborated here.

[0149] S165. Obtain the monitoring causes of the faulty associated power equipment, and based on the measured data of the monitoring causes of the faulty associated power equipment and the historical monitoring cause data of the faulty associated power equipment, obtain the monitoring result of the faulty associated power equipment.

[0150] The situation faced in this step is that for faulty associated power equipment, although the fault data of the power equipment will generate a fault transfer relationship, whether this faulty associated power equipment actually fails will also be affected by other factors. For example, when the faulty power equipment has transmitted faulty power data to other power equipment, the occurrence of corresponding faults in other power equipment will only show fault manifestations under the coupling effect of this faulty power data and the operating environment data. Therefore, in the monitoring of faulty associated power equipment, it is also necessary to determine the monitoring results therein to obtain the specific fault occurrence situation of the faulty associated power equipment.

[0151] Among them, the monitoring result of the faulty associated power equipment is the same as the methods in steps S120 - S150, which will not be elaborated here.

[0152] In addition, this application also discloses a power equipment monitoring system based on machine learning for executing the methods corresponding to all the above-mentioned steps, as Figure 2 shown, a schematic diagram of a power equipment monitoring system based on machine learning provided by an embodiment of this application, including: a data acquisition module, a data preprocessing module, an association analysis module, a data analysis module, a power equipment monitoring module, and a database; The data acquisition module is connected to the power equipment monitoring module, the data preprocessing module, the association analysis module, and the database, and is used to obtain the measured values of the monitoring causes of the power equipment, the historical operation data groups of the power equipment, and send the measured values of the monitoring causes of the power equipment to the data preprocessing module, the association analysis module, and the database; The association analysis module is also connected to the data analysis module, and is used to analyze the power connection relationship between power equipment and obtain the association data groups, the association relationship between high - impact fault causes and high - impact environmental causes; The data analysis module obtains the monitoring results of the power equipment based on the monitoring incentives of the power equipment.

[0153] It further includes: The data preprocessing module is also connected to the database and is used to preprocess the measured data of the monitoring incentives of the power equipment obtained by the power equipment monitoring module, and then send it to the database for storage. The data preprocessing module is also connected to the correlation analysis module and is used to preprocess the measured data of the monitoring incentives of the power equipment obtained by the power equipment monitoring module, and then send it to the correlation analysis module to analyze the correlation between high-impact fault incentives and high-impact environmental incentives.

[0154] It further includes: The data acquisition module also obtains the power connection relationship of the power equipment from the database to obtain the fault correlation relationship between the power equipment.

[0155] The technical solutions in the embodiments of the present application at least have the following technical effects or advantages: 1. It realizes the reduction of the number of monitoring objects. In the technical solution of the present application, for the operation faults of the power equipment, both the faults of the power equipment itself and the faults caused by environmental factors are analyzed, and then the correlation between these two types of factors and the occurrence of power equipment faults is determined. The fault influencing factors with higher influence degree are obtained from them, and then these influencing factors are used as the monitoring objects in the monitoring process of the power equipment. In this way, it is not necessary to obtain and process all the operation data, and only a part of the monitoring data needs to be obtained to achieve high-precision monitoring of the operation state of the power equipment.

[0156] 2. It realizes the determination of the fault correlation relationship of the power equipment. In the technical solution of the present application, the connection relationship between the power equipment is analyzed, especially the fault manifestations of the power equipment with a connection relationship are analyzed, and then the fault incentives with this connection relationship are determined and also used in the monitoring, so as to realize the fault correlation analysis of the power equipment, and thus ensure that based on the connection relationship of the power equipment, all possible faults in the current power system can be monitored.

[0157] 3. The early determination of the operation fault monitoring objects of power equipment is realized. In the technical solution of this application, based on the historical operation data of power equipment, the fault monitoring objects therein are analyzed in advance. Then, in actual monitoring, based on the early determination of such early-determined fault monitoring objects, the parameters of such determined monitoring objects can be obtained in actual monitoring. By the early determination of the operation fault monitoring objects, the fault determination efficiency in the monitoring operation of power equipment is improved. In addition, in specific monitoring, instead of using discrete data for processing, the data within a time period is used for monitoring, which improves the monitoring accuracy.

[0158] Those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to computer program instructions. The aforementioned computer program can be stored in a non-volatile storage medium. When the computer program is executed, it executes the steps including the above method embodiments. Alternatively, if the above integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a non-volatile storage medium. Based on such an understanding, the technical solution of the embodiments of the present invention, in essence, or the part that contributes to the prior art can be embodied in the form of a software product. This computer software product is stored in a non-volatile storage medium and includes several instructions for causing an electronic device (which can be a personal computer, a server, a network device, etc.) to execute all or part of the methods described in the various embodiments of the present invention.

[0159] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention.

Claims

1. A power equipment monitoring method based on machine learning, characterized in that, The monitoring method includes: Obtaining the historical operation data of power equipment and establishing a historical operation data group for each power equipment; Obtaining the time points of the historical operation data of power equipment and obtaining the corresponding environmental data at the time points; Based on the operation environment data and the historical operation data, obtaining the sorting of the fault causes of power equipment; Based on the sorting of the fault causes of power equipment, obtaining the monitoring causes of power equipment; Obtaining the historical data of the monitoring causes of power equipment in the time period before the occurrence of a fault, and obtaining the measured values of the monitoring causes of power equipment in the same length time period, and obtaining the monitoring result based on the historical data and the measured values; Based on the power connection relationship between power equipment, obtaining the sorting of the fault causes of the fault-related power equipment and obtaining the monitoring results of the fault-related power equipment.

2. The method for monitoring power equipment based on machine learning according to claim 1, characterized in that The obtaining the historical operation data of power equipment and establishing a historical operation data group for each power equipment includes: Obtaining the unique identification code of each power equipment in the power system and establishing an equipment data group based on the unique identification code; Adding the historical operation data of power equipment into the equipment data group to obtain the historical operation data group of power equipment; It also includes: Obtaining the acquisition time node of the historical operation data of power equipment to obtain the data acquisition time point; In the historical operation data group, establishing the corresponding relationship between the data acquisition time point and the historical operation data of the power equipment.

3. The machine learning-based power equipment monitoring method according to claim 1, characterized in that The obtaining the time points of the historical operation data of power equipment and obtaining the corresponding environmental data at the time points includes: Based on the unique identification code of power equipment, obtaining the corresponding environmental data of power equipment; Based on the data acquisition time point, obtaining the corresponding environmental data of the power equipment corresponding to the data acquisition time point and establishing a corresponding relationship; Supplementing the corresponding environmental data of the power equipment corresponding to the data acquisition time point into the historical operation data group corresponding to the power equipment to obtain the historical data group of power equipment.

4. The method for monitoring power equipment based on machine learning according to claim 1, characterized in that The obtaining the sorting of the fault causes of power equipment based on the operation environment data and the historical operation data includes: Based on the power data acquisition time point corresponding to the power equipment fault data, obtaining the historical operation data of the power equipment in the previous time period to obtain the operation data before the fault; Based on the operation data before the fault, obtaining the change rate of the historical operation data; Based on the change rate, obtaining the regression coefficients corresponding to each change rate, and the equation for determining the regression coefficients corresponding to the change rate is: ； Among them, represents the regression coefficient corresponding to the change rate, represents the change rate, m represents the category index of the historical operation data, n represents the total amount of categories of the historical operation data; Sorting the regression coefficients corresponding to the change rate to obtain the sorting of the fault causes of power equipment; It also includes: Based on the sorting of the fault causes of power equipment, obtaining the types of fault causes and obtaining the sorting of the self-fault causes and the operation environment causes of power equipment.

5. The method for monitoring power equipment based on machine learning according to claim 1, wherein The obtaining the monitoring causes of power equipment based on the sorting of the fault causes of power equipment includes: Based on the sorting of the self-fault causes of power equipment, comparing the regression coefficients corresponding to the self-fault causes with the preset regression coefficients and obtaining the fault causes not lower than the preset regression coefficients to obtain the high-impact fault causes; Based on the sorting of the environmental incentives, compare the regression coefficients corresponding to the environmental incentives with the preset environmental regression coefficients, and obtain the regression coefficients of the environmental incentives not lower than the preset environmental regression coefficients to obtain the high-impact environmental incentives; Based on the high-impact environmental incentives, obtain the power equipment failure probability caused by the high-impact environmental incentives. The power equipment failure probability determination equation is: ； Among them, represents the failure probability caused by high-impact environmental incentives; represents the intercept term, which is the log odds when all independent variables are 0; represents the regression coefficient of high-impact environmental incentives; represents the specific parameter of high-impact environmental incentives; i represents the index of high-impact environmental incentives; Compare the power equipment failure probability caused by the corresponding environmental data with the preset power equipment failure probability. If the power equipment failure probability caused by the corresponding environment is not lower than the preset power equipment failure probability, determine that the high-impact environmental incentive is the cause of the power equipment failure; Determine the high-impact failure incentive, or the high-impact failure incentive and the high-impact environmental incentive as the monitoring incentives for the power equipment.

6. The method for monitoring power equipment based on machine learning according to claim 1, wherein, The historical data of the monitoring incentives of the power equipment in the time period before the failure occurs is obtained, and the measured values of the monitoring incentives of the power equipment in the time period of the same length are obtained. The monitoring results are obtained based on the historical data and the measured values, including: Based on the pre-failure operation data and the monitoring incentives of the power equipment, obtain the historical monitoring incentive values of the power equipment; Based on the monitoring incentives of the power equipment, obtain the measured values of the monitoring incentives of the power equipment; Compare the historical monitoring incentive values of the power equipment with the measured values of the monitoring incentives of the power equipment to obtain the comparison result of the incentive values; Based on the comparison result of the incentive values and the regression coefficients corresponding to the monitoring incentives of the power equipment, obtain the monitoring results of the power equipment.

7. The method for monitoring power equipment based on machine learning according to claim 1, characterized in that Based on the power connection relationship between power equipment, obtain the sorting of the failure incentives of the fault-associated power equipment, and obtain the monitoring results of the fault-associated power equipment, including: Based on the power connection relationship between power equipment, obtain the historical operation data groups of the power equipment connected to the faulty power equipment to obtain the associated data groups; Obtain the power equipment failure data of the associated data group, and obtain the occurrence time node of the power equipment failure data of the associated data group; If the deviation amount between the occurrence time node of the power equipment failure data of the associated data group and the power data acquisition time point corresponding to the power equipment failure data is not higher than the preset deviation amount, determine that there is a fault association relationship between the power equipment to obtain the fault-associated power equipment; Obtain the monitoring incentives of the fault-associated power equipment, and obtain the monitoring results of the fault-associated power equipment based on the measured data of the monitoring incentives of the fault-associated power equipment and the historical monitoring incentive data of the fault-associated power equipment.

8. A machine learning-based power equipment monitoring system for performing the power equipment monitoring method according to any one of claims 1 to 7, characterized in that, Including: A data acquisition module, a data preprocessing module, an association analysis module, a data analysis module, a power equipment monitoring module, and a database; The data acquisition module is connected to the power equipment monitoring module, the data preprocessing module, the association analysis module, and the database, and is used to obtain the measured values of the monitoring incentives of the power equipment, the historical operation data of the power equipment, and send the measured values of the monitoring incentives of the power equipment to the data preprocessing module, the association analysis module, and the database; The correlation analysis module is also connected to the data analysis module, and is used to analyze the power connection relationship between power equipment and obtain the correlation data group, the correlation relationship between high-impact fault causes and high-impact environmental causes; The data analysis module obtains the monitoring results of power equipment based on the monitoring causes of power equipment.

9. The machine learning-based power equipment monitoring system according to claim 8, wherein, It further includes: The data preprocessing module is also connected to the database, and is used to preprocess the measured data of the monitoring causes of power equipment obtained by the power equipment monitoring module, and then send it to the database for storage; The data preprocessing module is also connected to the correlation analysis module, and is used to preprocess the measured data of the monitoring causes of power equipment obtained by the power equipment monitoring module, and then send it to the correlation analysis module to analyze the correlation between high-impact fault causes and high-impact environmental causes.

10. The power equipment monitoring system based on multi-region learning according to claim 8, characterized in that, It further includes: The data acquisition module also obtains the power connection relationship of power equipment from the database to obtain the fault correlation relationship between power equipment.

Images