A data-driven intelligent identification method and system for power equipment defects

Abstract

The present invention discloses a data-driven intelligent defect identification method and system for power equipment, which relates to the technical field of power equipment defect identification, including obtaining historical defect records of power equipment from a power grid platform, analyzing the state parameters in the defect identification data to determine the abnormality of the power equipment defects; analyzing the operating parameters in the equipment operation records to determine the parameter influence on the abnormal state parameters in the parameter abnormality data; predicting the change trend of the state parameters of the power equipment in the current cycle, obtaining parameter prediction data, and combining the defect identification data to identify the equipment defects of the power equipment; obtaining characteristic defect root cause data of the defective power equipment, obtaining parameter prediction data, analyzing the defect root cause of the defective power equipment, sending the defect root cause data of the defective power equipment to staff through the power grid platform, and prompting the staff to repair the defective power equipment.

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment defect identification, and in particular to a data-driven intelligent power equipment defect identification method and system. Background Art

[0002] In the power grid, power equipment ensures the safe and coordinated operation of the power generation, transformation, transmission and distribution links of the power grid, thereby realizing safe and stable power supply to users. Therefore, judging and identifying defects in power equipment is an indispensable part of ensuring the normal and efficient operation of the power grid. The traditional method of judging defects in power equipment is mainly for professionals to judge the equipment status of power equipment and find defective power equipment through on-site inspection and data collected by sensors. However, this method not only requires a lot of manual analysis, but also leads to errors in the judgment of defects in power equipment due to the lack of professional skills of professionals. At present, there is no good method for analyzing the root causes of defects in power equipment, which makes it impossible to discover and repair equipment defects in time, which has a great negative impact on the normal use of the power grid. Summary of the Invention

[0003] The purpose of the present invention is to provide a data-driven intelligent identification method and system for power equipment defects to solve the problems raised in the prior art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a data-driven intelligent identification method for power equipment defects, the method comprising:

[0005] Step S100: Obtain historical defect records of power equipment from the power grid platform, obtain defect identification data of the power equipment, analyze the state parameters in the defect identification data to determine the abnormality of the power equipment defect, and obtain parameter abnormality data;

[0006] Step S200: obtaining historical equipment operation records of the power equipment, analyzing the operating parameters in the historical equipment operation records, and determining the parameter impact degree of the abnormal state parameters in the parameter abnormality data to obtain parameter impact data;

[0007] Step S300: Acquire parameter impact data, monitor the power equipment in the current cycle, predict the change trend of the power equipment status parameters in the current cycle, obtain parameter prediction data, and combine the defect identification data to identify the power equipment defects and obtain defective power equipment;

[0008] Step S400: Acquire characteristic defect root cause data of the defective power equipment, acquire parameter prediction data, analyze the defect root cause of the defective power equipment, obtain the defect root cause data, send the defect root cause data of the defective power equipment to the staff through the power grid platform, and prompt the staff to repair the defective power equipment.

[0009] Furthermore, step S100 includes:

[0010] Step S101: Acquire various historical defect records of the power equipment, and obtain data of various status parameters of the power equipment from the historical defect records, wherein all the power equipment in the historical defect records has equipment defects;

[0011] Step S102: Obtaining defect identification data of the power equipment from the power grid platform, where the defect identification data is the range of various state parameters when performing defect identification on the power equipment;

[0012] When the maximum and minimum values ​​of various state parameters in a certain historical defect record are both within the range of the defect discrimination data, it is determined that the defect discrimination data has incorrectly discriminated the defect of the certain historical defect record, and the certain historical defect record is marked and recorded as an abnormal historical defect record;

[0013] Get the total number C of several unmarked historical defect records in the power equipment sum , obtaining several status parameters whose maximum or minimum values ​​in the unmarked historical defect records are outside the range of the defect discrimination data, and aggregating them to obtain a feature parameter set of the unmarked historical defect records;

[0014] Step S103: Calculate the defect judgment ratio of each state parameter, where the defect judgment ratio L of the ath state parameter in the power equipment is a =C (a,sum) / C sum , C (a,sum) is the total number of unmarked historical defect records containing the a-th state parameter in the feature parameter set;

[0015] Calculate the average value L' and standard deviation σ' of the defect discrimination ratio of each state parameter, and obtain the lower limit ratio threshold B of each state parameter min =L´-k×σ´, where k is a preset coefficient. When the defect discrimination ratio of a certain state parameter is less than the lower limit ratio threshold, the state parameter is marked and recorded as a suspected abnormal state parameter;

[0016] Step S104: Obtain the median D of the suspected abnormal state parameters in the defect identification data within the corresponding range △, respectively calculate the maximum value, minimum value and median D of the suspected abnormal state parameters in the abnormal historical defect records of the power equipment △ The absolute value of the difference, and take the maximum absolute value, divide the maximum value by the median D △ , obtain the parameter abnormal value e of the suspected abnormal state parameter in the abnormal historical defect record;

[0017] When the parameter abnormal value e is greater than a preset threshold value e´, the abnormal historical defect record is recorded as a target abnormal historical defect record of the suspected abnormal state parameter, and the ratio of the total number of target abnormal historical defect records of the suspected abnormal state parameter to the total number of abnormal historical defect records of each abnormal state parameter in the power equipment is obtained to obtain the abnormal ratio of the suspected abnormal state parameter;

[0018] When the abnormality ratio is greater than a preset threshold, it is determined that the suspected abnormal state parameter in the defect discrimination data is abnormal when identifying defects in the power equipment, and the suspected abnormal state parameter is recorded as the abnormal state parameter;

[0019] Acquire and aggregate several abnormal status parameters of the power equipment to obtain abnormal parameter data of the power equipment.

[0020] Furthermore, step S200 includes:

[0021] Step S201: Acquire each historical device operation record of the power equipment, pre-process the data in each historical device operation record, set a unit time, and acquire the average value of each operating parameter in the historical device operation record every unit time to obtain a parameter set of each operating parameter;

[0022] Step S202: Acquire abnormal parameter data of the power equipment, and obtain various abnormal state parameters from the abnormal parameter data;

[0023] Obtain the average value of each abnormal state parameter from the historical equipment operation record every unit time to obtain a parameter set of each abnormal state parameter;

[0024] Calculate the parameter influence degree between each operating parameter and each abnormal state parameter in the historical equipment operation record, where the parameter influence degree S between the fth operating parameter and the gth abnormal state parameter in the historical equipment operation record is (f,g) ;

[0025] Step S203: Calculate the data impact value R of the fth operating parameter on the gth abnormal state parameter (f,g) :

[0026] ,

[0027] Where z is the total number of historical equipment operation records of power equipment; S z (f,g) is the parameter influence degree between the fth operating parameter and the gth abnormal state parameter in the zth historical equipment operation record of the power equipment;

[0028] Step S204: Setting the impact threshold R, when |R (f,g) |≥R, then it is determined that the fth operating parameter has an impact on the data of the gth abnormal state parameter, and the fth operating parameter is recorded as the influencing operating parameter of the power equipment. The various influencing operating parameters of the power equipment are obtained and aggregated to obtain parameter impact data.

[0029] Furthermore, step S300 includes:

[0030] Step S301: Acquire parameter impact data, acquire various parameters affecting the operation of the power equipment, acquire various historical operation records of the power equipment, pre-process the data in each historical operation record, and sort each historical operation record in chronological order;

[0031] Build a data prediction model for power equipment and divide historical operation records into training sets and test sets according to preset ratios;

[0032] Obtain input data and output data of a model data set in the model, where the input data is various state parameters and various influencing operation parameters in a certain historical operation record, and the output data is various state parameters in the next historical operation record of a certain historical operation record;

[0033] Step S302: Use the training set to train the data prediction model, and use the test set to calculate the model prediction accuracy of the data prediction model. When the model prediction accuracy is greater than a preset threshold, it is determined that the data prediction model is constructed.

[0034] Step S303: monitoring the power equipment in the current cycle, collecting various state parameters of the power equipment and values ​​of various operating parameters affecting the power equipment at unit time intervals, and inputting these into a data prediction model to predict the various state parameters of the power equipment to obtain parameter prediction data, which includes a parameter set for predicting various state parameters;

[0035] When the maximum or minimum value of the state parameter in the parameter prediction data is out of the range of the state parameter in the defect judgment data, it is determined that there is a risk of equipment defect in the power equipment in the current cycle, and the state parameter and the power equipment are recorded as defective state parameter and defective power equipment respectively.

[0036] Furthermore, step S400 includes:

[0037] Step S401: Obtain defective power equipment in the power grid, obtain characteristic defect root cause data of the defective power equipment from the power grid platform, obtain root cause identification data for each defect root cause of the defective power equipment from the characteristic defect root cause data, obtain state parameters required for identifying a particular defect root cause from the characteristic defect root cause data, and record them as target state parameters for the particular defect root cause. The root cause identification data for a particular defect root cause is the data range within which each target state parameter for the particular defect root cause falls when the particular defect root cause occurs in the defective power equipment.

[0038] Step S402: Obtain a dataset of predicted defect state parameters of the defective power equipment. When a certain defect state parameter is the target state parameter of a certain defect root cause, and the value of an element in the dataset of the predicted defect state parameter is within the data range of the root cause identification data, it is recorded as a parameter match of the certain defect root cause. The total number of parameter matches of the certain defect root cause, M, is obtained. sum , obtain the total number of target state parameters of a certain defect root cause M sum ;

[0039] Calculate the probability of occurrence of a certain defect root cause P=M´ sum / M sum When the occurrence probability P is greater than the preset probability threshold, it is determined that the defective power equipment has a certain defect root cause in the current cycle, and the certain defect root cause is recorded as the target defect root cause;

[0040] Step S403: Obtain and aggregate the root causes of each target defect of the defective power equipment in the current cycle to obtain defect root cause data. The defect root cause data is sent to power grid staff via the power grid platform, prompting the staff to dispatch personnel to repair the defective power equipment.

[0041] The purpose of obtaining the target defect root cause of the defective power equipment in the above steps is to better repair the power equipment. Even if a defect in the power equipment is predicted, the cause of the defect in the power equipment, i.e., the target defect root cause, is unclear. Since the power equipment in the power grid is widely distributed, different defect root causes require different maintenance personnel, tools, and equipment, which makes it difficult to repair the power equipment. Therefore, obtaining the target defect root cause of the power equipment can not only save manpower and material resources, but also greatly speed up the maintenance efficiency and reduce the maintenance time, effectively ensuring the normal operation of the power grid.

[0042] In order to better implement the above method, a data-driven intelligent identification system for power equipment defects is also proposed. The system includes an anomaly analysis module, a parameter impact analysis module, an equipment defect identification module, and a defect root cause analysis module.

[0043] The abnormality analysis module is used to obtain the historical defect records of the power equipment, obtain the defect judgment data of the power equipment, analyze the state parameters in the defect judgment data to judge the abnormality degree of the defect of the power equipment, and obtain parameter abnormality data;

[0044] The parameter impact analysis module is used to obtain the historical equipment operation records of the power equipment, analyze the operating parameters in the equipment operation records, and determine the parameter impact degree of the abnormal state parameters in the parameter abnormality data to obtain parameter impact data;

[0045] The equipment defect identification module is used to monitor the power equipment in the current cycle, predict the change trend of the status parameters of the power equipment in the current cycle, obtain parameter prediction data, and combine the defect identification data to identify the equipment defects of the power equipment and obtain defective power equipment;

[0046] The defect root cause analysis module is used to analyze the defect root cause of the defective power equipment based on the characteristic defect root cause data, obtain the defect root cause data, send the defect root cause data of the defective power equipment to the staff through the power grid platform, and prompt the staff to repair the defective power equipment.

[0047] Furthermore, the anomaly analysis module includes an anomaly recording analysis unit and an anomaly analysis unit;

[0048] The record abnormality analysis unit is used to obtain each historical defect record of the power equipment, and analyze the record abnormalities of each historical defect record in combination with the defect discrimination data of the power equipment to obtain abnormal historical defect records;

[0049] The abnormality analysis unit is used to analyze the abnormal state of the power equipment based on the abnormal historical defect records and the various state parameters in the defect judgment data to obtain the parameter abnormality data.

[0050] Furthermore, the parameter impact analysis module includes a parameter set acquisition unit and a parameter impact analysis unit;

[0051] A parameter set acquisition unit is used to acquire historical equipment operation records of the power equipment and obtain parameter sets of various operating parameters from the historical equipment operation records;

[0052] The parameter impact analysis unit is used to obtain the parameter set of abnormal state parameters in the historical equipment operation record, and analyze the parameter impact degree of the operating parameters on the abnormal state parameters based on the parameter set of each operating parameter in the historical equipment operation record of the power equipment to obtain parameter impact data.

[0053] Furthermore, the equipment defect identification module includes a model building unit and an equipment defect identification unit;

[0054] A model building unit, configured to build a data prediction model for the power equipment based on various historical operation records of the power equipment;

[0055] The equipment defect identification unit is used to use the data prediction model to predict various state parameters of the power equipment in the current cycle, and to identify the equipment defects of the power equipment in combination with the defect identification data of the power equipment to obtain defective power equipment.

[0056] Further, the defect root cause analysis module includes a defect root cause analysis unit;

[0057] The defect root cause analysis unit is used to obtain the characteristic defect root cause data of the defective power equipment, and combine it with the parameter prediction data to analyze the defect root cause of the defective power equipment in the current cycle, and send it to the staff of the power grid platform to prompt the staff to repair the defective power equipment.

[0058] Compared with the prior art, the beneficial effects of the present invention are: the present invention realizes intelligent identification of defects in power equipment, and by obtaining historical defect records of inaccurate defect identification of power equipment, and combining defect identification data used for defect identification of power equipment, the state parameters of inaccurate identification of power equipment, that is, abnormal state parameters, are found, and the operating parameters that cause the abnormal state parameters to inaccurately identify defects in power equipment are analyzed, thereby optimizing and constructing a prediction model for the state parameters, improving the accuracy of model prediction and power equipment defect identification, and also analyzing the root cause of the defect in the power equipment, obtaining the root cause of the defect, and sending the root cause of the defect to the staff, which greatly improves the efficiency of power equipment maintenance and enables the safe and stable operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 This is a method flow chart of a data-driven intelligent identification method for power equipment defects according to the present invention;

[0060] Figure 2 This is a module diagram of a data-driven intelligent identification system for power equipment defects according to the present invention. DETAILED DESCRIPTION

[0061] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0062] Example: Figure 1-Figure 2 As shown, the present invention provides a technical solution, a data-driven intelligent identification method for power equipment defects, the method comprising:

[0063] Step S100: Obtain historical defect records of power equipment from the power grid platform, obtain defect identification data of the power equipment, analyze the state parameters in the defect identification data to determine the abnormality of the power equipment defect, and obtain parameter abnormality data;

[0064] Wherein, step S100 includes:

[0065] Step S101: Acquire various historical defect records of the power equipment, and obtain data of various status parameters of the power equipment from the historical defect records, wherein all the power equipment in the historical defect records has equipment defects;

[0066] For example, various state parameters include voltage, current, and harmonic content;

[0067] Step S102: Obtaining defect identification data of the power equipment from the power grid platform, where the defect identification data is the range of various state parameters when performing defect identification on the power equipment;

[0068] When the maximum and minimum values ​​of various state parameters in a certain historical defect record are both within the range of the defect discrimination data, it is determined that the defect discrimination data has incorrectly discriminated the defect of the certain historical defect record, and the certain historical defect record is marked and recorded as an abnormal historical defect record;

[0069] Get the total number C of several unmarked historical defect records in the power equipment sum , obtaining several status parameters whose maximum or minimum values ​​in the unmarked historical defect records are outside the range of the defect discrimination data, and aggregating them to obtain a feature parameter set of the unmarked historical defect records;

[0070] Step S103: Calculate the defect judgment ratio of each state parameter, where the defect judgment ratio L of the ath state parameter in the power equipment is a =C (a,sum) / C sum , C (a,sum) is the total number of unmarked historical defect records containing the a-th state parameter in the feature parameter set;

[0071] Calculate the average value L' and standard deviation σ' of the defect discrimination ratio of each state parameter, and obtain the lower limit ratio threshold B of each state parameter min=L´-k×σ´, where k is a preset coefficient. When the defect discrimination ratio of a certain state parameter is less than the lower limit ratio threshold, the state parameter is marked and recorded as a suspected abnormal state parameter;

[0072] Step S104: Obtain the median D of the suspected abnormal state parameters in the defect identification data within the corresponding range △ , respectively calculate the maximum value, minimum value and median D of the suspected abnormal state parameters in the abnormal historical defect records of the power equipment △ The absolute value of the difference, and take the maximum absolute value, divide the maximum value by the median D △ , obtain the parameter abnormal value e of the suspected abnormal state parameter in the abnormal historical defect record;

[0073] When the parameter abnormal value e is greater than a preset threshold value e´, the abnormal historical defect record is recorded as a target abnormal historical defect record of the suspected abnormal state parameter, and the ratio of the total number of target abnormal historical defect records of the suspected abnormal state parameter to the total number of abnormal historical defect records of each abnormal state parameter in the power equipment is obtained to obtain the abnormal ratio of the suspected abnormal state parameter;

[0074] When the abnormality ratio is greater than a preset threshold, it is determined that the suspected abnormal state parameter in the defect discrimination data is abnormal when identifying defects in the power equipment, and the suspected abnormal state parameter is recorded as the abnormal state parameter;

[0075] Acquire and aggregate several abnormal state parameters of the power equipment to obtain abnormal parameter data of the power equipment;

[0076] Step S200: obtaining historical equipment operation records of the power equipment, analyzing the operating parameters in the historical equipment operation records, and determining the parameter impact degree of the abnormal state parameters in the parameter abnormality data to obtain parameter impact data;

[0077] Wherein, step S200 includes:

[0078] Step S201: Acquire each historical device operation record of the power equipment, pre-process the data in each historical device operation record, set a unit time, and acquire the average value of each operating parameter in the historical device operation record every unit time to obtain a parameter set of each operating parameter;

[0079] For example, various operating parameters include temperature, noise, humidity, etc.

[0080] For example, the preprocessing of each historical equipment operation record includes cleaning the data, using interpolation (linear / spline) to handle missing values, and finally normalizing or standardizing;

[0081] Step S202: Acquire abnormal parameter data of the power equipment, and obtain various abnormal state parameters from the abnormal parameter data;

[0082] Obtain the average value of each abnormal state parameter from the historical equipment operation record every unit time to obtain a parameter set of each abnormal state parameter;

[0083] Calculate the parameter influence degree between each operating parameter and each abnormal state parameter in the historical equipment operation record, where the parameter influence degree S between the fth operating parameter and the gth abnormal state parameter in the historical equipment operation record is (f,g) ;

[0084] For example, the parameter influence S (f,g) The specific calculation formula is:

[0085] ,

[0086] Among them, x （f,i） is the value of the i-th element in the parameter set of the f-th operating parameter in the historical equipment operation record; y （g,i） is the value of the i-th element in the parameter set of the g-th abnormal status parameter in the historical equipment operation record; x´ f is the average value of each element in the parameter set of the fth operating parameter in the historical equipment operation record; g is the average value of each element in the parameter set of the g-th abnormal status parameter in the historical equipment operation record;

[0087] Step S203: Calculate the data impact value R of the fth operating parameter on the gth abnormal state parameter (f,g) :

[0088] ,

[0089] Where z is the total number of historical equipment operation records of power equipment; S z (f,g) is the parameter influence degree between the fth operating parameter and the gth abnormal state parameter in the zth historical equipment operation record of the power equipment;

[0090] Step S204: Setting the impact threshold R, when |R (f,g) |≥R, then it is determined that the fth operating parameter has an impact on the data of the gth abnormal state parameter, and the fth operating parameter is recorded as the influencing operating parameter of the power equipment. The various influencing operating parameters of the power equipment are obtained and aggregated to obtain parameter impact data;

[0091] Step S300: Acquire parameter impact data, monitor the power equipment in the current cycle, predict the change trend of the power equipment status parameters in the current cycle, obtain parameter prediction data, and combine the defect identification data to identify the power equipment defects and obtain defective power equipment;

[0092] Wherein, step S300 includes:

[0093] Step S301: Acquire parameter impact data, acquire various parameters affecting the operation of the power equipment, acquire various historical operation records of the power equipment, pre-process the data in each historical operation record, and sort each historical operation record in chronological order;

[0094] Build a data prediction model for power equipment and divide historical operation records into training sets and test sets according to preset ratios;

[0095] For example, the data prediction model for power equipment can be constructed using linear models such as linear regression models, time series models such as ARIMA and exponential smoothing models, machine learning models such as decision trees, random forests, and vector machines, and deep learning models such as LSTM models.

[0096] Obtain input data and output data of a model data set in the model, where the input data is various state parameters and various influencing operation parameters in a certain historical operation record, and the output data is various state parameters in the next historical operation record of a certain historical operation record;

[0097] For example, the input data in the training set is the state parameters and the parameter sets that affect the operation parameters in a certain historical operation record, and the output data is the parameter sets of the state parameters in the next historical operation record of a certain historical operation record;

[0098] Step S302: Use the training set to train the data prediction model, and use the test set to calculate the model prediction accuracy of the data prediction model. When the model prediction accuracy is greater than a preset threshold, it is determined that the data prediction model is constructed.

[0099] For example, when the data prediction model is a linear regression model, the model prediction accuracy can be the mean square error, root mean square error, and determination coefficient R 2 ;

[0100] Among them, when the model prediction accuracy is the mean square error MSE, the specific formula is:

[0101] ,

[0102] Where n is the total number of model data sets in the test set; y iis the true value of the average value of each state parameter after normalization in the i-th model data set in the test set; i is the predicted value of the average value of each state parameter after normalization in the i-th model data set in the test set;

[0103] Step S303: monitoring the power equipment in the current cycle, collecting various state parameters of the power equipment and values ​​of various operating parameters affecting the power equipment at unit time intervals, and inputting these into a data prediction model to predict the various state parameters of the power equipment to obtain parameter prediction data, which includes a parameter set for predicting various state parameters;

[0104] When the maximum or minimum value of the state parameter in the parameter prediction data is outside the range of the state parameter in the defect judgment data, it is determined that the power equipment has a risk of equipment defect in the current cycle, and the state parameter and the power equipment are recorded as defective state parameter and defective power equipment respectively;

[0105] Step S400: Acquire characteristic defect root cause data of the defective power equipment, acquire parameter prediction data, analyze the defect root cause of the defective power equipment, obtain defect root cause data, send the defect root cause data of the defective power equipment to staff via the power grid platform, and prompt the staff to repair the defective power equipment;

[0106] Wherein, step S400 includes:

[0107] Step S401: Obtain defective power equipment in the power grid, obtain characteristic defect root cause data of the defective power equipment from the power grid platform, obtain root cause identification data for each defect root cause of the defective power equipment from the characteristic defect root cause data, obtain state parameters required for identifying a particular defect root cause from the characteristic defect root cause data, and record them as target state parameters for the particular defect root cause. The root cause identification data for a particular defect root cause is the data range within which each target state parameter for the particular defect root cause falls when the particular defect root cause occurs in the defective power equipment.

[0108] Step S402: Obtain a dataset of predicted defect state parameters of the defective power equipment. When a certain defect state parameter is the target state parameter of a certain defect root cause, and the value of an element in the dataset of the predicted defect state parameter is within the data range of the root cause identification data, it is recorded as a parameter match of the certain defect root cause. The total number of parameter matches of the certain defect root cause, M, is obtained. sum , obtain the total number of target state parameters of a certain defect root cause M sum ;

[0109] Calculate the probability of occurrence of a certain defect root cause P=M´sum / M sum When the occurrence probability P is greater than the preset probability threshold, it is determined that the defective power equipment has a certain defect root cause in the current cycle, and the certain defect root cause is recorded as the target defect root cause;

[0110] Step S403: Obtain and aggregate the root causes of defects of each target item of defective power equipment in the current cycle to obtain defect root cause data, send the defect root cause data to power grid staff through the power grid platform, and prompt the staff to dispatch personnel to repair the defective power equipment.

[0111] In order to better implement the above method, a data-driven intelligent identification system for power equipment defects is also proposed. The system includes an anomaly analysis module, a parameter impact analysis module, an equipment defect identification module, and a defect root cause analysis module.

[0112] The abnormality analysis module is used to obtain the historical defect records of the power equipment, obtain the defect judgment data of the power equipment, analyze the state parameters in the defect judgment data to judge the abnormality degree of the defect of the power equipment, and obtain parameter abnormality data;

[0113] The parameter impact analysis module is used to obtain the historical equipment operation records of the power equipment, analyze the operating parameters in the equipment operation records, and determine the parameter impact degree of the abnormal state parameters in the parameter abnormality data to obtain parameter impact data;

[0114] The equipment defect identification module is used to monitor the power equipment in the current cycle, predict the change trend of the status parameters of the power equipment in the current cycle, obtain parameter prediction data, and combine the defect identification data to identify the equipment defects of the power equipment and obtain defective power equipment;

[0115] The defect root cause analysis module is used to analyze the defect root cause of the defective power equipment based on the characteristic defect root cause data, obtain the defect root cause data, send the defect root cause data of the defective power equipment to the staff through the power grid platform, and prompt the staff to repair the defective power equipment.

[0116] Among them, the abnormality analysis module includes a recording abnormality analysis unit and an abnormality analysis unit;

[0117] The record abnormality analysis unit is used to obtain each historical defect record of the power equipment, and analyze the record abnormalities of each historical defect record in combination with the defect discrimination data of the power equipment to obtain abnormal historical defect records;

[0118] The abnormality analysis unit is used to analyze the abnormal state of the power equipment based on the abnormal historical defect records and the various state parameters in the defect judgment data to obtain the parameter abnormality data.

[0119] Among them, the parameter impact analysis module includes a parameter set acquisition unit and a parameter impact analysis unit;

[0120] A parameter set acquisition unit is used to acquire historical equipment operation records of the power equipment and obtain parameter sets of various operating parameters from the historical equipment operation records;

[0121] The parameter impact analysis unit is used to obtain the parameter set of abnormal state parameters in the historical equipment operation record, and analyze the parameter impact degree of the operating parameters on the abnormal state parameters based on the parameter set of each operating parameter in the historical equipment operation record of the power equipment to obtain parameter impact data.

[0122] Among them, the equipment defect identification module includes a model building unit and an equipment defect identification unit;

[0123] A model building unit, configured to build a data prediction model for the power equipment based on various historical operation records of the power equipment;

[0124] The equipment defect identification unit is used to use the data prediction model to predict various state parameters of the power equipment in the current cycle, and to identify the equipment defects of the power equipment in combination with the defect identification data of the power equipment to obtain defective power equipment.

[0125] Wherein, the defect root cause analysis module includes a defect root cause analysis unit;

[0126] The defect root cause analysis unit is used to obtain the characteristic defect root cause data of the defective power equipment, and combine it with the parameter prediction data to analyze the defect root cause of the defective power equipment in the current cycle, and send it to the staff of the power grid platform to prompt the staff to repair the defective power equipment.

[0127] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A data-driven intelligent identification method for power equipment defects, characterized in that: The method comprises: Step S100: Obtain historical defect records of power equipment from a power grid platform, obtain defect identification data of the power equipment, analyze the state parameters in the defect identification data to determine the abnormality of the defect of the power equipment, and obtain parameter abnormality data; Step S200: obtaining historical equipment operation records of the power equipment, analyzing the operating parameters in the historical equipment operation records, and determining the degree of parameter influence on the abnormal state parameters in the parameter abnormality data to obtain parameter influence data; Step S300: Acquire the parameter impact data, monitor the power equipment in the current cycle, predict the change trend of the state parameters of the power equipment in the current cycle, obtain parameter prediction data, and combine the defect judgment data to judge the power equipment for equipment defects and obtain defective power equipment; Step S400: Acquire characteristic defect root cause data of the defective power equipment, acquire the parameter prediction data, analyze the defect root cause of the defective power equipment, obtain defect root cause data, send the defect root cause data of the defective power equipment to a staff member via the power grid platform, and prompt the staff member to repair the defective power equipment; The step S100 includes: Step S101: Acquire various historical defect records of the power equipment, and obtain data of various status parameters of the power equipment from the historical defect records, wherein the power equipment in the historical defect records all has equipment defects; Step S102: acquiring defect determination data of the power equipment from the power grid platform, wherein the defect determination data is the range of various state parameters when performing defect determination on the power equipment; When the maximum and minimum values ​​of various state parameters in a certain historical defect record are both within the range of the defect discrimination data, it is determined that the defect discrimination data has incorrectly discriminated the defect of the certain historical defect record, and the certain historical defect record is marked and recorded as an abnormal historical defect record; Obtain the total number C of several unmarked historical defect records in the power equipment sum , obtaining several status parameters whose maximum or minimum values ​​in the unmarked historical defect records are outside the range of the defect discrimination data, and aggregating them to obtain a feature parameter set of the unmarked historical defect records; Step S103: Calculate the defect judgment ratio of each state parameter, wherein the defect judgment ratio L of the ath state parameter in the power equipment is a =C (a,sum) / C sum , C (a,sum) is the total number of unmarked historical defect records containing the a-th state parameter in the feature parameter set; Calculate the average value L' and standard deviation σ' of the defect discrimination ratio of each state parameter, and obtain the lower limit ratio threshold B of each state parameter min =L´-k×σ´, where k is a preset coefficient. When the defect discrimination ratio of a certain state parameter is less than the lower limit ratio threshold, the state parameter is marked and recorded as a suspected abnormal state parameter; Step S104: Obtain the median D of the suspected abnormal state parameter in the defect identification data within the corresponding range △ , respectively calculate the maximum and minimum values ​​of the suspected abnormal state parameters in the abnormal historical defect records of the power equipment and the median D △ The absolute value of the difference, and take the maximum of the absolute values, and divide the maximum by the median D △ , obtaining a parameter abnormal value e of the suspected abnormal state parameter in the abnormal historical defect record; When the parameter abnormal value e is greater than a preset threshold value e′, the abnormal historical defect record is recorded as a target abnormal historical defect record of the suspected abnormal state parameter, and the ratio of the total number of target abnormal historical defect records of the suspected abnormal state parameter to the total number of abnormal historical defect records of each abnormal state parameter in the power equipment is obtained to obtain an abnormal ratio of the suspected abnormal state parameter; When the abnormality ratio is greater than a preset threshold, it is determined that the suspected abnormal state parameter in the defect discrimination data is abnormal when performing defect identification on the power equipment, and the suspected abnormal state parameter is recorded as an abnormal state parameter; A plurality of abnormal state parameters of the electric power equipment are acquired and collected to obtain parameter abnormality data of the electric power equipment.

2. The data-driven intelligent identification method for power equipment defects according to claim 1 is characterized in that: The step S200 includes: Step S201: Acquire each historical device operation record of the power equipment, pre-process the data in each historical device operation record, set a unit time, and acquire the average value of each operating parameter in the historical device operation record every unit time to obtain a parameter set of each operating parameter; Step S202: Acquire parameter abnormality data of the power equipment, and obtain various abnormal state parameters from the parameter abnormality data; Obtaining an average value of each abnormal state parameter from the historical device operation record every unit time to obtain a parameter set of each abnormal state parameter; Calculate the parameter influence degree between each operating parameter in the historical equipment operation record and each abnormal state parameter, wherein the parameter influence degree S between the fth operating parameter and the gth abnormal state parameter in the historical equipment operation record is (f,g) ; Step S203: Calculate the data impact value R of the fth operating parameter on the gth abnormal state parameter (f,g) : ， Wherein, Z is the total number of historical equipment operation records of the power equipment; S z (f,g) is the parameter influence degree between the f-th operating parameter and the g-th abnormal state parameter in the z-th historical equipment operation record of the power equipment; Step S204: Setting the impact threshold R, when |R (f,g) |≥R, it is determined that the f-th operating parameter has an impact on the data of the g-th abnormal state parameter, the f-th operating parameter is recorded as the influencing operating parameter of the power equipment, and the various influencing operating parameters of the power equipment are obtained and aggregated to obtain parameter impact data.

3. The data-driven intelligent identification method for power equipment defects according to claim 2 is characterized in that: The step S300 includes: Step S301: Acquire the parameter impact data, acquire various parameters affecting the operation of the power equipment, acquire various historical operation records of the power equipment, pre-process the data in the various historical operation records, and sort the various historical operation records in chronological order; Constructing a data prediction model for the power equipment, and dividing the historical operation records into a training set and a test set according to a preset ratio; Obtaining input data and output data of a model data set in the model, wherein the input data is various state parameters and various influencing operation parameters in a certain historical operation record, and the output data is various state parameters in a historical operation record next to the certain historical operation record; Step S302: using the training set to train the data prediction model, and using the test set to calculate the model prediction accuracy of the data prediction model, when the model prediction accuracy is greater than a preset threshold, it is determined that the data prediction model is constructed; Step S303: monitoring the power equipment in the current cycle, collecting various state parameters of the power equipment and values ​​of various operating parameters affecting the power equipment at unit time intervals, and inputting the collected data into the data prediction model to predict the various state parameters of the power equipment to obtain parameter prediction data, the parameter prediction data including a parameter set for predicting the various state parameters; When the maximum or minimum value of the state parameter exists in the parameter prediction data and is not within the range of the state parameter in the defect judgment data, it is determined that the power equipment has an equipment defect risk in the current cycle, and the state parameter and the power equipment are recorded as defective state parameters and defective power equipment, respectively.

4. The data-driven intelligent identification method for power equipment defects according to claim 3 is characterized in that: The step S400 includes: Step S401: Obtain defective power equipment in the power grid, obtain characteristic defect root cause data of the defective power equipment from the power grid platform, obtain root cause identification data of various defect root causes of the defective power equipment from the characteristic defect root cause data, obtain state parameters required for identification of a certain defect root cause from the characteristic defect root cause data, and record them as target state parameters of the certain defect root cause. The root cause identification data of the certain defect root cause is the data range within which the target state parameters of the certain defect root cause fall when the defective power equipment has the certain defect root cause. Step S402: Obtain a dataset of predicted defect state parameters of the defective power equipment. When a certain defect state parameter is the target state parameter of a certain defect root cause, and the value of an element in the dataset of the predicted defect state parameter is within the data range of the root cause identification data, it is recorded as a parameter match of the certain defect root cause. The total number of parameter matches of the certain defect root cause is obtained. sum , obtain the total number M of target state parameters of the defect root cause sum ; Calculate the probability of occurrence of a certain defect root cause P=M´ sum / M sum When the occurrence probability P is greater than a preset probability threshold, it is determined that the defective power equipment has the certain defect root cause in the current cycle, and the certain defect root cause is recorded as a target defect root cause; Step S403: Obtain and aggregate the root causes of each target defect of the defective power equipment in the current cycle to obtain defect root cause data, send the defect root cause data to the power grid staff through the power grid platform, and prompt the staff to dispatch personnel to repair the defective power equipment.

5. A data-driven intelligent identification system for power equipment defects, used to execute a data-driven intelligent identification method for power equipment defects according to any one of claims 1 to 4, characterized in that: The system includes an anomaly analysis module, a parameter impact analysis module, an equipment defect identification module, and a defect root cause analysis module; The abnormality analysis module is used to obtain historical defect records of the power equipment, obtain defect judgment data of the power equipment, analyze the state parameters in the defect judgment data to determine the abnormality degree of the defect judgment of the power equipment, and obtain parameter abnormality data; The parameter impact analysis module is used to obtain historical equipment operation records of the power equipment, analyze the operating parameters in the historical equipment operation records, and determine the degree of parameter influence on the abnormal state parameters in the parameter abnormality data to obtain parameter impact data; The equipment defect identification module is used to monitor the power equipment in the current cycle, predict the change trend of the state parameters of the power equipment in the current cycle, obtain parameter prediction data, and combine the defect identification data to perform equipment defect identification on the power equipment to obtain defective power equipment; The defect root cause analysis module is used to analyze the defect root cause of the defective power equipment based on the characteristic defect root cause data to obtain defect root cause data, send the defect root cause data of the defective power equipment to the staff through the power grid platform, and prompt the staff to repair the defective power equipment.

6. The data-driven intelligent identification system for power equipment defects according to claim 5 is characterized in that: The abnormality analysis module includes a recording abnormality analysis unit and an abnormality analysis unit; The record abnormality analysis unit is used to obtain each historical defect record of the power equipment, and analyze the record abnormalities of each historical defect record in combination with the defect discrimination data of the power equipment to obtain abnormal historical defect records; The abnormality analysis unit is used to analyze each state parameter in the defect discrimination data according to the abnormal historical defect record to perform defect discrimination on the abnormal state of the power equipment and obtain parameter abnormality data.

7. The data-driven intelligent identification system for power equipment defects according to claim 5 is characterized in that: The parameter impact analysis module includes a parameter set acquisition unit and a parameter impact analysis unit; The parameter set acquisition unit is configured to acquire historical equipment operation records of the power equipment, and acquire parameter sets of various operating parameters from the historical equipment operation records; The parameter impact analysis unit is used to obtain the parameter set of the abnormal state parameters in the historical equipment operation record, and analyze the degree of parameter influence of the operating parameters on the abnormal state parameters based on the parameter set of each operating parameter in the historical equipment operation record of the power equipment to obtain parameter impact data.

8. The data-driven intelligent identification system for power equipment defects according to claim 5 is characterized in that: The equipment defect identification module includes a model building unit and an equipment defect identification unit; The model building unit is used to build a data prediction model for the power equipment based on various historical operation records of the power equipment; The equipment defect judgment unit is used to use the data prediction model to predict various state parameters of the power equipment in the current cycle, and to judge the equipment defects of the power equipment in combination with the defect judgment data of the power equipment to obtain defective power equipment.

9. The data-driven intelligent identification system for power equipment defects according to claim 5 is characterized in that: The defect root cause analysis module includes a defect root cause analysis unit; The defect root cause analysis unit is used to obtain characteristic defect root cause data of the defective power equipment, and combine it with the parameter prediction data to analyze the defect root cause of the defective power equipment in the current cycle, and send it to the staff of the power grid platform to prompt the staff to repair the defective power equipment.

Images