CN118094168A - Correction method and device for electric power data - Google Patents
Correction method and device for electric power data Download PDFInfo
- Publication number
- CN118094168A CN118094168A CN202410508752.2A CN202410508752A CN118094168A CN 118094168 A CN118094168 A CN 118094168A CN 202410508752 A CN202410508752 A CN 202410508752A CN 118094168 A CN118094168 A CN 118094168A
- Authority
- CN
- China
- Prior art keywords
- data
- abnormal
- power
- power data
- preset
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000012937 correction Methods 0.000 title claims abstract description 145
- 238000000034 method Methods 0.000 title claims abstract description 115
- 230000002159 abnormal effect Effects 0.000 claims abstract description 449
- 238000012545 processing Methods 0.000 claims abstract description 47
- 230000008859 change Effects 0.000 claims abstract description 43
- 238000012300 Sequence Analysis Methods 0.000 claims abstract description 23
- 238000012360 testing method Methods 0.000 claims abstract description 19
- 238000011156 evaluation Methods 0.000 claims description 55
- 238000005516 engineering process Methods 0.000 claims description 31
- 238000004140 cleaning Methods 0.000 claims description 30
- 238000012544 monitoring process Methods 0.000 claims description 29
- 230000001932 seasonal effect Effects 0.000 claims description 28
- 238000001514 detection method Methods 0.000 claims description 27
- 238000004458 analytical method Methods 0.000 claims description 24
- 230000007774 longterm Effects 0.000 claims description 16
- 230000008439 repair process Effects 0.000 claims description 16
- 238000012549 training Methods 0.000 claims description 14
- 238000007405 data analysis Methods 0.000 claims description 13
- 230000004927 fusion Effects 0.000 claims description 12
- 238000005457 optimization Methods 0.000 claims description 11
- 238000007781 pre-processing Methods 0.000 claims description 11
- 238000010276 construction Methods 0.000 claims description 10
- 238000006243 chemical reaction Methods 0.000 claims description 9
- 230000005477 standard model Effects 0.000 claims description 9
- 238000012731 temporal analysis Methods 0.000 claims description 9
- 238000000700 time series analysis Methods 0.000 claims description 9
- 238000000354 decomposition reaction Methods 0.000 claims description 8
- 238000000605 extraction Methods 0.000 claims description 8
- 230000010354 integration Effects 0.000 claims description 8
- 238000010801 machine learning Methods 0.000 claims description 8
- 238000003672 processing method Methods 0.000 claims description 7
- 238000010586 diagram Methods 0.000 claims description 6
- 230000005856 abnormality Effects 0.000 claims description 5
- 238000005259 measurement Methods 0.000 claims description 5
- 238000013441 quality evaluation Methods 0.000 claims description 5
- 125000004122 cyclic group Chemical group 0.000 claims description 4
- 238000003066 decision tree Methods 0.000 claims description 4
- 238000009499 grossing Methods 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 4
- 230000000737 periodic effect Effects 0.000 claims description 4
- 230000011218 segmentation Effects 0.000 claims description 4
- 238000007619 statistical method Methods 0.000 claims description 4
- 238000013179 statistical model Methods 0.000 claims description 4
- 238000002372 labelling Methods 0.000 abstract description 10
- 230000008569 process Effects 0.000 description 7
- 238000007726 management method Methods 0.000 description 5
- 230000001788 irregular Effects 0.000 description 4
- 238000013075 data extraction Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000013480 data collection Methods 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000007499 fusion processing Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000013439 planning Methods 0.000 description 1
- 238000003908 quality control method Methods 0.000 description 1
- 238000010223 real-time analysis Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Landscapes
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a correction method and device of power data, and relates to the field of power data processing. According to the method, the operation state of the equipment is monitored, the obtained initial abnormal power data is based on actual equipment operation data, the first more accurate abnormal power data is identified from multiple angles through the combination of the abnormal data characteristics and a time sequence analysis method, and finally, the accurate correction of the abnormal data is completed through deviation prediction of the abnormal data based on deviation information and trend change information of the abnormal data. By combining the characteristics of the abnormal data and the time sequence analysis method, accurate abnormal data is obtained, accurate correction of the abnormal data is realized based on deviation information and trend change information of the abnormal data, the correction of the abnormal data is ensured to conform to the actual rules of operation of the power system and the principle of power engineering, data quality guarantee is provided for test labeling of the power data, and comprehensive management of the power system is promoted.
Description
Technical Field
The present invention relates to the field of power data processing, and in particular, to a method and apparatus for correcting power data.
Background
With the increasing size of power systems, related power data in the power systems are increasingly complex, and the trial marking of the power data is an important pre-step in the processing, analysis and utilization of the power data. However, the quality of the data generated during the test labeling process is often affected by many factors, such as instrument faults, communication interruption, and human input errors, which may cause irregular data and abnormal samples, and the correct identification and correction of these abnormal values is a key to maintaining the quality of the labeling result of the power data.
The normal operation data of the power system shows the characteristic of regularity, and the irregular data may suggest that the system has a certain problem or errors are generated in the data collection process, so that the power data needs to be correctly identified; and the correction of the identified abnormal value is more in line with the actual operation rule of the power system. However, in the current correction of the power data, problems such as incapability of accurately identifying abnormal values and system fluctuation values in the power data occur, so that the correction effect of the abnormal values cannot reach the standard of normal operation required by the power system. Thus, there is a need for a new method or apparatus for modifying power data.
Disclosure of Invention
The invention provides a method and a device for correcting electric power data, which are used for solving the technical problem that the electric power data cannot be accurately corrected in the prior art.
In order to solve the above technical problems, an embodiment of the present invention provides a method for correcting electric power data, including:
collecting first power data corresponding to a plurality of power devices, and monitoring the running state of each power device according to a preset state monitoring method to obtain a plurality of abnormal power devices and corresponding initial abnormal power data;
Extracting features of the initial abnormal power data according to a preset abnormal data analysis method to obtain abnormal data features corresponding to the initial abnormal power data;
According to a preset time sequence analysis method, combining preset historical power data and abnormal data characteristics corresponding to initial abnormal power data, and identifying the initial abnormal power data to obtain first abnormal power data;
According to a preset data deviation prediction method, performing deviation prediction on the first abnormal power data to obtain deviation information and trend change information corresponding to the first abnormal power data;
And inputting the first abnormal power data, the corresponding deviation information and the corresponding trend change information into a preset abnormal value correction model so that the abnormal value correction model can correct the first abnormal power data to obtain corrected power data.
It can be understood that, compared with the prior art, the method and the device have the advantages that the obtained initial abnormal power data is based on the actual equipment operation data through monitoring the equipment operation state, the more accurate first abnormal power data is identified from multiple angles through the combination of the abnormal data characteristics and a time sequence analysis method, and finally, the accurate correction of the abnormal data is completed through deviation prediction of the abnormal data based on deviation information and trend change information of the abnormal data. By combining the characteristics of the abnormal data and the time sequence analysis method, accurate abnormal data is obtained, accurate correction of the abnormal data is realized on the basis of deviation information and trend change information of the abnormal data, the correction of the abnormal data is ensured to conform to the actual rules of operation of the power system and the principle of power engineering, and therefore, data quality guarantee is provided for test labeling of the power data, and comprehensive management of the power system is promoted.
As a preferred solution, the collecting first power data corresponding to a plurality of power devices, and monitoring an operation state of each power device according to a preset state monitoring method to obtain a plurality of abnormal power devices and corresponding initial abnormal power data, including:
acquiring initial power data corresponding to a plurality of power devices, and performing data preprocessing on the initial power data according to a preset data preprocessing method to obtain first power data;
Acquiring operation data corresponding to the plurality of electric powers according to a preset state monitoring method to obtain an initial state data set;
and processing the initial state data set according to a preset state data processing method to obtain a plurality of abnormal power devices and corresponding initial abnormal power data.
By collecting initial power data and extracting abnormal data according to the state of the equipment, the accuracy of the abnormal data is ensured.
As a preferred solution, the processing the initial state data set according to a preset state data processing method to obtain a plurality of abnormal power devices and corresponding initial abnormal power data includes:
According to a preset signal processing technology, signal processing is carried out on the initial state data set, and key characteristic signals of equipment operation in the initial state data set are obtained;
Acquiring preset historical characteristic signal data and a standard model, and performing cross comparison on the key characteristic signals of the operation of the equipment in the initial state data set, the preset historical characteristic signal data and the standard model according to a preset state evaluation algorithm to obtain a state evaluation result corresponding to the initial state data set;
And extracting the power data corresponding to the power devices in the abnormal state according to the state evaluation result corresponding to the initial state data set to obtain a plurality of abnormal power devices and corresponding initial abnormal power data.
And finally, the abnormal power data is extracted according to the state evaluation result, so that the diversity and accuracy of the abnormal power data are improved.
Preferably, the obtaining a state evaluation result corresponding to the initial state data set further includes:
When the state evaluation result shows that the state of the equipment is abnormal, starting a fault diagnosis program corresponding to the abnormal equipment, judging the fault type and the position of the abnormal equipment according to a preset machine learning algorithm and an analysis data mode, and obtaining the fault type and the fault position of the abnormal equipment;
Inputting operation data corresponding to the abnormal equipment into a preset state prediction model, performing prediction analysis on the operation trend of the abnormal equipment, and generating early warning information when an operation problem is predicted;
And respectively inputting the fault type, the fault position and the early warning information of the abnormal equipment into a preset fault tree analysis model and a repair system to obtain a repair scheme corresponding to the abnormal equipment, so that a worker can repair the abnormal equipment.
By judging the abnormal state of the equipment, the fault type and the fault position are obtained, and early warning information is generated, so that workers can quickly repair the abnormal equipment, and the stable operation of the equipment is ensured.
As a preferred solution, the feature extraction of the initial abnormal power data according to a preset abnormal data analysis method, to obtain an abnormal data feature corresponding to the initial abnormal power data, includes:
according to a preset frequency data acquisition technology, carrying out corresponding frequency data acquisition on initial abnormal power data to obtain frequency data of a plurality of data types, wherein the frequency data of the plurality of data types comprises: time series data and device log data;
performing data fusion on the frequency data of the data types based on a preset data fusion technology to obtain a device state view, wherein the data fusion comprises data cleaning, data conversion and mapping and data integration;
and detecting the deviated data characteristics in the equipment state view according to a preset abnormality detection algorithm to obtain abnormal data characteristics corresponding to the initial abnormal power data.
By integrating the data into one view, more comprehensive and comprehensive equipment state information can be provided, so that the extraction of abnormal data features is more accurate.
As a preferred solution, the identifying, according to a preset time sequence analysis method, the initial abnormal power data according to the preset historical power data and the abnormal data characteristics corresponding to the initial abnormal power data, to obtain first abnormal power data includes:
indexing initial time sequence data in the initial abnormal power data according to the time stamp, so that the initial time sequence data are arranged according to a preset time sequence, and constructing a time sequence diagram corresponding to the initial time sequence data;
Processing the time series data according to a preset time series analysis and prediction algorithm, and removing short-term fluctuation in the time series data to obtain first time series data, corresponding long-term trend data and seasonal pattern data, wherein the preset time series analysis and prediction algorithm comprises: a moving average algorithm or an exponential smoothing method;
respectively carrying out data analysis on the long-term trend data and the seasonal pattern data according to a trend decomposition algorithm and a seasonal decomposition algorithm to obtain trend components corresponding to the long-term trend data and seasonal components corresponding to the seasonal pattern data;
According to the periodic detection method, identifying and separating the cyclic fluctuation in the first time sequence data to obtain second time sequence data; carrying out abnormal value identification on the second time sequence data according to a box graph statistical method to obtain abnormal fluctuation data;
performing stability detection on the second time series data by using an ADF (automatic frequency correction) detection method, performing differential processing or conversion processing on the second time series data to obtain stable second time series data when the detection result is unstable, and fitting the stable second time series data according to a preset statistical model to obtain third time series data;
According to a preset detection algorithm, the third time series data is analyzed for prediction errors by combining a time series chart, trend components corresponding to long-term trend data, seasonal components corresponding to seasonal mode data and abnormal fluctuation data, and when the prediction errors exceed a preset error range, the corresponding data points are marked abnormally until all the data points in the third time series data complete the analysis of the prediction errors, so that first abnormal power data are obtained.
By the time sequence analysis method, system fluctuation and abnormal data can be identified, the identified abnormal data is ensured to be accurate and real, and abnormal value correction errors caused by confusing the system fluctuation into the abnormal data are avoided.
As a preferred solution, the performing deviation prediction on the first abnormal power data according to a preset data deviation prediction method to obtain deviation information and trend change information corresponding to the first abnormal power data includes:
Evaluating the data points in the first abnormal power data according to a weighted least square method, when the evaluation result of the data points exceeds a preset evaluation range, marking the corresponding data points until all the data points in the first abnormal power data are evaluated, and sorting all the marked data points to obtain corresponding second abnormal power data;
Inputting the second abnormal power data into a preset power system model so that the power system model can evaluate the influence degree of the second abnormal power data to obtain an influence degree evaluation result of the second abnormal power data;
According to the influence degree evaluation result, the state variable data of the first abnormal power data are combined, and the state estimation data in the first abnormal power data are subjected to pre-correction to obtain an improved state estimation value corresponding to the first abnormal power data;
Optimizing the improved state estimation value corresponding to the first abnormal power data according to a preset optimization criterion to obtain an optimized improved state estimation value;
And according to a preset trend analysis tool and historical power data, combining the optimized improved state estimation value, and judging deviation information and trend change information of the first abnormal power data by using a preset state estimation technology to obtain the deviation information and trend change information corresponding to the first abnormal power data.
By acquiring the abnormal data deviation information and the trend change information, a theoretical basis can be provided for correcting the abnormal value, and the correction accuracy of the abnormal value is improved.
As a preferred solution, the constructing a first outlier correction model according to a preset correction model construction method includes:
acquiring a preset initial abnormal value correction model and real-time power data in a power system, wherein the real-time power data in the power system comprises: generating capacity, load demand, line flow, etc.;
after noise removal and missing value processing are carried out on the real-time power data according to a preset data cleaning algorithm, the statistical features and the frequency domain features of the real-time power data subjected to data cleaning are extracted according to the preset operation features of the power system, and the statistical features and the frequency domain features of the real-time power data subjected to data cleaning are obtained;
according to the acquired meteorological data and the equipment state of the power equipment to which the real-time power data belongs, converting the statistical characteristics and the frequency domain characteristics of the real-time power data subjected to data cleaning to obtain input characteristics;
According to an isolated forest algorithm, analyzing the relevance of the attributes in the real-time power data to obtain abnormal values and abnormal reasons in the real-time power data, wherein the attributes in the real-time power data comprise: load change, topology and running log;
Inputting the input characteristics, the abnormal value in the real-time power data and the abnormal cause into a preset initial abnormal value correction model, combining the power data, the current and the voltage measurement characteristics in the real-time power data, taking the correction of the abnormal value as a training target of the initial abnormal value correction model, training the initial abnormal value correction model by using a decision tree machine learning algorithm, and generating a corresponding correction rule;
evaluating a correction rule generated by the initial abnormal value correction model according to a time sequence segmentation technology to obtain an evaluation result of the initial abnormal value correction model; and according to the evaluation result, adjusting the parameters of the initial outlier correction model by using a grid search parameter optimization method until the evaluation result of the initial outlier correction model meets a preset standard, and completing training of the initial outlier correction model to obtain a first outlier correction model.
The abnormal value correction model is trained and optimized through the training data set, so that the correction capability of the abnormal value correction model can be improved, and the effect of repairing abnormal values in the electric power data is ensured.
As a preferred embodiment, the method further comprises the following steps:
Carrying out data integration on the corrected power data obtained after the first abnormal value correction model corrects the first abnormal power data and the first power data to obtain second power data; according to an automatic data cleaning algorithm, performing data cleaning on the second power data to obtain third power data;
After the statistical characteristics in the third electric power data are extracted, judging the equilibrium degree of the statistical characteristics, and when the equilibrium degree is not in a preset equilibrium range, increasing data samples in the third electric power data according to a data increasing technology so that the equilibrium degree of the statistical characteristics in the third electric power data meets the preset requirement to obtain fourth electric power data;
Processing the fourth power data according to ApacheHadoop or APACHESPARK distributed computing frames, and then performing check point processing on the fourth power data according to a preset check point and a data backup strategy to obtain fifth power data;
and performing test marking on the fifth electric power data according to a preset marking method, and evaluating the test marking result based on a preset log record and a data quality evaluation method to obtain a test marking evaluation result.
The quality and accuracy of the trial marking process can be improved through the optimization processing of the corrected power data.
Correspondingly, the embodiment of the invention provides a device for correcting electric power data, which comprises the following components: the system comprises a power data acquisition module, a data characteristic acquisition module, a data identification module, a data prediction module and a data correction module;
The power data acquisition module is used for acquiring first power data corresponding to a plurality of power devices, and monitoring the running state of each power device according to a preset state monitoring method to obtain a plurality of abnormal power devices and corresponding initial abnormal power data;
The data characteristic acquisition module is used for carrying out characteristic extraction on the initial abnormal power data according to a preset abnormal data analysis method to obtain abnormal data characteristics corresponding to the initial abnormal power data;
The data identification module is used for identifying the initial abnormal power data according to a preset time sequence analysis method by combining the preset historical power data and the abnormal data characteristics corresponding to the initial abnormal power data to obtain first abnormal power data;
The data prediction module is used for performing deviation prediction on the first abnormal power data according to a preset data deviation prediction method to obtain deviation information and trend change information corresponding to the first abnormal power data;
The data correction module is used for constructing a first abnormal value correction model according to a preset correction model construction method, and inputting the first abnormal power data, the corresponding deviation information and the trend change information into the first abnormal value correction model so that the first abnormal value correction model can correct the first abnormal power data to obtain corrected power data.
The device monitors the running state of the equipment, so that the obtained initial abnormal power data is based on the actual equipment running data, more accurate first abnormal power data is identified from multiple angles through the combination of the abnormal data characteristics and a time sequence analysis method, and finally, the accurate correction of the abnormal data is completed through the deviation prediction of the abnormal data and based on the deviation information and trend change information of the abnormal data. By combining the characteristics of the abnormal data and the time sequence analysis method, accurate abnormal data is obtained, accurate correction of the abnormal data is realized on the basis of deviation information and trend change information of the abnormal data, the correction of the abnormal data is ensured to conform to the actual rules of operation of the power system and the principle of power engineering, and therefore, data quality guarantee is provided for test labeling of the power data, and comprehensive management of the power system is promoted.
Drawings
Fig. 1: the method for correcting the power data comprises the following steps of a flow chart;
Fig. 2: the step flow chart of another electric power data correction method provided by the embodiment of the invention;
fig. 3: the step flow chart of the correction method of the power data provided by the embodiment of the invention is provided;
fig. 4: the embodiment of the invention provides a structural schematic diagram of a power data correction device.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
In the power industry, the quality of data generated in the process of labeling is often affected by a plurality of factors, such as instrument faults, communication interruption and human input errors, which can cause irregular data and abnormal samples, and the correct identification and correction of the abnormal values are key to maintaining the quality of the labeling result of the power data. The normal operation data of the power system shows a regular characteristic, and the irregular data may suggest that the system has a certain problem or an error is generated in the data collection process, and the complexity and dynamic variability of the operation of the power system need to be considered in the anomaly detection, including load fluctuation, equipment state switching and the like, so that the normal system fluctuation and real data anomaly can be distinguished. Correction of outliers is not simply a deletion or replacement, but requires that the data closest to the true value can be extrapolated with reference to the actual rules of operation of the power system, in combination with the principles of power engineering and historical data information. Therefore, the embodiment of the invention provides a method and a device for correcting electric power data.
Example 1
Referring to fig. 1, a flowchart of steps of a method for correcting power data according to an embodiment of the present invention includes steps S101-S105.
Step S101: and acquiring first power data corresponding to the plurality of power equipment, and monitoring the running state of each power equipment according to a preset state monitoring method to obtain a plurality of abnormal power equipment and corresponding initial abnormal power data.
In this embodiment, the collecting first power data corresponding to a plurality of power devices, and monitoring, according to a preset state monitoring method, an operation state of each power device to obtain a plurality of abnormal power devices and corresponding initial abnormal power data includes:
acquiring initial power data corresponding to a plurality of power devices, and performing data preprocessing on the initial power data according to a preset data preprocessing method to obtain first power data;
Acquiring operation data corresponding to the plurality of electric powers according to a preset state monitoring method to obtain an initial state data set;
and processing the initial state data set according to a preset state data processing method to obtain a plurality of abnormal power devices and corresponding initial abnormal power data.
By collecting initial power data and extracting abnormal data according to the state of the equipment, the accuracy of the abnormal data is ensured.
In an optional embodiment, the collecting initial power data corresponding to the plurality of power devices includes real-time data, historical data, planning data, meteorological data, and device data; before preprocessing the initial power data, the method further comprises: the data source of the initial power data is identified, and the data range and characteristics to be integrated are determined.
In an alternative embodiment, the data preprocessing of the initial power data includes: extracting data by adopting a data extraction tool according to the source and the type of the data; wherein, the data source is a relational database, and SQL inquiry is applied to extract data; when the data source is a time sequence database, performing data extraction by using API call; when the data source is a text file, the text parsing technology is used for data extraction.
In an alternative embodiment, the data preprocessing of the initial power data further includes: cleaning the data; correcting the data with the problem of data accuracy; supplementing or deleting data with incomplete record; the inconsistent data is adjusted by comparing the main data; and performing data elimination on the repeatedly recorded data by using a deduplication algorithm.
In an alternative embodiment, the cleaned data formats, units and codes are unified by a data conversion tool; when the date formats of the data are inconsistent, unifying the data to be in a standard format ISO8601; converting the measurement unit of the data into a preset international unit; converting the coding of the classified data according to a preset unified coding scheme; and processing and converting the data through a data fusion processing technology, an entity identification technology, a logic verification technology and a data synchronization technology.
In an alternative embodiment, after the data is processed and converted, the data with frequent change is configured with an automatic maintenance flow, and the quality of the data is diagnosed and repaired by using a monitoring tool; and evaluating the data according to a preset data quality control tool, and cleaning, converting and fusing the corresponding data again when the data quality evaluation index is lower than a preset threshold value.
In this embodiment, the processing the initial state data set according to the preset state data processing method to obtain a plurality of abnormal power devices and corresponding initial abnormal power data includes:
According to a preset signal processing technology, signal processing is carried out on the initial state data set, and key characteristic signals of equipment operation in the initial state data set are obtained;
Acquiring preset historical characteristic signal data and a standard model, and performing cross comparison on the key characteristic signals of the operation of the equipment in the initial state data set, the preset historical characteristic signal data and the standard model according to a preset state evaluation algorithm to obtain a state evaluation result corresponding to the initial state data set;
And extracting the power data corresponding to the power devices in the abnormal state according to the state evaluation result corresponding to the initial state data set to obtain a plurality of abnormal power devices and corresponding initial abnormal power data.
And finally, the abnormal power data is extracted according to the state evaluation result, so that the diversity and accuracy of the abnormal power data are improved.
In this embodiment, the obtaining the state evaluation result corresponding to the initial state data set further includes:
When the state evaluation result shows that the state of the equipment is abnormal, starting a fault diagnosis program corresponding to the abnormal equipment, judging the fault type and the position of the abnormal equipment according to a preset machine learning algorithm and an analysis data mode, and obtaining the fault type and the fault position of the abnormal equipment;
Inputting operation data corresponding to the abnormal equipment into a preset state prediction model, performing prediction analysis on the operation trend of the abnormal equipment, and generating early warning information when an operation problem is predicted;
And respectively inputting the fault type, the fault position and the early warning information of the abnormal equipment into a preset fault tree analysis model and a repair system to obtain a repair scheme corresponding to the abnormal equipment, so that a worker can repair the abnormal equipment.
In an alternative embodiment, the equipment state, the fault analysis result and the early warning information are subjected to data presentation through a preset human-computer interface, and a worker acquires key information and monitors the running condition of the equipment through the human-computer interface; after the early warning or fault signal is confirmed by the staff, the maintenance team is scheduled to respond through the automatic flow of the preset state monitoring system, and the processing process is recorded in a preset database.
By judging the abnormal state of the equipment, the fault type and the fault position are obtained, and early warning information is generated, so that workers can quickly repair the abnormal equipment, and the stable operation of the equipment is ensured.
Step S102: and extracting the characteristics of the initial abnormal power data according to a preset abnormal data analysis method to obtain the abnormal data characteristics corresponding to the initial abnormal power data.
In this embodiment, the feature extraction of the initial abnormal power data according to a preset abnormal data analysis method, to obtain an abnormal data feature corresponding to the initial abnormal power data, includes:
according to a preset frequency data acquisition technology, carrying out corresponding frequency data acquisition on initial abnormal power data to obtain frequency data of a plurality of data types, wherein the frequency data of the plurality of data types comprises: time series data and device log data;
performing data fusion on the frequency data of the data types based on a preset data fusion technology to obtain a device state view, wherein the data fusion comprises data cleaning, data conversion and mapping and data integration;
and detecting the deviated data characteristics in the equipment state view according to a preset abnormality detection algorithm to obtain abnormal data characteristics corresponding to the initial abnormal power data.
By integrating the data into one view, more comprehensive and comprehensive equipment state information can be provided, so that the extraction of abnormal data features is more accurate.
In an alternative embodiment, the state change of the equipment is captured according to a predictive analysis technology, and future load fluctuation and equipment performance trend of the equipment are predicted; when the risk is predicted, an early warning signal is sent to a worker through a preset warning and notifying mechanism, and specific problems and fault reasons of the equipment are judged according to fault diagnosis and root cause analysis technology.
Step S103: and according to a preset time sequence analysis method, combining the preset historical power data and abnormal data characteristics corresponding to the initial abnormal power data, and identifying the initial abnormal power data to obtain first abnormal power data.
In this embodiment, the identifying, according to a preset time sequence analysis method, the initial abnormal power data according to the preset historical power data and the abnormal data characteristics corresponding to the initial abnormal power data, to obtain first abnormal power data includes:
indexing initial time sequence data in the initial abnormal power data according to the time stamp, so that the initial time sequence data are arranged according to a preset time sequence, and constructing a time sequence diagram corresponding to the initial time sequence data;
Processing the time series data according to a preset time series analysis and prediction algorithm, and removing short-term fluctuation in the time series data to obtain first time series data, corresponding long-term trend data and seasonal pattern data, wherein the preset time series analysis and prediction algorithm comprises: a moving average algorithm or an exponential smoothing method;
respectively carrying out data analysis on the long-term trend data and the seasonal pattern data according to a trend decomposition algorithm and a seasonal decomposition algorithm to obtain trend components corresponding to the long-term trend data and seasonal components corresponding to the seasonal pattern data;
According to the periodic detection method, identifying and separating the cyclic fluctuation in the first time sequence data to obtain second time sequence data; carrying out abnormal value identification on the second time sequence data according to a box graph statistical method to obtain abnormal fluctuation data;
performing stability detection on the second time series data by using an ADF (automatic frequency correction) detection method, performing differential processing or conversion processing on the second time series data to obtain stable second time series data when the detection result is unstable, and fitting the stable second time series data according to a preset statistical model to obtain third time series data;
According to a preset detection algorithm, the third time series data is analyzed for prediction errors by combining a time series chart, trend components corresponding to long-term trend data, seasonal components corresponding to seasonal mode data and abnormal fluctuation data, and when the prediction errors exceed a preset error range, the corresponding data points are marked abnormally until all the data points in the third time series data complete the analysis of the prediction errors, so that first abnormal power data are obtained.
By the time sequence analysis method, system fluctuation and abnormal data can be identified, the identified abnormal data is ensured to be accurate and real, and abnormal value correction errors caused by confusing the system fluctuation into the abnormal data are avoided.
Step S104: and according to a preset data deviation prediction method, performing deviation prediction on the first abnormal power data to obtain deviation information and trend change information corresponding to the first abnormal power data.
In this embodiment, the performing deviation prediction on the first abnormal power data according to a preset data deviation prediction method to obtain deviation information and trend change information corresponding to the first abnormal power data includes:
Evaluating the data points in the first abnormal power data according to a weighted least square method, when the evaluation result of the data points exceeds a preset evaluation range, marking the corresponding data points until all the data points in the first abnormal power data are evaluated, and sorting all the marked data points to obtain corresponding second abnormal power data;
Inputting the second abnormal power data into a preset power system model so that the power system model can evaluate the influence degree of the second abnormal power data to obtain an influence degree evaluation result of the second abnormal power data;
According to the influence degree evaluation result, the state variable data of the first abnormal power data are combined, and the state estimation data in the first abnormal power data are subjected to pre-correction to obtain an improved state estimation value corresponding to the first abnormal power data;
Optimizing the improved state estimation value corresponding to the first abnormal power data according to a preset optimization criterion to obtain an optimized improved state estimation value;
And according to a preset trend analysis tool and historical power data, combining the optimized improved state estimation value, and judging deviation information and trend change information of the first abnormal power data by using a preset state estimation technology to obtain the deviation information and trend change information corresponding to the first abnormal power data.
By acquiring the abnormal data deviation information and the trend change information, a theoretical basis can be provided for correcting the abnormal value, and the correction accuracy of the abnormal value is improved.
Step S105: and constructing a first abnormal value correction model according to a preset correction model construction method, and inputting the first abnormal power data, the corresponding deviation information and the trend change information into the first abnormal value correction model so that the first abnormal value correction model can correct the first abnormal power data to obtain corrected power data.
In this embodiment, the constructing a first outlier correction model according to a preset correction model construction method includes:
acquiring a preset initial abnormal value correction model and real-time power data in a power system, wherein the real-time power data in the power system comprises: generating capacity, load demand, line flow, etc.;
after noise removal and missing value processing are carried out on the real-time power data according to a preset data cleaning algorithm, the statistical features and the frequency domain features of the real-time power data subjected to data cleaning are extracted according to the preset operation features of the power system, and the statistical features and the frequency domain features of the real-time power data subjected to data cleaning are obtained;
according to the acquired meteorological data and the equipment state of the power equipment to which the real-time power data belongs, converting the statistical characteristics and the frequency domain characteristics of the real-time power data subjected to data cleaning to obtain input characteristics;
According to an isolated forest algorithm, analyzing the relevance of the attributes in the real-time power data to obtain abnormal values and abnormal reasons in the real-time power data, wherein the attributes in the real-time power data comprise: load change, topology and running log;
Inputting the input characteristics, the abnormal value in the real-time power data and the abnormal cause into a preset initial abnormal value correction model, combining the power data, the current and the voltage measurement characteristics in the real-time power data, taking the correction of the abnormal value as a training target of the initial abnormal value correction model, training the initial abnormal value correction model by using a decision tree machine learning algorithm, and generating a corresponding correction rule;
evaluating a correction rule generated by the initial abnormal value correction model according to a time sequence segmentation technology to obtain an evaluation result of the initial abnormal value correction model; and according to the evaluation result, adjusting the parameters of the initial outlier correction model by using a grid search parameter optimization method until the evaluation result of the initial outlier correction model meets a preset standard, and completing training of the initial outlier correction model to obtain a first outlier correction model.
The abnormal value correction model is trained and optimized through the training data set, so that the correction capability of the abnormal value correction model can be improved, and the effect of repairing abnormal values in the electric power data is ensured.
In this embodiment, after obtaining the corrected power data, the method further includes:
Carrying out data integration on the corrected power data obtained after the first abnormal value correction model corrects the first abnormal power data and the first power data to obtain second power data; according to an automatic data cleaning algorithm, performing data cleaning on the second power data to obtain third power data;
After the statistical characteristics in the third electric power data are extracted, judging the equilibrium degree of the statistical characteristics, and when the equilibrium degree is not in a preset equilibrium range, increasing data samples in the third electric power data according to a data increasing technology so that the equilibrium degree of the statistical characteristics in the third electric power data meets the preset requirement to obtain fourth electric power data;
Processing the fourth power data according to ApacheHadoop or APACHESPARK distributed computing frames, and then performing check point processing on the fourth power data according to a preset check point and a data backup strategy to obtain fifth power data;
and performing test marking on the fifth electric power data according to a preset marking method, and evaluating the test marking result based on a preset log record and a data quality evaluation method to obtain a test marking evaluation result.
The quality and accuracy of the trial marking process can be improved through the optimization processing of the corrected power data.
In an alternative embodiment, please refer to fig. 2, which is a flowchart illustrating steps of another method for correcting electric power data according to an embodiment of the present invention. The method comprises the steps of collecting data of the power equipment, carrying out integration and real-time analysis, monitoring the state of the power equipment, and starting abnormality detection of the corresponding power equipment when the state of the power equipment is abnormal; carrying out time sequence analysis on the monitored abnormal data and carrying out state estimation on equipment to obtain data needing correction, and correcting the data; and finally, continuously iterating the steps of monitoring and converting the abnormal data and the model so that the model has more accurate abnormal data detection capability.
In an alternative embodiment, please refer to fig. 3, which is a flowchart illustrating a step of a method for correcting power data according to an embodiment of the present invention. The method comprises the steps of collecting data of the electric equipment, transmitting the data to a data processing center, carrying out encryption processing on the data, analyzing the data by using Fourier transformation, and extracting key characteristic signals of equipment operation; and comparing the characteristic signals, the historical data and the standard model, finally obtaining an evaluation result of the current running state of the equipment through an algorithm, and starting a fault diagnosis program to correct the abnormal value of the equipment according to the evaluation result.
According to the embodiment, the initial abnormal power data is based on actual equipment operation data through monitoring of the equipment operation state, the first more accurate abnormal power data is identified from multiple angles through combination of abnormal data characteristics and a time sequence analysis method, and finally, through deviation prediction of the abnormal data, accurate correction of the abnormal data is completed based on deviation information and trend change information of the abnormal data. By combining the characteristics of the abnormal data and the time sequence analysis method, accurate abnormal data is obtained, accurate correction of the abnormal data is realized on the basis of deviation information and trend change information of the abnormal data, the correction of the abnormal data is ensured to conform to the actual rules of operation of the power system and the principle of power engineering, and therefore, data quality guarantee is provided for test labeling of the power data, and comprehensive management of the power system is promoted.
Example two
Referring to fig. 4, a schematic structural diagram of a power data correction device according to an embodiment of the present invention includes: a power data acquisition module 201, a data feature acquisition module 202, a data identification module 203, a data prediction module 204, and a data correction module 205.
The power data acquisition module 201 is configured to collect first power data corresponding to a plurality of power devices, and monitor an operation state of each power device according to a preset state monitoring method, so as to obtain a plurality of abnormal power devices and corresponding initial abnormal power data.
In this embodiment, the power data acquisition module 201 includes: a power data acquisition unit;
The power data acquisition unit is used for acquiring initial power data corresponding to a plurality of power devices, and performing data preprocessing on the initial power data according to a preset data preprocessing method to obtain first power data;
Acquiring operation data corresponding to the plurality of electric powers according to a preset state monitoring method to obtain an initial state data set;
and processing the initial state data set according to a preset state data processing method to obtain a plurality of abnormal power devices and corresponding initial abnormal power data.
In this embodiment, the power data acquisition unit includes: an abnormal data acquisition subunit;
The abnormal data acquisition subunit is used for carrying out signal processing on the initial state data set according to a preset signal processing technology to obtain key characteristic signals of equipment operation in the initial state data set;
Acquiring preset historical characteristic signal data and a standard model, and performing cross comparison on the key characteristic signals of the operation of the equipment in the initial state data set, the preset historical characteristic signal data and the standard model according to a preset state evaluation algorithm to obtain a state evaluation result corresponding to the initial state data set;
And extracting the power data corresponding to the power devices in the abnormal state according to the state evaluation result corresponding to the initial state data set to obtain a plurality of abnormal power devices and corresponding initial abnormal power data.
In this embodiment, the abnormal data acquisition subunit further includes: an abnormal state repair subunit;
The abnormal state repair subunit is used for starting a fault diagnosis program corresponding to the abnormal equipment when the state evaluation result shows that the equipment is abnormal, judging the fault type and the position of the abnormal equipment according to a preset machine learning algorithm and an analysis data mode, and obtaining the fault type and the fault position of the abnormal equipment;
Inputting operation data corresponding to the abnormal equipment into a preset state prediction model, performing prediction analysis on the operation trend of the abnormal equipment, and generating early warning information when an operation problem is predicted;
And respectively inputting the fault type, the fault position and the early warning information of the abnormal equipment into a preset fault tree analysis model and a repair system to obtain a repair scheme corresponding to the abnormal equipment, so that a worker can repair the abnormal equipment.
The data feature obtaining module 202 is configured to perform feature extraction on the initial abnormal power data according to a preset abnormal data analysis method, so as to obtain an abnormal data feature corresponding to the initial abnormal power data.
In this embodiment, the data feature acquiring module 202 includes: a data feature acquisition unit;
the data characteristic acquisition unit is used for acquiring corresponding frequency data of the initial abnormal power data according to a preset frequency data acquisition technology to obtain frequency data of a plurality of data types, wherein the frequency data of the plurality of data types comprises: time series data and device log data;
performing data fusion on the frequency data of the data types based on a preset data fusion technology to obtain a device state view, wherein the data fusion comprises data cleaning, data conversion and mapping and data integration;
and detecting the deviated data characteristics in the equipment state view according to a preset abnormality detection algorithm to obtain abnormal data characteristics corresponding to the initial abnormal power data.
The data identification module 203 is configured to identify, according to a preset time sequence analysis method, the initial abnormal power data by combining the preset historical power data and the abnormal data characteristics corresponding to the initial abnormal power data, so as to obtain first abnormal power data.
In this embodiment, the data identifying module 203 includes: a data identification unit;
The data identification unit is used for indexing initial time sequence data in the initial abnormal power data according to the time stamp so as to enable the initial time sequence data to be arranged according to a preset time sequence and construct a time sequence diagram corresponding to the initial time sequence data;
Processing the time series data according to a preset time series analysis and prediction algorithm, and removing short-term fluctuation in the time series data to obtain first time series data, corresponding long-term trend data and seasonal pattern data, wherein the preset time series analysis and prediction algorithm comprises: a moving average algorithm or an exponential smoothing method;
respectively carrying out data analysis on the long-term trend data and the seasonal pattern data according to a trend decomposition algorithm and a seasonal decomposition algorithm to obtain trend components corresponding to the long-term trend data and seasonal components corresponding to the seasonal pattern data;
According to the periodic detection method, identifying and separating the cyclic fluctuation in the first time sequence data to obtain second time sequence data; carrying out abnormal value identification on the second time sequence data according to a box graph statistical method to obtain abnormal fluctuation data;
performing stability detection on the second time series data by using an ADF (automatic frequency correction) detection method, performing differential processing or conversion processing on the second time series data to obtain stable second time series data when the detection result is unstable, and fitting the stable second time series data according to a preset statistical model to obtain third time series data;
According to a preset detection algorithm, the third time series data is analyzed for prediction errors by combining a time series chart, trend components corresponding to long-term trend data, seasonal components corresponding to seasonal mode data and abnormal fluctuation data, and when the prediction errors exceed a preset error range, the corresponding data points are marked abnormally until all the data points in the third time series data complete the analysis of the prediction errors, so that first abnormal power data are obtained.
The data prediction module 204 is configured to perform deviation prediction on the first abnormal power data according to a preset data deviation prediction method, so as to obtain deviation information and trend change information corresponding to the first abnormal power data.
In this embodiment, the data prediction module 204 includes: a data prediction unit;
The data prediction unit is used for evaluating the data points in the first abnormal power data according to a weighted least square method, when the evaluation result of the data points exceeds a preset evaluation range, the corresponding data points are identified until all the data points in the first abnormal power data are evaluated, and all the identified data points are arranged to obtain corresponding second abnormal power data;
Inputting the second abnormal power data into a preset power system model so that the power system model can evaluate the influence degree of the second abnormal power data to obtain an influence degree evaluation result of the second abnormal power data;
According to the influence degree evaluation result, the state variable data of the first abnormal power data are combined, and the state estimation data in the first abnormal power data are subjected to pre-correction to obtain an improved state estimation value corresponding to the first abnormal power data;
Optimizing the improved state estimation value corresponding to the first abnormal power data according to a preset optimization criterion to obtain an optimized improved state estimation value;
And according to a preset trend analysis tool and historical power data, combining the optimized improved state estimation value, and judging deviation information and trend change information of the first abnormal power data by using a preset state estimation technology to obtain the deviation information and trend change information corresponding to the first abnormal power data.
The data correction module 205 is configured to construct a first abnormal value correction model according to a preset correction model construction method, and input the first abnormal power data and the corresponding deviation information and trend change information into the first abnormal value correction model, so that the first abnormal value correction model can correct the first abnormal power data to obtain corrected power data.
In this embodiment, the data correction module 205 includes: a correction model construction unit;
the correction model construction unit is used for acquiring a preset initial abnormal value correction model and real-time power data in a power system, wherein the real-time power data in the power system comprises: generating capacity, load demand, line flow, etc.;
after noise removal and missing value processing are carried out on the real-time power data according to a preset data cleaning algorithm, the statistical features and the frequency domain features of the real-time power data subjected to data cleaning are extracted according to the preset operation features of the power system, and the statistical features and the frequency domain features of the real-time power data subjected to data cleaning are obtained;
according to the acquired meteorological data and the equipment state of the power equipment to which the real-time power data belongs, converting the statistical characteristics and the frequency domain characteristics of the real-time power data subjected to data cleaning to obtain input characteristics;
According to an isolated forest algorithm, analyzing the relevance of the attributes in the real-time power data to obtain abnormal values and abnormal reasons in the real-time power data, wherein the attributes in the real-time power data comprise: load change, topology and running log;
Inputting the input characteristics, the abnormal value in the real-time power data and the abnormal cause into a preset initial abnormal value correction model, combining the power data, the current and the voltage measurement characteristics in the real-time power data, taking the correction of the abnormal value as a training target of the initial abnormal value correction model, training the initial abnormal value correction model by using a decision tree machine learning algorithm, and generating a corresponding correction rule;
evaluating a correction rule generated by the initial abnormal value correction model according to a time sequence segmentation technology to obtain an evaluation result of the initial abnormal value correction model; and according to the evaluation result, adjusting the parameters of the initial outlier correction model by using a grid search parameter optimization method until the evaluation result of the initial outlier correction model meets a preset standard, and completing training of the initial outlier correction model to obtain a first outlier correction model.
In this embodiment, the data correction module 205 further includes: a data optimizing unit;
The data optimization unit is used for integrating the corrected power data obtained after the first abnormal value correction model corrects the first abnormal power data with the first power data to obtain second power data; according to an automatic data cleaning algorithm, performing data cleaning on the second power data to obtain third power data;
After the statistical characteristics in the third electric power data are extracted, judging the equilibrium degree of the statistical characteristics, and when the equilibrium degree is not in a preset equilibrium range, increasing data samples in the third electric power data according to a data increasing technology so that the equilibrium degree of the statistical characteristics in the third electric power data meets the preset requirement to obtain fourth electric power data;
Processing the fourth power data according to ApacheHadoop or APACHESPARK distributed computing frames, and then performing check point processing on the fourth power data according to a preset check point and a data backup strategy to obtain fifth power data;
and performing test marking on the fifth electric power data according to a preset marking method, and evaluating the test marking result based on a preset log record and a data quality evaluation method to obtain a test marking evaluation result.
According to the embodiment, the initial abnormal power data is based on actual equipment operation data through monitoring of the equipment operation state, the first more accurate abnormal power data is identified from multiple angles through combination of abnormal data characteristics and a time sequence analysis method, and finally, through deviation prediction of the abnormal data, accurate correction of the abnormal data is completed based on deviation information and trend change information of the abnormal data. By combining the characteristics of the abnormal data and the time sequence analysis method, accurate abnormal data is obtained, accurate correction of the abnormal data is realized on the basis of deviation information and trend change information of the abnormal data, the correction of the abnormal data is ensured to conform to the actual rules of operation of the power system and the principle of power engineering, and therefore, data quality guarantee is provided for test labeling of the power data, and comprehensive management of the power system is promoted.
In summary, according to the embodiment of the invention, through monitoring the operation state of the device, the obtained initial abnormal power data is based on the actual operation data of the device, through the combination of the abnormal data characteristics and the time sequence analysis method, the more accurate first abnormal power data is identified from multiple angles, and finally, through deviation prediction of the abnormal data, the accurate correction of the abnormal data is completed based on the deviation information and the trend change information of the abnormal data. By combining the characteristics of the abnormal data and the time sequence analysis method, accurate abnormal data is obtained, accurate correction of the abnormal data is realized on the basis of deviation information and trend change information of the abnormal data, the correction of the abnormal data is ensured to conform to the actual rules of operation of the power system and the principle of power engineering, and therefore, data quality guarantee is provided for test labeling of the power data, and comprehensive management of the power system is promoted.
The foregoing embodiments have been provided for the purpose of illustrating the general principles of the present invention, and are not to be construed as limiting the scope of the invention. It should be noted that any modifications, equivalent substitutions, improvements, etc. made by those skilled in the art without departing from the spirit and principles of the present invention are intended to be included in the scope of the present invention.
Claims (10)
1. A method of correcting power data, comprising:
collecting first power data corresponding to a plurality of power devices, and monitoring the running state of each power device according to a preset state monitoring method to obtain a plurality of abnormal power devices and corresponding initial abnormal power data;
Extracting features of the initial abnormal power data according to a preset abnormal data analysis method to obtain abnormal data features corresponding to the initial abnormal power data;
According to a preset time sequence analysis method, combining preset historical power data and abnormal data characteristics corresponding to initial abnormal power data, and identifying the initial abnormal power data to obtain first abnormal power data;
According to a preset data deviation prediction method, performing deviation prediction on the first abnormal power data to obtain deviation information and trend change information corresponding to the first abnormal power data;
And constructing a first abnormal value correction model according to a preset correction model construction method, and inputting the first abnormal power data, the corresponding deviation information and the trend change information into the first abnormal value correction model so that the first abnormal value correction model can correct the first abnormal power data to obtain corrected power data.
2. The method for correcting power data according to claim 1, wherein the collecting the first power data corresponding to the plurality of power devices, and monitoring the operation state of each power device according to a preset state monitoring method, to obtain a plurality of abnormal power devices and corresponding initial abnormal power data, includes:
acquiring initial power data corresponding to a plurality of power devices, and performing data preprocessing on the initial power data according to a preset data preprocessing method to obtain first power data;
Acquiring operation data corresponding to the plurality of electric powers according to a preset state monitoring method to obtain an initial state data set;
and processing the initial state data set according to a preset state data processing method to obtain a plurality of abnormal power devices and corresponding initial abnormal power data.
3. The method for correcting power data according to claim 2, wherein the processing the initial state data set according to the preset state data processing method to obtain a plurality of abnormal power devices and corresponding initial abnormal power data includes:
According to a preset signal processing technology, signal processing is carried out on the initial state data set, and key characteristic signals of equipment operation in the initial state data set are obtained;
Acquiring preset historical characteristic signal data and a standard model, and performing cross comparison on the key characteristic signals of the operation of the equipment in the initial state data set, the preset historical characteristic signal data and the standard model according to a preset state evaluation algorithm to obtain a state evaluation result corresponding to the initial state data set;
And extracting the power data corresponding to the power devices in the abnormal state according to the state evaluation result corresponding to the initial state data set to obtain a plurality of abnormal power devices and corresponding initial abnormal power data.
4. The method for modifying power data according to claim 3, wherein said obtaining a state evaluation result corresponding to the initial state data set further comprises:
When the state evaluation result shows that the state of the equipment is abnormal, starting a fault diagnosis program corresponding to the abnormal equipment, judging the fault type and the position of the abnormal equipment according to a preset machine learning algorithm and an analysis data mode, and obtaining the fault type and the fault position of the abnormal equipment;
Inputting operation data corresponding to the abnormal equipment into a preset state prediction model, performing prediction analysis on the operation trend of the abnormal equipment, and generating early warning information when an operation problem is predicted;
And respectively inputting the fault type, the fault position and the early warning information of the abnormal equipment into a preset fault tree analysis model and a repair system to obtain a repair scheme corresponding to the abnormal equipment, so that a worker can repair the abnormal equipment.
5. The method for correcting electric power data according to claim 1, wherein the feature extraction of the initial abnormal electric power data according to a preset abnormal data analysis method to obtain the abnormal data feature corresponding to the initial abnormal electric power data comprises:
according to a preset frequency data acquisition technology, carrying out corresponding frequency data acquisition on initial abnormal power data to obtain frequency data of a plurality of data types, wherein the frequency data of the plurality of data types comprises: time series data and device log data;
performing data fusion on the frequency data of the data types based on a preset data fusion technology to obtain a device state view, wherein the data fusion comprises data cleaning, data conversion and mapping and data integration;
and detecting the deviated data characteristics in the equipment state view according to a preset abnormality detection algorithm to obtain abnormal data characteristics corresponding to the initial abnormal power data.
6. The method for correcting power data according to claim 1, wherein the identifying the initial abnormal power data according to the predetermined time-series analysis method in combination with the predetermined historical power data and the abnormal data characteristics corresponding to the initial abnormal power data to obtain the first abnormal power data includes:
indexing initial time sequence data in the initial abnormal power data according to the time stamp, so that the initial time sequence data are arranged according to a preset time sequence, and constructing a time sequence diagram corresponding to the initial time sequence data;
Processing the time series data according to a preset time series analysis and prediction algorithm, and removing short-term fluctuation in the time series data to obtain first time series data, corresponding long-term trend data and seasonal pattern data, wherein the preset time series analysis and prediction algorithm comprises: a moving average algorithm or an exponential smoothing method;
respectively carrying out data analysis on the long-term trend data and the seasonal pattern data according to a trend decomposition algorithm and a seasonal decomposition algorithm to obtain trend components corresponding to the long-term trend data and seasonal components corresponding to the seasonal pattern data;
According to the periodic detection method, identifying and separating the cyclic fluctuation in the first time sequence data to obtain second time sequence data; carrying out abnormal value identification on the second time sequence data according to a box graph statistical method to obtain abnormal fluctuation data;
performing stability detection on the second time series data by using an ADF (automatic frequency correction) detection method, performing differential processing or conversion processing on the second time series data to obtain stable second time series data when the detection result is unstable, and fitting the stable second time series data according to a preset statistical model to obtain third time series data;
According to a preset detection algorithm, the third time series data is analyzed for prediction errors by combining a time series chart, trend components corresponding to long-term trend data, seasonal components corresponding to seasonal mode data and abnormal fluctuation data, and when the prediction errors exceed a preset error range, the corresponding data points are marked abnormally until all the data points in the third time series data complete the analysis of the prediction errors, so that first abnormal power data are obtained.
7. The method for correcting electric power data according to claim 1, wherein the performing deviation prediction on the first abnormal electric power data according to a preset data deviation prediction method to obtain deviation information and trend change information corresponding to the first abnormal electric power data includes:
Evaluating the data points in the first abnormal power data according to a weighted least square method, when the evaluation result of the data points exceeds a preset evaluation range, marking the corresponding data points until all the data points in the first abnormal power data are evaluated, and sorting all the marked data points to obtain corresponding second abnormal power data;
Inputting the second abnormal power data into a preset power system model so that the power system model can evaluate the influence degree of the second abnormal power data to obtain an influence degree evaluation result of the second abnormal power data;
According to the influence degree evaluation result, the state variable data of the first abnormal power data are combined, and the state estimation data in the first abnormal power data are subjected to pre-correction to obtain an improved state estimation value corresponding to the first abnormal power data;
Optimizing the improved state estimation value corresponding to the first abnormal power data according to a preset optimization criterion to obtain an optimized improved state estimation value;
And according to a preset trend analysis tool and historical power data, combining the optimized improved state estimation value, and judging deviation information and trend change information of the first abnormal power data by using a preset state estimation technology to obtain the deviation information and trend change information corresponding to the first abnormal power data.
8. The method for correcting power data according to claim 1, wherein the constructing a first abnormal value correction model according to a preset correction model construction method includes:
acquiring a preset initial abnormal value correction model and real-time power data in a power system, wherein the real-time power data in the power system comprises: generating capacity, load demand, line flow, etc.;
after noise removal and missing value processing are carried out on the real-time power data according to a preset data cleaning algorithm, the statistical features and the frequency domain features of the real-time power data subjected to data cleaning are extracted according to the preset operation features of the power system, and the statistical features and the frequency domain features of the real-time power data subjected to data cleaning are obtained;
according to the acquired meteorological data and the equipment state of the power equipment to which the real-time power data belongs, converting the statistical characteristics and the frequency domain characteristics of the real-time power data subjected to data cleaning to obtain input characteristics;
According to an isolated forest algorithm, analyzing the relevance of the attributes in the real-time power data to obtain abnormal values and abnormal reasons in the real-time power data, wherein the attributes in the real-time power data comprise: load change, topology and running log;
Inputting the input characteristics, the abnormal value in the real-time power data and the abnormal cause into a preset initial abnormal value correction model, combining the power data, the current and the voltage measurement characteristics in the real-time power data, taking the correction of the abnormal value as a training target of the initial abnormal value correction model, training the initial abnormal value correction model by using a decision tree machine learning algorithm, and generating a corresponding correction rule;
evaluating a correction rule generated by the initial abnormal value correction model according to a time sequence segmentation technology to obtain an evaluation result of the initial abnormal value correction model; and according to the evaluation result, adjusting the parameters of the initial outlier correction model by using a grid search parameter optimization method until the evaluation result of the initial outlier correction model meets a preset standard, and completing training of the initial outlier correction model to obtain a first outlier correction model.
9. The method for modifying power data according to claim 1, further comprising:
Carrying out data integration on the corrected power data obtained after the first abnormal value correction model corrects the first abnormal power data and the first power data to obtain second power data; according to an automatic data cleaning algorithm, performing data cleaning on the second power data to obtain third power data;
After the statistical characteristics in the third electric power data are extracted, judging the equilibrium degree of the statistical characteristics, and when the equilibrium degree is not in a preset equilibrium range, increasing data samples in the third electric power data according to a data increasing technology so that the equilibrium degree of the statistical characteristics in the third electric power data meets the preset requirement to obtain fourth electric power data;
Processing the fourth power data according to ApacheHadoop or APACHESPARK distributed computing frames, and then performing check point processing on the fourth power data according to a preset check point and a data backup strategy to obtain fifth power data;
and performing test marking on the fifth electric power data according to a preset marking method, and evaluating the test marking result based on a preset log record and a data quality evaluation method to obtain a test marking evaluation result.
10. A correction device for electric power data, comprising: the system comprises a power data acquisition module, a data characteristic acquisition module, a data identification module, a data prediction module and a data correction module;
The power data acquisition module is used for acquiring first power data corresponding to a plurality of power devices, and monitoring the running state of each power device according to a preset state monitoring method to obtain a plurality of abnormal power devices and corresponding initial abnormal power data;
The data characteristic acquisition module is used for carrying out characteristic extraction on the initial abnormal power data according to a preset abnormal data analysis method to obtain abnormal data characteristics corresponding to the initial abnormal power data;
The data identification module is used for identifying the initial abnormal power data according to a preset time sequence analysis method by combining the preset historical power data and the abnormal data characteristics corresponding to the initial abnormal power data to obtain first abnormal power data;
The data prediction module is used for performing deviation prediction on the first abnormal power data according to a preset data deviation prediction method to obtain deviation information and trend change information corresponding to the first abnormal power data;
The data correction module is used for constructing a first abnormal value correction model according to a preset correction model construction method, and inputting the first abnormal power data, the corresponding deviation information and the trend change information into the first abnormal value correction model so that the first abnormal value correction model can correct the first abnormal power data to obtain corrected power data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410508752.2A CN118094168B (en) | 2024-04-26 | 2024-04-26 | Correction method and device for electric power data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410508752.2A CN118094168B (en) | 2024-04-26 | 2024-04-26 | Correction method and device for electric power data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118094168A true CN118094168A (en) | 2024-05-28 |
| CN118094168B CN118094168B (en) | 2024-07-23 |
Family
ID=91157666
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410508752.2A Active CN118094168B (en) | 2024-04-26 | 2024-04-26 | Correction method and device for electric power data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118094168B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016101690A1 (en) * | 2014-12-22 | 2016-06-30 | 国家电网公司 | Time sequence analysis-based state monitoring data cleaning method for power transmission and transformation device |
| CN111695620A (en) * | 2020-06-08 | 2020-09-22 | 中国电力科学研究院有限公司 | Method and system for detecting and correcting abnormal data of time sequence of power system |
| CN112527788A (en) * | 2020-12-17 | 2021-03-19 | 北京中恒博瑞数字电力科技有限公司 | Method and device for detecting and cleaning abnormal value of transformer monitoring data |
| CN117370744A (en) * | 2023-12-07 | 2024-01-09 | 国网天津市电力公司营销服务中心 | Dynamic cleaning method and system for abnormal power consumption data of power consumer |
| CN117436004A (en) * | 2023-12-20 | 2024-01-23 | 深圳市鸿明机电有限公司 | Motor performance real-time monitoring system and method |
| CN117762913A (en) * | 2023-12-14 | 2024-03-26 | 广西电网有限责任公司 | Automatic identification method and system for abnormal power data of offshore wind power |
-
2024
- 2024-04-26 CN CN202410508752.2A patent/CN118094168B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016101690A1 (en) * | 2014-12-22 | 2016-06-30 | 国家电网公司 | Time sequence analysis-based state monitoring data cleaning method for power transmission and transformation device |
| CN111695620A (en) * | 2020-06-08 | 2020-09-22 | 中国电力科学研究院有限公司 | Method and system for detecting and correcting abnormal data of time sequence of power system |
| CN112527788A (en) * | 2020-12-17 | 2021-03-19 | 北京中恒博瑞数字电力科技有限公司 | Method and device for detecting and cleaning abnormal value of transformer monitoring data |
| CN117370744A (en) * | 2023-12-07 | 2024-01-09 | 国网天津市电力公司营销服务中心 | Dynamic cleaning method and system for abnormal power consumption data of power consumer |
| CN117762913A (en) * | 2023-12-14 | 2024-03-26 | 广西电网有限责任公司 | Automatic identification method and system for abnormal power data of offshore wind power |
| CN117436004A (en) * | 2023-12-20 | 2024-01-23 | 深圳市鸿明机电有限公司 | Motor performance real-time monitoring system and method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118094168B (en) | 2024-07-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10444746B2 (en) | Method for managing subsystems of a process plant using a distributed control system | |
| CN111340250A (en) | Equipment maintenance device, method and computer readable storage medium | |
| CN110727531A (en) | Fault prediction and processing method and system for online system | |
| CN112949874B (en) | Power distribution terminal defect characteristic self-diagnosis method and system | |
| CN116992346A (en) | Enterprise production data processing system based on artificial intelligence big data analysis | |
| CN115392812B (en) | Abnormal root cause positioning method, device, equipment and medium | |
| CN116388402B (en) | Abnormality alarm analysis method applied to power transformation equipment | |
| CN117474357A (en) | Power distribution room operation and maintenance management method and system based on deep learning | |
| CN116739829A (en) | Big data-based power data analysis method, system and medium | |
| CN117311295B (en) | Production quality improving method and system based on wireless network equipment | |
| CN109523030B (en) | Telemetering parameter abnormity monitoring system based on machine learning | |
| CN117331790A (en) | Machine room fault detection method and device for data center | |
| CN118094168B (en) | Correction method and device for electric power data | |
| CN112381242A (en) | Nuclear power station equipment maintenance project data processing method and system | |
| CN117436569A (en) | Nuclear power equipment fault prediction and intelligent calibration method and system based on random forest | |
| CN116595353A (en) | Remote fault diagnosis and intelligent decision system for camera | |
| CN117074852A (en) | Power distribution network electric energy monitoring and early warning management method and system | |
| CN116308295A (en) | Industrial production data management method and system | |
| CN115587312B (en) | Fault point positioning method and system based on multidimensional big data analysis | |
| WO2022130789A1 (en) | Cause inference system and cause inference method | |
| CN115270982A (en) | Switch cabinet fault prediction method based on multi-data neural network | |
| CN113836203A (en) | Network data diagnosis detection analysis system | |
| CN113553352A (en) | Method and system for identifying familial defects of power secondary equipment | |
| CN113591909A (en) | Abnormality detection method, abnormality detection device, and storage medium for power system | |
| CN117172139B (en) | Performance test method and system for copper-clad aluminum alloy cable for communication |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |