CN114626433A - Fault prediction and classification method, device and system for intelligent electric energy meter - Google Patents
Fault prediction and classification method, device and system for intelligent electric energy meter Download PDFInfo
- Publication number
- CN114626433A CN114626433A CN202210097394.1A CN202210097394A CN114626433A CN 114626433 A CN114626433 A CN 114626433A CN 202210097394 A CN202210097394 A CN 202210097394A CN 114626433 A CN114626433 A CN 114626433A
- Authority
- CN
- China
- Prior art keywords
- data
- prediction
- fault
- classification
- electric energy
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000012549 training Methods 0.000 claims abstract description 48
- 238000012706 support-vector machine Methods 0.000 claims abstract description 45
- 238000003066 decision tree Methods 0.000 claims abstract description 37
- 230000002159 abnormal effect Effects 0.000 claims abstract description 36
- 238000013499 data model Methods 0.000 claims description 71
- 238000004422 calculation algorithm Methods 0.000 claims description 22
- 238000012795 verification Methods 0.000 claims description 12
- 238000005457 optimization Methods 0.000 claims description 10
- 238000004458 analytical method Methods 0.000 claims description 6
- 238000009434 installation Methods 0.000 claims description 5
- 238000012545 processing Methods 0.000 abstract description 10
- 238000007635 classification algorithm Methods 0.000 description 11
- 238000012360 testing method Methods 0.000 description 8
- 230000008901 benefit Effects 0.000 description 7
- 230000005856 abnormality Effects 0.000 description 6
- 230000008569 process Effects 0.000 description 5
- 238000013145 classification model Methods 0.000 description 4
- 230000007547 defect Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000005611 electricity Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000012423 maintenance Methods 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000012790 confirmation Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000013138 pruning Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000033772 system development Effects 0.000 description 1
- 238000013024 troubleshooting Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Human Resources & Organizations (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Water Supply & Treatment (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a method, a device and a system for predicting and classifying faults of an intelligent electric energy meter, and belongs to the technical field of intelligent electric energy. The method utilizes the electric energy meter fault historical data as prediction sample data to carry out decision tree and support vector machine self-supervision learning training, and obtains a model reaching certain accuracy. And then, carrying out suspicious abnormal data classification on the newly acquired electric energy meter data by using a decision tree model and a support vector machine, and predicting the fault type. By applying the system, intelligent abnormity prediction and classification of the intelligent electric energy meter can be realized, the fault of the terminal equipment can be recognized as early as possible, the fault type can be positioned, the time for processing the fault of the electric power terminal equipment can be shortened, and the fault solving efficiency can be improved.
Description
Technical Field
The invention relates to the technical field of intelligent electric energy, in particular to a method, a device and a system for predicting and classifying faults of an intelligent electric energy meter.
Background
The power industry is the basic industry of national economy in China, the running condition of the power industry is related to the development conditions of energy safety and national economy in China, and a complete information management and intelligent state monitoring service system is an important guarantee for efficient and orderly operation of a power system. However, the current user side power distribution network has a complex operation environment, and the intelligent electric energy meter has various faults, so that the fault phenomenon and the reason analysis process are more and more difficult. In many areas, faults of the electric energy meters are still solved by periodically sampling and inspecting the electric energy meters and alternately replacing unqualified meters, the troubleshooting mode is low in efficiency, many hidden problems are not found timely, and serious challenges are brought to the safety, stability and controllability of power utilization of users.
In order to guarantee safe operation of a power grid and stable power supply, the fault of terminal equipment is predicted, the fault type is positioned, the time for processing the fault of the power terminal equipment is shortened, and the efficiency of solving the fault is improved, so that the problem that the resident power supply needs to be solved urgently by technical personnel in the field is solved.
Disclosure of Invention
In order to solve the problems in the prior art, the embodiment of the invention provides a method, a device and a system for predicting and classifying faults of an intelligent electric energy meter. The technical scheme is as follows:
in a first aspect, a method for predicting and classifying faults of an intelligent electric energy meter is provided, and the method includes:
setting a fault prediction and classification data model;
acquiring historical fault data of the intelligent electric energy meter, and generating prediction sample data after determining the fault type;
training the fault prediction data model through the prediction sample data to obtain a trained fault prediction data model;
collecting electric energy meter data, and performing fault prediction and abnormal data classification on the electric energy meter data through the prediction data model;
and continuously optimizing the fault prediction and classification data model according to the prediction result and the classification result.
Optionally, the setting the fault prediction data model further includes:
and setting the fault prediction data model according to a decision tree model and a Support Vector Machine (SVM) algorithm.
Optionally, after acquiring historical fault data of the intelligent electric energy meter, the method further includes:
deleting repeated values and noise points from historical fault data, and then complementing missing items by using empirical values;
the historical fault data comprises electric energy meter load data, electric energy quality data, event acquisition data, electric energy meter version, running time, installation place and abnormal phenomena.
Optionally, after generating the prediction sample data, the method further includes
Setting 60% of prediction sample data as training samples;
setting 40% of prediction sample data as verification data, wherein the verification data is used for verifying whether the trained fault prediction data model meets the precision requirement through cross verification.
Optionally, the method further includes:
and displaying the fault prediction and classification data model, and displaying a prediction result and a splitting result.
Optionally, the continuously optimizing the fault prediction and classification data model according to the prediction result and the classification result includes:
after confirming the prediction result and the classification result on site by maintenance personnel, adding the confirmed prediction result and classification result as abnormal characteristic data to the prediction sample data;
and continuously optimizing the fault prediction and classification data model through the prediction sample data.
In a second aspect, a method and a device for predicting and classifying faults of an intelligent electric energy meter are provided,
the setting module is used for setting a fault prediction and classification data model;
the sample generation module is used for acquiring historical fault data of the intelligent electric energy meter and generating prediction sample data after the fault type is determined;
the training module is used for training the fault prediction data model through the prediction sample data to obtain a trained fault prediction data model;
the prediction classification module is used for acquiring electric energy meter data, and performing fault prediction and abnormal data classification on the electric energy meter data through the prediction data model;
and the optimization module is used for continuously optimizing the fault prediction and classification data model according to the prediction result and the classification result.
Optionally, the setting module is specifically configured to:
and setting the fault prediction data model according to a decision tree model and a Support Vector Machine (SVM) algorithm.
In a third aspect, an intelligent electric energy meter fault prediction and classification device is provided, where the device includes a processor and a memory, the memory is used to store a set of program codes, and the processor calls the program codes stored in the memory to execute the method according to any one of the first aspect.
In a fourth aspect, a system for predicting and classifying faults of an intelligent electric energy meter is provided, the system comprising:
the setting module is used for setting a fault prediction and classification data model;
the sample generation module is used for acquiring historical fault data of the intelligent electric energy meter and generating prediction sample data after the fault type is determined;
the training module is used for training the fault prediction data model through the prediction sample data to obtain a trained fault prediction data model;
the prediction classification module is used for acquiring electric energy meter data, and performing fault prediction and abnormal data classification on the electric energy meter data through the prediction data model;
and the optimization module is used for continuously optimizing the fault prediction and classification data model according to the prediction result and the classification result.
The technical scheme provided by the embodiment of the invention has the following beneficial effects:
by adopting decision trees combined with SVM classification algorithms in the aspect of classification algorithms, the advantages of each algorithm are fully exerted, the defect of a single classification algorithm is avoided, and the calculation capability and the prediction accuracy of the system are improved;
and (4) performing self-supervision learning training on a decision tree and a support vector machine by using the electric energy meter fault historical data as prediction sample data to obtain a model reaching certain accuracy. And then, carrying out suspicious abnormal data classification on the newly acquired electric energy meter data by using a decision tree model and a support vector machine, and predicting the fault type. By applying the system, intelligent abnormity prediction and classification of the intelligent electric energy meter can be realized, the fault of the terminal equipment can be recognized as early as possible, the fault type can be positioned, the time for processing the fault of the electric power terminal equipment can be shortened, and the fault solving efficiency can be improved.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
FIG. 1 is a schematic diagram of a model construction provided by an embodiment of the invention;
FIG. 2 is a diagram illustrating an example SVM model result provided by an embodiment of the present invention;
fig. 3 is a schematic diagram of a model according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In a first aspect, a method for predicting and classifying faults of an intelligent electric energy meter is provided, and the method includes:
setting a fault prediction and classification data model;
acquiring historical fault data of the intelligent electric energy meter, and generating prediction sample data after determining the fault type;
training the fault prediction data model through the prediction sample data to obtain a trained fault prediction data model;
collecting electric energy meter data, and performing fault prediction and abnormal data classification on the electric energy meter data through the prediction data model;
and continuously optimizing the fault prediction and classification data model according to the prediction result and the classification result.
Optionally, the setting the fault prediction data model further includes:
and setting the fault prediction data model according to a decision tree model and a Support Vector Machine (SVM) algorithm.
Optionally, after obtaining the historical fault data of the intelligent electric energy meter, the method further includes:
deleting repeated values and noise points from historical fault data, and then complementing missing items by using empirical values;
the historical fault data comprises electric energy meter load data, electric energy quality data, event acquisition data, electric energy meter version, running time, installation place and abnormal phenomena.
Optionally, after generating the prediction sample data, the method further includes
Setting 60% of prediction sample data as training samples;
setting 40% of prediction sample data as verification data, wherein the verification data is used for verifying whether the trained fault prediction data model meets the precision requirement through cross verification.
Optionally, the method further includes:
and displaying the fault prediction and classification data model, and displaying a prediction result and a splitting result.
Optionally, the continuously optimizing the fault prediction and classification data model according to the prediction result and the classification result includes:
after confirming the prediction result and the classification result on site by a maintainer, adding the confirmed prediction result and classification result as abnormal characteristic data to the prediction sample data;
and continuously optimizing the fault prediction and classification data model through the prediction sample data.
In a second aspect, a method and apparatus for predicting and classifying faults of an intelligent electric energy meter are provided,
the setting module is used for setting a fault prediction and classification data model;
the intelligent electric energy meter comprises a sample generation module, a fault analysis module and a fault analysis module, wherein the sample generation module is used for acquiring historical fault data of the intelligent electric energy meter and generating prediction sample data after determining the fault type;
the training module is used for training the fault prediction data model through the prediction sample data to obtain a trained fault prediction data model;
the prediction classification module is used for acquiring electric energy meter data, and performing fault prediction and abnormal data classification on the electric energy meter data through the prediction data model;
and the optimization module is used for continuously optimizing the fault prediction and classification data model according to the prediction result and the classification result.
Optionally, the setting module is specifically configured to:
and setting the fault prediction data model according to a decision tree model and a Support Vector Machine (SVM) algorithm.
In a third aspect, an intelligent electric energy meter fault predicting and classifying device is provided, where the device includes a processor and a memory, the memory is used for storing a set of program codes, and the processor calls the program codes stored in the memory to execute the method according to any one of the first aspect.
In a fourth aspect, a system for predicting and classifying faults of an intelligent electric energy meter is provided, the system comprising:
the setting module is used for setting a fault prediction and classification data model;
the sample generation module is used for acquiring historical fault data of the intelligent electric energy meter and generating prediction sample data after the fault type is determined;
the training module is used for training the fault prediction data model through the prediction sample data to obtain a trained fault prediction data model;
the prediction classification module is used for acquiring electric energy meter data, and performing fault prediction and abnormal data classification on the electric energy meter data through the prediction data model;
and the optimization module is used for continuously optimizing the fault prediction and classification data model according to the prediction result and the classification result.
The technical scheme provided by the embodiment of the invention has the following beneficial effects:
by adopting decision trees combined with SVM classification algorithm in the aspect of classification algorithm, the advantages of each algorithm are fully exerted, the defect of a single classification algorithm is avoided, and the computing power and the prediction accuracy of the system are improved;
and performing decision tree and support vector machine self-supervision learning training by using the electric energy meter fault historical data as prediction sample data to obtain a model reaching certain accuracy. And then, carrying out suspicious abnormal data classification on the newly acquired electric energy meter data by using a decision tree model and a support vector machine, and predicting the fault type. By applying the system, intelligent abnormity prediction and classification of the intelligent electric energy meter can be realized, the fault of the terminal equipment can be recognized as early as possible, the fault type can be positioned, the time for processing the fault of the electric power terminal equipment can be shortened, and the fault solving efficiency can be improved.
To further illustrate the methods described in the examples of the present invention, further,
in order to achieve the purpose, the method and the system for classifying the collected data of the electric energy meter are provided, a decision tree is combined with an SVM classification algorithm in the aspect of classification algorithm, the advantages of each algorithm are fully exerted, the defect of a single classification algorithm is overcome, and the computing capacity and the prediction accuracy of the system are improved.
The system is based on an electric energy meter data acquisition system and is used for collecting electric energy meter abnormal events acquired on line, abnormal data detected by the system, fault data confirmed on site and the like, and the method mainly comprises the following steps:
determining fault type and predicting data
Historical data of the electric energy meter at the fault time period is extracted from the electric energy meter data acquisition system, the fault data is preprocessed, and repeated data and noise points in the fault data are removed. Completing the missing items by using the empirical values to obtain a prediction data set;
optionally, the load data, the power quality data, the event acquisition data, the version, the running time, the installation place and the abnormal phenomenon of the electric energy meter acquired by the system are used as input for analyzing the training of the abnormal model of the electric energy meter.
Analyzing the prediction data set, summarizing the characteristic data of the fault type of the electric energy meter, determining the corresponding relation between the fault type and the characteristic data, and determining the evaluation standard of the prediction result of the algorithm model according to the summarized corresponding relation.
Model training
60% of the data in the prediction dataset is extracted for creating training samples for the decision tree model, the rest of the data is used as the test dataset, the training samples represent categories by classification attributes of the training samples, and each record belongs to a determined category. And training by using a decision tree C5.0 algorithm to obtain a decision tree reaching the lowest accuracy, then predicting a test set by using the test set, comparing a classification result with an actual fault type, and analyzing the prediction accuracy of the decision tree model.
After the training of the decision tree model is finished, pruning can be carried out on the decision tree, and the node subtrees with insufficient classification precision are replaced by leaf nodes, so that overfitting of the decision tree model is avoided. The decision model generated by the decision tree is simple and visual, the structure is clear, the time consumption of the classification process is low, the continuous value processing is supported, and the visual processing of the classification result is convenient. However, the analysis and identification precision is not sufficient in the aspect of jointly deciding faults by multi-dimensional characteristic data, for example, the electricity stealing of a user needs the identification of the multi-characteristic data such as the times of opening the meter cover of the electric energy meter, the current of a zero line, the electricity utilization trend and the like.
The SVM algorithm is good, the classification precision is high, after most faults are separated through the decision tree, the remaining faults are subjected to one-by-one two-classification model training processing on the faults of the multi-dimensional data decision by the SVM, and the time consumption of the SVM classification algorithm is greatly reduced.
Specifically, the system adopts a Support Vector Machine (SVM) to carry out two-classification on the prediction data set, and because the SVM can process a large-scale feature space, the interaction effect between nonlinear features is obvious, the whole data does not need to be relied on, and the like, and normal and abnormal data two-classification model training is carried out on the prediction data set data. The system extracts data collected by the electric energy meter in a fault time period and data collected by the electric energy meter in normal operation and used for creating a training sample of the model, respectively sets labels on a normal data sample and an abnormal sample data sample, and performs sample data training by using an SVM algorithm to obtain an optimal hyperplane equation. And then, performing model verification by using the test data set, comparing the result set with an actual value, adjusting algorithm parameters, and continuously optimizing the data model to obtain the SVM model according with accuracy.
And performing exception classification and fault prediction on newly acquired data through a decision tree and an SVM (support vector machine) by storing and using the data model.
And after a group of models with good performance are obtained, the models are stored for use. The method comprises the steps of reading data models of real-time electric energy meter data collected by a system, predicting most of abnormalities through a decision tree model, detecting complex abnormalities such as electricity stealing of users in current data by using a Support Vector Machine (SVM) data model, and returning the detected data with faults to the system. And then generating work order information in the system, and sending early warning information and a processing scheme to system maintenance personnel to rush to the system as soon as possible to confirm and solve the field problems.
Predictive model optimization
And after field fault confirmation processing, inputting new fault characteristic data into an abnormal library, continuously enriching an abnormal data prediction data set of the electric meter, and optimizing a data prediction model through continuous training of the model. Meanwhile, on-site solutions are input into the system, and a fault solution knowledge base is enriched.
The core of the application is to provide a method for analyzing the data collected by the mass electric energy meters based on the combination of a neural network and a graphical algorithm, and help system maintenance personnel to solve related problems as early as possible before the influence is really generated due to abnormality, so that the stable operation of a power grid system is promoted. Meanwhile, the decision tree algorithm is combined with the SVM, so that the advantages of clear decision tree structure, less time consumption in the classification process and easiness in understanding are brought into play. The SVM has the advantages of two classifications and the advantage of classifying multi-dimensional characteristic data, and the defects that the SVM is used independently for long time and decision logic is difficult to show are overcome. Meanwhile, the system realizes a corresponding relation library of fault types, fault data characteristics and solutions, and provides support for continuous optimization of the system and treatment of field problems.
The Python language is adopted for system development, a machine learning library provided by a third party is used, and model training sample data can read the database and also support reading of file data.
Predictive data cleansing
Data cleansing deals mainly with the following data.
Removing noise points of abnormal data suddenly generated due to abnormality of the acquisition equipment;
removing repeated data;
eliminating abrupt change data of the user power consumption caused by field meter disassembly and meter replacement;
completing the missing items by using empirical values to obtain a standard prediction data set;
model training
And the data model training module uses a Python machine learning sklern packet to introduce an SVM module and a tree module.
The load data, the power quality data, the event acquisition data, the version, the running time, the installation place and the abnormal phenomenon of the electric energy meter acquired by the system are used as input for analyzing the training of the abnormal model of the electric energy meter.
Analyzing the prediction data set, summarizing characteristic data of the fault type of the electric energy meter, determining the corresponding relation between the fault type and the characteristic data, and determining the summarized corresponding relation as an important evaluation standard of the prediction result of the algorithm model.
60% of data in the prediction data set is extracted to be used for creating a training sample of the decision tree model, the rest data is used as a test data set, the training sample represents a category by using classification attributes, and each record belongs to a determined category. The decision tree C5.0 algorithm is used for training, and the decision tree C5.0 algorithm adopts the information gain rate to select the features, so that the feature selection subjectively biased to a plurality of sub-categories is weakened. The algorithm formula is as follows:
step1 information entropy value
Step2 information gain g (D, A) ═ H (D) — H (D | A)
Step3 information gain Rate
And performing cross validation by using the residual 40% of data in the prediction data set as a test set, and calculating the classification accuracy of the decision tree by a program. And storing the model with the accuracy of more than 80%, comparing the classification result with the actual fault type, and analyzing and judging the prediction accuracy of the decision tree model. And comparing the result set with the actual value to obtain the optimal decision tree model.
The SVM binary classification algorithm needs data acquired during fault and data acquired by a normally-operated electric energy meter to perform model training, partial data are taken from a historical database to serve as a prediction data set, 60% of electric energy meter data are extracted to serve as training samples, a label is set to be 1 for a normal data sample, and a label is set to be-1 for an abnormal sample data sample.
Generalizing from a two-dimensional plane to a hyperplane, and separating normal data from abnormal data to obtain a hyperplane equation f (x) w*·x+b*. According to the formula of the distance from a point to a hyperplaneThe shortest distance formula is solved and the distance is calculated,
get w*x+b*The minimum value is 1, the optimal distance formula is
The electric energy meter has more data collection dimensions, the training data is not linearly separable, the current linearly inseparable data is mapped to a high-dimensional space by combining a relaxation variable with a kernel function so as to achieve the purpose of linearly separable, and the classification correctness of the SVM is improved by repeatedly adjusting a normalized expression and the kernel function, as shown in an exemplary diagram of a classification result by using a Gaussian kernel in FIG. 2.
And then, performing test set verification by using the residual historical records, and comparing the result prediction result with an actual value to obtain the SVM model with the accuracy of more than or equal to 80%.
A large amount of training data are classified purposefully through a decision tree and a Support Vector Machine (SVM) algorithm, and valuable and potential abnormal information is found from the training data for classification.
And carrying out visual mapping on the model training result by using a pyplot module in a matplotlib package and a grapeviz module in a pydotplus package.
Fault classification and prediction
Before the fault classification model is used for prediction, collected data are pre-screened to identify deterministic anomalies, and the data can be identified without a neural network algorithm, wherein the fault classification model comprises the following contents:
communication abnormity, namely abnormity of unsuccessful data acquisition, incomplete data and the like
The abnormality is easy to recognize: the abnormal conditions which can be confirmed by means of simple calculation, data comparison and the like, such as unequal sum of the total display number and each rate, data backward walking and the like.
And storing the decision tree model and the SVM training model which pass the test set verification, reading the corresponding models when in use, obtaining an abnormal data list by passing the latest acquired data through the decision tree model and the SVM model, and pushing related responsible persons in time by the system in a mode of sending short messages, mails and the like to confirm and process the abnormality as soon as possible.
4. Model optimization
Maintaining the corresponding relation among the fault types, the fault characteristic data and the fault solutions in the system database, recording new fault characteristic data into an abnormal database after field fault confirmation processing, continuously enriching an ammeter abnormal data prediction data set, and simultaneously recording field solutions into the system to perfect the fault solutions. And continuously training the model, and optimizing the data prediction model.
All the above-mentioned optional technical solutions can be combined arbitrarily to form the optional embodiments of the present invention, and are not described herein again.
It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program instructing relevant hardware, where the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (10)
1. A fault prediction and classification method for an intelligent electric energy meter is characterized by comprising the following steps:
setting a fault prediction and classification data model;
acquiring historical fault data of the intelligent electric energy meter, and generating prediction sample data after determining the fault type;
training the fault prediction data model through the prediction sample data to obtain a trained fault prediction data model;
collecting electric energy meter data, and performing fault prediction and abnormal data classification on the electric energy meter data through the prediction data model;
and continuously optimizing the fault prediction and classification data model according to the prediction result and the classification result.
2. The method of claim 1, wherein said setting a fault prediction data model further comprises:
and setting the fault prediction data model according to a decision tree model and a Support Vector Machine (SVM) algorithm.
3. The method according to claim 2, wherein after acquiring the historical fault data of the intelligent electric energy meter, the method further comprises:
deleting repeated values and noise points from historical fault data, and then complementing missing items by using empirical values;
the historical fault data comprises electric energy meter load data, electric energy quality data, event acquisition data, electric energy meter version, running time, installation place and abnormal phenomena.
4. The method of claim 3, wherein after generating the prediction sample data, the method further comprises
Setting 60% of prediction sample data as training samples;
setting 40% of prediction sample data as verification data, wherein the verification data is used for verifying whether the trained fault prediction data model meets the precision requirement through cross verification.
5. The method of claims 1 to 4, further comprising:
and displaying the fault prediction and classification data model, and displaying a prediction result and a splitting result.
6. The method of claims 1 to 4, wherein said continuously optimizing said fault prediction and classification data model based on prediction results and classification results comprises:
after confirming the prediction result and the classification result on site by a maintainer, adding the confirmed prediction result and classification result as abnormal characteristic data to the prediction sample data;
and continuously optimizing the fault prediction and classification data model through the prediction sample data.
7. A method and a device for predicting and classifying faults of an intelligent electric energy meter are characterized in that,
the setting module is used for setting a fault prediction and classification data model;
the sample generation module is used for acquiring historical fault data of the intelligent electric energy meter and generating prediction sample data after the fault type is determined;
the training module is used for training the fault prediction data model through the prediction sample data to obtain a trained fault prediction data model;
the prediction classification module is used for acquiring electric energy meter data, and performing fault prediction and abnormal data classification on the electric energy meter data through the prediction data model;
and the optimization module is used for continuously optimizing the fault prediction and classification data model according to the prediction result and the classification result.
8. The apparatus of claim 7, wherein the setup module is specifically configured to:
and setting the fault prediction data model according to a decision tree model and a Support Vector Machine (SVM) algorithm.
9. An intelligent electric energy meter fault prediction and classification method device, characterized in that the device comprises a processor and a memory, the memory is used for storing a group of program codes, and the processor calls the program codes stored in the memory to execute the method of any one of claims 1 to 6.
10. An intelligent electric energy meter fault prediction and classification system, characterized in that the system comprises:
the setting module is used for setting a fault prediction and classification data model;
the intelligent electric energy meter comprises a sample generation module, a fault analysis module and a fault analysis module, wherein the sample generation module is used for acquiring historical fault data of the intelligent electric energy meter and generating prediction sample data after determining the fault type;
the training module is used for training the fault prediction data model through the prediction sample data to obtain a trained fault prediction data model;
the prediction classification module is used for acquiring electric energy meter data, and performing fault prediction and abnormal data classification on the electric energy meter data through the prediction data model;
and the optimization module is used for continuously optimizing the fault prediction and classification data model according to the prediction result and the classification result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210097394.1A CN114626433A (en) | 2022-01-27 | 2022-01-27 | Fault prediction and classification method, device and system for intelligent electric energy meter |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210097394.1A CN114626433A (en) | 2022-01-27 | 2022-01-27 | Fault prediction and classification method, device and system for intelligent electric energy meter |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114626433A true CN114626433A (en) | 2022-06-14 |
Family
ID=81899043
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210097394.1A Pending CN114626433A (en) | 2022-01-27 | 2022-01-27 | Fault prediction and classification method, device and system for intelligent electric energy meter |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114626433A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114781760A (en) * | 2022-06-17 | 2022-07-22 | 四川观想科技股份有限公司 | Fault prediction method based on big data |
| CN118091283A (en) * | 2024-02-27 | 2024-05-28 | 青岛乾程科技股份有限公司 | Smart electric meter fault detection method based on electric energy quality analysis |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106054104A (en) * | 2016-05-20 | 2016-10-26 | 国网新疆电力公司电力科学研究院 | Intelligent ammeter fault real time prediction method based on decision-making tree |
| CN106503439A (en) * | 2016-10-21 | 2017-03-15 | 国网福建省电力有限公司 | A kind of method of the collection fault early warning system based on data mining |
| WO2017136489A1 (en) * | 2016-02-03 | 2017-08-10 | Caspo, Llc | Smart cooking system that produces and uses hydrogen fuel |
| CN110059357A (en) * | 2019-03-19 | 2019-07-26 | 中国电力科学研究院有限公司 | A kind of intelligent electric energy meter failure modes detection method and system based on autoencoder network |
| CN110794360A (en) * | 2019-10-22 | 2020-02-14 | 中国电力科学研究院有限公司 | Method and system for predicting fault of intelligent electric energy meter based on machine learning |
| CN111612019A (en) * | 2020-05-15 | 2020-09-01 | 国网河北省电力有限公司电力科学研究院 | Method for identifying and analyzing fault abnormality of intelligent electric meter based on big data model |
| CN112464995A (en) * | 2020-11-06 | 2021-03-09 | 广东电网有限责任公司东莞供电局 | Power grid distribution transformer fault diagnosis method and system based on decision tree algorithm |
| CN112966879A (en) * | 2021-04-02 | 2021-06-15 | 阳光电源股份有限公司 | Environmental test chamber fault prediction method and device, computer equipment and storage medium |
| CN113011530A (en) * | 2021-04-29 | 2021-06-22 | 国网新疆电力有限公司营销服务中心(资金集约中心、计量中心) | Intelligent ammeter fault prediction method based on multi-classifier fusion |
| CN113420849A (en) * | 2021-08-25 | 2021-09-21 | 深圳市信润富联数字科技有限公司 | Model online incremental training method, device and medium based on active learning |
| CN113792825A (en) * | 2021-11-17 | 2021-12-14 | 国网江苏省电力有限公司营销服务中心 | Fault classification model training method and device for electricity information acquisition equipment |
| CN113962299A (en) * | 2021-10-15 | 2022-01-21 | 上海核工程研究设计院有限公司 | Intelligent operation monitoring and fault diagnosis general model for nuclear power equipment |
-
2022
- 2022-01-27 CN CN202210097394.1A patent/CN114626433A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017136489A1 (en) * | 2016-02-03 | 2017-08-10 | Caspo, Llc | Smart cooking system that produces and uses hydrogen fuel |
| CN106054104A (en) * | 2016-05-20 | 2016-10-26 | 国网新疆电力公司电力科学研究院 | Intelligent ammeter fault real time prediction method based on decision-making tree |
| CN106503439A (en) * | 2016-10-21 | 2017-03-15 | 国网福建省电力有限公司 | A kind of method of the collection fault early warning system based on data mining |
| CN110059357A (en) * | 2019-03-19 | 2019-07-26 | 中国电力科学研究院有限公司 | A kind of intelligent electric energy meter failure modes detection method and system based on autoencoder network |
| CN110794360A (en) * | 2019-10-22 | 2020-02-14 | 中国电力科学研究院有限公司 | Method and system for predicting fault of intelligent electric energy meter based on machine learning |
| CN111612019A (en) * | 2020-05-15 | 2020-09-01 | 国网河北省电力有限公司电力科学研究院 | Method for identifying and analyzing fault abnormality of intelligent electric meter based on big data model |
| CN112464995A (en) * | 2020-11-06 | 2021-03-09 | 广东电网有限责任公司东莞供电局 | Power grid distribution transformer fault diagnosis method and system based on decision tree algorithm |
| CN112966879A (en) * | 2021-04-02 | 2021-06-15 | 阳光电源股份有限公司 | Environmental test chamber fault prediction method and device, computer equipment and storage medium |
| CN113011530A (en) * | 2021-04-29 | 2021-06-22 | 国网新疆电力有限公司营销服务中心(资金集约中心、计量中心) | Intelligent ammeter fault prediction method based on multi-classifier fusion |
| CN113420849A (en) * | 2021-08-25 | 2021-09-21 | 深圳市信润富联数字科技有限公司 | Model online incremental training method, device and medium based on active learning |
| CN113962299A (en) * | 2021-10-15 | 2022-01-21 | 上海核工程研究设计院有限公司 | Intelligent operation monitoring and fault diagnosis general model for nuclear power equipment |
| CN113792825A (en) * | 2021-11-17 | 2021-12-14 | 国网江苏省电力有限公司营销服务中心 | Fault classification model training method and device for electricity information acquisition equipment |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114781760A (en) * | 2022-06-17 | 2022-07-22 | 四川观想科技股份有限公司 | Fault prediction method based on big data |
| CN118091283A (en) * | 2024-02-27 | 2024-05-28 | 青岛乾程科技股份有限公司 | Smart electric meter fault detection method based on electric energy quality analysis |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111738462B (en) | Fault first-aid repair active service early warning method for electric power metering device | |
| CN112199367B (en) | Long-time scale power balance check and big data analysis method and system | |
| CN110930198A (en) | Electric energy substitution potential prediction method and system based on random forest, storage medium and computer equipment | |
| CN108537394B (en) | Real-time safety early warning method and device for smart power grid | |
| CN109409444B (en) | Multivariate power grid fault type discrimination method based on prior probability | |
| CN111539585B (en) | Random forest-based power customer appeal sensitivity supervision and early warning method | |
| CN114626433A (en) | Fault prediction and classification method, device and system for intelligent electric energy meter | |
| CN108304567B (en) | Method and system for identifying working condition mode and classifying data of high-voltage transformer | |
| CN108345670B (en) | Service hotspot discovery method for 95598 power work order | |
| KR20230030542A (en) | AI-based facility data anomaly detection system and method using random cut forest algorithm | |
| CN115760484A (en) | Method, device and system for improving potential danger identification capability of power distribution station area and storage medium | |
| CN111126820A (en) | Electricity stealing prevention method and system | |
| CN111612149A (en) | Main network line state detection method, system and medium based on decision tree | |
| CN114201374A (en) | Operation and maintenance time sequence data anomaly detection method and system based on hybrid machine learning | |
| CN116739829B (en) | Big data-based power data analysis method, system and medium | |
| Xue et al. | Adaptive ultra-short-term wind power prediction based on risk assessment | |
| CN117390546A (en) | Multimode database fusion calculation model for instant anti-electricity-theft detection | |
| CN110807014B (en) | Cross validation based station data anomaly discrimination method and device | |
| CN115859191A (en) | Fault diagnosis method and device, computer readable storage medium and computer equipment | |
| CN110968703A (en) | Method and system for constructing abnormal metering point knowledge base based on LSTM end-to-end extraction algorithm | |
| CN114169998A (en) | Financial big data analysis and mining algorithm | |
| CN112595918A (en) | Low-voltage meter reading fault detection method and device | |
| CN116452054A (en) | Method and device for managing material spot check of electric power system | |
| CN112819182B (en) | Training method, device and equipment for banknote box state prediction model | |
| CN115147242A (en) | Power grid data management system based on data mining |
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 |