CN115017971A

CN115017971A - Self-learning-based energy consumption meter for power utilization abnormity diagnosis

Info

Publication number: CN115017971A
Application number: CN202210448455.4A
Authority: CN
Inventors: 刘洋; 刘伟; 王义汉
Original assignee: Hefei Jinren Technology Co ltd
Current assignee: Hefei Jinren Technology Co ltd
Priority date: 2022-04-26
Filing date: 2022-04-26
Publication date: 2022-09-06

Abstract

The invention discloses an energy consumption meter for power utilization abnormity diagnosis based on self-learning, and relates to the technical field of energy consumption management. The system comprises an information storage module, an information preprocessing module, a learning training module, a judging module and an early warning module, wherein the information storage module is used for preprocessing historical electricity utilization original information of a user on the basis of an edge computing node to obtain optimized electricity utilization data; receiving optimized power consumption data, training a training model, and generating an XGboost detection model; inputting data to be detected into an XGboost detection model, and judging whether the electricity data is abnormal or not; the early warning module is connected with the judging module and feeds back abnormal information to the power worker. The invention utilizes the improved genetic algorithm to carry out the XGboost algorithm of the hyper-parameter optimization, carries out the detection of the abnormal behavior of the electricity consumption for the terminal user, has low equipment cost, is not easy to be interfered by the external environment, and improves the detection efficiency and the accuracy.

Description

Self-learning-based energy consumption meter for power utilization abnormity diagnosis

Technical Field

The invention belongs to the technical field of energy consumption management, and particularly relates to an energy consumption meter for diagnosing power consumption abnormity based on self-learning.

Background

With the development of power technology, smart meters have become popular, and as one of basic devices for collecting data of smart grids, smart meters have intelligent functions such as bidirectional multi-rate metering function, user side control function, bidirectional data communication function with multiple data transmission modes, electricity larceny prevention function and the like in addition to the metering function of basic electricity consumption of traditional electric energy meters, thereby bringing great convenience to the metering data collection.

The intelligent electric meter is used for detecting abnormal electricity consumption, and becomes an important means for researching abnormal consumption behaviors of customers and timely discovering unexpected electricity consumption events. During the operation of the power grid, no matter the metering device is in failure or the user steals electricity, the real electricity utilization data of the user cannot be collected, and the false electricity utilization data is called abnormal electricity utilization data. If the situations cannot be found and processed in time, serious interference and influence can be generated on the power supply order of normal users and power supply companies, so that economic loss is caused, and hidden dangers in the aspect of power supply and utilization safety are caused more seriously, so that the method has important practical significance for automatically monitoring abnormal power supply and utilization of the users.

Currently, monitoring abnormal electricity consumption behavior is mainly implemented by hardware: external monitoring devices such as a complex detection system and the like which are composed of a camera, a sensor and a networking device are used for monitoring whether power supply equipment is damaged or not and whether power utilization behaviors are normal or not in real time. However, the monitoring method requires high equipment cost, hardware equipment is easily interfered by external factors such as weather, equipment maintenance is difficult, and abnormal power utilization behaviors such as software power stealing and remote control power stealing are difficult to identify.

Disclosure of Invention

The invention aims to provide an energy consumption meter for diagnosing abnormal electricity consumption based on self-learning, which detects abnormal electricity consumption behaviors of a terminal user by using an XGboost algorithm for carrying out super-parameter optimization by using an improved genetic algorithm, and solves the problems that the conventional method for monitoring abnormal electricity consumption behaviors through hardware needs higher equipment cost, hardware equipment is easily interfered by external factors such as weather, equipment maintenance is difficult, and the abnormal electricity consumption behaviors such as software electricity stealing and remote control electricity stealing are difficult to identify.

In order to solve the technical problems, the invention is realized by the following technical scheme:

the invention relates to an energy consumption meter based on self-learning electricity utilization abnormity diagnosis, which comprises an information storage module and is used for storing historical electricity utilization information of a user. And the information preprocessing module is used for preprocessing the historical electricity utilization original information of the user based on the edge computing node to obtain optimized electricity utilization data. And the learning training module is used for receiving the optimized power utilization data, training the training model and generating the detection model. And the judging module is used for inputting the data to be detected into the detection model and judging whether the power utilization data is abnormal or not. And the early warning module is connected with the judging module and feeds back abnormal information to the electric power working personnel.

As a preferential technical scheme of the invention, the information stored by the information storage module comprises historical power consumption information of a user and abnormal power consumption information of terminal equipment.

As a preferential technical scheme of the invention, the information preprocessing module comprises data cleaning, missing value processing and data dimension reduction.

As a preferred technical solution of the present invention, the missing value processing in the information preprocessing module includes processing the missing value by using an expectation maximization interpolation method.

As a preferential technical scheme of the invention, the detection model adopts an XGboost model.

As a preferred technical scheme of the invention, the learning training module comprises a parameter optimization unit, and the parameter optimization unit performs parameter optimization on the XGboost model based on an improved genetic algorithm, so that the XGboost detection model is determined to be an optimal hyper-parameter combination; the judging module judges whether the electricity utilization data are abnormal or not based on the optimal hyper-parameter combination.

As a preferred embodiment of the present invention, the improved genetic algorithm package in the parameter optimization unitThe cross probability and the mutation probability containing exponential decay, wherein the expression of the cross probability containing exponential decay is as follows:

where pc is the crossover probability, f _ave Is the average value of the fitness of the population, f' is the fitness of the larger fitness of the two crossed individuals, a ₁ 、a ₂ Are all constant, and 0<a ₁ <1,0<a ₂ <1, beta is attenuation coefficient, and n is current genetic iteration number. The expression of the exponentially decaying mutation probability is:

pm is the variation probability, f is the fitness of the variant individual, b ₁ 、b ₂ Are all constant, and 0<b ₁ <1,0<b ₂ <1。

As a preferred technical solution of the present invention, the improved genetic algorithm in the parameter optimization unit further includes a fitness function, and an expression of the fitness function is:

wherein f is _fitness The method is a fitness function, omega is a weight coefficient, precision is an accuracy rate, recall is a recall rate, and k is the number of the hyper-parameters.

As a preferential technical scheme of the invention, the learning training module comprises a communication unit connected with an external server, and is used for remotely monitoring the operation condition of the system and carrying out manual processing and correction on abnormal information of the system.

As a preferential technical scheme of the invention, the meter is provided with the GPS positioner, so that the position of the terminal meter is convenient to inquire, and the electricity consumption abnormal condition is convenient to be checked by power workers.

The invention has the following beneficial effects:

according to the invention, the energy consumption meter based on self-learning electricity consumption abnormity diagnosis is designed, electricity consumption data of a user is locally stored, analyzed and processed, abnormal electricity consumption conditions are uploaded, an improved genetic algorithm is used for carrying out an XGboost algorithm with super-parameter optimization, abnormal electricity consumption behavior detection is carried out on a terminal user, the equipment cost is low, the interference of an external environment is not easy to occur, and the detection efficiency and the accuracy are improved.

Of course, it is not necessary for any product to practice the invention to achieve all of the above-described advantages at the same time.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings 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 that other drawings can be obtained according to the drawings without creative efforts.

FIG. 1 is a schematic diagram of a connection structure of modules inside an energy consumption meter based on self-learning power utilization abnormality diagnosis according to the present invention;

FIG. 2 is a flow chart of a self-learning power consumption anomaly diagnosis based energy consumption meter detection method according to the present invention;

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the 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 the description of the present invention, it is to be understood that the terms "opening," "upper," "lower," "thickness," "top," "middle," "length," "inner," "peripheral," and the like are used in an orientation or positional relationship that is merely for convenience in describing and simplifying the description, and do not indicate or imply that the referenced component or element must have a particular orientation, be constructed and operated in a particular orientation, and thus should not be considered as limiting the present invention.

Example one

Referring to fig. 1, the invention relates to an energy consumption meter based on self-learning power consumption abnormality diagnosis, which comprises an information storage module, an information preprocessing module, a learning training module, a judgment module and an early warning module. The detection steps of the electricity utilization information among the modules specifically comprise:

referring to fig. 2, in a first step, an information storage module stores historical electricity consumption information of a user, where the electricity consumption information includes historical electricity consumption information of the user and abnormal electricity consumption information of a terminal device.

And secondly, the information preprocessing module performs preprocessing such as data cleaning, missing value processing and data dimension reduction on historical electricity utilization original information of the user based on the edge computing node to obtain optimized electricity utilization data. And processing the missing value by adopting an expectation maximization interpolation method. The data dimensionality reduction process is used to normalize the user data, with a mean of 0 and a variance of 1. Calculating a covariance matrix, an eigenvalue of the covariance matrix and an eigenvector corresponding to the eigenvalue; sorting the eigenvalues according to the sizes, selecting the largest m of the eigenvalues, and taking the corresponding eigenvectors as row vectors respectively to form an eigenvector matrix. The user data is transformed into a new space consisting of m feature vectors. The data preprocessing method of data cleaning, missing value processing and data dimension reduction is adopted to carry out training sample construction on the original data collected by the terminal, so that the original data is more complete, the training data volume is greatly reduced, the training time and the detection accuracy of the model are improved, the dimension of the data is reduced by a principal component analysis method, new characteristic data without losing the information volume of the original data is generated, the calculation time is greatly shortened, and the detection capability of the model is improved.

And thirdly, receiving the optimized power utilization data by the learning training module, training the training model and generating an XGboost detection model. The learning training module comprises a parameter optimization unit, and the parameter optimization unit performs parameter optimization on the XGboost model based on an improved genetic algorithm so as to determine the optimal hyper-parameter combination of the XGboost detection model. The improved genetic algorithm process in the parameter optimization unit comprises a fitness function, and the expression of the fitness function is as follows:

In addition, in the present embodiment, the genetic algorithm improved in the parameter optimization unit includes exponentially decaying cross probability and mutation probability. And optimizing the hyper-parameters of the monitoring model by using the cross probability and the variation probability of the exponential decay. The cross probability expression is:

where pc is the crossover probability, f _ave Is the average value of the fitness of the population, f' is the fitness of the larger fitness of the two crossed individuals, a ₁ 、a ₂ Are all constant, and 0<a ₁ <1,0<a ₂ <1, beta is attenuation coefficient, and n is current genetic iteration number. The expression for the exponentially decaying mutation probability is:

pm is the variation probability, f is the fitness of the individual, b ₁ 、b ₂ Are all constant, and 0<b ₁ <1,0<b ₂ <1，f _ave And the fitness average value of the population. The optimization efficiency of the algorithm is improved, the cross probability and the variation probability are higher in the early stage of generation selection optimization, the cross probability and the variation probability are lower in the later stage of generation selection optimization, the improvement of the algorithm efficiency is facilitated, the values of two control parameters are reasonably designed, the optimal super-parameter combination of the XGboost detection model is found by the genetic algorithm, and the local optimal solution can be skipped to find the global optimal solution.

And the learning training module also comprises a communication unit connected with an external server and used for remotely monitoring the running condition of the system and periodically carrying out manual processing and correction on the abnormal information of the system.

And fourthly, inputting the data to be detected into the XGboost detection model, and judging whether the optimized electricity data is abnormal or not by the judgment module based on the optimal hyper-parameter combination.

The fifth step: if the judging module judges that the electricity utilization data is abnormal, the early warning module feeds back the abnormal electricity utilization data to the power staff to remind the staff to process and confirm the abnormal information.

In conclusion, the energy consumption meter based on self-learning electricity consumption abnormity diagnosis of the embodiment stores, analyzes and processes electricity consumption data of a user locally, uploads abnormal electricity consumption conditions, utilizes an improved genetic algorithm to perform an XGboost algorithm with optimized hyper-parameters, and performs electricity consumption abnormal behavior detection on a terminal user, thereby greatly improving accuracy and rapidity of detection.

Example two

Based on the first embodiment, the second embodiment is different in that:

install the GPS locator on the strapping table, when electric power staff verified the power consumption data to the scene, be convenient for find terminal strapping table position, promote work efficiency.

In the description herein, references to the description of "one embodiment," "an example," "a specific example," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

The preferred embodiments of the invention disclosed above are intended to be illustrative only. The preferred embodiments are not intended to be exhaustive or to limit the invention to the precise embodiments disclosed. Obviously, many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the invention and the practical application, to thereby enable others skilled in the art to best utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. An energy consumption meter based on self-learning electricity consumption abnormity diagnosis, which is characterized by comprising:

the information storage module is used for storing historical electricity utilization information of a user;

the information preprocessing module is used for preprocessing historical electricity utilization original information of the user based on the edge computing node to obtain optimized electricity utilization data;

the learning training module is used for receiving the optimized power utilization data, training the training model and generating a detection model;

the judging module is used for inputting the data to be detected into the detection model and judging whether the electricity utilization data are abnormal or not;

and the early warning module is connected with the judging module and feeds back abnormal information to the electric power working personnel.

2. The energy consumption meter based on self-learning power utilization abnormity diagnosis as recited in claim 1, wherein the information stored by the information storage module comprises historical power utilization information of a user and abnormal power utilization information of terminal equipment.

3. The energy consumption meter based on self-learning electricity abnormality diagnosis of claim 1, wherein the information preprocessing module comprises data cleaning, missing value processing and data dimension reduction.

4. The self-learning based power consumption abnormality diagnosis energy consumption meter according to claim 3, wherein the missing value processing in the information preprocessing module comprises processing the missing value by adopting an expectation maximization interpolation method.

5. The energy consumption meter based on self-learning electricity consumption abnormity diagnosis according to claim 1, wherein the detection model adopts an XGboost model.

6. The energy consumption meter based on self-learning electricity abnormality diagnosis is characterized in that the learning training module comprises a parameter optimization unit, and the parameter optimization unit performs parameter optimization on an XGboost model based on an improved genetic algorithm so as to determine that the XGboost detection model is the optimal super-parameter combination;

the judging module judges whether the electricity utilization data is abnormal or not based on the optimal hyper-parameter combination.

7. The energy consumption meter based on self-learning electricity abnormality diagnosis as claimed in claim 6, wherein the improved genetic algorithm in the parameter optimization unit comprises exponentially decaying cross probability and mutation probability, and the expression of the exponentially decaying cross probability is as follows:

where pc is the crossover probability, f _ave Is the average value of the fitness of the population, f' is the fitness of the larger fitness of the two crossed individuals, a ₁ 、a ₂ Are all constant, and 0<a ₁ <1,0<a ₂ <1, beta is an attenuation coefficient, and n is the current genetic iteration number;

the expression of the exponentially decaying mutation probability is:

pm is the variation probability, f is the fitness of the individual, b ₁ 、b ₂ Are all constant, and 0<b ₁ <1,0<b ₂ <1。

8. The energy consumption meter based on self-learning electricity consumption abnormity diagnosis as recited in claim 7, wherein the improved genetic algorithm in the parameter optimization unit further comprises a fitness function, and the expression of the fitness function is as follows:

9. The energy consumption meter based on self-learning electricity abnormality diagnosis as recited in claim 1, wherein the learning and training module comprises a communication unit connected with an external server.

10. The energy consumption meter based on self-learning electricity abnormality diagnosis as claimed in claim 1, wherein the meter is provided with a GPS locator.