Internet of Things integrated control system based on electric energy meter

Abstract

The invention relates to an Internet of Things integrated control system based on an electric energy meter, and belongs to the technical field of electric energy integrated control. The system comprises a signal acquisition module, a signal preprocessing module, a feature extraction module and an error detection module. The signal acquisition module is used for acquiring an electric energy signal through an electric energy meter and obtaining a lightning protection electric energy signal through the lightning protection module according to the electric energy signal; the signal preprocessing module is used for obtaining a complex dynamic electric energy signal data set through an electric energy signal preprocessing model according to the lightning protection electric energy signal; the feature extraction module is used for obtaining complex dynamic signal features through a dynamic feature extraction model according to the complex dynamic electric energy signal data set; the error detection module is used for obtaining abnormal features through an error detection model according to the dynamic signal features. The electric energy signal abnormity detection based on the electric energy meter is realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of electric energy integrated control, and in particular relates to an Internet of Things integrated control system based on an electric energy meter. Background Art

[0002] Traditionally, the operation management of the same batch of electric energy meters is generally to infer the operating status of the entire electric energy meter through a certain proportion of sampling inspection results; At present, with the promotion and application of remote data acquisition technology and the development of big data analysis technology, the accuracy of electric energy meter status monitoring technology is also constantly improving. The main research method is to remotely conduct quantitative and qualitative analysis on the data such as power consumption, current, voltage, harmonic frequency and other data before and after the fault of the faulty electric energy meter based on the data obtained by the electricity consumption information collection system, and determine the operating status and abnormal classification of the electric energy meter, including screening of abnormal metering electric energy meters based on regular non-continuous algorithm and classification continuous difference algorithm; establish an electric energy meter abnormal fault library based on k-means clustering analysis method to realize the judgment of various types of faults; realize fault information correlation matching based on data mining association analysis; These methods do not take into account the characteristics of complex dynamic electric energy signals when performing dynamic error detection of electric energy meters. Summary of the invention

[0003] In order to solve the above problems existing in the prior art, the present invention provides an Internet of Things integrated control system based on an electric energy meter. The purpose of the present invention can be achieved through the following technical solutions: An Internet of Things integrated control system based on an electric energy meter, comprising a signal acquisition module, a signal preprocessing module, a feature extraction module, and an error detection module; The signal acquisition module is used to collect electric energy signals through the electric energy meter, and obtain lightning protection electric energy signals through the lightning protection module according to the electric energy signals; The signal preprocessing module is used to obtain a complex dynamic electric energy signal data set according to the lightning protection electric energy signal through an electric energy signal preprocessing model; The feature extraction module is used to obtain complex dynamic signal features through a dynamic feature extraction model according to the complex dynamic electric energy signal data set; The error detection module is used to obtain abnormal features through an error detection model according to the dynamic signal features.

[0004] Preferably, the lightning protection module includes a communication signal lightning protection circuit and a power supply lightning protection circuit, the communication signal lightning protection circuit includes a primary signal protection unit, a No. 1 decoupling unit, a secondary signal protection unit and a No. 1 reactance module; the power supply lightning protection circuit includes a primary power supply protection unit, a No. 2 decoupling unit, a secondary power supply protection unit and a No. 2 reactance module.

[0005] Preferably, obtaining a complex dynamic electric energy signal data set through an electric energy signal preprocessing model according to the lightning protection electric energy signal comprises: Obtaining a lightning protection power energy signal data set by storing the lightning protection power energy signal in an ES cluster; According to the lightning protection electric energy signal data set, a cleaned lightning protection electric energy signal data set is obtained by data cleaning; According to the cleaned lightning protection electric energy signal data set, missing values ​​are supplemented by linear interpolation to obtain a complete cleaned lightning protection electric energy signal data set; The complete cleaning lightning protection power signal data set includes voltage signal, current signal, signal amplitude, and signal phase; A sampling frequency is preset, and the complex dynamic electric energy signal data set is obtained by constructing a complex dynamic electric energy signal model according to the voltage signal, the current signal, the signal amplitude, the signal phase, and the sampling frequency.

[0006] Preferably, obtaining complex dynamic signal features through a dynamic feature extraction model according to the complex dynamic power signal data set includes: According to the complex dynamic electric energy signal data set, a harmonic period phase compensation matrix is ​​obtained by improving the short-time Fourier transform algorithm model; The complex dynamic signal characteristics are obtained according to the harmonic period phase compensation matrix through a time-frequency domain global feature extraction model.

[0007] Preferably, obtaining a harmonic period phase compensation matrix by improving a short-time Fourier transform algorithm model according to the complex dynamic power signal data set comprises: Calculating the periodic signal length using a period inversion algorithm according to the complex dynamic electric energy signal data set; Obtaining the real part of the harmonics and the imaginary part of the harmonics by Fourier transform according to the complex dynamic electric energy signal data set and the periodic signal length; The harmonic period phase compensation matrix is ​​calculated according to the real part of the harmonic and the imaginary part of the harmonic by using a phase compensation algorithm.

[0008] Preferably, the calculating the periodic signal length by a period inversion algorithm according to the complex dynamic electric energy signal data set comprises: According to the complex dynamic electric energy signal data set, the signal is intercepted by the window function length to obtain a signal subsequence; Taking the final value of the signal subsequence as the search center, presetting the search radius, and obtaining the signal cycle end point search interval according to the search center and the search radius; According to the search interval at the end point of the signal cycle, a near-zero point is obtained by the zero value theorem; The periodic signal length is obtained by performing a difference calculation based on the near-zero point.

[0009] Preferably, the calculating the harmonic period phase compensation matrix according to the real part of the harmonic and the imaginary part of the harmonic by a phase compensation algorithm comprises: Obtaining a harmonic phase by phase calculation according to the real part of the harmonic and the imaginary part of the harmonic; Acquire the ordinate of the sampling point, and obtain the harmonic phase compensation value through secondary compensation calculation according to the ordinate of the sampling point and the harmonic phase; The harmonic periodic phase compensation matrix is ​​obtained by storing the harmonic phase compensation value in a matrix.

[0010] Preferably, the complex dynamic signal feature is obtained by using a global feature extraction model in the time-frequency domain according to the harmonic period phase compensation matrix, including: Compensating the complex dynamic electric energy signal data set according to the harmonic period phase compensation matrix to obtain complex dynamic electric energy compensation signal data; A time-frequency domain harmonic characteristic matrix is ​​preset, and the complex dynamic signal characteristics are obtained by feature extraction according to the complex dynamic electric energy compensation signal data and the time-frequency domain harmonic characteristic matrix.

[0011] Preferably, obtaining the abnormal feature through an error detection model according to the dynamic signal feature includes: S401: Acquire historical dynamic signal features, wherein the historical dynamic signal features carry an outlier label, preset a basic prediction value, and calculate a target function through a loss function according to the outlier label and the basic prediction value; S402: Calculate the optimal splitting point by second-order Taylor expansion and derivation according to the objective function, and construct a decision tree; S403: Obtaining an abnormal prediction value through the decision tree according to the historical dynamic signal characteristics; S404: Calculating an updated objective function according to the abnormal prediction value and the abnormal value label through the loss function; S405: According to the update objective function, the update optimal splitting point is calculated by the second-order Taylor expansion and derivation, and an update decision tree is constructed; S406: Preset an abnormal threshold, and obtain an abnormal difference value by performing a difference calculation based on the abnormal prediction value and the abnormal value label; S407: Obtaining an abnormality judgment decision tree through abnormality judgment according to the abnormal difference and the abnormal threshold; S408: Obtain abnormal features through the abnormality judgment decision tree according to the dynamic signal features.

[0012] Preferably, the abnormality judgment includes judging whether the abnormal difference is less than the abnormal threshold value, and if so, obtaining the abnormality judgment decision tree; if not, repeating the step S404 and the step S405 until the abnormal difference is less than the abnormal threshold value.

[0013] The beneficial effects of the present invention are: (1) The electric energy signal is protected by the communication signal lightning protection circuit and the power supply lightning protection circuit, which increases the application range of the lightning protection module and integrates it into one module, thereby reducing the volume of the lightning protection module, avoiding the influence of the environment on the electric energy signal, and improving the stability of the electric energy signal.

[0014] (2) Through the period inversion algorithm, Fourier transform algorithm and phase compensation algorithm, the periodic fluctuation and harmonic interference in the power signal are eliminated, thereby enhancing the robustness of the model in signal fluctuation scenarios.

[0015] (3) Through phase compensation and time-frequency domain global feature extraction model, the deep-level features of complex dynamic signals can be extracted, providing more comprehensive data support for subsequent error detection.

[0016] (4) Through the optimization mechanism of the decision tree model and the objective function, an anomaly judgment decision tree is constructed by gradually splitting and updating the splitting points to achieve high efficiency and dynamic adjustment capability of anomaly detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0018] Figure 1 The present invention is a flowchart of an Internet of Things integrated control system based on an electric energy meter. DETAILED DESCRIPTION

[0019] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects of the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments.

[0020] See also Figure 1 , an Internet of Things integrated control system based on an electric energy meter, including a signal acquisition module, a signal preprocessing module, a feature extraction module, and an error detection module: The signal acquisition module is used to collect electric energy signals through the electric energy meter, and obtain lightning protection electric energy signals through the lightning protection module according to the electric energy signals; The signal preprocessing module is used to obtain a complex dynamic electric energy signal data set according to the lightning protection electric energy signal through an electric energy signal preprocessing model; The feature extraction module is used to obtain complex dynamic signal features through a dynamic feature extraction model according to the complex dynamic electric energy signal data set; The error detection module is used to obtain abnormal features through an error detection model according to the dynamic signal features.

[0021] In step S1, the lightning protection module includes a communication signal lightning protection circuit and a power supply lightning protection circuit, the communication signal lightning protection circuit includes a primary signal protection unit, a No. 1 decoupling unit, a secondary signal protection unit and a No. 1 reactance module, one end of the primary signal protection unit serves as the input end of the communication signal lightning protection circuit, the output end of the primary signal protection unit is connected to the input end of the No. 1 decoupling unit, the output end of the No. 1 decoupling unit is connected to the input end of the secondary signal protection unit, the output end of the No. 2 signal protection unit is connected to the input end of the No. 1 reactance module, and the output end of the No. 1 reactance module serves as the output end of the communication signal lightning protection circuit. The power supply lightning protection circuit includes a primary power supply protection unit, a secondary decoupling unit, a secondary power supply protection unit and a secondary reactance module. One end of the primary power supply protection unit serves as the input end of the power supply lightning protection circuit, the output end of the primary power supply protection unit is connected to the input end of the secondary decoupling unit, the output end of the secondary decoupling unit is connected to the input end of the secondary power supply protection unit, the output end of the secondary power supply protection unit is connected to the input end of the secondary reactance module, and the output end of the secondary reactance module serves as the output end of the power supply lightning protection circuit.

[0022] In step S2, the electric energy signal preprocessing model includes: S201: obtaining a lightning protection power signal data set by storing the lightning protection power signal in an ES cluster; S202: obtaining a cleaned lightning protection power signal data set by performing data cleaning according to the lightning protection power signal data set; S203: supplementing missing values ​​by linear interpolation according to the cleaned lightning protection electric energy signal data set to obtain a complete cleaned lightning protection electric energy signal data set; S204: The complete cleaning lightning protection power signal data set includes a voltage signal, a current signal, a signal amplitude, and a signal phase; S205: Preset a sampling frequency, and obtain the complex dynamic power signal data set by constructing a complex dynamic power signal model according to the voltage signal, the current signal, the signal amplitude, the signal phase, and the sampling frequency; The complex dynamic electric energy signal model is constructed as follows: , in, Represents the data in the complex dynamic power signal dataset, ~ represents a non-stationary random signal, Indicates Sub-signal harmonic components, L is the harmonic component order of the signal, represents the signal amplitude, represents the signal phase, sin is the sine function calculation, represents the time domain sampling point, f 1 are the harmonic frequencies of the signal, f s is the sampling frequency, is Gaussian white noise.

[0023] In step S3, the dynamic feature extraction model includes: S301: Obtaining a harmonic period phase compensation matrix by improving a short-time Fourier transform algorithm model according to the complex dynamic power signal data set; S302: Obtain the complex dynamic signal features through a global feature extraction model in the time-frequency domain according to the harmonic period phase compensation matrix.

[0024] Specifically, in step S301, the improved short-time Fourier transform algorithm model includes: S301-1: Calculating the periodic signal length by using a period inversion algorithm according to the complex dynamic electric energy signal data set; S301-2: Obtaining a real harmonic part and an imaginary harmonic part by Fourier transform according to the complex dynamic electric energy signal data set and the periodic signal length; The Fourier transform expression is: , in, a(n,l) represents the real part of the harmonic, f 0 is the rated frequency of the power grid, E n is the starting sampling point, is the periodic signal length, Represents the data in the complex dynamic power signal dataset, cos Represents the cosine function calculation, represents the harmonic components, represents the time domain sampling point, h represents the window function calculation, j is an imaginary unit, b(n,l) represents the harmonic imaginary part, sin Represents the calculation of sine function; S301 - 3 : Calculate the harmonic period phase compensation matrix according to the real part of the harmonic and the imaginary part of the harmonic by using a phase compensation algorithm.

[0025] Specifically, the period inversion algorithm calculation includes: According to the complex dynamic electric energy signal data set, the signal is intercepted by the window function length to obtain a signal subsequence; The window function length expression is: , in, M s is the length of the window function, f 1 are the harmonic frequencies of the signal, f s is the sampling frequency; Taking the final value of the signal subsequence as the search center, presetting the search radius, and obtaining the signal cycle end point search interval according to the search center and the search radius; According to the search interval at the end point of the signal cycle, a near-zero point is obtained by the zero value theorem; The periodic signal length is obtained by performing a difference calculation based on the near-zero point.

[0026] Specifically, the phase compensation algorithm calculation includes: Obtaining a harmonic phase by phase calculation according to the real part of the harmonic and the imaginary part of the harmonic; The phase calculation expression is: , in, is the harmonic phase, arctan Indicates the calculation of the inverse tangent function, a(n,l) represents the real part of the harmonic, b(n,l) represents the imaginary part of the harmonic; Acquire the ordinate of the sampling point, and obtain the harmonic phase compensation value through secondary compensation calculation according to the ordinate of the sampling point and the harmonic phase; The quadratic compensation calculation expression is: , in, is the harmonic phase compensation value, , Represents the absolute value of the ordinate of the sampling point, , represents the harmonic phase of sampling point 1 and sampling point 2, | | represents absolute value calculation, 1 and 2 Indicates different sampling points; The harmonic period phase compensation matrix is ​​obtained by storing the harmonic phase compensation value in a matrix; The harmonic period phase compensation matrix is ​​expressed as: , in, is the harmonic period phase compensation matrix, is the harmonic phase compensation value, L represents the number of rows of the harmonic period phase compensation matrix, and N represents the number of columns of the harmonic period phase compensation matrix.

[0027] Specifically, in step S302, the time-frequency domain global feature extraction model includes: S302-1: Compensating the complex dynamic power signal data set according to the harmonic period phase compensation matrix to obtain complex dynamic power compensation signal data; S302-2: Preset a time-frequency domain harmonic feature matrix, and obtain the complex dynamic signal feature through feature extraction based on the complex dynamic power compensation signal data and the time-frequency domain harmonic feature matrix.

[0028] In this embodiment, the time-frequency domain harmonic characteristic matrix includes seven time-frequency domain harmonic characteristic matrices, including a time-frequency domain voltage amplitude matrix and a time-frequency domain current amplitude matrix, a time-frequency domain voltage phase matrix and a time-frequency domain current phase matrix, a harmonic relative amplitude matrix, a time-frequency domain active power matrix and a reactive power matrix, a current harmonic amplitude and phase joint distribution matrix, a harmonic current amplitude probability density matrix and a phase probability density function, and a current amplitude impact intensity matrix.

[0029] In step S4, the error detection model includes: S401: Acquire historical dynamic signal features, wherein the historical dynamic signal features carry an outlier label, wherein the outlier label includes 0 and 1, preset a basic prediction value, and calculate a target function through a loss function according to the outlier label and the basic prediction value; The loss function calculation expression is: , in, Obj (t) is the objective function, represents the loss function, i represents the i-th historical dynamic signal feature, n is the number of historical dynamic signal features, is the regularization term, C is a constant term, y i is the outlier label, is the basic prediction value, y i (t-1) represents the predicted value of the t-1th iteration, ft (x i ) represents the predicted value of the t-th tree for the i-th historical dynamic signal feature; S402: Calculate the optimal splitting point by second-order Taylor expansion and derivation according to the objective function, and construct a decision tree; S403: Obtaining an abnormal prediction value through the decision tree according to the historical dynamic signal characteristics; S404: Calculating an updated objective function according to the abnormal prediction value and the abnormal value label through the loss function; S405: According to the update objective function, the update optimal splitting point is calculated by the second-order Taylor expansion and derivation, and an update decision tree is constructed; S406: Preset an abnormal threshold, and obtain an abnormal difference value by performing a difference calculation based on the abnormal prediction value and the abnormal value label; S407: Determine whether the abnormal difference is less than the abnormal threshold. If yes, obtain an abnormality judgment decision tree; if no, repeat step S404 and step S405 until the abnormal difference is less than the abnormal threshold; S408: Obtain abnormal features through the abnormality judgment decision tree according to the dynamic signal features.

[0030] The computer storage medium of the embodiment of the present invention may adopt any combination of one or more computer-readable media. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device.

[0031] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0032] The program code included on the computer readable medium can be transmitted with any appropriate medium, including but not limited to wireless, electric wire, optical cable, RF, etc., or any suitable combination of the above. The computer program code for performing the operation of the present invention can be written in one or more programming languages ​​or their combinations, and the programming language includes object-oriented programming languages-such as Java, Smalltalk, C++, and also includes conventional procedural programming languages-such as "C" language or similar programming languages. The program code can be executed completely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on the remote computer, or completely on the remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0033] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. An Internet of Things integrated control system based on electric energy meter, characterized in that: It includes a signal acquisition module, a signal preprocessing module, a feature extraction module, and an error detection module; The signal acquisition module is used to collect electric energy signals through the electric energy meter, and obtain lightning protection electric energy signals through the lightning protection module according to the electric energy signals; The signal preprocessing module is used to obtain a complex dynamic electric energy signal data set according to the lightning protection electric energy signal through an electric energy signal preprocessing model; The feature extraction module is used to obtain complex dynamic signal features through a dynamic feature extraction model according to the complex dynamic electric energy signal data set; The dynamic feature extraction model includes: According to the complex dynamic electric energy signal data set, a harmonic period phase compensation matrix is ​​obtained by improving the short-time Fourier transform algorithm model; The improved short-time Fourier transform algorithm model includes: Calculating the periodic signal length using a period inversion algorithm according to the complex dynamic electric energy signal data set; Obtaining the real part of the harmonics and the imaginary part of the harmonics by Fourier transform according to the complex dynamic electric energy signal data set and the periodic signal length; The harmonic period phase compensation matrix is ​​calculated by a phase compensation algorithm according to the real part of the harmonic and the imaginary part of the harmonic; Obtaining the complex dynamic signal feature through a global feature extraction model in the time-frequency domain according to the harmonic period phase compensation matrix; The time-frequency domain global feature extraction model includes: Compensating the complex dynamic electric energy signal data set according to the harmonic period phase compensation matrix to obtain complex dynamic electric energy compensation signal data; Preset a time-frequency domain harmonic feature matrix, and obtain the complex dynamic signal feature by feature extraction based on the complex dynamic power compensation signal data and the time-frequency domain harmonic feature matrix; The error detection module is used to obtain abnormal features through an error detection model according to the dynamic signal features.

2. The Internet of Things integrated control system based on electric energy meter according to claim 1 is characterized in that: The lightning protection module includes a communication signal lightning protection circuit and a power supply lightning protection circuit. The communication signal lightning protection circuit includes a primary signal protection unit, a first decoupling unit, a secondary signal protection unit and a first reactance module; the power supply lightning protection circuit includes a primary power supply protection unit, a second decoupling unit, a secondary power supply protection unit and a second reactance module.

3. The Internet of Things integrated control system based on electric energy meter according to claim 1 is characterized in that: The obtaining of a complex dynamic electric energy signal data set through an electric energy signal preprocessing model according to the lightning protection electric energy signal comprises: Obtaining a lightning protection power energy signal data set by storing the lightning protection power energy signal in an ES cluster; According to the lightning protection electric energy signal data set, a cleaned lightning protection electric energy signal data set is obtained by data cleaning; According to the cleaned lightning protection electric energy signal data set, missing values ​​are supplemented by linear interpolation to obtain a complete cleaned lightning protection electric energy signal data set; The complete cleaning lightning protection power signal data set includes voltage signal, current signal, signal amplitude, and signal phase; A sampling frequency is preset, and the complex dynamic electric energy signal data set is obtained by constructing a complex dynamic electric energy signal model according to the voltage signal, the current signal, the signal amplitude, the signal phase, and the sampling frequency.

4. The Internet of Things integrated control system based on electric energy meter according to claim 1 is characterized in that: The calculating the periodic signal length by using a period inversion algorithm according to the complex dynamic electric energy signal data set comprises: According to the complex dynamic electric energy signal data set, the signal is intercepted by the window function length to obtain a signal subsequence; Taking the final value of the signal subsequence as the search center, presetting the search radius, and obtaining the signal cycle end point search interval according to the search center and the search radius; According to the search interval at the end point of the signal cycle, a near-zero point is obtained by the zero value theorem; The periodic signal length is obtained by performing a difference calculation based on the near-zero point.

5. The Internet of Things integrated control system based on electric energy meter according to claim 1 is characterized in that: The step of calculating the harmonic period phase compensation matrix according to the real part of the harmonic and the imaginary part of the harmonic by a phase compensation algorithm comprises: Obtaining a harmonic phase by phase calculation according to the real part of the harmonic and the imaginary part of the harmonic; Obtaining the ordinate of the sampling point, and obtaining a harmonic phase compensation value through secondary compensation calculation according to the ordinate of the sampling point and the harmonic phase; The harmonic periodic phase compensation matrix is ​​obtained by storing the harmonic phase compensation value in a matrix.

6. The Internet of Things integrated control system based on electric energy meter according to claim 1 is characterized in that: The obtaining of abnormal features through an error detection model according to the dynamic signal features comprises: S401: Acquire historical dynamic signal features, wherein the historical dynamic signal features carry an outlier label, preset a basic prediction value, and calculate a target function through a loss function according to the outlier label and the basic prediction value; S402: Calculate the optimal splitting point by second-order Taylor expansion and derivation according to the objective function, and construct a decision tree; S403: Obtaining an abnormal prediction value through the decision tree according to the historical dynamic signal characteristics; S404: Calculating an updated objective function according to the abnormal prediction value and the abnormal value label through the loss function; S405: According to the update objective function, the update optimal splitting point is calculated by the second-order Taylor expansion and derivation, and an update decision tree is constructed; S406: Preset an abnormal threshold, and obtain an abnormal difference value by performing a difference calculation based on the abnormal prediction value and the abnormal value label; S407: Obtaining an abnormality judgment decision tree through abnormality judgment according to the abnormal difference and the abnormal threshold; S408: Obtain abnormal features through the abnormality judgment decision tree according to the dynamic signal features.

7. The Internet of Things integrated control system based on electric energy meter according to claim 6 is characterized in that: The abnormality judgment includes judging whether the abnormal difference is less than the abnormal threshold value, and if so, obtaining the abnormality judgment decision tree; if not, repeating the step S404 and the step S405 until the abnormal difference is less than the abnormal threshold value.

Images