Malignant load identification method suitable for electric energy meter and use and acquisition terminal, electric energy meter and use and acquisition terminal

Abstract

The invention discloses a malignant load identification method suitable for an electric energy meter and an acquisition terminal, the electric energy meter and the acquisition terminal, and belongs to the technical field of load identification. The method comprises the steps that a target malignant load set is established, current data of the target malignant load set are collected, and after frequency domain features are extracted, a reference frequency domain combination feature vector of each load is constructed; obtaining target working condition current data in actual operation, and constructing a to-be-identified frequency domain combination feature vector in the same mode; and judging whether a malignant load exists or not by calculating the cosine similarity of the included angle between the two vectors. The invention also discloses an electric energy meter and a use and acquisition terminal based on the method. According to the method, the amplitude of the harmonic waves and the phase information are synthesized to construct the combined feature vector and perform similarity matching, so that the limitation that the prior art only depends on the amplitude of the harmonic waves is overcome, finer and more reliable identification of the malignant load is realized, and the misjudgment rate is effectively reduced.

Description

Technical Field

[0001] This invention belongs to the field of load identification technology, specifically relating to a malicious load identification method applicable to electricity meters and data collection terminals, as well as the electricity meter and data collection terminal. Background Technology

[0002] In collective electrical environments such as dormitories, office buildings, and factories, to prevent safety accidents such as fires and circuit overloads caused by the use of unauthorized high-power or substandard electrical appliances (collectively referred to as "malicious loads"), a method of limiting maximum power consumption is typically adopted. Once the power is detected to exceed a preset threshold, the system takes intervention measures such as power cut-off to prevent circuit overload or fire risks. This power threshold-based identification method is simple in principle and easy to implement, and has become the mainstream means of preventing malicious loads.

[0003] However, traditional power limiting methods cannot identify malicious loads. For example, devices such as "power sockets" or "energy-saving devices" on the market use diodes and other components to perform half-wave rectification of AC power, shaping the current waveform and significantly reducing the effective power of the load in the grid, thus circumventing power threshold monitoring. These devices not only bypass safety regulations, leading to the continued use of unauthorized electrical appliances, but also inject a large amount of harmonics into the grid due to current waveform distortion, polluting the grid and potentially damaging other equipment on the same line, thus posing new safety hazards.

[0004] To address the aforementioned issues, existing technologies have further introduced identification schemes based on harmonic analysis. This method leverages the characteristic that malicious loads (especially those with waveform shaping) exhibit a significant increase in current harmonic content during operation. It identifies potentially malicious loads by monitoring the amplitude of specific harmonics (such as the 3rd and 5th harmonics) and setting corresponding thresholds. However, this method still has the following drawbacks in practical applications: Firstly, many compliant nonlinear loads, such as computers, LED lighting equipment, and switching power supplies, also generate abundant harmonics during normal operation, easily leading to system misjudgments and interfering with normal power consumption. Secondly, relying solely on the amplitude information of a few harmonic components makes it difficult to comprehensively capture load characteristics. When facing loads with complex types or varied shaping strategies, both the identification accuracy and system robustness are insufficient. Summary of the Invention

[0005] This invention proposes a method for identifying malicious loads applicable to electricity meters and data acquisition terminals. The purpose is to provide a method for identifying time-domain waveforms by matching the amplitude and phase vector relationship of the spectrum and utilizing frequency domain features, thereby overcoming the high misjudgment rate problem caused by existing technologies that rely solely on harmonic component analysis.

[0006] The technical solution of this invention is as follows:

[0007] A method for identifying malicious loads applicable to electricity meters and user terminals includes the following steps:

[0008] Step S1. Create a set of target malignant loads. For each type of target malignant load, establish a malignant load condition and obtain current sampling data under the malignant load condition.

[0009] Step S2. Traverse the set of target malignant loads. For each target malignant load, extract frequency domain feature information from the corresponding current sampling data. Then, based on the frequency domain feature information, determine the amplitude-related harmonic order sequence and the phase-related harmonic order sequence corresponding to the target malignant load. Then, based on the amplitude-related harmonic order sequence, the phase-related harmonic order sequence, and the frequency domain feature information, obtain the reference amplitude feature vector and the reference phase feature vector corresponding to the target malignant load. Finally, construct a reference frequency domain combined feature vector based on the reference amplitude feature vector and the reference phase feature vector.

[0010] Step S3. Obtain current sampling data under the target operating condition, extract frequency domain feature information from it, and then construct the amplitude feature vector and phase feature vector to be identified based on the frequency domain feature information of the current sampling data obtained in this step and the amplitude-related harmonic order sequence and phase-related harmonic order sequence obtained in step S2. Based on the amplitude feature vector and phase feature vector to be identified, construct the frequency domain combination feature vector to be identified.

[0011] Step S4. Calculate the similarity between the frequency domain combined feature vector to be identified and the reference frequency domain combined feature vector corresponding to each type of target malignant load, and determine whether the target operating condition has malignant load based on the similarity.

[0012] As a further improvement to the malicious load identification method applicable to electricity meters and data collection terminals, the step of extracting frequency domain feature information from current sampling data is as follows:

[0013] Step A1. Calculate the DC component of the current sampling data;

[0014] Let the signal frequency be The sampling frequency is Then the number of sampling points per wave for:

[0015] ;

[0016] Accumulate the current sampling data by adding integer multiples of the cycle, and calculate the DC component according to the following formula:

[0017] ;

[0018] ;

[0019] in, Represents the DC component of the current path. Indicates the first current sampling data Current sampling value at each moment, This indicates the number of points that need to be accumulated to calculate the DC component. This represents the accumulated integer number of cycles, which is a preset value. This indicates a round-down operation;

[0020] Step A2. Subtract the DC component from the current sample value at each sampling point in the current sampling data to obtain the de-DC sampling data;

[0021] The first DC sampling data The current value at each moment is ; The first in the original current sampling data Current sample value at each moment;

[0022] Step A3. Perform Fourier analysis on the DC-to-DC sampling data to obtain the amplitude percentage and phase of each harmonic; the specific steps are as follows:

[0023] Take from DC sampling data Harmonic analysis was performed on each current sample value, and the measured values ​​were calculated using the following formula. Current waveform spectrum information of each current sample value The subharmonic is represented as:

[0024] , ;

[0025] in: For imaginary units, It is the rotation factor;

[0026] but The magnitude of the second harmonic is:

[0027] ;

[0028] In the formula, for Subharmonic The real part, for Subharmonic The imaginary part;

[0029] Amplitude of subharmonics for:

[0030] ;

[0031] The amplitude percentage of the subharmonic is:

[0032] ;

[0033] Phase of subharmonics for:

[0034] ;

[0035] In the formula: It is the arctangent function.

[0036] As a further improvement to the malignant load identification method applicable to electricity meters and user terminals, in step S2, the method for determining the amplitude-related harmonic number sequence and phase-related harmonic number sequence corresponding to the target malignant load based on frequency domain feature information is as follows: set amplitude-related harmonic number judgment conditions and phase-related harmonic number judgment conditions for the target malignant load, form an amplitude-related harmonic number sequence from the harmonic numbers in the frequency domain feature information of the target malignant load that meet the amplitude-related harmonic number judgment conditions, and form a phase-related harmonic number sequence from the harmonic numbers in the frequency domain feature information of the target malignant load that meet the phase-related harmonic number judgment conditions.

[0037] As a further improvement to the malicious load identification method applicable to electricity meters and user terminals, in step S2, a reference amplitude feature vector is constructed. The method is as follows:

[0038] ;

[0039] in, Represents the amplitude-dependent harmonic order sequence The first in There is one element, which represents the harmonic order. Indicates the length of the amplitude-dependent harmonic order sequence. Indicates the current target's malicious load. Percentage of the amplitude of the second harmonic;

[0040] Constructing reference phase eigenvectors The method is as follows:

[0041] ;

[0042] in, Represents the phase-correlated harmonic order sequence The first in There is one element, which represents the harmonic order. Indicates the length of the phase-correlated harmonic order sequence. Indicates the current target's malicious load. Phase of the subharmonic.

[0043] As a further improvement to the malicious load identification method applicable to electricity meters and user terminals, in step S2, the method for constructing a reference frequency domain combined feature vector based on the reference amplitude feature vector and the reference phase feature vector is as follows: the elements of the reference amplitude feature vector and the reference phase feature vector are concatenated in sequence to obtain the reference frequency domain combined feature vector.

[0044] ;

[0045] In the above formula, This is the reference frequency domain combined feature vector.

[0046] As a further improvement to the malicious load identification method applicable to electricity meters and user terminals, in step S3, the method for constructing the amplitude feature vector to be identified is as follows:

[0047] ;

[0048] In the above formula, Represents the amplitude-dependent harmonic order sequence The first in There is one element, which represents the harmonic order. Indicates the length of the amplitude-dependent harmonic order sequence. This indicates the frequency domain feature information under the current target operating condition. Percentage of the amplitude of the second harmonic;

[0049] The method for constructing the phase feature vector to be identified is as follows:

[0050] ;

[0051] in, Represents the phase-correlated harmonic order sequence The first in There is one element, which represents the harmonic order. Indicates the length of the phase-correlated harmonic order sequence. Indicates the current target's malicious load. The relative phase of the subharmonics;

[0052] for The relative phase of subharmonics The calculation method is as follows:

[0053] ;

[0054] In the formula: The first harmonic phase of the current target malignant load is the theoretical fundamental phase. This indicates the frequency domain feature information under the current target operating condition. Phase of the subharmonic, This indicates the modulo operation.

[0055] As a further improvement to the malicious load identification method applicable to electricity meters and user terminals, in step S3, the method for constructing the frequency domain combined feature vector to be identified based on the amplitude feature vector to be identified and the phase feature vector to be identified is as follows: the elements of the amplitude feature vector to be identified and the phase feature vector to be identified are concatenated in sequence to obtain the frequency domain combined feature vector to be identified, that is:

[0056] ;

[0057] In the above formula, The frequency domain combined feature vector to be identified.

[0058] As a further improvement to the malicious load identification method applicable to electricity meters and user terminals, in step S4, the similarity between the frequency domain combined feature vector to be identified and the reference frequency domain combined feature vector is expressed as the angle between the combined feature vectors. express:

[0059] ;

[0060] In the formula, For vector dot product operation, For vector 2-norm operations;

[0061] The method for determining whether a target operating condition has a malignant load based on similarity is as follows: if the angle between the combined feature vectors of the target operating condition and the reference frequency domain combined feature vector corresponding to a certain malignant load is... If the load is less than the preset threshold, it is determined that there is a malignant load on the target in the target operating condition.

[0062] The present invention also discloses an energy meter, including a processor and a memory, wherein a program stored in the memory is configured to be executed by the processor to implement the malicious load identification method.

[0063] The present invention also discloses a terminal for acquiring data, including a processor and a memory, wherein a program stored in the memory is configured to be executed by the processor, and when executed, implements the malicious load identification method.

[0064] Compared with the prior art, the present invention has the following beneficial effects:

[0065] 1. This invention overcomes the limitations of traditional malicious load identification methods that rely solely on a few harmonic amplitude thresholds. By introducing multi-dimensional feature combinations of amplitude and phase in the frequency domain, it constructs a reference and target frequency domain combined feature vector, and matches them based on the similarity of the vector angles. This achieves a more comprehensive and accurate identification of the load current waveform. This method not only effectively distinguishes between compliant nonlinear loads and malicious loads but also significantly reduces the risk of misjudgment due to similar harmonic content, improving identification accuracy and system robustness.

[0066] 2. This invention constructs a combined feature vector containing amplitude and phase features, and uses the angle measurement method in the vector space for similarity judgment, enabling the recognition process to capture the overall features of the waveform more effectively. Compared to traditional methods that only focus on a single harmonic component, this invention can extract more discriminative features from the amplitude percentage and relative phase of multiple harmonic orders, thus maintaining stable recognition performance even when faced with complex load types or changes in waveform shaping strategies.

[0067] 3. In the frequency domain feature extraction stage, this invention accurately obtains the amplitude percentage and phase information of each harmonic by combining DC removal processing with Fourier analysis. Furthermore, based on preset conditions, it filters out key harmonic order sequences, effectively eliminating unrepresentative harmonic components and improving the information density and discriminative ability of the feature vector. This method enhances the ability to express the essential characteristics of malicious load waveforms while ensuring computational efficiency.

[0068] 4. This invention also proposes a phase feature construction method based on relative phase. By associating the measured harmonic phase with the theoretical fundamental phase, the phase offset caused by different sampling starting points is eliminated, so that the phase features are still comparable and stable under different measurement conditions, further improving the adaptability and reliability of the identification method. Attached Figure Description

[0069] Figure 1 The graph shows the time-domain waveform of the diode load, with the horizontal axis representing the sampling point index and the vertical axis representing the normalized sampling value. Detailed Implementation

[0070] The technical solution of the present invention will now be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0071] Example 1

[0072] This embodiment discloses a malicious load identification method applicable to electricity meters and data collection terminals, including the following steps:

[0073] Step S1. Create a set of target malignant loads. For each type of target malignant load, establish a malignant load condition and obtain current sampling data under the malignant load condition.

[0074] For ease of description in this embodiment, the term "diode load" is used to refer to any malicious load that uses a diode as a rectifier, and the diode load is used as an example of the target malicious load. The time-domain waveform of the diode load is referenced below. Figure 1 For the sake of simplicity, the amplitude values ​​in the diagram have been normalized. Clearly, the waveform in the diagram only shows the positive half-cycle.

[0075] Severe load conditions can be constructed using real electrical equipment or through simulation.

[0076] The current signal is continuously sampled, and multiple sets of current signals are stored in a buffer. According to the Nyquist sampling theorem, the sampling frequency should be greater than twice the highest cutoff frequency of the signal. Based on engineering experience, the sampling frequency is preferably more than 10 times the signal frequency. For a 50Hz power frequency signal, a sampling frequency of 6.4kHz or 12.8kHz can be considered. To simplify the analysis, continuous sampling for two cycles is generally performed starting from the zero-crossing point.

[0077] Step S2. Traverse the set of target malignant loads. For each target malignant load, extract frequency domain feature information from the corresponding current sampling data. Then, based on the frequency domain feature information, determine the amplitude-related harmonic order sequence and the phase-related harmonic order sequence corresponding to the target malignant load. Then, based on the amplitude-related harmonic order sequence, the phase-related harmonic order sequence, and the frequency domain feature information, obtain the reference amplitude feature vector and the reference phase feature vector corresponding to the target malignant load. Finally, construct a reference frequency domain combined feature vector based on the reference amplitude feature vector and the reference phase feature vector.

[0078] Furthermore, the steps for extracting frequency domain feature information from the current sampling data are as follows:

[0079] Step A1. Calculate the DC component of the current sampling data.

[0080] Assume the signal frequency is The sampling frequency is Then the number of sampling points per wave for:

[0081] .

[0082] Accumulate the current sampling data by adding integer multiples of the cycle, and calculate the DC component according to the following formula:

[0083] ;

[0084] ;

[0085] in, Represents the DC component of the current path. Indicates the first current sampling data Current sampling value at each moment, This indicates the number of points that need to be accumulated to calculate the DC component. This indicates the cumulative integer number of cycles, which is a preset value (e.g., 2 cycles). This indicates the floor function.

[0086] Step A2. Subtract the DC component from the current sample value at each sampling point in the current sampling data to obtain the de-DC sampling data.

[0087] The first DC sampling data The current value at each moment is . The first in the original current sampling data The current sampling value at each moment.

[0088] Step A3. Perform Fourier analysis on the DC sampling data to obtain the amplitude percentage and phase of each harmonic.

[0089] Based on engineering experience, two cycles of current samples are taken from the DC sampling data for harmonic analysis. The length of the selected current samples is denoted as... Then we have: .

[0090] Calculate the amount according to the following formula. Current waveform spectrum information of each current sample value The subharmonic is represented as:

[0091] , ;

[0092] in: For imaginary units, is the rotation factor.

[0093] but The magnitude of the second harmonic is:

[0094] ;

[0095] In the formula, for Subharmonic The real part, for Subharmonic The imaginary part.

[0096] Amplitude of subharmonics for:

[0097] ;

[0098] The amplitude percentage of the subharmonic is:

[0099] .

[0100] Phase of subharmonics for:

[0101] ;

[0102] In the formula: It is the arctangent function.

[0103] Furthermore, the method for determining the amplitude-related harmonic order sequence and phase-related harmonic order sequence corresponding to the target malignant load based on frequency domain feature information is as follows: set amplitude-related harmonic order judgment conditions and phase-related harmonic order judgment conditions for the target malignant load, form an amplitude-related harmonic order sequence from the harmonic orders in the frequency domain feature information of the target malignant load that meet the amplitude-related harmonic order judgment conditions, and form a phase-related harmonic order sequence from the harmonic orders in the frequency domain feature information of the target malignant load that meet the phase-related harmonic order judgment conditions.

[0104] This embodiment takes a diode load as an example, and its corresponding frequency domain characteristic information is shown in Table 1.

[0105] Table 1: Amplitude percentage and phase of each harmonic of the diode load waveform.

[0106]

[0107] For diode loads, the condition for determining amplitude-dependent harmonic orders is that the following two conditions must be met simultaneously:

[0108] Condition 1: The amplitude percentage is less than 0.5% or greater than 2%;

[0109] Condition 2: The harmonic order is less than or equal to 10.

[0110] Therefore, the amplitude-dependent harmonic order sequence of the diode load is: .

[0111] For diode loads, the condition for determining the phase-related harmonic order is that the following two conditions must be met simultaneously:

[0112] Condition 1: Amplitude percentage greater than 2%;

[0113] Condition 2: The harmonic order is less than or equal to 10.

[0114] Therefore, the phase-dependent harmonic order sequence of the diode load is: .

[0115] Furthermore, construct the reference amplitude feature vector. The method is as follows:

[0116] ;

[0117] in, Represents the amplitude-dependent harmonic order sequence The first in There is one element, which represents the harmonic order. Indicates the length of the amplitude-dependent harmonic order sequence. Indicates the current target's malicious load. The percentage of the amplitude of the subharmonic.

[0118] Furthermore, a reference phase eigenvector is constructed. The method is as follows:

[0119] ;

[0120] in, Represents the phase-correlated harmonic order sequence The first in There is one element, which represents the harmonic order. Indicates the length of the phase-correlated harmonic order sequence. Indicates the current target's malicious load. Phase of the subharmonic.

[0121] Furthermore, the method for constructing a reference frequency domain combined feature vector based on the reference amplitude feature vector and the reference phase feature vector is as follows: the elements of the reference amplitude feature vector and the reference phase feature vector are concatenated in order to obtain the reference frequency domain combined feature vector. That is:

[0122] ;

[0123] In the above formula, This is the reference frequency domain combined feature vector.

[0124] Step S3. Obtain current sampling data under the target operating condition, extract frequency domain feature information from it, and then construct the amplitude feature vector and phase feature vector to be identified based on the frequency domain feature information of the current sampling data obtained in this step and the amplitude-related harmonic order sequence and phase-related harmonic order sequence obtained in step S2. Based on the amplitude feature vector and phase feature vector to be identified, construct the frequency domain combination feature vector to be identified.

[0125] Specifically, the method for extracting frequency domain feature information from the current sampling data under the target operating condition is the same as the method for extracting frequency domain feature information in step S2, and will not be elaborated here.

[0126] Furthermore, the method for constructing the feature vector of the amplitude to be identified is as follows:

[0127] ;

[0128] In the above formula, Represents the amplitude-dependent harmonic order sequence The first in There is one element, which represents the harmonic order. Indicates the length of the amplitude-dependent harmonic order sequence. This indicates the frequency domain feature information under the current target operating condition. The percentage of the amplitude of the subharmonic.

[0129] Furthermore, the method for constructing the phase feature vector to be identified is as follows:

[0130] ;

[0131] in, Represents the phase-correlated harmonic order sequence The first in There is one element, which represents the harmonic order. Indicates the length of the phase-correlated harmonic order sequence. Indicates the current target's malicious load. The relative phase of the subharmonics.

[0132] for The relative phase of subharmonics The calculation method is as follows:

[0133] ;

[0134] In the formula: The first harmonic phase of the current target malignant load is the theoretical fundamental phase. This indicates the frequency domain feature information under the current target operating condition. Phase of the subharmonic, This indicates the modulo operation.

[0135] Furthermore, the method for constructing the combined frequency domain feature vector based on the amplitude feature vector and the phase feature vector to be identified is as follows: the elements of the amplitude feature vector and the phase feature vector to be identified are concatenated in order to obtain the combined frequency domain feature vector to be identified. That is:

[0136] ;

[0137] In the above formula, The frequency domain combined feature vector to be identified.

[0138] Step S4. Calculate the similarity between the frequency domain combined feature vector to be identified and the reference frequency domain combined feature vector corresponding to each type of target malignant load, and determine whether the target operating condition has malignant load based on the similarity.

[0139] Furthermore, the similarity between the frequency domain combined feature vector to be identified and the reference frequency domain combined feature vector is expressed as the angle between the combined feature vectors. express:

[0140] ;

[0141] In the formula, For vector dot product operation, This refers to the 2-norm operation for vectors.

[0142] Furthermore, the method for determining whether a target operating condition has a malignant load based on similarity is as follows: if the angle between the combined feature vectors of the target operating condition and the reference combined feature vector corresponding to a certain malignant load in the frequency domain is... If the value is less than the preset threshold (3° in this embodiment), it is determined that there is a malignant load on the target in the target working condition.

[0143] Taking a diode load as an example, if the angle between the combined feature vector of the diode load waveform and the combined feature vector of the detected waveform... If the value is less than 3°, it is determined that the current time-domain waveform being detected matches the time-domain waveform of the diode load, confirming the presence of a diode load.

[0144] Example 2

[0145] This embodiment discloses an energy meter, including a processor and a memory. A program stored in the memory is configured to be executed by the processor, and when executed, it implements the malicious load identification method described in Embodiment 1.

[0146] Example 3

[0147] This embodiment discloses a terminal for data collection, including a processor and a memory. A program stored in the memory is configured to be executed by the processor, and when executed, it implements the malicious load identification method described in Embodiment 1.

[0148] It should be noted that, as will be apparent to those skilled in the art, the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics thereof. The scope of the present invention is defined by the claims rather than the foregoing description.

Claims

1. A method for identifying malicious loads applicable to electricity meters and user terminals, characterized in that: Includes the following steps: Step S1. Create a set of target malignant loads. For each type of target malignant load, establish a malignant load condition and obtain current sampling data under the malignant load condition. Step S2. Traverse the set of target malignant loads. For each target malignant load, extract frequency domain feature information from the corresponding current sampling data. Then, based on the frequency domain feature information, determine the amplitude-related harmonic order sequence and the phase-related harmonic order sequence corresponding to the target malignant load. Then, based on the amplitude-related harmonic order sequence, the phase-related harmonic order sequence, and the frequency domain feature information, obtain the reference amplitude feature vector and the reference phase feature vector corresponding to the target malignant load. Finally, construct a reference frequency domain combined feature vector based on the reference amplitude feature vector and the reference phase feature vector. Step S3. Obtain current sampling data under the target operating condition, extract frequency domain feature information from it, and then construct the amplitude feature vector and phase feature vector to be identified based on the frequency domain feature information of the current sampling data obtained in this step and the amplitude-related harmonic order sequence and phase-related harmonic order sequence obtained in step S2. Based on the amplitude feature vector and phase feature vector to be identified, construct the frequency domain combination feature vector to be identified. Step S4. Calculate the similarity between the frequency domain combined feature vector to be identified and the reference frequency domain combined feature vector corresponding to each type of target malignant load, and determine whether the target operating condition has malignant load based on the similarity.

2. The malicious load identification method applicable to electricity meters and user terminals as described in claim 1, characterized in that: The steps for extracting frequency domain feature information from current sampling data are as follows: Step A1. Calculate the DC component of the current sampling data; Let the signal frequency be... The sampling frequency is Then the number of sampling points per wave for: ； Accumulate the current sampling data by adding integer multiples of the cycle, and calculate the DC component according to the following formula: ； ； in, Represents the DC component of the current path. Indicates the first current sampling data Current sampling value at each moment, This indicates the number of points that need to be accumulated to calculate the DC component. This represents the accumulated integer number of cycles, which is a preset value. This indicates a round-down operation; Step A2. Subtract the DC component from the current sample value at each sampling point in the current sampling data to obtain the de-DC sampling data; The first DC sampling data The current value at each moment is ; The first in the original current sampling data Current sample value at each moment; Step A3. Perform Fourier analysis on the DC-to-DC sampling data to obtain the amplitude percentage and phase of each harmonic; the specific steps are as follows: Take from DC sampling data Harmonic analysis was performed on each current sample value, and the measured values ​​were calculated using the following formula. Current waveform spectrum information of each current sample value The subharmonic is represented as: ， ； in: For virtual part units, It is the rotation factor; but The magnitude of the second harmonic is: ； In the formula, for Subharmonic The real part, for Subharmonic The imaginary part; Amplitude of subharmonics for: ； The amplitude percentage of the subharmonic is: ； Phase of subharmonics for: ； In the formula: It is the arctangent function.

3. The malicious load identification method applicable to electricity meters and user terminals as described in claim 1, characterized in that, In step S2, the method for determining the amplitude-related harmonic order sequence and phase-related harmonic order sequence corresponding to the target malignant load based on frequency domain feature information is as follows: set amplitude-related harmonic order judgment conditions and phase-related harmonic order judgment conditions for the target malignant load, form an amplitude-related harmonic order sequence from the harmonic orders in the frequency domain feature information of the target malignant load that meet the amplitude-related harmonic order judgment conditions, and form a phase-related harmonic order sequence from the harmonic orders in the frequency domain feature information of the target malignant load that meet the phase-related harmonic order judgment conditions.

4. The malicious load identification method applicable to electricity meters and user terminals as described in claim 1, characterized in that, In step S2, a reference amplitude feature vector is constructed. The method is as follows: ； in, Represents the amplitude-dependent harmonic order sequence The first in There is one element, which represents the harmonic order. Indicates the length of the amplitude-dependent harmonic order sequence. Indicates the current target's malicious load. Percentage of the amplitude of the second harmonic; Constructing reference phase eigenvectors The method is as follows: ； in, Represents the phase-correlated harmonic order sequence The first in There is one element, which represents the harmonic order. Indicates the length of the phase-correlated harmonic order sequence. Indicates the current target's malicious load. Phase of the subharmonic.

5. The malicious load identification method applicable to electricity meters and user terminals as described in claim 4, characterized in that, In step S2, the method for constructing the reference frequency domain combined feature vector based on the reference amplitude feature vector and the reference phase feature vector is as follows: the elements of the reference amplitude feature vector and the reference phase feature vector are concatenated in order to obtain the reference frequency domain combined feature vector. ； In the above formula, This is the reference frequency domain combined feature vector.

6. The malicious load identification method applicable to electricity meters and user terminals as described in claim 1, characterized in that, In step S3, the method for constructing the amplitude feature vector to be identified is as follows: ； In the above formula, Represents the amplitude-dependent harmonic order sequence The first in There is one element, which represents the harmonic order. Indicates the length of the amplitude-dependent harmonic order sequence. This indicates the frequency domain feature information under the current target operating condition. Percentage of the amplitude of the second harmonic; The method for constructing the phase feature vector to be identified is as follows: ； in, Represents the phase-correlated harmonic order sequence The first in There is one element, which represents the harmonic order. Indicates the length of the phase-correlated harmonic order sequence. Indicates the current target's malicious load. The relative phase of the subharmonics; for The relative phase of subharmonics The calculation method is as follows: ； In the formula: The first harmonic phase of the current target malignant load is the theoretical fundamental phase. This indicates the frequency domain feature information under the current target operating condition. Phase of the subharmonic, This indicates the modulo operation.

7. The malicious load identification method applicable to electricity meters and user terminals as described in claim 6, characterized in that, In step S3, the method for constructing the frequency domain combined feature vector to be identified based on the amplitude feature vector and the phase feature vector to be identified is as follows: the elements of the amplitude feature vector and the phase feature vector to be identified are concatenated in order to obtain the frequency domain combined feature vector to be identified, that is: ； In the above formula, The frequency domain combined feature vector to be identified.

8. The malicious load identification method applicable to electricity meters and user terminals as described in claim 1, characterized in that, In step S4, the similarity between the frequency domain combined feature vector to be identified and the reference frequency domain combined feature vector is expressed as the angle between the combined feature vectors. express: ； In the formula, For vector dot product operation, For vector 2-norm operations; The method for determining whether a target operating condition has a malignant load based on similarity is as follows: if the angle between the combined feature vectors of the target operating condition and the reference frequency domain combined feature vector corresponding to a certain malignant load is... If the load is less than the preset threshold, it is determined that there is a malignant load on the target in the target operating condition.

9. An electricity meter, characterized in that: It includes a processor and a memory, wherein a program stored in the memory is configured to be executed by the processor, and when executed, implements the malicious load identification method as described in any one of claims 1 to 8.

10. A terminal for data acquisition, characterized in that: It includes a processor and a memory, wherein a program stored in the memory is configured to be executed by the processor, and when executed, implements the malicious load identification method as described in any one of claims 1 to 8.

Images