Method and system for automatically detecting and judging abnormal phenomenon of electric energy meter

Abstract

The invention discloses a method and a system for automatically detecting and judging an abnormal phenomenon of an electric energy meter. The method adopted by the invention comprises the following steps: acquiring multi-dimensional electric power parameter data in real time, mapping the data to a nonlinear space, constructing a discrimination matrix, effectively distinguishing normal and abnormal data by minimizing an objective function and optimizing a projection matrix, and performing multi-stage dynamic filtering on a feature vector of the electric power parameter data to obtain a purified feature vector; the purified feature vectors are processed through an adaptive convolution inverse transformation algorithm to reconstruct time and space features of the signals and generate frequency domain signals, complete time feature signals are recovered through inverse transformation, optimal abnormal feature vectors are obtained through a multi-iteration convergence algorithm, the optimal abnormal feature vectors are classified, and the optimal abnormal feature vectors are obtained. And automatic detection and determination are completed. According to the invention, the accuracy of anomaly detection can be improved, the signal-to-noise ratio can be improved, and different types of anomaly phenomena can be accurately identified.

Description

Technical Field

[0001] The invention belongs to the technical field of electric power metering and monitoring, and in particular to a method and system for automatically detecting and determining abnormal phenomena of an electric energy meter. Background Art

[0002] With the increasing complexity of the power system and the continuous growth of electricity demand, the working environment of the energy meter has become more and more diverse and complex. Especially in the context of smart grids and distributed energy systems, the energy meter not only needs to accurately measure the user's electricity consumption, but also needs to monitor and transmit multi-dimensional power parameter data in real time, such as voltage, current, active power, reactive power, power factor, harmonic components, etc. These data are of great significance for power companies to optimize power grid operation, improve power supply quality and reduce power loss. However, due to changes in the external environment, aging of power equipment, external interference and other factors, various abnormal phenomena may occur in the actual operation of the energy meter, such as measurement errors, data transmission interruptions, abnormal fluctuations in power parameters, etc. Therefore, how to effectively and accurately detect and determine the abnormal phenomena of the energy meter has become the focus of attention of power companies and equipment manufacturers.

[0003] The existing technology has at least the following technical problems: it is unable to dynamically adapt to spectrum changes, resulting in misjudgment or missed detection, resulting in a low signal-to-noise ratio, unable to fully reflect the complexity of electricity meter data, resulting in inaccurate abnormality judgment, lack of intelligent classification models, unable to effectively respond to complex and changeable power system anomalies, and obviously insufficient judgment capabilities. During long-term use, the detection performance gradually decreases, making it difficult to maintain a high level of detection and judgment capabilities. Summary of the invention

[0004] The technical problem to be solved by the present invention is to overcome the defects of the above-mentioned prior art and provide a method and system for automatically detecting and determining abnormal phenomena of an electric energy meter, so as to improve the accuracy of abnormality detection, improve the signal-to-noise ratio, and accurately identify different types of abnormal phenomena.

[0005] In a first aspect, the present invention provides a method for automatically detecting and determining abnormal phenomena of an electric energy meter, which comprises the steps of: S1. Real-time collection of multi-dimensional power parameter data, mapping the data to a nonlinear space, constructing a discriminant matrix, optimizing the projection matrix by minimizing the objective function to effectively distinguish normal and abnormal data, and performing multi-stage dynamic filtering on the feature vector of the power parameter data to obtain a purified feature vector; S2. The purified feature vector is processed by an adaptive inverse convolution transform algorithm to reconstruct the time and space characteristics of the signal and generate a frequency domain signal. The complete time feature signal is then restored through inverse transform. The optimal abnormal feature vector is obtained through a multiple iterative convergence algorithm. The optimal abnormal feature vector is classified to complete automatic detection and judgment.

[0006] Furthermore, in S1, the collected multi-dimensional power parameter data is preprocessed, and nonlinear mapping is performed on the preprocessed multi-dimensional data in order to extract deep nonlinear features in the data.

[0007] Furthermore, in S1, the mapped data set is used to perform adaptive weighted discriminant analysis, and a weighted discriminant matrix is ​​constructed to effectively distinguish normal data from abnormal data in a nonlinear space.

[0008] Furthermore, in S1, an objective function is established to obtain an optimal projection matrix, the intra-class scatter matrix is ​​eigenvalue decomposed to find its corresponding eigenvector, and the eigenvectors are connected to correspond to the direction in which the intra-class scatter is minimized; by solving the generalized eigenvalue problem, the direction in which the inter-class scatter is maximized is determined; and the eigenvectors with the largest first eigenvalues ​​are selected as column vectors of the projection matrix.

[0009] Furthermore, in S1, in order to purify the feature vector, a multi-stage dynamic filtering method is adopted, a frequency domain filter is designed based on the spectral characteristics of the abnormal data, and each feature vector is frequency-domain filtered to remove high-frequency noise and low-frequency fluctuations. After applying the frequency domain filter, the feature vector is converted into a purified feature vector.

[0010] Furthermore, in S2, the adaptive convolution inverse transform algorithm introduces multiple modulations and adaptive adjustment mechanisms in the convolution kernel design and inverse transform process to better adapt to the complex abnormal signals in the electricity meter data; the purified feature vector is input into the convolution kernel function, and the design of the convolution kernel takes into account the mixed modulation of time and frequency characteristics.

[0011] Furthermore, in S2, the purified feature vector is processed into a time domain signal through a convolution operation. After obtaining the frequency domain signal, an inverse transform operation is performed by combining joint modulation of the frequency domain and the time domain to restore the complete time characteristics of the signal.

[0012] Furthermore, in S2, features are extracted from the reconstructed signal, and the extracted feature vector is input into a multiple iterative convergence algorithm of the feature space, and a potential optimal abnormal feature vector in the signal is extracted through multiple iterative optimizations.

[0013] Furthermore, in S2, each feature is analyzed according to the value and feature distribution of the optimal abnormal feature vector to identify possible abnormal features; by calculating the statistical distribution of each feature under normal operating conditions and comparing it with the currently extracted features, if the value of a certain feature deviates from the normal distribution to a greater extent than a preset threshold, it is considered that the current feature is abnormal; in order to further determine the type of abnormal phenomenon, the optimal abnormal feature vector is classified using a pre-trained classification model.

[0014] In a second aspect, the present invention provides an automatic detection and determination system for abnormal phenomena of an electric energy meter, comprising: Feature vector purification unit: collects multi-dimensional power parameter data in real time, maps the data into nonlinear space, constructs a discriminant matrix, optimizes the projection matrix by minimizing the objective function to effectively distinguish normal and abnormal data, and performs multi-stage dynamic filtering on the feature vector of the power parameter data to obtain the purified feature vector; Automatic detection and judgment unit: The purified feature vector is processed by an adaptive inverse convolution transform algorithm to reconstruct the time and space characteristics of the signal, generate a frequency domain signal, and then restore the complete time feature signal through inverse transform. The optimal abnormal feature vector is obtained through a multiple iterative convergence algorithm, and the optimal abnormal feature vector is classified to complete automatic detection and judgment.

[0015] The present invention has the following beneficial effects: 1. By performing deep nonlinear mapping on the multi-dimensional features of the original data, the present invention enables the abnormal pattern to be fully displayed in the mapped high-dimensional space, and further effectively distinguishes normal and abnormal data through the optimized discriminant matrix, thereby significantly improving the accuracy of anomaly detection.

[0016] 2. Through the multi-stage dynamic filtering method and the adaptive inverse convolution transform algorithm, the present invention can effectively remove high-frequency noise and low-frequency interference information in the feature vector, making the retained signal features purer and more representative, and effectively improving the signal-to-noise ratio.

[0017] 3. In the process of signal reconstruction, the adaptive inverse convolution transform algorithm not only retains the original abnormal characteristics, but also enhances the temporal resolution; it can more accurately capture and analyze the occurrence time and development process of abnormal phenomena, which is helpful for the accurate judgment of abnormal phenomena.

[0018] 4. The present invention fully reflects the complex characteristics of the electric energy meter data by extracting multiple features from the reconstructed signal, including time domain features, frequency domain features and nonlinear features; then, through a multiple iterative convergence algorithm in the feature space, the features are optimized multiple times to finally obtain the optimal abnormal feature vector. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a flow chart of a method for automatically detecting and determining abnormal phenomena of an electric energy meter according to the present invention; Figure 2 The present invention is a structural diagram of an automatic detection and determination system for abnormal phenomena of an electric energy meter. DETAILED DESCRIPTION

[0020] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0022] The specific scheme of the method and system for automatically detecting and determining abnormal phenomena of an electric energy meter provided by the present invention is described in detail below with reference to the accompanying drawings.

[0023] See attached Figure 1 , which shows a flow chart of a method for automatically detecting and determining abnormal phenomena of an electric energy meter provided by an embodiment of the present invention, the method comprising the following steps: S1. Collect multi-dimensional power parameter data in real time, map the data to nonlinear space, construct a discriminant matrix, optimize the projection matrix by minimizing the objective function to effectively distinguish normal and abnormal data, and perform multi-stage dynamic filtering on the feature vector of the power parameter data to obtain the purified feature vector.

[0024] The electric energy meter collects multi-dimensional power parameter data in real time through sensors, including voltage, current, active power, reactive power, power factor, harmonic components, etc. Each sensor is strictly calibrated to ensure the accuracy and stability of the data, and the collected raw data is pre-processed, including signal filtering, denoising and outlier removal. The pre-processing methods used are all existing technologies.

[0025] The preprocessed multidimensional data set is recorded as , Represents a multidimensional parameter vector at the i-th time point, each It contains all the power parameters at that moment, and synchronizes them through a unified timestamp to ensure the time consistency of the data. In order to extract the deep nonlinear features in the data, the preprocessed multi-dimensional data is nonlinearly mapped, and the mapping function is defined as:

[0026] in, Represents the mapped data set, are the coefficients of the polynomial terms, is the power exponent of the polynomial, is the highest power of the polynomial, are the coefficients of the trigonometric terms, controlling the weight of each sine term, is the power exponent of the sine function, is the highest power of the sine function. By combining polynomial and logarithmic functions, the original linear data is mapped to a space with complex nonlinear features, so that the subsequent detection steps can better identify potential abnormal patterns.

[0027] The mapped data set is used for adaptive weighted discriminant analysis, and the normal data and abnormal data are effectively distinguished in the nonlinear space by constructing a weighted discriminant matrix. The discriminant matrix is ​​defined as:

[0028] in, is the discriminant matrix, and is the projection matrix, which controls the distribution of data in the nonlinear space so as to maximize the difference between classes and minimize the difference within classes. is the parameter for adjusting the weight of the i-th data, is the transpose of the matrix, is the number of data after mapping. The optimal projection matrix is ​​obtained by optimizing the following objective function:

[0029] in, represents the trace of the matrix, and are the intra-class and inter-class scatter matrices, respectively, which are defined as follows:

[0030]

[0031] in, Represents the mean vector of the k-th class of data, which represents the average position of all samples of this class in the feature space; Represents the global mean vector of all data, indicating the average position of all samples in the feature space; is the number of categories of data, is the number of samples of the kth class, indicating the total number of samples of this class.

[0032] The within-class scatter matrix Perform eigenvalue decomposition to find the corresponding eigenvector. The features are connected to correspond to the direction that minimizes the intra-class scatter. Then, by solving the generalized eigenvalue problem, determine the direction that maximizes the inter-class scatter. The larger the eigenvalue, the more important the corresponding eigenvector is. Select the eigenvectors with the largest eigenvalues ​​as the column vectors of the projection matrix. The column vectors of the projection matrix represent the best projection direction, which can maximize the separation between different classes while maintaining the compactness of the intra-class data, and distinguish normal data from abnormal data. By minimizing the objective function, the optimized discriminant matrix can effectively distinguish normal and abnormal data.

[0033] Multiply the optimized discriminant matrix with the mapped data to obtain the eigenvector of each data point in the nonlinear space, that is:

[0034] Eigenvector The feature vector contains the abnormal feature information of each data point in the nonlinear space, but the feature vector still contains noise and irrelevant interference information. Therefore, further purification is required to extract the most representative abnormal signal features.

[0035] In order to purify the feature vector, a multi-stage dynamic filtering method is used. Specifically, a frequency domain filter is designed based on the spectral characteristics of the abnormal data, and each feature vector is filtered in the frequency domain to remove high-frequency noise and low-frequency fluctuations. The frequency response function of the filter is defined as:

[0036] in, is the frequency response function of the filter, indicating that the filter has different frequencies. The response amplitude under is the bandwidth of the control filter, are the Fourier coefficients, represents the frequency of the signal, is the index variable of the Fourier series, which is used to represent the order of different frequency components in the filter frequency response. is the factorial. Through the frequency response function, the filter can dynamically adjust its frequency response characteristics to adapt to the spectrum characteristics of different eigenvectors, thereby removing unnecessary noise components.

[0037] After applying the frequency domain filter, the eigenvector is converted into a purified feature vector. The specific purification process can be described by the following convolution formula:

[0038] in, represents the purified feature vector, is the filter response function in the time domain, which is transformed by Fourier transform and frequency response function associated; is a time variable; is the offset in the convolution operation. Through the convolution operation, the filter can remove irrelevant high-frequency and low-frequency components in the time domain, retain the representative intermediate frequency features, and thus obtain the purified feature vector.

[0039] S2. The purified feature vector is processed by an adaptive inverse convolution transform algorithm to reconstruct the time and space characteristics of the signal and generate a frequency domain signal. The complete time feature signal is then restored through inverse transform. The optimal abnormal feature vector is obtained through a multiple iterative convergence algorithm. The optimal abnormal feature vector is classified to complete automatic detection and judgment.

[0040] The purified feature vector has a higher signal-to-noise ratio and better reflects the abnormal characteristics in the electricity meter data. The purified feature vector will be further processed and the time and space characteristics of the signal will be reconstructed through the adaptive inverse convolution transform algorithm. The reconstructed signal not only retains the original abnormal characteristics, but also enhances the time resolution, which helps to more accurately determine and analyze abnormal phenomena.

[0041] Specifically, the adaptive inverse convolution transform algorithm introduces multiple modulation and adaptive adjustment mechanisms in the convolution kernel design and inverse transformation process to better adapt to the complex abnormal signals in the electric energy meter data; the purified feature vector is input into the convolution kernel function, and the design of the convolution kernel takes into account the mixed modulation of time and frequency characteristics. The convolution kernel function is defined as:

[0042] in, represents the convolution kernel function, is the center position of the convolution kernel, Control the expansion factor of the convolution kernel, is the periodic parameter of the convolution kernel. The convolution kernel combines the smoothness of the Gaussian function and the periodicity of the sine and cosine functions, and can be adaptively adjusted to capture different time and frequency characteristics in the signal.

[0043] Through the convolution operation, the purified feature vector is processed into a time domain signal. The convolution process formula is as follows:

[0044] in, It is a frequency domain signal, which represents the representation of the convolved signal in the frequency domain. It is the shift parameter in the product operation, which indicates the amount of signal translation. The frequency domain signal contains the multi-level time characteristics of the original signal. Thanks to the adaptive modulation of the convolution kernel, it can suppress noise and irrelevant information while retaining the local abnormal characteristics of the signal.

[0045] After obtaining the frequency domain signal, in order to restore the complete time characteristics of the signal, the inverse transform operation is performed by combining the joint modulation of the frequency domain and the time domain. The formula is as follows:

[0046] in, Represents the reconstructed signal, the modulation term in the inverse transform It is used to dynamically adjust the frequency components to ensure that the time characteristics of the signal are preserved and enhanced during the reconstruction process. The reconstructed signal contains the complete time characteristics of the abnormal phenomena in the power meter data.

[0047] Extract features such as mean, variance, skewness, kurtosis, etc. from the reconstructed signal, input the extracted feature vector into the multiple iterative convergence algorithm of the feature space, and extract the potential optimal abnormal feature vector in the signal through multiple iterative optimization. In order to achieve this goal, the objective function is defined as follows:

[0048] in, is the objective function, It is the iteration coefficient matrix, which is used to control the weight transfer during the iteration process. is a regularization parameter used to avoid overfitting, It is The eigenvectors in the iterations, is the number of iterations. By optimizing the objective function multiple times, the optimal abnormal feature vector can be gradually approached.

[0049] According to the value and feature distribution of the optimal abnormal feature vector, each feature is analyzed to identify possible abnormal features. Specifically, the statistical distribution (such as mean and standard deviation) of each feature under normal operation is calculated and compared with the currently extracted features. If the value of a feature deviates from the normal distribution by more than a preset threshold, it can be considered that the current feature is abnormal.

[0050] In order to further determine the type of abnormal phenomenon, the optimal abnormal feature vector is classified using a pre-trained classification model, which can be a machine learning-based model, such as a support vector machine (SVM), random forest, or neural network. The input of the classification model is the optimal abnormal feature vector, and the output is the possible abnormal type. The final abnormal type is determined by maximizing the output probability of the classifier.

[0051] See attached Figure 2 , which shows a structural diagram of an electric energy meter abnormal phenomenon automatic detection and determination system provided by another embodiment of the present invention, the system comprising: Feature vector purification unit: collects multi-dimensional power parameter data in real time, maps the data into nonlinear space, constructs a discriminant matrix, optimizes the projection matrix by minimizing the objective function to effectively distinguish normal and abnormal data, and performs multi-stage dynamic filtering on the feature vector of the power parameter data to obtain the purified feature vector; Automatic detection and judgment unit: The purified feature vector is processed by an adaptive inverse convolution transform algorithm to reconstruct the time and space characteristics of the signal, generate a frequency domain signal, and then restore the complete time feature signal through inverse transform. The optimal abnormal feature vector is obtained through a multiple iterative convergence algorithm, and the optimal abnormal feature vector is classified to complete automatic detection and judgment.

[0052] It should be noted that each unit in the above-mentioned system for automatic detection and determination of abnormal phenomena of electric energy meters can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned units can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned units. For the specific definition of an automatic detection and determination system for abnormal phenomena of electric energy meters, please refer to the definition of an automatic detection and determination method for abnormal phenomena of electric energy meters mentioned above. The two have the same functions and effects, which will not be repeated here.

[0053] The present invention has the following beneficial effects: 1. The present invention realizes high-precision detection of potential anomalies in electric energy meter data through complex nonlinear mapping functions and adaptive weighted discriminant analysis. By performing deep nonlinear mapping on the multi-dimensional features of the original data, the abnormal patterns are fully displayed in the mapped high-dimensional space. The normal and abnormal data are further effectively distinguished through the optimized discriminant matrix, which significantly improves the accuracy of anomaly detection.

[0054] 2. Through the multi-stage dynamic filtering method and the adaptive inverse convolution transform algorithm, the present invention can effectively remove high-frequency noise and low-frequency interference information in the feature vector. It not only dynamically adjusts the frequency response, but also combines the multiple modulation functions of the convolution kernel, so that the retained signal characteristics are purer and more representative, effectively improving the signal-to-noise ratio.

[0055] 3. In the process of signal reconstruction, the adaptive inverse convolution transform algorithm combines the joint modulation of time domain and frequency domain, so that the reconstructed signal not only retains the original abnormal characteristics, but also enhances the temporal resolution; it helps the system to more accurately capture and analyze the occurrence time and development process of abnormal phenomena, and contributes to the accurate judgment of abnormal phenomena.

[0056] 4. The present invention fully reflects the complex characteristics of the electric energy meter data by extracting multiple features from the reconstructed signal, including time domain features, frequency domain features and nonlinear features; then, the features are optimized multiple times through a multiple iterative convergence algorithm in the feature space to finally obtain the optimal abnormal feature vector; in order to further refine and determine the type of abnormal phenomenon, a classification model based on machine learning is introduced. Through the trained model, the optimal abnormal feature vector can be intelligently classified to accurately identify different types of abnormal phenomena; this classification mechanism helps power companies quickly understand the specific types of abnormalities and take corresponding countermeasures.

[0057] The above-mentioned method for automatic detection and determination of abnormal phenomena in electric energy meters is used as follows.

[0058] 1. Data collection and preprocessing The data collected by the electric energy meter for power system monitoring includes parameters such as voltage, current, power factor, and power. Assume that the collected data set includes six dimensions of power parameters: voltage , Current , Active Power , reactive power , Power Factor , Harmonic Components , the data at each time point is a 6-dimensional vector.

[0059] For example, the power parameters collected at 5 time points are as follows:

[0060] The data at each time point is:

[0061]

[0062]

[0063]

[0064]

[0065] These data are collected and stored in real time by sensors. Then they are pre-processed by using simple filtering and denoising algorithms (such as low-pass filtering) to remove high-frequency noise and outliers.

[0066] 2. Nonlinear Mapping In order to extract deep nonlinear features from the original data, each data point Mapped to a new nonlinear space, the mapping function used is:

[0067] in, and , that is, use quadratic polynomial and linear sine function for mapping. Select , , .

[0068] For the first data point , the mapping result is:

[0069] right Perform the same mapping and get:

[0070] This results in the mapped dataset:

[0071] 3. Weighted Discriminant Analysis Use weighted discriminant analysis (WDA) to optimize the projection matrix W to distinguish normal and abnormal data. First, you need to build the intra-class scatter matrix and the between-class scatter matrix .

[0072] 3.1 Calculating the intra-class scatter matrix

[0073] Classify the data into 2 categories (normal and abnormal). Assumptions: Class 1 (Normal) contains data points

[0074] Class 2 (anomaly) contains data points

[0075] Compute the within-class scatter matrix The formula is:

[0076] in, and are the means of class 1 and class 2 respectively.

[0077] Mean of class 1:

[0078] Mean of class 2:

[0079] Then calculate the intra-class scatter matrix : First calculate the contribution of class 1:

[0080] Then calculate the contribution of class 2:

[0081] Therefore, the within-class scatter matrix for:

[0082] 3.2 Calculating the inter-class scatter matrix

[0083] Between-class scatter matrix The calculation formula is:

[0084] in, is the mean of all data points:

[0085] Therefore, the between-class scatter matrix is:

[0086] 3.3 Optimizing the projection matrix

[0087] Optimizing the projection matrix , that is, solving the objective function:

[0088] The calculation of the objective function is simplified to:

[0089] The objective function becomes:

[0090] By simplifying we get:

[0091] This means that the projection matrix needs to be adjusted The size of the within-class scatter matrix and the between-class scatter matrix The ratio of is as small as possible, so as to maximize the inter-class spread and minimize the intra-class spread.

[0092] Optimized projection matrix After iterative adjustment, a suitable projection direction is finally obtained:

[0093] 4. Feature purification (detailed calculation process) After optimizing the projection matrix After that, the feature vector needs to be purified. The purification process adopts a multi-stage dynamic filtering method.

[0094] 4.1 Design of frequency domain filter Choose a simple Gaussian filter as the frequency domain filter, the frequency response function is:

[0095] choose , , these two parameters are obtained through experimental adjustment. The specific calculation process of the frequency response function is as follows: Assume that the frequency of the signal (i.e. a sine wave with a period of 1 second), we get:

[0096] 4.2 Filtering Operation For the feature vector For filtering, assume that there is the following feature vector before purification:

[0097] The filtered feature vector This can be calculated using a frequency domain filter:

[0098] Assumptions is a constant signal, the response of the frequency domain filter is frequency-dependent, so it can be simply calculated:

[0099] For all feature vectors Perform the same frequency domain filtering to obtain:

[0100] In this way, the purified feature vector is:

[0101] 4.3 Inverse transformation from frequency domain to time domain The purified feature vector also needs to be reconstructed in the time domain through the adaptive inverse convolution transform algorithm. Assuming the selected convolution kernel function As shown below:

[0102] Assumption Selection , , and , then the convolution kernel function becomes:

[0103] Through the convolution operation, the reconstructed signal is obtained:

[0104] 5. Anomaly Detection and Classification 5.1 Abnormal feature extraction and multiple iterative optimization Through multiple iterative convergence algorithms, the features in the reconstructed signal are further optimized. The objective function is:

[0105] Assume the following experimental data and parameters: Initial eigenvector

[0106] Iteration coefficient

[0107] Regularization parameter

[0108] Iterations

[0109] 5.2 Calculation of the objective function First, for each iteration step, the first part of the objective function is used to calculate the difference with the previous iteration result. Consider the first iteration step, Substitute into the formula to calculate .

[0110] The first term in the objective function is:

[0111] In the first step, substitute and :

[0112] Compute the first term of the objective function:

[0113] 5.3 Part II: Regularization Term The second part of the objective function is the regularization term, which is calculated as follows:

[0114] First, calculate each 2-norm of . Assume that for the first eigenvector :

[0115] for calculate function:

[0116] Next, for each (Assumption is a constant value, and similarly calculated and Calculate the regularization term. The sum of the regularization terms is:

[0117] 5.4 Calculate the total value of the objective function Combining the first and second terms, the total value of the objective function is:

[0118] 5.5 Iterative Optimization By adjusting the feature vectors through multiple iterations, we gradually approach the optimal abnormal feature vector. In each iteration, we update By minimizing the objective function To obtain more accurate abnormal features. In actual operation, the gradient descent method or other optimization algorithms are used to solve these iterative steps. Assume that after multiple iterations, the optimal abnormal feature vector is obtained. :

[0119] 5.6 Anomaly Detection and Judgment After the abnormal feature extraction is completed, the optimal abnormal feature vector Compare with the statistical distribution of normal data. Assume that the statistical distribution of normal data is: The mean of normal data is 0.1969 The standard deviation of normal data is 0.0035 The value of each optimal eigenvector is compared with the distribution difference of normal data, and the following judgment criteria are used: if the value of a feature deviates from the normal mean by more than 3 standard deviations, the feature is judged to be abnormal.

[0120] for , and its deviation is , which is less than 3 standard deviations (3*0.0035 = 0.0105), so the feature is normal.

[0121] for , and its deviation is , which is greater than 3 standard deviations, so this feature is abnormal.

[0122] for , and its deviation is , which is less than 3 standard deviations, so the feature is normal.

[0123] for , and its deviation is , which is less than 3 standard deviations, so the feature is normal.

[0124] for , and its deviation is , which is less than 3 standard deviations, so the feature is normal.

[0125] According to the above judgment, only It is determined to be an abnormal feature. Finally, combined with the analysis of various abnormal features, it can be concluded that the electric energy meter has abnormal phenomena at certain times. The sequence of the embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0126] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0127] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A method for automatically detecting and determining abnormal phenomena of an electric energy meter, characterized in that: Includes steps: S1. Real-time collection of multi-dimensional power parameter data, mapping the data to a nonlinear space, constructing a discriminant matrix, optimizing the projection matrix by minimizing the objective function to effectively distinguish normal and abnormal data, and performing multi-stage dynamic filtering on the feature vector of the power parameter data to obtain a purified feature vector; S2. The purified feature vector is processed by an adaptive inverse convolution transform algorithm to reconstruct the time and space characteristics of the signal and generate a frequency domain signal. The complete time feature signal is then restored through inverse transform. The optimal abnormal feature vector is obtained through a multiple iterative convergence algorithm. The optimal abnormal feature vector is classified to complete automatic detection and judgment.

2. The method for automatic detection and determination of abnormal phenomena of electric energy meters according to claim 1 is characterized in that: In S1, the collected multi-dimensional power parameter data is preprocessed, and nonlinear mapping is performed on the preprocessed multi-dimensional data to extract deep nonlinear features in the data.

3. The method for automatic detection and determination of abnormal phenomena of electric energy meters according to claim 1 is characterized in that: In S1, adaptive weighted discriminant analysis is performed using the mapped data set, and a weighted discriminant matrix is ​​constructed to effectively distinguish normal data from abnormal data in a nonlinear space.

4. The method for automatically detecting and determining abnormal phenomena of an electric energy meter according to claim 1, characterized in that: In S1, an objective function is set up to obtain an optimal projection matrix, the intra-class scatter matrix is ​​subjected to eigenvalue decomposition to find its corresponding eigenvector, and the eigenvectors are connected to correspond to the direction in which the intra-class scatter is minimized; by solving the generalized eigenvalue problem, the direction in which the inter-class scatter is maximized is determined; and the eigenvectors with the largest first eigenvalues ​​are selected as column vectors of the projection matrix.

5. The method for automatically detecting and determining abnormal phenomena of an electric energy meter according to claim 1, characterized in that: In S1, in order to purify the feature vector, a multi-stage dynamic filtering method is adopted, a frequency domain filter is designed based on the spectral characteristics of the abnormal data, and each feature vector is frequency-domain filtered to remove high-frequency noise and low-frequency fluctuations. After applying the frequency domain filter, the feature vector is converted into a purified feature vector.

6. The method for automatically detecting and determining abnormal phenomena of an electric energy meter according to claim 1, characterized in that: In S2, the adaptive convolution inverse transform algorithm introduces multiple modulations and adaptive adjustment mechanisms in the convolution kernel design and inverse transform process to better adapt to the complex abnormal signals in the electric energy meter data; the purified feature vector is input into the convolution kernel function, and the design of the convolution kernel takes into account the mixed modulation of time and frequency characteristics.

7. The method for automatically detecting and determining abnormal phenomena of an electric energy meter according to claim 1, characterized in that: In S2, the purified feature vector is processed into a time domain signal through a convolution operation. After obtaining the frequency domain signal, an inverse transform operation is performed by combining joint modulation of the frequency domain and the time domain to restore the complete time characteristics of the signal.

8. The method for automatic detection and determination of abnormal phenomena of electric energy meters according to claim 1, characterized in that: In S2, features are extracted from the reconstructed signal, and the extracted feature vector is input into a multiple iterative convergence algorithm of the feature space. Through multiple iterative optimizations, the potential optimal abnormal feature vector in the signal is extracted.

9. The method for automatically detecting and determining abnormal phenomena of an electric energy meter according to claim 8, characterized in that: In S2, each feature is analyzed according to the value and feature distribution of the optimal abnormal feature vector to identify possible abnormal features; the statistical distribution of each feature under normal operating conditions is calculated and compared with the currently extracted features. If the value of a certain feature deviates from the normal distribution to a greater extent than a preset threshold, the current feature is considered to be abnormal; in order to further determine the type of abnormal phenomenon, the optimal abnormal feature vector is classified using a pre-trained classification model.

10. An automatic detection and determination system for abnormal phenomena of electric energy meters, used to implement the automatic detection and determination method for abnormal phenomena of electric energy meters according to any one of claims 1 to 9, characterized in that: include: Feature vector purification unit: collects multi-dimensional power parameter data in real time, maps the data into nonlinear space, constructs a discriminant matrix, optimizes the projection matrix by minimizing the objective function to effectively distinguish normal and abnormal data, and performs multi-stage dynamic filtering on the feature vector of the power parameter data to obtain the purified feature vector; Automatic detection and judgment unit: The purified feature vector is processed by an adaptive inverse convolution transform algorithm to reconstruct the time and space characteristics of the signal, generate a frequency domain signal, and then restore the complete time feature signal through inverse transform. The optimal abnormal feature vector is obtained through a multiple iterative convergence algorithm, and the optimal abnormal feature vector is classified to complete automatic detection and judgment.

Images