Fault electric energy meter nondestructive testing method and device based on broadband impedance spectroscopy and computer equipment

Abstract

The embodiment of the invention provides a fault electric energy meter nondestructive testing method and device based on a broadband impedance spectrum, computer equipment, a storage medium and a computer program product, and relates to the technical field of electric power detection. The method comprises the following steps: acquiring target impedance data of a to-be-measured electric energy meter; according to the target impedance data, obtaining an impedance grey-scale map of the to-be-measured electric energy meter; obtaining the fault probability of the to-be-detected electric energy meter based on the impedance grey-scale map by using a pre-trained fault recognition model; and obtaining a fault detection result of the to-be-detected electric energy meter according to the fault probability. According to the method, the intelligence of electric energy meter fault detection is improved, the operation cost of electric energy meter fault detection is reduced, and then the efficiency, simplicity and accuracy of electric energy meter fault detection can be improved.

Description

Technical Field

[0001] The present application relates to the field of electric power detection technology, and in particular to a method, device, computer equipment, storage medium and computer program product for nondestructive detection of faulty electric energy meters based on broadband impedance spectrum. Background Art

[0002] With the rapid development of smart grids, energy meters, as core equipment for power metering and information collection, have been widely used in power systems. Energy meters not only undertake the basic function of accurately measuring electricity, but also transmit power consumption data in real time through communication networks, providing data support for power management, user demand response, and energy optimization configuration. The realization of this function has greatly improved the automation level of the power system and effectively improved the utilization efficiency of power resources. However, with the extension of the service life of energy meters and the complexity of their working environment, their fault types have gradually increased and their manifestations have become increasingly diverse.

[0003] Especially in the context of smart grids, energy meters are widely installed in various environments, operating under harsh conditions such as high temperature and humidity, low temperature and severe cold, and strong electromagnetic interference, which significantly increases the risk of failure of energy meters. At the same time, due to differences in equipment components and manufacturing processes from different suppliers, the failure performance of installed energy meters varies. Some energy meters may have potential problems only after long-term operation, while others may show obvious signs of failure in a short period of time. This complex situation undoubtedly increases the difficulty and complexity of fault detection of retired energy meters.

[0004] In the existing research on electric energy meters, relevant literature has proposed a variety of methods to deal with the problems of electric energy meter fault detection and quality control, such as using association models to analyze the correlation between the fault phenomenon of the electric energy meter and the failure of components, or using image processing technology to realize the automatic detection of the quality of the electric energy meter display panel. However, most studies are still limited to open-cover detection or appearance detection, and cannot comprehensively and effectively identify the faults of the internal components of the electric energy meter. Although the open-cover detection can intuitively observe the internal components of the electric energy meter, the operation is complicated and there are high labor costs and safety risks. The appearance detection method can only identify surface damage or defects, and cannot detect internal circuit faults. Although penetrating detection methods such as X-ray imaging technology can identify the damage of the internal circuit of the electric energy meter, its equipment cost is high, the operation is complicated, and it is difficult to apply in large-scale detection. There is a problem of low efficiency in electric energy meter fault detection. Summary of the invention

[0005] Based on this, it is necessary to provide a method, device, computer equipment, storage medium and computer program product for non-destructive detection of faulty electric energy meters based on broadband impedance spectrum in order to solve the above technical problems.

[0006] In a first aspect, the present application provides a method for nondestructive detection of a faulty electric energy meter based on a broadband impedance spectrum. The method comprises:

[0007] Obtain target impedance data of the electric energy meter to be tested;

[0008] According to the target impedance data, an impedance grayscale image of the electric energy meter to be tested is obtained;

[0009] Using a pre-trained fault identification model, based on the impedance grayscale image, the fault probability of the electric energy meter to be tested is obtained;

[0010] According to the fault probability, a fault detection result of the electric energy meter to be tested is obtained.

[0011] In one of the embodiments, the step of obtaining target impedance data of the electric energy meter to be tested includes: obtaining initial impedance data of the electric energy meter to be tested; and performing data preprocessing on the initial impedance data to obtain the target impedance data.

[0012] In one embodiment, the data preprocessing of the initial impedance data to obtain the target impedance data includes: data cleaning of the initial impedance data to obtain the cleaned initial impedance data; and normalizing the cleaned initial impedance data to obtain the target impedance data.

[0013] In one of the embodiments, obtaining the impedance grayscale map of the electric energy meter to be tested based on the target impedance data includes: obtaining the grayscale value of each of the multiple data points of the target impedance data; and according to the grayscale values, plotting the multiple data points on a two-dimensional plane according to a preset arrangement rule to obtain the impedance grayscale map of the electric energy meter to be tested.

[0014] In one embodiment, the method of using a pre-trained fault identification model to obtain the fault probability of the electric energy meter to be tested based on the impedance grayscale map includes: inputting the impedance grayscale map into the pre-trained fault identification model, determining and outputting the fault probability of the electric energy meter to be tested.

[0015] In one embodiment, the step of inputting the impedance grayscale map into the pre-trained fault identification model to determine and output the fault probability of the electric energy meter to be tested comprises: inputting the impedance grayscale map into the pre-trained fault identification model, and having a feature extraction module of the fault identification model determine and output the impedance characteristics of the electric energy meter to be tested based on the impedance grayscale map; and having a classification module of the fault identification model determine and output the fault probability of the electric energy meter to be tested based on the impedance characteristics output by the feature extraction module.

[0016] In one embodiment, before obtaining the target impedance data of the electric energy meter to be tested, the method further includes:

[0017] Obtain sample impedance data of a sample electric energy meter under each of the electric energy meter fault types in a plurality of electric energy meter fault types; obtain a sample impedance grayscale map corresponding to each of the electric energy meter fault types based on the sample impedance data; divide the sample impedance grayscale map into a training sample impedance grayscale map and a test sample impedance grayscale map; obtain an initial fault recognition model using the training sample impedance grayscale map; optimize the initial fault recognition model based on the test sample impedance grayscale map to obtain the pre-trained fault recognition model.

[0018] In a second aspect, the present application provides a nondestructive detection device for a faulty electric energy meter based on a broadband impedance spectrum. The device comprises:

[0019] A data measurement module, used to obtain target impedance data of the electric energy meter to be tested;

[0020] A drawing module, used for obtaining an impedance grayscale image of the electric energy meter to be tested according to the target impedance data;

[0021] A calculation module, used for obtaining the fault probability of the electric energy meter to be tested based on the impedance grayscale map by using a pre-trained fault identification model;

[0022] The result acquisition module is used to acquire the fault detection result of the electric energy meter to be tested according to the fault probability.

[0023] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the following steps when executing the computer program:

[0024] Obtain target impedance data of the electric energy meter to be tested;

[0025] According to the target impedance data, an impedance grayscale image of the electric energy meter to be tested is obtained;

[0026] Using a pre-trained fault identification model, based on the impedance grayscale image, the fault probability of the electric energy meter to be tested is obtained;

[0027] According to the fault probability, a fault detection result of the electric energy meter to be tested is obtained.

[0028] In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0029] Obtain target impedance data of the electric energy meter to be tested;

[0030] According to the target impedance data, an impedance grayscale image of the electric energy meter to be tested is obtained;

[0031] Using a pre-trained fault identification model, based on the impedance grayscale image, the fault probability of the electric energy meter to be tested is obtained;

[0032] According to the fault probability, a fault detection result of the electric energy meter to be tested is obtained.

[0033] In a fifth aspect, the present application further provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0034] Obtain target impedance data of the electric energy meter to be tested;

[0035] According to the target impedance data, an impedance grayscale image of the electric energy meter to be tested is obtained;

[0036] Using a pre-trained fault identification model, based on the impedance grayscale image, the fault probability of the electric energy meter to be tested is obtained;

[0037] According to the fault probability, a fault detection result of the electric energy meter to be tested is obtained.

[0038] In the above-mentioned method, device, computer equipment, storage medium and computer program product for nondestructive detection of faulty electric energy meters based on broadband impedance spectrum, first, the target impedance data of the electric energy meter to be tested can be obtained; next, the impedance grayscale map of the electric energy meter to be tested can be obtained based on the target impedance data; then, the fault probability of the electric energy meter to be tested can be obtained based on the impedance grayscale map using a pre-trained fault identification model; finally, the fault detection result of the electric energy meter to be tested can be obtained based on the fault probability. In the method provided in the embodiment of the present application, the fault detection result of the electric energy meter to be tested can be obtained based on the impedance data of the electric energy meter to be tested in combination with a pre-trained fault identification model, thereby avoiding the need to open the cover for detection of the electric energy meter to be tested, improving the intelligence of the electric energy meter fault detection, reducing the operating cost of the electric energy meter fault detection, and further, improving the efficiency, simplicity and accuracy of the electric energy meter fault detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0040] FIG1 is a schematic flow chart of a method for nondestructive detection of a faulty electric energy meter based on broadband impedance spectroscopy provided in an embodiment of the present application;

[0041] FIG2 is a schematic diagram of a convolutional neural network model provided in an embodiment of the present application;

[0042] FIG3 is a schematic diagram of a process for acquiring target impedance data provided by an embodiment of the present application;

[0043] FIG4 is a schematic diagram of a data preprocessing process provided by an embodiment of the present application;

[0044] FIG5 is a schematic diagram of a process for obtaining an impedance grayscale image provided in an embodiment of the present application;

[0045] FIG6 is a schematic diagram of a process for obtaining a fault identification model provided in an embodiment of the present application;

[0046] FIG7 is a schematic diagram of a grayscale image provided in an embodiment of the present application;

[0047] FIG8 is a graph showing the accuracy of a CNN model on a training set and a test set provided by an embodiment of the present application;

[0048] FIG9 is a loss function curve diagram of a CNN model provided in an embodiment of the present application on a training set and a test set;

[0049] FIG10 is a structural block diagram of a nondestructive detection device for a faulty electric energy meter based on a broadband impedance spectrum provided in an embodiment of the present application;

[0050] FIG. 11 is an internal structure diagram of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0052] In an exemplary embodiment, as shown in FIG1 , a method for nondestructive detection of a faulty electric energy meter based on a broadband impedance spectrum is provided. This embodiment is illustrated by applying the method to a server. It is understandable that the method can also be applied to a terminal, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0053] Step 102, obtaining target impedance data of the electric energy meter to be tested.

[0054] The electric energy meter to be tested can be any electric energy meter that needs fault detection. In this step, first, the initial impedance data of the electric energy meter to be tested can be obtained; then, the initial impedance data can be preprocessed to obtain the target impedance data corresponding to the initial impedance data. The acquisition of the initial impedance data of the electric energy meter to be tested can include: the impedance of the electric energy meter to be tested can be measured. Specifically, an AC voltage signal of different frequencies can be applied to the port of the electric energy meter through an impedance tester, and the port current can be measured to obtain the port impedance, and the port impedance includes the amplitude and phase angle of the impedance. The amplitude of the impedance reflects the impedance capacity of the internal circuit of the electric energy meter to the current, and the phase angle reveals the phase response of the internal components of the electric energy meter to the current and voltage. The data preprocessing can include data cleaning and data normalization processing, wherein data cleaning: remove invalid or noise data to ensure data quality. The cleaned data only contains the real impedance measurement value. Normalization processing: the dimension and range of the cleaned initial impedance data may be greatly different, and it needs to be normalized to the interval of 0 to 1. In this step, the minimum-maximum normalization (Min-Max Scaling) method is used to normalize each of the multiple data points of the cleaned initial impedance data to obtain the target impedance data of the electric energy meter to be tested, which can be seen from formula (1):

[0055]

[0056] in, Represents the normalized value of a single data. is the minimum value of the column where the data is located. is the maximum value of the column where the data is located. The actual value of a single data.

[0057] Step 104: Obtain an impedance grayscale image of the electric energy meter to be tested according to the target impedance data.

[0058] Among them, the target impedance data can be converted into a grayscale image, which is an image that represents numerical data in grayscale, and the grayscale value range is usually between 0 and 255. Specifically, in the first step, each of the multiple data points of the target impedance data is mapped to a grayscale value between 0 and 255. In the second step, each target impedance data is plotted on a two-dimensional plane according to a certain arrangement rule to generate an impedance grayscale image of the electric energy meter to be tested. Each pixel of the impedance grayscale image corresponds to an impedance measurement value, black (grayscale value of 0) represents low impedance, and white (grayscale value of 255) represents high impedance, as shown in formula (2):

[0059]

[0060] in, for The normalized impedance measurement value, i.e. the target impedance data, The grayscale value converted to the target impedance data.

[0061] Step 106, using the pre-trained fault identification model, based on the impedance grayscale image, obtain the fault probability of the electric energy meter to be tested.

[0062] As shown in FIG2 , the pre-trained fault recognition model may be a convolutional neural network (CNN). The pre-trained fault recognition model may include an input layer, at least one convolutional layer, at least one pooling layer, a fully connected layer, and an output layer. The input layer may receive an input grayscale image of an electric energy meter, and the size of the input layer is set to 300×400×1 considering the size of the model; the convolutional layer may use multiple convolution kernels to extract features from the input image, with depths of 32 and 64, respectively. The size of the convolution kernel is 5×5, and the moving step is 1. The convolution operation can capture local features in the grayscale image, such as the mutation point of impedance or the abnormal response at a specific frequency; the pooling layer can downsample the feature map output by the convolution layer to reduce the dimension of the data and the complexity of the model while maintaining the main features. The filter size of the pooling layer is 2×2, the moving step is 2, and the model adopts the maximum pooling layer (Max pooling); in a possible implementation, the pre-trained fault recognition model may include a flattening layer, which can flatten the pooled feature map into a one-dimensional vector to facilitate the subsequent processing of the fully connected layer; the fully connected layer can be responsible for further nonlinear transformation of the flattened features to form the final classification result. The number of nodes in the fully connected layer is 64. During the training process, the Dropout layer is added to prevent overfitting; the output layer can use a nonlinear activation function to convert the output of the CNN into a probability distribution of faults and non-faults.

[0063] The convolution layer is used to extract the features of the input data; the pooling layer is used to sample the input data. Both the convolution layer and the pooling layer include an activation function. The impedance grayscale map may include multiple impedance data.

[0064] Specifically, the impedance grayscale image of the electric energy meter to be tested can be input into the pre-trained fault identification model, and the convolution layer can be used to extract the initial features of multiple impedance data. The first step is to convert multiple impedance data into vectors to obtain multiple impedance data vectors, and the multiple impedance data vectors can be combined into an impedance data vector matrix; the second step is to input the impedance data vector matrix into the convolution layer, and use the convolution kernel to perform convolution operation with the impedance data vector matrix, that is, to perform inner product operation on the impedance data vector matrix and the convolution kernel to obtain the convolution result corresponding to the impedance data vector matrix; next, based on the activation function, the convolution result is nonlinearly transformed, and the bias vector is added to obtain the initial feature vector; the third step is to input the initial feature vector into the pooling layer, and the initial feature vector can be sampled; then, based on the activation function, the feature sampling result is nonlinearly transformed, and the bias vector is added to obtain the impedance characteristics of the electric energy meter to be tested.

[0065] The fully connected layer can classify the impedance features and obtain the fault label of the electric energy meter to be tested. In addition, the fully connected layer can include an activation function, and the activation function includes a weight matrix and a bias constant.

[0066] Specifically, the impedance feature can be input into the fully connected layer, and the impedance feature can be nonlinearly transformed based on the weight matrix and bias vector of the activation function, and then normalized to obtain the fault label of the electric energy meter to be tested. The fault label can indicate whether the electric energy meter to be tested has a fault.

[0067] Step 108: Obtain the fault detection result of the electric energy meter to be tested according to the fault probability.

[0068] Among them, the fault probability can be used to characterize whether the electric energy meter to be tested has a fault. The fault probability includes the probability of the presence of a fault and the probability of the absence of a fault, and it can be determined whether the electric energy meter to be tested has a fault based on the probability. If the probability of the presence of a fault is greater than the probability of the absence of a fault, it can be determined that the electric energy meter to be tested has a fault.

[0069] In the method of this embodiment, first, the target impedance data of the electric energy meter to be tested can be obtained; next, the impedance grayscale map of the electric energy meter to be tested can be obtained based on the target impedance data; then, the fault probability of the electric energy meter to be tested can be obtained based on the impedance grayscale map using a pre-trained fault identification model; finally, the fault detection result of the electric energy meter to be tested can be obtained based on the fault probability. In the method provided by the embodiment of the present application, the fault detection result of the electric energy meter to be tested can be obtained based on the impedance data of the electric energy meter to be tested in combination with a pre-trained fault identification model, thereby avoiding the need to open the cover of the electric energy meter to be tested, improving the intelligence of the electric energy meter fault detection, reducing the operating cost of the electric energy meter fault detection, and further, improving the efficiency, simplicity and accuracy of the electric energy meter fault detection.

[0070] In an exemplary embodiment, as shown in FIG3 , step 102 may include steps 302 to 304. Among them:

[0071] Step 302: Acquire initial impedance data of the electric energy meter to be tested.

[0072] Step 304: preprocess the initial impedance data to obtain target impedance data.

[0073] The data preprocessing may include data cleaning and data normalization.

[0074] In an exemplary embodiment, as shown in FIG4 , step 304 may include steps 402 to 404. Among them:

[0075] Step 402, cleaning the initial impedance data to obtain cleaned initial impedance data.

[0076] Data cleaning: removes invalid or noisy data to ensure data quality. The cleaned data only contains true impedance measurement values.

[0077] Step 404, normalize the initial impedance data after cleaning to obtain target impedance data.

[0078] Among them, the dimension and range of the initial impedance data after cleaning may be greatly different, and it needs to be normalized to the range of 0 to 1. In this step, the minimum-maximum normalization (Min-Max Scaling) method is used to normalize each of the multiple data points of the initial impedance data after cleaning to obtain the target impedance data of the electric energy meter to be tested, which can be seen in formula (1).

[0079] In the method of this embodiment, when obtaining the initial impedance data of the electric energy meter to be tested, the initial impedance data can be cleaned and normalized to obtain the normalized impedance measurement value of the electric energy meter to be tested, thereby improving the accuracy of data processing and, further, improving the accuracy of fault detection.

[0080] In an exemplary embodiment, as shown in FIG5 , step 104 may include steps 502 to 504. Among them:

[0081] Step 502: Obtain the grayscale value of each data point among a plurality of data points of target impedance data.

[0082] Step 504 , based on the grayscale values, multiple data points are plotted on a two-dimensional plane according to a preset arrangement rule to obtain an impedance grayscale map of the electric energy meter to be tested.

[0083] The target impedance data can be converted into a grayscale image, which is an image that represents numerical data in grayscale, and the grayscale value range is usually between 0 and 255. Specifically, in the first step, each of the multiple data points of the target impedance data is mapped to a grayscale value between 0 and 255. In the second step, each target impedance data is plotted on a two-dimensional plane according to a certain arrangement rule to generate an impedance grayscale image of the electric energy meter to be tested. Each pixel of the impedance grayscale image corresponds to an impedance measurement value, black (grayscale value 0) represents low impedance, and white (grayscale value 255) represents high impedance, as shown in formula (2).

[0084] In the method of this embodiment, the target impedance data of the electric energy meter to be tested can be converted into a grayscale image, and then the grayscale image is convolved, thereby avoiding a complex feature extraction process, improving calculation efficiency, and further improving the detection efficiency of fault detection.

[0085] In an exemplary embodiment, step 106 may include:

[0086] The impedance grayscale image is input into the pre-trained fault identification model to determine and output the fault probability of the electric energy meter to be tested.

[0087] In an exemplary embodiment, the step of inputting the impedance grayscale image into a pre-trained fault identification model to determine and output the fault probability of the electric energy meter to be tested may include:

[0088] The impedance grayscale image is input into a pre-trained fault recognition model, and the feature extraction module of the fault recognition model determines and outputs the impedance characteristics of the electric energy meter to be tested based on the impedance grayscale image; the classification module of the fault recognition model determines and outputs the fault probability of the electric energy meter to be tested based on the impedance characteristics output by the feature extraction module.

[0089] The pre-trained fault recognition model can be a convolutional neural network (CNN). The pre-trained fault recognition model may include an input layer, at least one convolutional layer, at least one pooling layer, a fully connected layer, and an output layer. The feature extraction module of the fault recognition model may include at least one convolutional layer and at least one pooling layer; the classification module of the fault recognition model may include at least one fully connected layer. Among them, the input layer can receive the input grayscale image of the electric energy meter, and the size of the input layer is set to 300×400×1 considering the size of the model; the convolution layer can use multiple convolution kernels to extract features from the input image, with depths of 32 and 64 respectively. The convolution kernel size is 5×5, and the moving step is 1. The convolution operation can capture local features in the grayscale image, such as the mutation point of impedance or the abnormal response at a specific frequency; the pooling layer can downsample the feature map output by the convolution layer to reduce the dimension of the data and the complexity of the model while maintaining the main features. The filter size of the pooling layer is 2×2, the moving step is 2, and the model adopts the maximum pooling layer (Max pooling); in a possible implementation, the pre-trained fault recognition model may include a flattening layer, which can flatten the pooled feature map into a one-dimensional vector to facilitate the subsequent processing of the fully connected layer; the fully connected layer can be responsible for further nonlinear transformation of the flattened features to form the final classification result. The number of nodes in the fully connected layer is 64. During the training process, the Dropout layer is added to prevent overfitting; the output layer can use a nonlinear activation function to convert the output of the CNN into a probability distribution of faults and non-faults.

[0090] The convolution layer is used to extract the features of the input data; the pooling layer is used to sample the input data. Both the convolution layer and the pooling layer include an activation function. The impedance grayscale map may include multiple impedance data.

[0091] Specifically, the impedance grayscale image of the electric energy meter to be tested can be input into the pre-trained fault identification model, and the convolution layer can be used to extract the initial features of multiple impedance data. The first step is to convert multiple impedance data into vectors to obtain multiple impedance data vectors, and the multiple impedance data vectors can be combined into an impedance data vector matrix; the second step is to input the impedance data vector matrix into the convolution layer, and use the convolution kernel to perform convolution operation with the impedance data vector matrix, that is, to perform inner product operation on the impedance data vector matrix and the convolution kernel to obtain the convolution result corresponding to the impedance data vector matrix; next, based on the activation function, the convolution result is nonlinearly transformed, and the bias vector is added to obtain the initial feature vector; the third step is to input the initial feature vector into the pooling layer, and the initial feature vector can be sampled; then, based on the activation function, the feature sampling result is nonlinearly transformed, and the bias vector is added to obtain the impedance characteristics of the electric energy meter to be tested.

[0092] The fully connected layer can classify the impedance features and obtain the fault label of the electric energy meter to be tested. In addition, the fully connected layer can include an activation function, and the activation function includes a weight matrix and a bias constant.

[0093] Specifically, the impedance feature can be input into the fully connected layer, and the impedance feature can be nonlinearly transformed based on the weight matrix and bias vector of the activation function, and then normalized to obtain the fault label of the electric energy meter to be tested. The fault label can indicate whether the electric energy meter to be tested has a fault.

[0094] In an exemplary embodiment, as shown in FIG6 , the embodiment of the present application further includes the following steps, which are used to provide an implementation method for obtaining a fault identification model, specifically including steps 602 to 610. Among them:

[0095] Step 602: Obtain sample impedance data of a sample electric energy meter under each electric energy meter fault type among multiple electric energy meter fault types.

[0096] Among them, multiple fault types of the sample electric energy meter can be determined, and the multiple fault types include but are not limited to capacitors (power modules), LCD screens, optocouplers (metering circuits), main chips, Bluetooth modules, thermistors, metering chips, transformers, and voltage regulator chips (main circuits). Next, multiple sample impedance data under each of the multiple fault types can be obtained. In addition, multiple sample impedance data of the sample electric energy meter under a fault-free condition are obtained as reference samples. In a possible implementation, as shown in the fault type classification table in Table 1, the multiple fault types can be numbered to obtain the fault numbers corresponding to each fault type, and the number of sample impedance data corresponding to each fault type can be determined. For example, the capacitor (power module) corresponds to 3 sample impedance data and the LCD screen corresponds to 4 sample impedance data. As shown in Table 1 and Figure 7, there are 10 types of fault types for the sample electric energy meter (including normal meters), and the specific name and corresponding number of samples of each fault type are shown in Figure 1, and the 10 fault types are numbered 0 to 9 respectively. Among the normal electric energy meters, two are from the same manufacturer and the other is from a different manufacturer. For the faulty electric energy meters, except for the three types of electric energy meters of modules 6-8, which come from the same manufacturer, the other electric energy meters of faults 1-5 and 9 come from different manufacturers. This distribution shows that although some faulty meters come from the same manufacturer, the distribution of fault types is relatively wide, involving multiple different manufacturers. In this step, it can also be understood as obtaining the sample impedance data of the sample electric energy meter and the fault type label of the sample impedance data, which is used to mark the fault type of the corresponding sample impedance data.

[0097]

[0098] Table 1 Fault type classification table

[0099] As shown in Figure 7, (a) and (b) in the figure give grayscale images of normal energy meters from two different manufacturers, and (c) and (d) in the figure give grayscale images of faulty energy meters caused by different faulty components of the same manufacturer (where (c) and (d) are faults caused by thermistor module and metering chip module respectively). (e) and (f) in the figure show grayscale images of faulty energy meters caused by different faulty components of different manufacturers (where (e) and (f) are faults caused by main chip module and Bluetooth module respectively). These images show that regardless of whether the energy meters are from the same manufacturer or which specific faulty component causes it, as long as it is a faulty energy meter, its grayscale image is significantly different from that of a normal energy meter. This difference enables the convolutional neural network to accurately identify the faulty energy meter, regardless of its source and specific faulty component. The mechanism behind this is that although the energy meters produced by different manufacturers may differ in design and manufacturing, the internal equivalent circuit of a normal energy meter is basically the same; similarly, the internal circuit of a faulty energy meter will also show significantly different characteristics from that of a normal energy meter due to the presence of a fault. Therefore, convolutional neural networks can effectively distinguish normal electricity meters from faulty electricity meters by capturing these feature differences.

[0100] Step 604: Obtain a sample impedance grayscale image corresponding to each electric energy meter fault type according to the sample impedance data.

[0101] Among them, data preprocessing can be performed on at least one sample impedance data of each type of electric energy meter fault to obtain processed sample impedance data of each type of electric energy meter fault. The data preprocessing may include data cleaning and data normalization processing, wherein data cleaning: removing invalid or noise data to ensure data quality, and the cleaned data only contains real impedance measurement values. Normalization processing: the dimension and range of the cleaned sample impedance data may be greatly different, and it is necessary to normalize it to the interval of 0 to 1. In this step, the minimum-maximum normalization (Min-MaxScaling) method is adopted to normalize each of the multiple data points of the cleaned sample impedance data to obtain the processed sample impedance data of each type of electric energy meter fault, which can be seen in formula (1). Furthermore, according to at least one processed sample impedance data of each type of electric energy meter fault, the sample impedance grayscale map corresponding to each electric energy meter fault type can be obtained. Specifically, in the first step, each of the multiple data points of at least one processed sample impedance data is mapped to a grayscale value between 0 and 255. In the second step, the grayscale value corresponding to at least one processed sample impedance data is plotted on a two-dimensional plane according to a certain arrangement rule to generate a sample impedance grayscale map corresponding to each electric energy meter fault type. Each pixel of the sample impedance grayscale map corresponds to a sample impedance measurement value, black (grayscale value 0) represents low impedance, and white (grayscale value 255) represents high impedance, as shown in formula (3):

[0102]

[0103] in, for The normalized impedance measurement value, that is, the processed sample impedance data, is the grayscale value of the processed sample impedance data.

[0104] Step 606: divide the sample impedance grayscale map into a training sample impedance grayscale map and a test sample impedance grayscale map.

[0105] Among them, multiple sample impedance grayscale images corresponding to multiple electric energy meter fault types can be divided into training sample impedance grayscale images and test sample impedance grayscale images. The training sample impedance grayscale images can be used to train the fault recognition model to be trained to obtain a trained fault recognition model. Furthermore, the test sample impedance grayscale images can be used to optimize the trained fault recognition model to obtain an optimized fault recognition model, that is, the pre-trained fault recognition model.

[0106] Step 608: using the training sample impedance grayscale image, obtain an initial fault recognition model.

[0107] The training sample impedance grayscale image can be used to adjust the weight of the convolution kernel of the fault recognition model to be trained, so that the model can accurately identify the fault state of the electric energy meter. Specifically, the training sample impedance grayscale image and the actual fault label of the training sample impedance grayscale image can be input into the fault recognition model to be trained, and the feature extraction module in the fault recognition model to be trained can obtain the sample impedance feature corresponding to the training sample impedance grayscale image according to the training sample impedance grayscale image, and the classification module in the fault recognition model to be trained can obtain the predicted fault label of the training sample impedance grayscale image according to the sample impedance feature output by the feature extraction module. Then, the fault recognition model to be trained can be trained according to the difference between the actual fault label and the predicted fault label of the training sample impedance grayscale image to obtain a trained fault recognition model. Among them, the feature extraction module of the fault recognition model can include at least one convolution layer and at least one pooling layer; the classification module of the fault recognition model can include at least one fully connected layer.

[0108] Step 610, optimizing the initial fault recognition model according to the test sample impedance grayscale image to obtain a pre-trained fault recognition model.

[0109] Among them, in the test phase, the impedance grayscale image of the test sample can be input into the trained fault identification model, and the trained fault identification model determines and outputs the predicted fault label of the impedance grayscale image of the test sample. Then, the classification performance of the trained fault identification model can be judged according to the difference between the actual fault label and the predicted fault label of the impedance grayscale image of the test sample. For example, the classification accuracy and error performance index can be calculated, and the classification performance of the trained fault identification model can be evaluated according to the classification accuracy and error performance index to obtain the evaluation result of the trained fault identification model. Then, according to the evaluation result, the trained fault identification model can be optimized to obtain the optimized fault identification model, that is, the final version of the pre-trained fault identification model mentioned above. In a possible implementation method, the accuracy of the trained fault identification model, which can be called the initial fault identification model, can be obtained, and the prediction accuracy of the initial fault identification model can be evaluated according to the accuracy, which can be seen in formula (4):

[0110]

[0111] From the accuracy curves of the CNN model on the training set and the test set in Figure 8, it can be seen that the CNN model performs quite well in fault meter identification, with high accuracy on both the training set and the test set, and the difference between the two is not large. This shows that the model did not show obvious overfitting during the training process and can be effectively generalized to unseen data. Especially in distinguishing between normal meters and faulty meters, no matter which component caused the fault, the model showed high accuracy and robustness. These results show that the CNN model is very suitable for the fault identification problem of returned electric energy meters, and has strong predictive ability and stability.

[0112] The difference between the actual fault label and the predicted fault label of the training sample impedance grayscale image, and the difference between the actual fault label and the predicted fault label of the test sample impedance grayscale image can be measured by the loss function. The loss function is an important indicator for measuring the prediction error of the model. It represents the difference between the model prediction value, that is, the predicted fault label, and the actual value, that is, the actual fault label. The smaller the value, the more accurate the model prediction. In a possible implementation method, the binary cross-entropy loss function can be used to measure the difference between the actual fault label and the predicted fault label, as shown in formula (5):

[0113]

[0114] in, is the actual fault label (0 or 1), is the predicted fault label (model output), which can be the predicted probability.

[0115] In addition, it can be seen from the loss function curve of the CNN model on the training set and the test set in Figure 9 that the figure shows that as the number of training iterations increases, the training loss and the test loss both drop rapidly in the first few iterations, indicating that the model quickly learns and adjusts the weights at the beginning of training to reduce the value of the loss function. As the training progresses, the rate of decline of the loss function gradually slows down and eventually stabilizes, indicating that the model gradually approaches the optimal state and the loss no longer decreases significantly. Overall, the performance of the CNN neural network in this task is robust, and the steady decline of the loss function indicates that the model can effectively learn and reduce errors in the process of continuous optimization. This performance shows that the model has a good training effect and can maintain a low error when applied to test data.

[0116] The classification success rates for different types of electric energy meter faults can be shown in Table 2.

[0117]

[0118] Table 2 Summary of model classification success rate

[0119] Among them, there are 45 impedance grayscale images in the samples of each electric energy meter fault type, which comprehensively represent the internal structure of its circuit. Under different fault types, the CNN model has achieved a high level of classification success rate for faulty electric energy meters, especially in categories such as "LCD screen", "optical coupler (metering circuit)", and "main chip", the recognition success rate has reached 100%. Although the recognition success rate is slightly lower for some specific fault types (such as "capacitor (power module)" and "transformer"), overall, the model can accurately distinguish between normal electric energy meters and faulty electric energy meters. This shows that the CNN model can not only identify faulty electric energy meters, but also effectively ignore the specific component type of the fault and make accurate two-classification judgments.

[0120] In the method of this embodiment, a pre-trained fault identification model can be combined to obtain the fault detection result of the electric energy meter to be tested based on the impedance data of the electric energy meter to be tested, thereby avoiding opening the cover of the electric energy meter to be tested, improving the intelligence of the electric energy meter fault detection, and reducing the operating cost of the electric energy meter fault detection. Furthermore, the efficiency, simplicity and accuracy of the electric energy meter fault detection can be improved.

[0121] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0122] Based on the same inventive concept, the embodiment of the present application also provides a nondestructive detection device for a faulty electric energy meter based on a broadband impedance spectrum for realizing the nondestructive detection method for a faulty electric energy meter based on a broadband impedance spectrum. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in the embodiments of one or more nondestructive detection devices for faulty electric energy meters based on a broadband impedance spectrum provided below can refer to the limitations of the nondestructive detection method for faulty electric energy meters based on a broadband impedance spectrum in the above text, and will not be repeated here.

[0123] In one embodiment, as shown in FIG10 , a nondestructive detection device for a faulty electric energy meter based on a broadband impedance spectrum is provided, comprising: a data measurement module 1002 , a drawing module 1004 , a calculation module 1006 and a result acquisition module 1008 , wherein:

[0124] The data measurement module 1002 is used to obtain the target impedance data of the electric energy meter to be tested;

[0125] A drawing module 1004 is used to obtain an impedance grayscale map of the electric energy meter to be tested according to the target impedance data;

[0126] A calculation module 1006 is used to obtain the fault probability of the electric energy meter to be tested based on the impedance grayscale map by using a pre-trained fault identification model;

[0127] The result acquisition module 1008 is used to acquire the fault detection result of the electric energy meter to be tested according to the fault probability.

[0128] In one embodiment, the data measurement module 1002 is further used to: obtain initial impedance data of the electric energy meter to be measured; and perform data preprocessing on the initial impedance data to obtain the target impedance data.

[0129] In one embodiment, the data measurement module 1002 is further used to: perform data cleaning on the initial impedance data to obtain the cleaned initial impedance data; and perform normalization processing on the cleaned initial impedance data to obtain the target impedance data.

[0130] In one embodiment, the calculation module 1006 is also used to: obtain the grayscale value of each of the multiple data points of the target impedance data; based on the grayscale value, draw the multiple data points on a two-dimensional plane according to a preset arrangement rule to obtain the impedance grayscale map of the electric energy meter to be tested.

[0131] In one embodiment, the calculation module 1006 is further used to: input the impedance grayscale image into the pre-trained fault identification model to determine and output the fault probability of the electric energy meter to be tested.

[0132] In one embodiment, the calculation module 1006 is also used to: input the impedance grayscale map into the pre-trained fault identification model, and the feature extraction module of the fault identification model determines and outputs the impedance characteristics of the electric energy meter to be tested based on the impedance grayscale map; and the classification module of the fault identification model determines and outputs the fault probability of the electric energy meter to be tested based on the impedance characteristics output by the feature extraction module.

[0133] In one embodiment, the calculation module 1006 is further used to: obtain sample impedance data of a sample electric energy meter under each of the electric energy meter fault types among multiple electric energy meter fault types; obtain a sample impedance grayscale map corresponding to each of the electric energy meter fault types based on the sample impedance data; divide the sample impedance grayscale map into a training sample impedance grayscale map and a test sample impedance grayscale map; use the training sample impedance grayscale map to obtain an initial fault recognition model; optimize the initial fault recognition model based on the test sample impedance grayscale map to obtain the pre-trained fault recognition model.

[0134] Each module in the above-mentioned nondestructive detection device for faulty electric energy meters based on broadband impedance spectrum can be implemented in whole or in part by software, hardware and their combination. Each of the above-mentioned modules 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 corresponding operations of each of the above modules.

[0135] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as shown in FIG11. The computer device includes a processor, a memory, and a network interface connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data related to non-destructive detection of faulty electric energy meters based on broadband impedance spectra. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for non-destructive detection of faulty electric energy meters based on broadband impedance spectra is implemented.

[0136] Those skilled in the art will understand that the structure shown in FIG. 11 is merely a block diagram of a partial structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0137] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0138] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0139] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0140] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0141] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0142] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0143] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A nondestructive detection method for faulty electric energy meters based on broadband impedance spectroscopy, characterized in that: The method comprises: Obtain target impedance data of the electric energy meter to be tested; According to the target impedance data, an impedance grayscale image of the electric energy meter to be tested is obtained; Using a pre-trained fault identification model, based on the impedance grayscale image, the fault probability of the electric energy meter to be tested is obtained; According to the fault probability, a fault detection result of the electric energy meter to be tested is obtained.

2. The method according to claim 1, characterized in that: The step of obtaining target impedance data of the electric energy meter to be tested includes: Acquiring initial impedance data of the electric energy meter to be tested; The initial impedance data is preprocessed to obtain the target impedance data.

3. The method according to claim 2, characterized in that The performing data preprocessing on the initial impedance data to obtain the target impedance data includes: Performing data cleaning on the initial impedance data to obtain the cleaned initial impedance data; The initial impedance data after cleaning is normalized to obtain the target impedance data.

4. The method according to claim 1, characterized in that: The step of obtaining an impedance grayscale image of the electric energy meter to be tested according to the target impedance data includes: Acquire a grayscale value of each of a plurality of data points of the target impedance data; According to the grayscale values, the plurality of data points are plotted on a two-dimensional plane according to a preset arrangement rule to obtain the impedance grayscale map of the electric energy meter to be tested.

5. The method according to claim 1, characterized in that The method of using a pre-trained fault identification model to obtain the fault probability of the electric energy meter to be tested based on the impedance grayscale map includes: The impedance grayscale image is input into the pre-trained fault identification model to determine and output the fault probability of the electric energy meter to be tested.

6. The method according to claim 5, characterized in that The step of inputting the impedance grayscale image into the pre-trained fault identification model to determine and output the fault probability of the electric energy meter to be tested comprises: The impedance grayscale image is input into the pre-trained fault identification model, and the feature extraction module of the fault identification model determines and outputs the impedance feature of the electric energy meter to be tested according to the impedance grayscale image; The classification module of the fault identification model determines and outputs the fault probability of the electric energy meter to be tested according to the impedance feature output by the feature extraction module.

7. The method according to any one of claims 1 to 6, characterized in that Before obtaining the target impedance data of the electric energy meter to be tested, the method further includes: Acquire sample impedance data of a sample electric energy meter under each of the electric energy meter fault types among a plurality of electric energy meter fault types; According to the sample impedance data, a sample impedance grayscale image corresponding to each fault type of the electric energy meter is obtained; Dividing the sample impedance grayscale map into a training sample impedance grayscale map and a test sample impedance grayscale map; Using the training sample impedance grayscale image, an initial fault recognition model is obtained; The initial fault identification model is optimized according to the test sample impedance grayscale image to obtain the pre-trained fault identification model.

8. A nondestructive detection device for faulty electric energy meters based on broadband impedance spectroscopy, characterized in that: The device comprises: A data measurement module, used to obtain target impedance data of the electric energy meter to be tested; A drawing module, used for obtaining an impedance grayscale image of the electric energy meter to be tested according to the target impedance data; A calculation module, used for obtaining the fault probability of the electric energy meter to be tested based on the impedance grayscale map by using a pre-trained fault identification model; The result acquisition module is used to acquire the fault detection result of the electric energy meter to be tested according to the fault probability.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images