Electric quantity portrait-based nondestructive testing method and device for disassembled electric energy meter, computer equipment and storage medium

Abstract

The embodiment of the invention provides a non-destructive testing method and device for a disassembled electric energy meter based on an electric quantity portrait, 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 the collected electric quantity data of a to-be-tested disassembled electric energy meter; screening out target electric quantity data of the to-be-tested dismantled electric energy meter from the acquired electric quantity data according to the relevancy between the acquired electric quantity data and the to-be-tested dismantled electric energy meter; according to the target electric quantity data, obtaining an electric quantity portrait grey-scale map of the to-be-tested disassembled electric energy meter; and obtaining a fault detection result of the to-be-detected dismantled electric energy meter based on the electric quantity portrait grey-scale map by using a pre-trained fault detection model. According to the method, the problems of complex operation and high operation cost caused by uncovering detection are avoided, the intelligence of fault detection of the disassembled electric energy meter is improved, the operation cost of fault detection of the disassembled electric energy meter is reduced, and then the efficiency, simplicity and accuracy of fault detection of the disassembled electric energy meter 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 non-destructive detection of dismantled electric energy meters based on electric quantity portrait. Background Art

[0002] With the rapid advancement of the construction of new power systems, the energy meter, as the core equipment for power metering and information collection, plays a vital role in the entire power system. With the increase in the number of power users and the continuous popularization of smart grids, the installation volume of energy meters has also grown rapidly. However, since the energy meter needs to be exposed to a variety of complex working environments for a long time during operation, including high temperature, low temperature, high humidity and strong electromagnetic interference, these factors greatly affect the reliability and life of the energy meter. Therefore, as time goes by, a large number of long-used energy meters are dismantled and these retired energy meters need to be fully inspected and analyzed to ensure whether they can continue to be used or must be scrapped. This makes how to efficiently and accurately detect the faults of these dismantled energy meters an important and urgent research topic.

[0003] In existing research, fault detection of electric energy meters mainly focuses on appearance inspection or cover opening inspection. Appearance inspection can identify surface damage or defects, but cannot detect internal circuit faults. Cover opening inspection requires manual operation, which has high costs and safety risks. In addition, although penetrating detection methods such as X-ray imaging technology can identify damage to 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, and there is a problem of low fault detection efficiency. Summary of the invention

[0004] Based on this, it is necessary to provide a method, device, computer equipment, storage medium and computer program product for non-destructive testing of dismantled electricity meters based on electricity quantity profiling to address the above technical problems.

[0005] In a first aspect, the present application provides a method for non-destructive testing of a disassembled electric energy meter based on an electric quantity profile. The method comprises:

[0006] Obtain the collected electricity data of the electric energy meter to be tested;

[0007] According to the correlation between the collected power data and the removed power meter to be tested, the target power data of the removed power meter to be tested is screened out from the collected power data;

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

[0009] By using a pre-trained fault detection model, based on the grayscale image of the electrical quantity portrait, the fault detection result of the disassembled electric energy meter to be tested is obtained.

[0010] In one of the embodiments, the method of obtaining a grayscale image of the electric power portrait of the electric energy meter to be tested based on the target electric power data includes: normalizing the target electric power data to obtain normalized electric power data of the electric energy meter to be tested; and obtaining the grayscale image of the electric power portrait of the electric energy meter to be tested based on the normalized electric power data.

[0011] In one embodiment, the normalized electricity data includes multiple normalized electricity data points, and obtaining the electricity portrait grayscale image of the electric energy meter to be tested based on the normalized electricity data includes: obtaining the grayscale value corresponding to each of the multiple normalized electricity data points; plotting the grayscale values ​​on a two-dimensional plane according to a preset arrangement rule to generate the electricity portrait grayscale image of the electric energy meter to be tested.

[0012] In one of the embodiments, the method of using a pre-trained fault detection model to obtain the fault detection result of the returned electric energy meter to be tested based on the electric energy portrait grayscale image includes: inputting the electric energy portrait grayscale image into the pre-trained fault detection model; determining and outputting the electric energy characteristics of the returned electric energy meter to be tested according to the electric energy portrait grayscale image by a feature extraction module in the fault detection model; obtaining the fault label of the returned electric energy meter to be tested according to the electric energy characteristics output by the feature extraction module by a classification module in the fault detection model; and obtaining the fault detection result of the returned electric energy meter to be tested according to the fault label.

[0013] In one of the embodiments, before obtaining the collected electricity data of the disassembled electricity meter to be tested, it also includes: obtaining multiple groups of sample electricity data of multiple sample disassembled electricity meters; the multiple sample disassembled electricity meters include disassembled electricity meters of multiple types of electricity meters; according to the multiple groups of sample electricity data, obtaining a sample electricity portrait grayscale image of each of the sample disassembled electricity meters; dividing the sample electricity portrait grayscale image into a training sample electricity portrait grayscale image and a test sample electricity portrait grayscale image; using the training sample electricity portrait grayscale image, obtaining an initial fault detection model; optimizing the initial fault detection model according to the test sample electricity portrait grayscale image to obtain the pre-trained fault detection model.

[0014] In one of the embodiments, the initial fault detection model is obtained by using the training sample electricity portrait grayscale image, including: obtaining the actual fault label of the sample disassembled electric energy meter corresponding to the training sample electricity portrait grayscale image; inputting the training sample electricity portrait grayscale image into the pre-trained fault detection model; obtaining the sample electricity feature of the sample disassembled electric energy meter according to the training sample electricity portrait grayscale; obtaining the predicted fault label of the sample disassembled electric energy meter according to the sample electricity feature; and training the pre-trained fault detection model according to the difference between the actual fault label and the predicted fault label to obtain the fault detection model.

[0015] In a second aspect, the present application provides a non-destructive detection device for dismantled electric energy meters based on electric quantity portrait. The device comprises:

[0016] A data acquisition module is used to obtain the collected power data of the electric energy meter to be tested;

[0017] A data cleaning module, used for filtering out target electric quantity data of the electric quantity meter to be tested from the collected electric quantity data according to the correlation between the collected electric quantity data and the electric quantity meter to be tested;

[0018] A grayscale image generation module is used to obtain a grayscale image of the electric quantity portrait of the electric energy meter to be tested according to the target electric quantity data;

[0019] The calculation module is used to use a pre-trained fault detection model to obtain the fault detection result of the disassembled electric energy meter to be tested based on the grayscale image of the electric quantity portrait.

[0020] 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:

[0021] Obtain the collected electricity data of the electric energy meter to be tested;

[0022] According to the correlation between the collected power data and the removed power meter to be tested, the target power data of the removed power meter to be tested is screened out from the collected power data;

[0023] According to the target power data, a grayscale image of the power profile of the electric energy meter to be tested is obtained;

[0024] By using a pre-trained fault detection model, based on the grayscale image of the electrical quantity portrait, the fault detection result of the disassembled electric energy meter to be tested is obtained.

[0025] 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:

[0026] Obtain the collected electricity data of the electric energy meter to be tested;

[0027] According to the correlation between the collected power data and the removed power meter to be tested, the target power data of the removed power meter to be tested is screened out from the collected power data;

[0028] According to the target power data, a grayscale image of the power profile of the electric energy meter to be tested is obtained;

[0029] By using a pre-trained fault detection model, based on the grayscale image of the electrical quantity portrait, the fault detection result of the disassembled electric energy meter to be tested is obtained.

[0030] 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:

[0031] Obtain the collected electricity data of the electric energy meter to be tested;

[0032] According to the correlation between the collected power data and the removed power meter to be tested, the target power data of the removed power meter to be tested is screened out from the collected power data;

[0033] According to the target power data, a grayscale image of the power profile of the electric energy meter to be tested is obtained;

[0034] By using a pre-trained fault detection model, based on the grayscale image of the electrical quantity portrait, the fault detection result of the disassembled electric energy meter to be tested is obtained.

[0035] In the above-mentioned non-destructive testing method, device, computer equipment, storage medium and computer program product for returned electric energy meters based on electricity quantity portrait, first, the collected electricity quantity data of the returned electric energy meter to be tested can be obtained; next, according to the correlation between the collected electricity quantity data and the returned electric energy meter to be tested, the target electricity quantity data of the returned electric energy meter to be tested can be screened out from the collected electricity quantity data; further, according to the target electricity quantity data, the grayscale image of the electricity quantity portrait of the returned electric energy meter to be tested can be obtained; finally, the pre-trained fault detection model can be used to obtain the fault detection result of the returned electric energy meter to be tested based on the grayscale image of the electricity quantity portrait. In the method provided in the embodiment of the present application, a pre-trained fault detection model is used to obtain the fault detection result of the returned electric energy meter to be tested based on the grayscale image of the electric power portrait. The neural network model and the electric power data of the returned electric energy meter to be tested are combined to perform fault detection on the returned electric energy meter to be tested, thereby avoiding the problems of complicated operation and high operating cost caused by opening the cover for detection, improving the intelligence of fault detection of the returned electric energy meter, and reducing the operating cost of fault detection of the returned electric energy meter. Furthermore, the efficiency, simplicity and accuracy of fault detection of the returned electric energy meter can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments of the present application or the 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.

[0037] FIG1 is a flow chart of a method for non-destructive testing of a disassembled electric energy meter based on an electric quantity profile provided in an embodiment of the present application;

[0038] FIG2 is a schematic diagram of a convolutional neural network model provided in an embodiment of the present application;

[0039] FIG3 is a schematic diagram of a process for obtaining a grayscale image of an electric power portrait provided in an embodiment of the present application;

[0040] FIG4 is a schematic diagram of a process for obtaining a grayscale image of an electric quantity portrait according to normalized electric quantity data provided by an embodiment of the present application;

[0041] FIG5 is a schematic diagram of a process for obtaining a fault detection result provided in an embodiment of the present application;

[0042] FIG6 is a schematic diagram of a process for obtaining a fault detection model provided in an embodiment of the present application;

[0043] FIG7 is a schematic diagram of a grayscale image provided in an embodiment of the present application;

[0044] FIG8 is a schematic diagram of a process for obtaining an initial fault detection model provided by an embodiment of the present application;

[0045] FIG9 is a structural block diagram of a non-destructive testing device for dismantled electric energy meters based on electric quantity portrait provided in an embodiment of the present application;

[0046] FIG. 10 is an internal structure diagram of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0047] 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.

[0048] In an exemplary embodiment, as shown in FIG1 , a non-destructive detection method for dismantled electric energy meters based on power profile is provided. This embodiment uses the method applied to a server as an example. 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:

[0049] Step 102, obtaining the collected electricity data of the removed electric energy meter to be tested.

[0050] Among them, the dismantled electric energy meter to be tested can be any retired electric energy meter. The types of electric energy meters that are dismantled can include multiple categories of electric energy meters, and the multiple categories of electric energy meters can include but are not limited to single-phase meters, three-phase four-wire fee control meters, three-phase three-wire multi-function meters, and three-phase four-wire multi-function meters, etc. The type of electric energy meter that is dismantled and tested can be any one of single-phase meters, three-phase four-wire fee control meters, three-phase three-wire multi-function meters, and three-phase four-wire multi-function meters. In this step, input signals of different frequencies and different waveforms can be injected into the dismantled electric energy meter to be tested, and the corresponding response power can be read at the communication port to obtain the collected power data of the dismantled electric energy meter to be tested. Among them, the input signal is in the form of a sine wave fundamental wave superimposed with harmonics of different frequencies and amplitudes. By adjusting the total harmonic distortion (THD), input signals of different waveforms can be obtained. Then, various types of transient power information stored in the microcontroller unit (MCU) can be read at the 485 communication port of the electric energy meter to be tested. Furthermore, the collected power data of the electric energy meter to be tested can be obtained based on the various types of transient power information. For example, for a single-phase meter, the transient power information includes one-phase data, such as phase A voltage, phase A current, instantaneous phase A active power and phase A power factor.

[0051] Step 104, based on the correlation between the collected power data and the removed power meter to be tested, the target power data of the removed power meter to be tested is screened out from the collected power data.

[0052] Wherein, in the case of obtaining the collected power data of the electric energy meter to be tested, the collected power data can be further preprocessed to obtain the target power data of the electric energy meter to be tested. The data preprocessing may include data cleaning, and the target power data may be the power data after data cleaning. In the process of data cleaning the collected power data, invalid or noise data in the collected power data may be removed to ensure data quality. For example, during the collection process, it is found that the waveform distortion of each phase voltage and the 1st to 21st harmonic content of each phase voltage in the transient response are not read in most of the removed electric energy meters. Therefore, the two parts of data may not be included in the target power data, and the target power data may include 30 transient responses other than the waveform distortion of each phase voltage and the 1st to 21st harmonic content of each phase voltage. In this step, the correlation between the collected power data and the fault detection of the electric energy meter to be tested can be obtained, and the collected power data with the correlation greater than the preset correlation threshold is determined as the target power data of the electric energy meter to be tested. The target power data may also be called the power profile of the electric energy meter to be tested.

[0053] Step 106, obtaining a grayscale image of the power consumption profile of the dismantled electric energy meter to be tested according to the target power consumption data.

[0054] In this step, since the dimension and range of the target power data may vary greatly, it is necessary to normalize it to the interval of 0 to 1, that is, to normalize the target power data. In a possible implementation, the minimum-maximum normalization (Min-Max Scaling) method can be used to normalize each of the multiple data points of the target power data to obtain the normalized power data of the disassembled electric energy meter to be tested, which can be seen in formula (1):

[0055]

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

[0057] Next, the normalized power data can be converted into a grayscale image. A grayscale image 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 normalized data points of the normalized power data can be mapped to a grayscale value between 0 and 255; in the second step, the grayscale value of each of the multiple normalized data points is plotted on a two-dimensional plane according to a certain arrangement rule to generate a grayscale image of the electric energy meter. Each pixel of the image corresponds to a power measurement value, black (grayscale value 0) represents low power, and white (grayscale value 255) represents high power, as shown in formula (2):

[0058]

[0059] in, is the normalized data value, The grayscale value converted to the normalized data value.

[0060] Step 108, using the pre-trained fault detection model, based on the grayscale image of the power profile, obtain the fault detection result of the disassembled electric energy meter to be tested.

[0061] As shown in FIG2 , the pre-trained fault detection model may be a convolutional neural network (CNN). The pre-trained fault detection 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 detection model may include at least one convolutional layer and at least one pooling layer; the classification module of the fault detection model may include at least one fully connected layer. The input layer may receive an input grayscale image of the power portrait, and the size of the input layer is set to 300×400×1 considering the size of the model; the convolution layer may 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; each convolution layer is connected to a pooling layer, which can downsample the feature map output by the convolution layer, reduce the dimension of the data, reduce the complexity of the model, and maintain 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 detection 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, a 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.

[0062] 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 activation functions. The power portrait grayscale image may include multiple power data.

[0063] Specifically, the grayscale image of the power portrait of the disassembled electric energy meter to be tested can be input into the pre-trained fault detection model, and the convolution layer can be used to extract the initial features of multiple power data. The first step is to convert multiple power data into vectors to obtain multiple power data vectors, and the multiple power data vectors can be combined into an electric data vector matrix; the second step is to input the electric data vector matrix into the convolution layer, and use the convolution kernel to perform convolution operation with the electric data vector matrix, that is, to perform inner product operation on the electric data vector matrix and the convolution kernel to obtain the convolution result corresponding to the electric 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 power characteristics of the disassembled electric energy meter to be tested.

[0064] The fully connected layer can be classified according to the electric quantity feature to obtain the fault label of the electric energy meter to be tested. In addition, the fully connected layer may include an activation function, and the activation function includes a weight matrix and a bias constant.

[0065] Specifically, the electric quantity feature can be input into the fully connected layer, and the electric quantity 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, and then, the fault detection result of the electric energy meter to be tested can be obtained according to the fault label.

[0066] In the method of this embodiment, first, the collected electricity data of the disassembled electric energy meter to be tested can be obtained; next, according to the correlation between the collected electricity data and the disassembled electric energy meter to be tested, the target electricity data of the disassembled electric energy meter to be tested can be screened out from the collected electricity data; then, according to the target electricity data, the electricity portrait grayscale image of the disassembled electric energy meter to be tested can be obtained; finally, the pre-trained fault detection model can be used to obtain the fault detection result of the disassembled electric energy meter to be tested based on the electricity portrait grayscale image. In the method provided by the embodiment of the present application, a pre-trained fault detection model is used to obtain the fault detection result of the disassembled electric energy meter to be tested based on the electricity portrait grayscale image, and the neural network model and the electricity data of the disassembled electric energy meter to be tested are combined to perform fault detection on the disassembled electric energy meter to be tested, thereby avoiding the problems of complex operation and high operating cost caused by opening the cover for detection, improving the intelligence of fault detection of disassembled electric energy meters, and reducing the operating cost of fault detection of disassembled electric energy meters, thereby improving the efficiency, simplicity and accuracy of fault detection of disassembled electric energy meters.

[0067] In an exemplary embodiment, as shown in FIG3 , step 106 may include steps 302 to 304. Among them:

[0068] Step 302, normalize the target power data to obtain normalized power data of the disassembled electric energy meter to be tested.

[0069] In this step, since the dimension and range of the target power data may vary greatly, it is necessary to normalize it to the interval of 0 to 1, that is, to normalize the target power data. In a possible implementation, the minimum-maximum normalization (Min-Max Scaling) method can be used to normalize each of the multiple data points of the target power data to obtain the normalized power data of the disassembled power meter to be tested, which can be seen in formula (1).

[0070] Step 304, obtaining a grayscale image of the electricity consumption profile of the dismantled electricity meter to be tested based on the normalized electricity consumption data.

[0071] In an exemplary embodiment, as shown in FIG4 , step 304 may include steps 402 to 404. Among them:

[0072] Step 402: Obtain the grayscale value corresponding to each normalized power data point among a plurality of normalized power data points.

[0073] Step 404, plotting the grayscale values ​​on a two-dimensional plane according to a preset arrangement rule, and generating a grayscale image of the electricity consumption of the dismantled electric energy meter to be tested.

[0074] The normalized power data can be converted into a grayscale image. A grayscale image 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 normalized data points of the normalized power data can be mapped to a grayscale value between 0 and 255; in the second step, the grayscale value of each of the multiple normalized data points is plotted on a two-dimensional plane according to a certain arrangement rule to generate a grayscale image of the electric energy meter. Each pixel of the image corresponds to a power measurement value, black (grayscale value 0) represents low power, and white (grayscale value 255) represents high power, as shown in formula (2).

[0075] In the method of this embodiment, the target electric quantity data can be normalized to obtain the normalized electric quantity measurement value of the electric energy meter to be tested, thereby improving the accuracy of data processing, and further, the accuracy of fault detection can be improved. In addition, the normalized electric quantity data of the electric energy meter to be tested can be converted into a grayscale image, and then the grayscale image can be convolved, thereby avoiding a complex feature extraction process, improving the calculation efficiency, and further, improving the detection efficiency of fault detection.

[0076] In an exemplary embodiment, as shown in FIG5 , step 108 may include steps 502 to 508. Among them:

[0077] Step 502: input the grayscale image of the power consumption profile into a pre-trained fault detection model.

[0078] Step 504, the feature extraction module in the fault detection model determines and outputs the power characteristics of the disassembled electric energy meter to be tested based on the grayscale image of the power portrait.

[0079] Step 506: The classification module in the fault detection model obtains the fault label of the disassembled electric energy meter to be tested according to the electric quantity feature output by the feature extraction module.

[0080] Step 508, obtaining the fault detection result of the disassembled electric energy meter to be tested according to the fault tag.

[0081] The pre-trained fault detection model may be a convolutional neural network (CNN). The pre-trained fault detection 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 detection model may include at least one convolutional layer and at least one pooling layer; the classification module of the fault detection model may include at least one fully connected layer.

[0082] 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 activation functions. The power portrait grayscale image may include multiple power data.

[0083] Specifically, the grayscale image of the power portrait of the disassembled electric energy meter to be tested can be input into the pre-trained fault detection model, and the convolution layer can be used to extract the initial features of multiple power data. The first step is to convert multiple power data into vectors to obtain multiple power data vectors, and the multiple power data vectors can be combined into an electric data vector matrix; the second step is to input the electric data vector matrix into the convolution layer, and use the convolution kernel to perform convolution operation with the electric data vector matrix, that is, to perform inner product operation on the electric data vector matrix and the convolution kernel to obtain the convolution result corresponding to the electric 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 power characteristics of the disassembled electric energy meter to be tested.

[0084] The fully connected layer can be classified according to the electric quantity feature to obtain the fault label of the electric energy meter to be tested. In addition, the fully connected layer may include an activation function, and the activation function includes a weight matrix and a bias constant.

[0085] Specifically, the electric quantity feature can be input into the fully connected layer, and the electric quantity 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 characterize the fault label of the electric energy meter to be tested, and the fault label can indicate whether the electric energy meter to be tested has a fault, and the fault detection result of the electric energy meter to be tested can be obtained according to the fault label.

[0086] In the method of this embodiment, the neural network model and the electricity data of the returned electric energy meter to be tested can be combined to perform fault detection on the returned electric energy meter to be tested, thereby avoiding the problems of complicated operation and high operating cost caused by opening the cover for detection, improving the intelligence of fault detection of the returned electric energy meter, and reducing the operating cost of fault detection of the returned electric energy meter. Furthermore, the efficiency, simplicity and accuracy of fault detection of the returned electric energy meter can be improved.

[0087] 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 detection model, specifically including steps 602 to 610. Among them:

[0088] Step 602, obtaining multiple groups of sample electricity data from multiple sampled electricity meters.

[0089] Among them, the multiple sample disassembled electric energy meters include disassembled electric energy meters of multiple electric energy meter types. The multiple electric energy meter types may include but are not limited to single-phase meters, three-phase four-wire fee control meters, three-phase three-wire multifunctional meters, and three-phase four-wire multifunctional meters, etc. Each meter type has normal meter samples and faulty meter samples, and comes from different manufacturers. The measurement data has a total of 978 samples, as shown in the sample power data summary table in Table 1. In this step, first, multiple groups of initial sample power data of multiple sample disassembled electric energy meters can be obtained, so as to inject input signals of different frequencies and different waveforms into the sample disassembled electric energy meters, and read the corresponding response power at the communication port, so as to obtain the initial sample power data of the sample disassembled electric energy meter. Among them, the input signal adopts the form of a sine wave fundamental wave superimposed with harmonics of different frequencies and different amplitudes. The input signals of different waveforms can be obtained by adjusting the total harmonic distortion (THD), and then, various types of transient power information stored in the microcontroller unit (MCU) can be read at the 485 communication port of the sample disassembled electric energy meter. Furthermore, the initial sample power data of the sampled electric energy meter can be obtained based on the various types of transient power information. For example, for a single-phase meter, the transient power information includes one-phase data, such as phase A voltage, phase A current, instantaneous phase A active power and phase A power factor. Next, the initial sample electricity data of the sample returned electricity meter can be preprocessed to obtain multiple groups of sample electricity data of the multiple sample returned electricity meters. The data preprocessing may include data cleaning and normalization. Specifically, in the first step, the initial sample electricity data of the sample returned electricity meter can be cleaned. In the process of data cleaning the initial sample electricity data, invalid or noise data in the initial sample electricity data can be removed to ensure data quality. For example, during the acquisition process, it is found that the distortion of each phase voltage waveform and the 1st to 21st harmonic content of each phase voltage in the transient response are not read in most of the returned electricity meters. Therefore, these two parts of data may not be included in the sample electricity data. The sample electricity data may include 30 transient responses except the distortion of each phase voltage waveform and the 1st to 21st harmonic content of each phase voltage. In the second step, the initial sample power data after cleaning can be normalized. In a possible implementation, the minimum-maximum normalization (Min-Max Scaling) method can be used to normalize each data point of the multiple data points of the initial sample power data after cleaning, and normalize them to the interval of 0 to 1 to obtain the sample power data of the sample removed from the electric energy meter, which can be seen in formula (1).

[0090]

[0091] Table 1 Summary of sample power data

[0092] Step 604, based on the multiple groups of sample electricity data, obtain a sample electricity portrait grayscale image of each sample removed from the electricity meter.

[0093] Each group of sample power data can be converted into a grayscale image. A grayscale image 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 normalized data point in the multiple normalized data points of the sample power data can be mapped to a grayscale value between 0 and 255; in the second step, the grayscale value of each normalized data point in the multiple normalized data points is plotted on a two-dimensional plane according to a certain arrangement rule to generate a grayscale image of the electric energy meter. Each pixel of the image corresponds to a power measurement value, black (grayscale value 0) represents low power, and white (grayscale value 255) represents high power, as shown in formula (2).

[0094] Step 606, dividing the sample power portrait grayscale image into a training sample power portrait grayscale image and a test sample power portrait grayscale image.

[0095] Among them, multiple samples can be disassembled back into multiple groups of sample power portrait grayscale images of the electric energy meter and divided into training sample power portrait grayscale images and test sample power portrait grayscale images. The training sample power portrait grayscale images can be used to train the fault detection model to be trained to obtain a trained fault detection model, that is, an initial fault detection model. Then, the test sample power portrait grayscale images can be used to optimize the initial fault detection model to obtain an optimized fault detection model, that is, the pre-trained fault recognition model.

[0096] As shown in Figure 7, Figure 7 provides grayscale images of two normal single-phase meters, grayscale images of two normal three-phase four-wire fee control meters, and two single-phase smart meters with different faults, where the faulty components of a) and b) are MCU and diode respectively, and the two same-category electric energy meters in the same control group are from different manufacturers. As can be seen from Figure 7, in the main feature columns, the grayscale image of the power portrait shows the characteristics of being similar in the same category (even if from different manufacturers) and dissimilar in different categories, which proves that the grayscale image of the power portrait can characterize the internal structure of the electric energy meter, distinguish normal meters of different categories, and then distinguish normal meters from faulty meters of the same category. At the same time, it can be seen that in some non-main feature columns, electric energy meters of the same category also show slight differences, which may be related to the fact that the internal circuits designed by different manufacturers are not completely the same.

[0097] Step 608, using the grayscale image of the power profile of the training sample to obtain an initial fault detection model.

[0098] The training sample power portrait grayscale image can be used to adjust the weight of the convolution kernel of the pre-trained fault detection model, so that the model can accurately identify the fault state of the electric energy meter. Specifically, the training sample power portrait grayscale image and the actual fault label of the training sample power portrait grayscale image can be input into the pre-trained fault detection model, and the feature extraction module in the pre-trained fault detection model can obtain the sample power feature corresponding to the training sample power portrait grayscale image according to the training sample power portrait grayscale image, and the classification module in the pre-trained fault detection model can obtain the predicted fault label of the training sample power portrait grayscale image according to the sample power feature output by the feature extraction module, and then, the pre-trained fault detection model can be trained according to the difference between the actual fault label and the predicted fault label of the training sample power portrait grayscale image to obtain a trained fault detection model. The feature extraction module of the fault detection model can include at least one convolution layer and at least one pooling layer; the classification module of the fault detection model can include at least one fully connected layer.

[0099] Step 610, optimizing the initial fault recognition model according to the grayscale image of the test sample power portrait to obtain a pre-trained fault detection model.

[0100] Among them, in the testing stage, the grayscale image of the test sample power portrait can be input into the trained fault detection model, and the trained fault detection model determines and outputs the predicted fault label of the test sample power portrait grayscale image. Then, the classification performance of the trained fault detection model can be judged according to the difference between the actual fault label and the predicted fault label of the test sample power portrait grayscale image. For example, the classification accuracy and error performance index can be calculated, and the classification performance of the trained fault detection model can be evaluated according to the classification accuracy and error performance index to obtain the evaluation result of the trained fault detection model. Then, according to the evaluation result, the trained fault detection model can be optimized to obtain the optimized fault detection model, that is, the final version of the pre-trained fault detection model mentioned above. In a possible implementation method, the accuracy of the trained fault detection model, which can be called the initial fault detection model, can be obtained, and the prediction accuracy of the initial fault detection model can be evaluated according to the accuracy.

[0101] The classification success rates for different types of electric energy meters can be shown in Table 2.

[0102]

[0103] Table 2 Summary of model classification success rate

[0104] Among them, when the disassembled electric energy meters of the four types of electric energy meters in Table 2 are predicted separately (the four categories all contain different manufacturers), 85% of the sample electric energy portrait grayscale images of the sample electric energy data are used as the training set, that is, the training sample electric energy portrait grayscale image, 5% are used as the test set, that is, the test sample electric energy portrait grayscale image, and 10% are used as the verification set. When the number of training times reaches about 40 times, the probability of successful classification is about 90%.

[0105] In an exemplary embodiment, as shown in FIG8 , step 608 may include steps 802 to 810. Among them:

[0106] Step 802, obtain the actual fault label of the electric energy meter corresponding to the sample of the training sample electricity portrait grayscale image.

[0107] Step 804: input the grayscale image of the training sample power profile into the pre-trained fault detection model.

[0108] Step 806, according to the grayscale of the training sample electricity portrait, obtain the sample electricity characteristics of the sample disassembled back into the electricity meter.

[0109] Step 808: Obtain the predicted fault label of the sample removed electric energy meter according to the sample power characteristics.

[0110] Step 810 , training the pre-trained fault detection model according to the difference between the actual fault label and the predicted fault label to obtain a fault detection model.

[0111] Among them, the difference between the actual fault label and the predicted fault label of the training sample power portrait grayscale image, as well as the difference between the actual fault label and the predicted fault label of the test sample power portrait 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 one 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.

[0112] In the method of this embodiment, the sample power portrait grayscale image can be divided into a training sample power portrait grayscale image and a test sample power portrait grayscale image, and the initial fault recognition model can be further optimized according to the test sample power portrait grayscale image to obtain a pre-trained fault detection model, thereby improving the accuracy of model training and, further, improving the accuracy of fault detection.

[0113] 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.

[0114] Based on the same inventive concept, the embodiment of the present application also provides a non-destructive detection device for dismantling electric energy meters based on electric quantity portraits for realizing the non-destructive detection method for dismantling electric energy meters based on electric quantity portraits involved above. 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 one or more embodiments of the non-destructive detection device for dismantling electric energy meters based on electric quantity portraits provided below can refer to the limitations of the non-destructive detection method for dismantling electric energy meters based on electric quantity portraits above, and will not be repeated here.

[0115] In one embodiment, as shown in FIG9 , a non-destructive detection device for dismantled electric energy meters based on electric quantity portrait is provided, including: a data acquisition module 902, a data cleaning module 904, a grayscale image generation module 906 and a calculation module 908, wherein:

[0116] The data acquisition module 902 is used to acquire the collected power data of the electric energy meter to be tested;

[0117] A data cleaning module 904 is used to filter out target power data of the removed power meter to be tested from the collected power data according to the correlation between the collected power data and the removed power meter to be tested;

[0118] A grayscale image generating module 906 is used to obtain a grayscale image of the electric quantity portrait of the electric energy meter to be tested according to the target electric quantity data;

[0119] The calculation module 908 is used to use a pre-trained fault detection model to obtain the fault detection result of the disassembled electric energy meter to be tested based on the grayscale image of the electric quantity portrait.

[0120] In one embodiment, the grayscale image generation module 906 is also used to: normalize the target power data to obtain normalized power data of the electric energy meter to be tested; and obtain the power portrait grayscale image of the electric energy meter to be tested based on the normalized power data.

[0121] In one embodiment, the grayscale image generation module 906 is also used to: obtain the grayscale value corresponding to each of the multiple normalized electric quantity data points; plot the grayscale values ​​on a two-dimensional plane according to a preset arrangement rule to generate the electric quantity portrait grayscale image of the electric energy meter to be tested.

[0122] In one embodiment, the calculation module 908 is also used to: input the electricity portrait grayscale image into the pre-trained fault detection model; the feature extraction module in the fault detection model determines and outputs the electricity characteristics of the electric energy meter to be tested based on the electricity portrait grayscale image; the classification module in the fault detection model obtains the fault label of the electric energy meter to be tested based on the electricity characteristics output by the feature extraction module; and obtains the fault detection result of the electric energy meter to be tested based on the fault label.

[0123] In one embodiment, the calculation module 908 is also used to: obtain multiple groups of sample electricity data from multiple sample returned electricity meters; the multiple sample returned electricity meters include returned electricity meters of multiple types of electricity meters; obtain a sample electricity portrait grayscale image of each sample returned electricity meter based on the multiple groups of sample electricity data; divide the sample electricity portrait grayscale image into a training sample electricity portrait grayscale image and a test sample electricity portrait grayscale image; use the training sample electricity portrait grayscale image to obtain an initial fault detection model; optimize the initial fault identification model based on the test sample electricity portrait grayscale image to obtain the pre-trained fault detection model.

[0124] In one embodiment, the calculation module 908 is also used to: obtain the actual fault label of the sample returned electric energy meter corresponding to the training sample electric energy portrait grayscale image; input the training sample electric energy portrait grayscale image into a pre-trained fault detection model; obtain the sample electric energy feature of the sample returned electric energy meter according to the training sample electric energy portrait grayscale; obtain the predicted fault label of the sample returned electric energy meter according to the sample electric energy feature; train the pre-trained fault detection model according to the difference between the actual fault label and the predicted fault label to obtain the fault detection model.

[0125] Each module in the above-mentioned non-destructive testing device for dismantling electric energy meters based on electric quantity portrait 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 operations corresponding to each of the above modules.

[0126] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as shown in FIG10. 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 testing of disassembled electric energy meters based on power profiles. 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 testing of disassembled electric energy meters based on power profiles is implemented.

[0127] Those skilled in the art will understand that the structure shown in FIG. 10 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.

[0128] 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.

[0129] 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.

[0130] 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.

[0131] 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.

[0132] 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.

[0133] The technical features of the above embodiments may be arbitrarily combined. 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.

[0134] 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 non-destructive testing method for dismantled electric energy meters based on electric quantity portrait, characterized in that: The method comprises: Obtain the collected electricity data of the electric energy meter to be tested; According to the correlation between the collected power data and the removed power meter to be tested, the target power data of the removed power meter to be tested is screened out from the collected power data; According to the target power data, a grayscale image of the power profile of the electric energy meter to be tested is obtained; By using a pre-trained fault detection model, based on the grayscale image of the electrical quantity portrait, the fault detection result of the disassembled electric energy meter to be tested is obtained.

2. The method according to claim 1, characterized in that The step of obtaining a grayscale image of the electric quantity of the electric energy meter to be tested according to the target electric quantity data includes: Normalizing the target electric quantity data to obtain normalized electric quantity data of the electric energy meter to be tested; According to the normalized electricity data, the electricity portrait grayscale image of the disassembled electricity meter to be tested is obtained.

3. The method according to claim 2, characterized in that The normalized electric quantity data includes a plurality of normalized electric quantity data points, and the step of obtaining the electric quantity portrait grayscale image of the electric energy meter to be tested according to the normalized electric quantity data includes: Obtaining a grayscale value corresponding to each of the normalized power data points; The grayscale values ​​are plotted on a two-dimensional plane according to a preset arrangement rule to generate a grayscale image of the electrical quantity of the disassembled electric energy meter to be tested.

4. The method according to claim 1, characterized in that: The method of using the pre-trained fault detection model to obtain the fault detection result of the disassembled electric energy meter to be tested based on the electric quantity portrait grayscale image includes: Inputting the electrical quantity portrait grayscale image into the pre-trained fault detection model; The feature extraction module in the fault detection model determines and outputs the electric quantity feature of the disassembled electric energy meter to be tested according to the electric quantity portrait grayscale image; The classification module in the fault detection model obtains the fault label of the disassembled electric energy meter to be tested according to the electric quantity feature output by the feature extraction module; According to the fault tag, a fault detection result of the removed electric energy meter to be tested is obtained.

5. The method according to any one of claims 1 to 4, characterized in that Before obtaining the collected electricity data of the electric energy meter to be tested, the method further includes: Acquire multiple groups of sample electricity data of multiple sample returned electric energy meters; the multiple sample returned electric energy meters include returned electric energy meters of multiple types of electric energy meters; According to the plurality of groups of sample electricity data, a grayscale image of a sample electricity profile of each sample disassembled electric energy meter is obtained; Dividing the sample power portrait grayscale image into a training sample power portrait grayscale image and a test sample power portrait grayscale image; Using the training sample power portrait grayscale image, an initial fault detection model is obtained; The initial fault detection model is optimized according to the test sample electrical quantity portrait grayscale image to obtain the pre-trained fault detection model.

6. The method according to claim 5, characterized in that The method of using the training sample power profile grayscale image to obtain an initial fault detection model includes: Obtaining the actual fault label of the electric energy meter removed from the sample corresponding to the grayscale image of the training sample electric quantity portrait; Inputting the training sample power image grayscale image into a pre-trained fault detection model; According to the grayscale of the training sample power image, the sample power characteristics of the sample disassembled back into the electric energy meter are obtained; According to the sample power characteristics, obtaining a predicted fault tag of the sample removed electric energy meter; The pre-trained fault detection model is trained according to the difference between the actual fault label and the predicted fault label to obtain the fault detection model.

7. A non-destructive testing device for dismantled electric energy meters based on electric quantity portrait, characterized in that: The device comprises: A data acquisition module is used to obtain the collected power data of the electric energy meter to be tested; A data cleaning module, used for filtering out target electric quantity data of the electric quantity meter to be tested from the collected electric quantity data according to the correlation between the collected electric quantity data and the electric quantity meter to be tested; A grayscale image generation module is used to obtain a grayscale image of the electric quantity portrait of the electric energy meter to be tested according to the target electric quantity data; The calculation module is used to use a pre-trained fault detection model to obtain the fault detection result of the disassembled electric energy meter to be tested based on the grayscale image of the electric quantity portrait.

8. 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 6 are implemented.

9. 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 6 are implemented.

10. A computer program product, comprising a computer program, 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 6 are implemented.

Images