Intelligent electric energy meter error prediction method, system and device and storage medium

Abstract

The invention relates to an intelligent electric energy meter error prediction method, system and device, and a storage medium. The method comprises the following steps: collecting and preprocessing the historical error and environment variable data of an intelligent electric energy meter; establishing an environmental influence error prediction model to output error prediction considering environmental factors in combination with data; meanwhile, a time influence error prediction model is constructed, time features are mined only through error history, and error prediction based on time factors is generated; and finally, performing multi-source fusion on the two types of prediction results to obtain a final error prediction value of the intelligent electric energy meter which comprehensively considers dual factors of environment and time, thereby realizing more accurate metering performance evaluation and early warning.

Description

Technical Field

[0001] The present invention relates to a method, system, device and storage medium for predicting errors in an intelligent electric energy meter, and belongs to the technical field of electric energy meter error prediction. Background Art

[0002] As a core metering device, smart electricity meters play a key role in electricity billing, energy consumption analysis and power dispatching. However, the measurement errors of smart electricity meters are affected by a variety of complex factors, especially showing significant long-term trends and seasonal cycle characteristics. Studies have shown that the measurement errors of smart electricity meters show obvious periodic fluctuation characteristics with seasonal changes. This phenomenon is usually caused by the periodic changes in ambient temperature and humidity caused by seasonal climate changes and the seasonal differences in grid load.

[0003] Experimental results show that in high-humidity and high-heat environments, smart energy meters exhibit significant error drift, primarily due to moisture in circuit boards and aging of electrical components caused by high humidity. In high-salt fog environments, long-term corrosion can affect the measurement accuracy of energy meter sensors, leading to increased error. Furthermore, experimental data indicates that smart energy meter errors exhibit nonlinear and periodic fluctuations, indicating that the errors exhibit a certain trend with environmental variables and are also affected by short-term grid load fluctuations.

[0004] The invention patent application with publication number "CN115542236A" discloses a method and device for estimating the operating error of an electric energy meter, the method comprising: establishing an electric energy meter error model; penalizing and solving the error coefficient of each electric energy meter in the electric energy meter error model based on a regularization coefficient to obtain a first error coefficient of each electric energy meter in the electric energy meter error model; determining candidate out-of-tolerance meters in the electric energy meter error model based on the regularization coefficient; penalizing and solving the error coefficient of other electric energy meters in the electric energy meter error model except the candidate out-of-tolerance meters in the electric energy meter error model based on a regularization coefficient to obtain a second error coefficient of each electric energy meter in the electric energy meter error model; determining the operating error of the electric energy meter in the electric energy meter error model based on the second error coefficient of each electric energy meter in the electric energy meter error model; the invention can screen out potential out-of-tolerance meters in the electric energy meter error model and reasonably assign regularization coefficients to them, thereby improving the calculation quality of the error coefficients of the electric energy meters in the electric energy meter error model and improving the accuracy of error estimation; However, the model in this solution is based solely on user electricity consumption and line loss data. It doesn't consider environmental variables like temperature and humidity, nor does it model seasonal or trend variations in errors over time. For meters significantly affected by the environment, or where operational errors exhibit significant periodicity or aging effects, the accuracy of the estimated results will be significantly reduced. Summary of the Invention

[0005] In order to solve the above problems existing in the prior art, the present invention proposes a smart electricity meter error prediction method, system, device and storage medium.

[0006] The technical solutions of the present invention are as follows: In one aspect, the present invention provides a method for predicting an error of a smart energy meter, comprising the following steps: Collect historical error data and historical environmental variable data of smart energy meters and perform preprocessing; An environmental impact error prediction model is constructed, and the pre-processed historical error data and historical environmental variable data of the smart energy meter are input into the environmental impact error prediction model to obtain the error prediction result of the smart energy meter considering the environmental impact; Construct a time-influenced error prediction model, input the preprocessed historical error data of the smart energy meter into the time-influenced error prediction model to analyze the time characteristics of the historical error data of the smart energy meter, and obtain the error prediction result of the smart energy meter considering the influence of the time characteristics based on the time characteristics; The error prediction results of the smart electricity meter considering environmental impact and time characteristics are integrated to obtain the final error prediction result of the smart electricity meter.

[0007] Preferably, the preprocessing steps of the historical error data and the historical environmental variable data of the smart electric energy meter are: The Z-score algorithm is used to identify outliers in the historical error data and environmental variable data of smart energy meters. The outliers are then filled in with an interpolation algorithm to construct a comprehensive error environment dataset. The error environment comprehensive data set includes multiple data sample groups, and each data sample group is composed of historical error data samples and historical environmental variable data samples of a smart electric energy meter.

[0008] Preferably, the structure of the error environment comprehensive data set is optimized by a KD-Tree algorithm.

[0009] Preferably, the environmental impact error prediction model is constructed based on a generalized regression neural network, and the error prediction result of the smart energy meter considering the environmental impact is specifically shown in the following formula: ; ; in: It represents the error prediction result of the smart energy meter considering the environmental impact; represents the number of data sample groups in the error environment synthetic dataset; Indicates the Environmental index weight of each data sample group; Indicates the Historical error data samples in a data sample group; Represents the current environment variable data; Indicates the Historical environmental variable data samples in a data sample group; Represents the smoothing factor of the generalized regression neural network.

[0010] Preferably, the time-influenced error prediction model is constructed based on an Autoformer model, and the preprocessed historical error data of the smart energy meter is constructed into a historical error time series. The Autoformer model uses an autocorrelation attention mechanism to extract the periodic features of the historical error time series, as shown in the following formula: ; in: express Periodic characteristics of the time series of historical errors; represents conjugate calculation; represents the self-attention function; Represents the historical error time series in The query vector at the moment; Represents the historical error time series in The key vector of the moment; represents Fourier transform; represents the inverse Fourier transform; The trend characteristics in the historical error time series are extracted through a time window of preset length, as shown in the following formula: ; in: express Trend characteristics of the time series of historical errors; Indicates the length of the time window; express Sample of the moment historical error time series; The error prediction results of smart electric energy meters considering the influence of time characteristics are obtained by integrating the periodic characteristics and trend characteristics of the historical error time series, as shown in the following formula: ; in: It represents the error prediction result of the smart energy meter considering the influence of time characteristics; represents random noise.

[0011] Preferably, the time impact error prediction model uses AdamW optimizer to optimize model parameters.

[0012] Preferably, the error prediction result of the smart energy meter considering the environmental impact and the error prediction result considering the time characteristic impact are fused, as shown in the following formula: ; ; in: Represents the final error prediction result of the smart energy meter; represents the fusion coefficient; represents the prediction error of the time-influenced error prediction model; Represents the prediction error of the environmental impact error prediction model.

[0013] On the other hand, the present invention also provides a smart energy meter error prediction system, including a data acquisition module, an environmental impact error prediction module, a time impact error prediction module and a fusion module; The data acquisition module is used to collect historical error data and historical environmental variable data of the smart energy meter and perform preprocessing; The environmental impact error prediction module is used to build an environmental impact error prediction model, input the pre-processed historical error data and historical environmental variable data of the smart electric energy meter into the environmental impact error prediction model, and obtain the error prediction result of the smart electric energy meter considering the environmental impact; The time-affected error prediction module is used to construct a time-affected error prediction model, input the pre-processed historical error data of the smart electric energy meter into the time-affected error prediction model to analyze the time characteristics of the historical error data of the smart electric energy meter, and obtain the error prediction result of the smart electric energy meter considering the influence of the time characteristics based on the time characteristics; The fusion module is used to fuse the error prediction result of the smart electric energy meter considering the environmental impact and the error prediction result considering the time characteristic impact to obtain the final error prediction result of the smart electric energy meter.

[0014] The method implemented in the architecture diagram of the present invention combines Z-score anomaly detection with multiple interpolation algorithms in the data preprocessing phase and introduces a KD-Tree data structure for acceleration, achieving efficient cleaning and reconstruction of historical error and environmental variable data. Secondly, the environmental impact error prediction module adopts a generalized regression neural network (GRNN) with a Gaussian kernel function to accurately characterize the nonlinear mapping between error and multiple environmental factors such as temperature, humidity, and salt spray concentration. Combined with Bayesian optimization, it automatically tunes the smoothing factor, significantly improving prediction accuracy in environmentally sensitive scenarios. The temporal impact error prediction module uses an Autoformer architecture, splitting the error sequence into two parallel paths: periodicity and trend. Autocorrelation attention and one-dimensional convolutional / linear layers are used to extract key features, respectively. Multi-head self-attention and cross-attention are fused in the decoder to effectively capture seasonal fluctuations and long-term trends in the error. Finally, the present invention designs a dynamic fusion strategy based on a weighted coefficient α, organically integrating the outputs of the two models through multiplication and summation nodes, taking into account both environmental and temporal driving factors, significantly enhancing the stability and robustness of the prediction results. This method not only breaks through the limitations of a single model in capturing environmental or time series features, but also demonstrates excellent generalization ability and online prediction efficiency under complex working conditions such as high humidity, heat, and high salt fog.

[0015] On the other hand, the present invention further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method of the present invention when executing the program.

[0016] In another aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, which implements the method of the present invention when the program is executed by a processor.

[0017] The present invention has the following beneficial effects: 1. The present invention uses GRNN to efficiently model the nonlinear relationship between environmental factors and errors, and combines the Autoformer model with the autocorrelation attention mechanism and time series decomposition method to accurately separate long-term trends and cyclical fluctuations. It can effectively make up for the shortcomings of seasonal error analysis in current research and significantly improve the stability and accuracy of electricity meter error prediction.

[0018] 2. The present invention adopts a modular dual-pathway architecture design. Each submodule (data preprocessing, GRNN environmental model, Autoformer time model, and fusion module) can be independently optimized and replaced, achieving high maintainability and scalability of the method. It is convenient for the subsequent rapid integration of more environmental variables, timing analysis technologies, or new models, and can meet the flexible deployment requirements of different power grid scenarios and business needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 Flow chart of the method of the present invention; Figure 2 This is a diagram of the electric energy meter error prediction model architecture in an embodiment; Figure 3 2 is a comparison chart of the prediction performance of different models in the embodiment. DETAILED DESCRIPTION

[0020] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0021] It should be understood that the step numbers used herein are only for convenience of description and are not intended to limit the order in which the steps are to be executed.

[0022] It should be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, the singular forms "a", "an" and "the" are intended to include the plural forms unless the context clearly indicates otherwise.

[0023] The terms “include” and “comprising” indicate the presence of described features, integers, steps, operations, elements and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0024] The term "and / or" refers to and includes any and all possible combinations of one or more of the associated listed items.

[0025] Example 1: See also Figure 1 , a smart energy meter error prediction method, comprising the following steps: S1. Collect historical error data and historical environmental variable data of smart energy meters and preprocess them; In this embodiment, error prediction is performed on smart energy meter data collected from a 10kV industrial and commercial distribution station in a certain area. The data sampling interval is once per hour, and 17,520 historical error data and 17,520 environmental variable data are collected. The environmental variable data include temperature, relative humidity, and salt spray concentration. The Z-score algorithm is used to identify outliers in the historical error data and environmental variable data of smart electricity meters. Then, interpolation algorithms (such as KNN and mean interpolation) are used to fill in the outliers and construct a comprehensive error environment dataset. The error environment comprehensive data set includes multiple data sample groups, and each data sample group is composed of historical error data samples and historical environmental variable data samples of a smart electric energy meter.

[0026] Since a large amount of sample data is involved and the computational complexity is high, the KD-Tree algorithm is used to optimize the structure of the error environment comprehensive data set. The specific steps are as follows: Calculate the distribution of sample points in the error environment comprehensive data set, select the best segmentation dimension and recursively construct the KD-Tree; At the same time, k-nearest neighbor search is used for fast query to improve computing efficiency.

[0027] S2. Build an environmental impact error prediction model, input the preprocessed historical error data and historical environmental variable data of the smart energy meter into the environmental impact error prediction model, and obtain the error prediction result of the smart energy meter considering the environmental impact; The environmental impact error prediction model is constructed based on a generalized regression neural network (GRNN). The error prediction result of the smart energy meter considering environmental impact is specifically shown in the following formula: ; ; in: It represents the error prediction result of the smart energy meter considering the environmental impact; represents the number of data sample groups in the error environment synthetic dataset; Indicates the Environmental index weight of each data sample group; Indicates the Historical error data samples in a data sample group; Represents the current environment variable data; Indicates the Historical environmental variable data samples in a data sample group; represents the smoothing factor of the generalized regression neural network; In this embodiment, the kernel function of the generalized regression neural network is Gaussian, and the division ratio of the training-validation-test set is: 7:2:1; In this embodiment, the smoothing factor of the generalized regression neural network is optimized based on a Bayesian optimization algorithm.

[0028] S3. Construct a time-affected error prediction model, input the preprocessed historical error data of the smart energy meter into the time-affected error prediction model to analyze the time characteristics of the historical error data of the smart energy meter, and obtain the error prediction results of the smart energy meter considering the influence of time characteristics based on the time characteristics; The time-influenced error prediction model is built based on the Autoformer model. The preprocessed historical error data of the smart energy meter is constructed into a historical error time series. The Autoformer model uses the autocorrelation attention mechanism to extract the periodic features of the historical error time series, as shown in the following formula: ; in: express Periodic characteristics of the time series of historical errors; Express Perform conjugate calculations; represents the self-attention function; Represents the historical error time series in The query vector at the moment; Represents the historical error time series in The key vector of the moment; represents Fourier transform; represents the inverse Fourier transform; The trend characteristics in the historical error time series are extracted through a time window of preset length, as shown in the following formula: ; in: express Trend characteristics of the time series of historical errors; Indicates the length of the time window; express Sample of the moment historical error time series; The error prediction results of smart electric energy meters considering the influence of time characteristics are obtained by integrating the periodic characteristics and trend characteristics of the historical error time series, as shown in the following formula: ; in: It represents the error prediction result of the smart energy meter considering the influence of time characteristics; represents random noise; In this embodiment, the hyperparameter settings of the Autoformer model are shown in the following table:

[0029] Preferably, the time impact error prediction model uses AdamW optimizer to optimize model parameters, and the parameter settings of the model are specifically shown in the following formula: ; in: Indicates that the time-influence error prediction model is Parameters of the moment; Indicates that the time-influence error prediction model is Parameters of the moment; represents the learning rate; represents a very small constant (usually taken as ); Represents the weight decay coefficient, which is used to regularize model parameters to avoid overfitting; Indicates that the time-influence error prediction model is The first-order moment of the gradient at the moment is shown in the following formula: ; in: represents the exponential decay coefficient of the first-order moment (usually 0.9); Indicates that the time-influence error prediction model is The first moment of the gradient at the moment; Represents the gradient of the loss function of the time-influence error prediction model with respect to the parameters; Indicates that the time-influence error prediction model is The second-order moment of the gradient at the moment is shown in the following formula: ; in: represents the exponential decay coefficient of the second-order moment (usually 0.999); Indicates that the time-influence error prediction model is Second moment of gradient; S4. Combine the error prediction results of the smart energy meter that consider environmental impact and the error prediction results that consider time characteristics, as shown in the following formula: ; ; in: Represents the final error prediction result of the smart energy meter; represents the fusion coefficient; represents the prediction error of the time-influenced error prediction model; represents the prediction error of the environmental impact error prediction model; Optionally, the prediction errors of the time impact error prediction model and the environment impact error prediction model can be evaluated by a mean square error loss function (MSE), a mean absolute error loss function (MAE), or a root mean square error loss function (RMSE); In this embodiment, the root mean square error loss function is selected for evaluation to obtain the prediction error of the time impact error prediction model. , the prediction error of the environmental impact error prediction model , then the fusion coefficient is calculated ; After experimental verification under complex environmental conditions of high salt fog, high humidity and heat, the prediction performance was compared and analyzed with traditional statistical models (ARIMA), machine learning models (SVR), deep learning models (LSTM), as well as common Bayesian regression (BNR) and BP neural network (BPNN). Figure 2 As shown, the GRNN+Autoformer method proposed in this paper has the lowest RMSE value (0.040) and the highest coefficient of determination (R²=0.97). Compared with the traditional statistical method ARIMA (RMSE=0.085, R²=0.85), the traditional machine learning method SVR (RMSE=0.074, R²=0.88), and the classic deep learning method LSTM (RMSE=0.065, R²=0.91), the prediction error of the proposed method is significantly lower and the prediction accuracy is higher. In addition, compared with the commonly used Bayesian regression (BNR, RMSE=0.060, R²=0.93) and BP neural network (BPNN, RMSE=0.055, R²=0.94), the proposed method still shows superior generalization performance and prediction stability in high salt fog and high humidity and heat environments, demonstrating better environmental adaptability.

[0030] S5. Calculate the measurement correction value of the smart energy meter based on the final error prediction result of the smart energy meter, as shown in the following formula: ; in: Indicates the measurement correction value of the smart energy meter; Indicates the actual measurement value of the smart energy meter.

[0031] Example 2: An intelligent electric energy meter error prediction system includes a data acquisition module, an environmental impact error prediction module, a time impact error prediction module and a fusion module; The data acquisition module is used to collect historical error data and historical environmental variable data of the smart energy meter and perform preprocessing; The environmental impact error prediction module is used to build an environmental impact error prediction model, input the pre-processed historical error data and historical environmental variable data of the smart electric energy meter into the environmental impact error prediction model, and obtain the error prediction result of the smart electric energy meter considering the environmental impact; The time-affected error prediction module is used to construct a time-affected error prediction model, input the pre-processed historical error data of the smart electric energy meter into the time-affected error prediction model to analyze the time characteristics of the historical error data of the smart electric energy meter, and obtain the error prediction result of the smart electric energy meter considering the influence of the time characteristics based on the time characteristics; The fusion module is used to fuse the error prediction results of the smart electric energy meter that take into account environmental impact and time characteristics to obtain the final error prediction result of the smart electric energy meter.

[0032] This system is used to implement the method in Example 1 and will not be described in detail here.

[0033] Example 3: This embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the method described in any embodiment of the present invention is implemented.

[0034] Example 4: This embodiment provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the method described in any embodiment of the present invention is implemented.

[0035] In the embodiments of the present application, "at least one" refers to one or more, and "more" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent the existence of A alone, the existence of A and B at the same time, and the existence of B alone. Among them, A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b and c can represent: a, b, c, a and b, a and c, b and c or a and b and c, where a, b, c can be single or multiple.

[0036] Those skilled in the art will appreciate that the various units and algorithm steps described in the embodiments disclosed herein can be implemented using a combination of electronic hardware, computer software, and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0037] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0038] In the several embodiments provided in this application, if any function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of this application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory; hereinafter referred to as: ROM), random access memory (Random Access Memory; hereinafter referred to as: RAM), magnetic disk or optical disk, and other media that can store program code.

[0039] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention's description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for predicting errors in smart electric energy meters, characterized in that: The following steps are involved: Collect historical error data and historical environmental variable data of smart energy meters and perform preprocessing; An environmental impact error prediction model is constructed, and the pre-processed historical error data and historical environmental variable data of the smart energy meter are input into the environmental impact error prediction model to obtain the error prediction result of the smart energy meter considering the environmental impact; Construct a time-influenced error prediction model, input the preprocessed historical error data of the smart energy meter into the time-influenced error prediction model to analyze the time characteristics of the historical error data of the smart energy meter, and obtain the error prediction result of the smart energy meter considering the influence of the time characteristics based on the time characteristics; The error prediction results of the smart electricity meter considering environmental impact and time characteristics are integrated to obtain the final error prediction result of the smart electricity meter.

2. The method for predicting an error of a smart electric energy meter according to claim 1, characterized in that: The preprocessing steps of the historical error data and the historical environmental variable data of the smart electric energy meter are as follows: The Z-score algorithm is used to identify outliers in the historical error data and environmental variable data of smart electricity meters. Then, an interpolation algorithm is used to fill in the outliers and construct a comprehensive error environment data set. The error environment comprehensive data set includes multiple data sample groups, and each data sample group is composed of historical error data samples and historical environmental variable data samples of a smart electric energy meter.

3. The method for predicting an error of a smart electric energy meter according to claim 2, wherein: The structure of the error environment comprehensive dataset is optimized using the KD-Tree algorithm.

4. The method for predicting an error of a smart electric energy meter according to claim 2, wherein: The environmental impact error prediction model is constructed based on a generalized regression neural network. The error prediction result of the smart energy meter considering the environmental impact is specifically shown in the following formula: ； ； in: It represents the error prediction result of the smart energy meter considering the environmental impact; represents the number of data sample groups in the error environment synthetic dataset; Indicates the Environmental index weight of each data sample group; Indicates the Historical error data samples in a data sample group; Represents the current environment variable data; Indicates the Historical environmental variable data samples in a data sample group; Represents the smoothing factor of the generalized regression neural network.

5. The method for predicting an error of a smart electric energy meter according to claim 4, characterized in that: The time-influenced error prediction model is built based on the Autoformer model. The preprocessed historical error data of the smart energy meter is constructed into a historical error time series. The Autoformer model uses the autocorrelation attention mechanism to extract the periodic features of the historical error time series, as shown in the following formula: ； in: express Periodic characteristics of the time series of historical errors; represents conjugate calculation; represents the self-attention function; Represents the historical error time series in The query vector at the moment; Represents the historical error time series in The key vector of the moment; represents Fourier transform; represents the inverse Fourier transform; The trend characteristics in the historical error time series are extracted through a time window of preset length, as shown in the following formula: ； in: express Trend characteristics of the time series of historical errors; Indicates the length of the time window; express Sample of the moment historical error time series; The error prediction results of smart electric energy meters considering the influence of time characteristics are obtained by integrating the periodic characteristics and trend characteristics of the historical error time series, as shown in the following formula: ； in: It represents the error prediction result of the smart energy meter considering the influence of time characteristics; represents random noise.

6. The method for predicting an error of a smart electric energy meter according to claim 5, characterized in that: The time impact error prediction model uses the AdamW optimizer to optimize model parameters.

7. The method for predicting an error of a smart electric energy meter according to claim 5, characterized in that: The error prediction results of the smart energy meter considering environmental impact and the error prediction results considering time characteristics are integrated, as shown in the following formula: ； ； in: Represents the final error prediction result of the smart energy meter; represents the fusion coefficient; represents the prediction error of the time-influenced error prediction model; Represents the prediction error of the environmental impact error prediction model.

8. A smart energy meter error prediction system, characterized in that: It includes data acquisition module, environmental impact error prediction module, time impact error prediction module and fusion module; The data acquisition module is used to collect historical error data and historical environmental variable data of the smart energy meter and perform preprocessing; The environmental impact error prediction module is used to build an environmental impact error prediction model, input the pre-processed historical error data and historical environmental variable data of the smart electric energy meter into the environmental impact error prediction model, and obtain the error prediction result of the smart electric energy meter considering the environmental impact; The time-affected error prediction module is used to construct a time-affected error prediction model, input the pre-processed historical error data of the smart electric energy meter into the time-affected error prediction model to analyze the time characteristics of the historical error data of the smart electric energy meter, and obtain the error prediction result of the smart electric energy meter considering the influence of the time characteristics based on the time characteristics; The fusion module is used to fuse the error prediction results of the smart electric energy meter considering environmental influences and time characteristics to obtain the final error prediction result of the smart electric energy meter.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the method according to any one of claims 1 to 7 is implemented.

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

Images