Electric energy meter error compensation method, equipment and medium

Abstract

The invention provides an electric energy meter error compensation method and device and a medium. The method comprises the steps that basic parameters of an electric energy meter are determined; the basic parameters are input into an error estimation model deployed in the local of the electric energy meter, local estimation errors output by the error estimation model are obtained, parameters of the error estimation model are remotely updated by the data master station, and the data master station updates the parameters of the error estimation model under the condition that the model errors of the error estimation model are larger than a preset threshold value; and performing error compensation on the metering result of the electric energy meter based on the local estimation error. According to the method, the device and the medium provided by the invention, the error estimation model is adopted to perform error compensation on the electric energy meter, and the reliability of error estimation can be effectively ensured. Model parameters are remotely updated through the data master station, that is, error estimation is carried out by applying the field electric energy meter, parameter updating is carried out by applying the remote data master station, and the reliability and accuracy of error compensation are ensured through a mode of combining field operation and master station operation.

Description

Technical Field

[0001] The present invention relates to the technical field of electric energy meter measurement, and in particular to an electric energy meter error compensation method, equipment and medium. Background Art

[0002] Smart IoT energy meters play a vital role in modern power systems, and their measurement accuracy is an important indicator of energy meter performance. However, during the operation of energy meters, the self-heating effect and changes in environmental factors will have a significant impact on the measurement accuracy.

[0003] There are currently two main solutions for error compensation of electric energy meters. Solution 1 is error compensation based on sampling element monitoring. Through the interaction between the metering chip and the sampling element, the change of the output signal value of the sampling element is monitored to perform error compensation. This solution requires that the electric energy meter is equipped with a metering chip with an error self-monitoring function. Since the constant electrical signal emitted by the metering chip is easily affected by environmental factors, it may cause a large compensation error. Solution 2 is error compensation based on temperature. By monitoring the relationship between the internal temperature of the electric energy meter and the active pulse error, the power gain is adjusted in real time. Although this solution solves some of the errors caused by temperature changes, it cannot eliminate the errors caused by load current differences. The nonlinear relationship between temperature and error increases the complexity of the fitting curve, which places a heavy computational burden on the microcontroller and may affect its normal operation.

[0004] Therefore, how to efficiently and reliably improve the measurement accuracy of electric energy meters remains an urgent problem to be solved. Summary of the invention

[0005] The present invention provides an electric energy meter error compensation method, device and medium, which are used to solve the defect of inaccurate measurement of electric energy meters in related technologies.

[0006] The present invention provides an electric energy meter error compensation method, which is applied to an electric energy meter. The electric energy meter error compensation method comprises: Determining basic parameters of the electric energy meter; The basic parameters are input into an error estimation model deployed locally at the electric energy meter to obtain a local estimated error output by the error estimation model, wherein the parameters of the error estimation model are remotely updated by a data master station, and the data master station updates the parameters of the error estimation model when the model error of the error estimation model is greater than a preset threshold value; Based on the local estimated error, error compensation is performed on the measurement result of the electric energy meter.

[0007] According to an electric energy meter error compensation method provided by the present invention, the basic parameters include the load current, single chip computer temperature and terminal seat temperature of the electric energy meter; The step of inputting the basic parameters into an error estimation model deployed locally at the electric energy meter to obtain a local estimation error output by the error estimation model includes: The load current, the temperature of the single-chip computer and the temperature of the terminal seat of the electric energy meter are input into an error estimation model deployed locally on the electric energy meter to obtain a local estimation error output by the error estimation model.

[0008] According to an electric energy meter error compensation method provided by the present invention, the model error is the difference between the local estimation error and the master station estimation error; The master station estimation error is obtained by the data master station performing error estimation on the electric energy meter based on a master station estimation model. The master station estimation model is deployed at the data master station, and the scale of the master station estimation model is greater than or equal to the scale of the error estimation model.

[0009] According to an electric energy meter error compensation method provided by the present invention, the input parameters of the master station estimation model include the basic parameters and the operating time, power factor and grid frequency of the electric energy meter.

[0010] According to an electric energy meter error compensation method provided by the present invention, the power factor includes an instantaneous value of the power factor and an average value of the power factor; The grid frequency includes an instantaneous grid frequency value and an average grid frequency value.

[0011] According to an electric energy meter error compensation method provided by the present invention, the error estimation model is a back propagation neural network.

[0012] The present invention also provides a parameter updating method of an error estimation model, which is applied to a data master station. The parameter updating method of the error estimation model comprises: Obtaining a model error of an error estimation model, where the error estimation model is deployed locally at the electric energy meter, the error estimation model is used to output a local estimation error based on basic parameters of the electric energy meter, and the local estimation error is used to perform error compensation on a metering result of the electric energy meter; When the model error is greater than a preset threshold, parameters of the error estimation model are updated, and the updated parameters are configured locally in the electric energy meter.

[0013] The present invention also provides 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, an electric energy meter error compensation method or a parameter updating method of an error estimation model as described above is implemented.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for compensating an electric energy meter error or the method for updating parameters of an error estimation model as described above is implemented.

[0015] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above-mentioned methods for compensating an electric energy meter error or methods for updating parameters of an error estimation model.

[0016] The electric energy meter error compensation method, device and medium provided by the present invention firstly adopt an error estimation model to perform error compensation on the electric energy meter, which can effectively ensure the reliability of error estimation. And on this basis, when the model error of the error estimation model is greater than a preset threshold, the model parameters are remotely updated through the data master station, that is, the electric energy meter on site is used for error estimation, and the remote data master station is used for parameter update. By combining on-site calculation with master station calculation, under the premise that the computing power resources of the electric energy meter itself are limited, the reliability and accuracy of error compensation are further ensured. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or related technologies, the following briefly introduces the drawings required for use in the embodiments or related technical descriptions. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 It is a flow chart of the electric energy meter error compensation method provided by the present invention.

[0019] Figure 2 It is a schematic diagram of the back propagation neural network structure provided by the present invention.

[0020] Figure 3 It is a flow chart of the parameter updating method of the error estimation model provided by the present invention.

[0021] Figure 4 It is a structural schematic diagram of the electric energy meter error compensation device provided by the present invention.

[0022] Figure 5 It is a structural schematic diagram of a parameter updating device for an error estimation model provided by the present invention.

[0023] Figure 6 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] During the operation of the smart IoT energy meter, due to the self-heating effect (internal manganese copper heating) and environmental factors, measurement errors will occur. Existing smart IoT energy meters generally use a single-chip microcomputer plus a metering chip to measure energy. The measurement accuracy of the metering chip depends on the correction coefficient written into it by the single-chip microcomputer program. Among the many correction coefficients, the power gain coefficient has the most direct impact on the measurement accuracy. Therefore, the key to error compensation lies in how the single-chip microcomputer program dynamically adjusts the power gain. There are two main solutions for existing error compensation technologies.

[0026] The first solution is error compensation based on sampling element monitoring. This solution requires the metering chip to have an error self-monitoring function. The basic principle is that when the electric energy meter leaves the factory, the metering chip actively outputs a constant electrical signal to the sampling element and records the output signal value of the sampling element as a reference signal. After the electric energy meter leaves the factory, this operation is repeated at regular intervals to monitor whether the signal value output by the sampling element deviates from the reference signal. If the signal value deviates, the metering chip adjusts the power gain coefficient according to the deviation. This solution can eliminate errors caused by environmental factors, but it requires a specific metering chip, and if the constant electrical signal is affected by the environment, the compensation will be inaccurate.

[0027] Another solution is temperature-based error compensation. The basic principle is to draw a curve of the relationship between the internal temperature of the energy meter and the active pulse error, and write its expression into the energy meter. When the energy meter is running, the microcontroller program monitors the internal temperature in real time and dynamically adjusts the power gain according to the curve expression to eliminate the impact of temperature changes on measurement accuracy. Although this solution solves some of the error problems caused by temperature changes, it cannot eliminate the errors caused by load current differences. The nonlinear relationship between temperature and error increases the computational complexity and requires high computing power for the microcontroller.

[0028] In view of the above problems, an embodiment of the present invention provides an electric energy meter error compensation method. Figure 1 FIG. 1 is a flow chart of the method for compensating an electric energy meter error provided by the present invention. Figure 1 As shown, the method is applied to an electric energy meter, and the method comprises: Step 110: determining basic parameters of the electric energy meter.

[0029] The electric energy meter here is the electric energy meter that needs error compensation. The electric energy meter itself has the function of collecting its own parameters, so it can obtain parameters associated with the pulse error generated by itself during metering. In the embodiment of the present invention, the parameters associated with the pulse error that the electric energy meter can collect by itself are recorded as basic parameters.

[0030] For example, the load current of the electric energy meter, the temperature of the electric energy meter, etc. can be used as basic parameters. Furthermore, the temperature of the electric energy meter can be specifically refined into the temperature of the single chip computer inside the electric energy meter and the temperature of the terminal seat of the electric energy meter as the basic parameters of the electric energy meter, and the embodiment of the present invention does not specifically limit this.

[0031] Step 120: input the basic parameters into the error estimation model deployed locally at the electric energy meter to obtain a local estimated error output by the error estimation model. The parameters of the error estimation model are remotely updated by a data master station. The data master station updates the parameters of the error estimation model when the model error of the error estimation model is greater than a preset threshold.

[0032] Specifically, in order to quickly and conveniently realize error compensation for the electric energy meter, an error estimation model can be deployed locally in the electric energy meter. Here, the error estimation model is a neural network model for error estimation. It can be understood that the error estimation model is deployed locally in the electric energy meter, specifically storing the parameters of the error estimation model locally in the electric energy meter, so that when error compensation is required for the electric energy meter, the electric energy meter can perform error estimation on the electric energy meter based on the basic parameters of the electric energy meter obtained by the collection and the parameters of the error estimation model stored locally, thereby obtaining the local estimated error of the electric energy meter.

[0033] During specific operation, the electric energy meter can use the basic parameters of the electric energy meter obtained in step 110 as input parameters of the locally deployed error estimation model, that is, input the basic parameters of the electric energy meter into the error estimation model, and implement error estimation based on the basic parameters of the electric energy meter by running the error estimation model locally, thereby using the output result of the error estimation model as the local estimation error. Here, the local estimation error is the positive active pulse error of the electric energy meter estimated locally in the electric energy meter.

[0034] On this basis, in order to ensure the reliability and accuracy of the error estimation model itself, it is necessary to monitor the model error of the error estimation model, and update and optimize the parameters of the error estimation model when the model error is greater than the preset threshold. The model error referred to here can be understood as the difference between the local estimated error based on the output of the error estimation model and the true error. The smaller the difference between the two, the smaller the model error, and the higher the reliability and accuracy of the error estimation model itself; conversely, the larger the difference between the two, the larger the model error, and the worse the reliability and accuracy of the error estimation model itself.

[0035] Considering that the computing power of the electric energy meter itself is limited, and updating and optimizing the parameters of the error estimation model usually requires a lot of computing power support, this type of complex calculation is obviously not suitable for running on the electric energy meter. In view of this situation, in an embodiment of the present invention, the parameters of the error estimation model are updated based on the data master station, and the data master station remotely configures the updated parameters to the electric energy meter.

[0036] Here, the data master station is a device with powerful computing power, and the data master station has the function of communicating with the electric energy meter. For example, at the equipment site, a portable computer equipped with a data transmission device can be used as a data master station. The data master station can monitor the model error of the error estimation model deployed locally at the electric energy meter, and when the model error is greater than a preset threshold, the parameters of the error estimation model are updated and optimized, and then the updated parameters are remotely configured to the electric energy meter, thereby updating the parameters of the error estimation model at the electric energy meter, and ensuring the reliability and accuracy of the error estimation performed locally at the electric energy meter.

[0037] Here, the preset threshold is a pre-set threshold, and the preset threshold can be understood as the maximum value of the tolerable model error.

[0038] Step 130: Based on the local estimation error, error compensation is performed on the measurement result of the electric energy meter.

[0039] Specifically, after the local estimated error is obtained, the power gain coefficient inside the electric energy meter can be dynamically adjusted based on the local estimated error to correct the measurement error of the electric energy meter.

[0040] In practical applications, the electric energy meter can monitor the basic parameters of the electric energy meter in real time (such as load current, microcontroller temperature and terminal block temperature), input the basic parameters into the locally deployed error compensation model, and then obtain the local estimated error. According to the local estimated error, the power gain coefficient is adjusted to make the output measurement result of the electric energy meter more accurate.

[0041] In addition, for application scenarios with high accuracy requirements, such as industrial electricity, the data master station operation can also optimize and update the error estimation model, and remotely configure and update the optimized model parameters to the on-site electric energy meter, thereby ensuring high accuracy and reliability of error compensation. Through this dynamic adjustment and parameter update method, the electric energy meter can maintain high-precision measurement performance in various complex environments.

[0042] The electric energy meter error compensation method provided by the present invention first adopts an error estimation model to perform error compensation on the electric energy meter, which can effectively ensure the reliability of error estimation. And on this basis, when the model error of the error estimation model is greater than a preset threshold, the model parameters are remotely updated through the data master station, that is, the electric energy meter on site is used for error estimation, and the remote data master station is used for parameter update. By combining on-site calculation with master station calculation, under the premise that the computing power resources of the electric energy meter itself are limited, the reliability and accuracy of error compensation are further guaranteed.

[0043] Based on the above embodiment, the basic parameters include the load current, the single chip computer temperature and the terminal seat temperature of the electric energy meter; Accordingly, in step 120, the step of inputting the basic parameters into an error estimation model deployed locally at the electric energy meter to obtain a local estimation error output by the error estimation model includes: The load current, the temperature of the single-chip computer and the temperature of the terminal seat of the electric energy meter are input into an error estimation model deployed locally on the electric energy meter to obtain a local estimation error output by the error estimation model.

[0044] Specifically, when applying the error estimation model deployed locally in the electric energy meter to perform error estimation, that is, performing on-site calculations, parameters including the load current, microcontroller temperature and terminal seat temperature of the electric energy meter can be used as basic parameters input into the error estimation model. Thus, the error estimation model can realize the estimation of the forward active pulse error of the electric energy meter locally based on the load current, microcontroller temperature and terminal seat temperature.

[0045] It can be understood that the load current, microcontroller temperature and terminal block temperature of the electric energy meter referred to here are all for the same moment.

[0046] Based on any of the above embodiments, the error estimation model is a back propagation neural network.

[0047] Specifically, the field operation part, that is, the error estimation model deployed locally in the electric energy meter, can be a back propagation neural network. In the embodiment of the present invention, the error of the electric energy meter can be accurately estimated through the trained back propagation neural network, thereby dynamically adjusting the compensation parameters to improve the metering accuracy.

[0048] Figure 2 Schematic diagram of the back propagation neural network structure provided by the present invention. Figure 2 As shown, the back propagation neural network adopts a four-layer structural model, including an input layer, two hidden layers and an output layer. Among them, the input layer includes 3 neurons, corresponding to the input load current, the temperature of the single-chip computer and the temperature of the terminal block respectively; the first hidden layer includes 6 neurons, using the rectified linear unit as the activation function; the second hidden layer includes 4 neurons, also using the rectified linear unit as the activation function; the output layer includes 1 neuron, and the output is the error value of the electric energy meter, that is, the local estimation error. When training the back propagation neural network, the learning rate and the maximum number of training times can be set. For example, the learning rate can be set to 0.001 and the maximum number of training times can be set to 100,000 times. The embodiment of the present invention does not limit this.

[0049] Assume that the neuron set of the network input layer is represented as:

[0050] in, x 1, x 2, x 3 are load current, microcontroller temperature and terminal block temperature respectively.

[0051] The first hidden layer neuron set is expressed as:

[0052] The second hidden layer neuron set is expressed as:

[0053] The output layer neuron set is expressed as:

[0054] The set of bias coefficients of the first hidden layer is expressed as:

[0055] The set of bias coefficients of the second hidden layer is expressed as:

[0056] The weight matrix between the input layer and the first hidden layer is expressed as:

[0057] The weight matrix between the first hidden layer and the second hidden layer is expressed as:

[0058] The weight matrix between the second hidden layer and the output layer is expressed as:

[0059] In the forward propagation process of the error estimation model, the input layer receives the basic parameters of the electric energy meter and passes the basic parameters to each neuron in the first hidden layer, calculates the weighted sum of each neuron and adjusts the output of the neuron through the modified linear unit. The output of the first hidden layer is further passed to the second hidden layer for similar calculations. Finally, the output of the second hidden layer is passed to the output layer, and the error estimate is calculated and output as the local estimated error. It can be understood that the forward propagation process of the error estimation model can be understood as the process of performing error estimation based on the error estimation model after obtaining the basic parameters.

[0060] Specifically, the first hidden layer j The input value of a neuron , output value ; The second hidden layer j The input value of a neuron , output value ; Input value of output layer neuron , output value o Equal to input value O , output value o That is the prediction result of the network, that is, the local estimation error.

[0061] In addition, before applying the error estimation model to perform error estimation, it is also necessary to train the error estimation model. The training process here can be performed on a host computer other than the electric energy meter.

[0062] Here, the process of training the error estimation model can also be understood as the process of back propagation for the error estimation model.

[0063] For the back propagation process of the error estimation model, the error between the actual output of the output layer and the target output (true error value) is calculated, the error is passed back to each layer through the back propagation algorithm, the gradient of each weight and bias is calculated, and the weights and biases are adjusted using the gradient descent method to minimize the error function.

[0064] Specifically, in the back-propagation process, the model parameters need to be initialized first, that is, the bias coefficient set BM , BN And the matrix WX , WM , WN Perform random initialization. Then, calculate the output value of each sample in the training set o Next, the error of the error estimation model is calculated.

[0065] Here, the error of the error estimation model can be expressed as ,in A represents the number of samples in the training set, o i Indicates i The output value of the sample, o i ' indicates the i Based on the calculated error of the error estimation model, the parameters of the error estimation model can be iteratively updated.

[0066] Here, the update process for parameters including bias coefficients and weight matrices in the error estimation model is as follows: BM The update algorithm is , where After the tth adjustment The value of After the t+1th adjustment The value of Error e Bias coefficient The partial derivative of .

[0067] BN The update algorithm is , where After the tth adjustment The value of After the t+1th adjustment The value of Error e Bias coefficient The partial derivative of .

[0068] WX The update algorithm is , where After the tth adjustment The value of After the t+1th adjustment The value of Error e Weight coefficient The partial derivative of .

[0069] WM The update algorithm is , where After the tth adjustment The value of After the t+1th adjustment The value of Error e Weight coefficient The partial derivative of .

[0070] WN The update algorithm is , where After the tth adjustment The value of After the t+1th adjustment The value of Error e Weight coefficient The partial derivative of .

[0071] In the embodiment of the present invention, the above training process first needs to obtain training samples. In order to train the back propagation neural network, a large number of training samples need to be collected. The training samples contain the operating data of the electric energy meter under different load currents, specifically including the temperature of the single chip computer, the temperature of the terminal block and the measured pulse error. In the embodiment of the present invention, the minimum current value I min , turning current value I tr , ten times the turning current value 10I tr , half of the maximum current value I max / 2 and the maximum current value I max , as the sampling point of the load current, that is, the value point of the load current in the training sample. Thus, the load current is recorded i , MCU temperature T 0.Terminal seat temperature T 1 and active pulse error o ′, and finally form training samples. For example, the training samples can be expressed as shown in the following table: Table 1 Training samples

[0072] It should be noted that under a current of 10000mA, the temperature of the microcontroller may be 23.5°C, or it may be any value in the range of 23.5°C to 55.0°C (including 55.0°C), and the corresponding terminal seat temperature value is the same. The ultimate goal is to calculate the pulse error based on the first three parameters, and then correct the measurement results of the electricity meter.

[0073] After obtaining the training samples, the back propagation neural network can be trained. The training aims to enable the back propagation neural network to accurately predict the error of the electric energy meter under different conditions through a large amount of training data and repeated optimization.

[0074] First, the acquired training samples are randomly arranged and divided into training set and test set. The first 80% of the samples are used for training, and the last 20% of the samples are used for testing. The training set is used to adjust the weights and biases of the neural network, while the test set is used to evaluate the generalization ability and accuracy of the model. Then, the weight matrix and bias coefficients of the neural network are randomly initialized. The weight matrix includes the weight matrix between the input layer and the first hidden layer. WX , the weight matrix between the first hidden layer and the second hidden layer WM , the weight matrix between the second hidden layer and the output layer WN The bias coefficient set includes the bias coefficient set of the first hidden layer BM and the bias coefficient set of the second hidden layer BN . Then, according to the forward propagation calculation method, the output value of each sample is calculated o Compare the actual output value 𝑜 of the output layer with the target output value (the pulse error in the training sample) o ′, calculate the error function. For example, the error function can be expressed as If the error e is greater than the set error threshold, repeat the forward propagation and back propagation process until the error e Less than the set threshold or reaching the maximum number of training times. Through repeated iterations, the back propagation neural network gradually adjusts the weights and biases of each layer, making the output error estimate closer and closer to the target output value.

[0075] In addition, after completing the back-propagation neural network training, the back-propagation neural network can also be tested to evaluate its performance and generalization ability to ensure that it can provide accurate error estimation in practical applications. The test set used here can be the last 20% of the test set obtained by the above division. The test set is not used to adjust the network weights and biases during the training process, so the generalization ability and actual performance of the back-propagation neural network as an error estimation model can be effectively evaluated.

[0076] For each sample data in the test set, its output value is calculated through forward propagation, and the error calculation function is used , calculate the test error. If the error e If the error is still less than the set error threshold, it means that the network parameters are correct. Otherwise, it means that the network has overfitting or underfitting problems, and the learning rate or the maximum number of training times needs to be adjusted to retrain the neural network.

[0077] After the test is passed, the trained back propagation neural network can be deployed as an error estimation model to the electric energy meter to realize on-site operation. In the specific implementation, only the above model parameters need to be BM , BN And the matrix WI , WM , WN After the data is written into the electric energy meter, the electric energy meter will predict the positive active pulse error at this time according to the model parameters currently stored in the meter, and then adjust the "power gain" coefficient of the metering chip in reverse according to this error to compensate for the error. Through the powerful learning ability and nonlinear fitting ability of the back propagation neural network, high-precision error estimation and compensation can be achieved, thus improving the metering accuracy of the electric energy meter.

[0078] Based on any of the above embodiments, the parameters of the error estimation model are stored locally in the electric energy meter.

[0079] Specifically, when the electric energy meter leaves the factory, the trained and optimized back propagation neural network model parameters (including weight matrix and bias coefficient) are pre-loaded and stored in the local memory of the electric energy meter. These model parameters cover all connection weights and bias values ​​between the input layer, hidden layer and output layer, ensuring that the electric energy meter can perform error estimation in real time during actual operation.

[0080] When the energy meter is put into use, the field operation part can monitor and process the input basic parameters (such as load current, microcontroller temperature and terminal block temperature) in real time, and use the stored model parameters to calculate the current local estimation error through forward propagation. According to the local estimation error, the energy meter will dynamically adjust the power gain coefficient of its metering chip to correct the metering error. This process is completely carried out locally in the energy meter, without relying on external data or computing resources, ensuring real-time and high efficiency.

[0081] The above implementation method enables the electric energy meter to quickly and accurately perform error compensation through on-site calculations to ensure basic metering accuracy and reliability. At the same time, the locally stored model parameters also facilitate the electric energy meter to respond and adjust quickly during operation to adapt to various basic environmental changes and load conditions, thereby meeting the needs of electric energy meter error calculation.

[0082] Based on any of the above embodiments, the model error is the difference between the local estimation error and the master station estimation error; The master station estimation error is obtained by the data master station performing error estimation on the electric energy meter based on a master station estimation model. The master station estimation model is deployed at the data master station, and the scale of the master station estimation model is greater than or equal to the scale of the error estimation model.

[0083] Specifically, a master station estimation model is deployed at the data master station. Similar to the error estimation model deployed locally at the electric energy meter, the master station estimation model is also used for error estimation.

[0084] In addition, since the data master station itself has powerful computing capabilities and can complete the processing of huge amounts of data, the master station estimation model deployed at the data master station can be larger in scale than the error estimation model deployed locally on the electric energy meter. Accordingly, the master station estimation model has higher accuracy and greater reliability in error estimation than the error estimation model.

[0085] It is understandable that the scale of the master station estimation model is greater than or equal to the scale of the error estimation model, which can be specifically reflected in that the master station estimation model has more model parameters and more types of input parameters of the master station estimation model, etc. The embodiment of the present invention does not make specific limitations on this.

[0086] Considering that the scale of the master station estimation model is greater than or equal to the scale of the error estimation model, that is, the master station estimation model has better accuracy in error estimation than the error estimation model, in the process of error compensation for the electric energy meter, while obtaining the local estimation error based on the error estimation model deployed locally at the electric energy meter, the master station estimation error can also be obtained based on the master station estimation model deployed at the data master station. It can be understood that the reliability of the master station estimation error is higher than that of the local estimation error.

[0087] Therefore, after respectively obtaining the local estimation error and the master station estimation error, the difference between the local estimation error and the master station estimation error can be calculated, and the difference can be used as the model error of the local estimation model.

[0088] Based on any of the above embodiments, the input parameters of the master station estimation model include the basic parameters and the operating time, power factor and grid frequency of the electric energy meter.

[0089] Specifically, since the master station operation implemented by the master station estimation model at the data master station adopts high-performance, high-computing power equipment, it can handle more complex computing tasks and can therefore effectively process richer input data, thereby improving the accuracy and reliability of error estimation.

[0090] Therefore, in the embodiment of the present invention, in addition to using the basic parameters including load current, single-chip computer temperature and terminal seat temperature as input parameters of the master station estimation model, more parameters that affect the measurement accuracy can also be obtained as input parameters of the master station estimation model. That is, the input parameters of the master station estimation model can also include the operation time, power factor and grid frequency of the electric energy meter. These parameters provide more comprehensive operating status and environmental information for the data master station to perform error estimation, which helps to estimate the error of the electric energy meter more accurately.

[0091] In practical applications, various parameters including load current, MCU temperature, terminal block temperature, operating time, power factor and grid frequency can be input into the master station estimation model at the data master station to achieve training and optimization for the master station estimation model. After training and optimization, the master station estimation model can handle more variables and more complex inputs and provide higher-precision error estimation compared to the error estimation model deployed locally on the energy meter.

[0092] Based on any of the above embodiments, the power factor includes an instantaneous value of the power factor and an average value of the power factor; the grid frequency includes an instantaneous value of the grid frequency and an average value of the grid frequency.

[0093] Specifically, among the input parameters of the master station estimation model, the power factor and grid frequency can be divided into two categories: instantaneous value and average value. The instantaneous value is a set of data and the corresponding error recorded by the system at a certain moment, while the average value is the arithmetic mean of the parameters from the time the energy meter is enabled / restarted or from a specific time point to the time when a set of data is recorded.

[0094] For example, the master station estimation error can be calculated at the data master station as follows: First, collect and prepare various operating parameters of the energy meter, including: load current i , MCU temperature 𝑇0, terminal block temperature T 1. Operating time, power factor, and grid frequency. The power factor and grid frequency can be further divided into instantaneous value and average value. Input the above parameters into the trained and optimized master station estimation model to obtain the master station estimation error estimated by the master station estimation model. o '',Right now:

[0095] In addition, the local estimation error calculated by the error estimation model at the local power meter is It can be expressed as:

[0096] Among them, k0, k1 and k2 are correction coefficients, , and All are model functions. PF , f grd is the instantaneous value of power factor and grid frequency, is the average value of power factor and grid frequency, t run is the running time.

[0097] Error compared to local estimates calculated under field conditions Compared with the master station, the master station estimation error obtained under the master station condition is o ''In the calculation, the input parameters are increased from the original 3 to 8 (at the same time, the correction coefficient is set according to the characteristics of each data k ), and introduced two variable averages to verify the cumulative error impact, and increased the number of neurons according to demand, so that the error relative to the local estimation can be obtained More accurate and reliable master station estimation error o′′.

[0098] In the electric energy meter error compensation method provided in the embodiment of the present invention, the output value of the error estimation model of the neural network structure is used as the compensation data of the electric energy meter operation error. Since a nonlinear activation function is added to the neural network, theoretically, as long as a sufficient number of neurons and layers are added, the neural network can fit any function. Therefore, the error estimation model used in the embodiment of the present invention has a strong generalization ability.

[0099] In addition, the parameters of the error estimation model in the embodiment of the present invention support dynamic updating. If higher compensation accuracy is desired, the parameters can be optimized and updated through the data master station. Moreover, more input features can be added to the error estimation model or the number of neurons can be increased, and the solution provided by the embodiment of the present invention is still applicable.

[0100] Furthermore, the error estimation model in the embodiment of the present invention is trained on the data master station and applied to the electric energy meter. Since the complex calculations of the error estimation model lie in the training process and the calculation amount of the application process is relatively simple, the compensation method provided in the embodiment of the present invention has no obvious impact on the performance requirements of the single chip microcomputer of the electric energy meter.

[0101] Based on any of the above embodiments, Figure 3 FIG. 1 is a flow chart of a parameter updating method of an error estimation model provided by the present invention. Figure 3 As shown, the method is applied to a data master station, and the method includes: Step 310, obtaining a model error of an error estimation model, wherein the error estimation model is deployed locally at the electric energy meter, and the error estimation model is used to output a local estimation error based on basic parameters of the electric energy meter, and the local estimation error is used to perform error compensation on a metering result of the electric energy meter.

[0102] Step 320, when the model error is greater than a preset threshold, update the parameters of the error estimation model and configure the updated parameters locally in the electric energy meter.

[0103] The electric energy meter here is the electric energy meter that needs error compensation. The electric energy meter itself has the function of collecting its own parameters, so it can obtain parameters associated with the pulse error generated by itself during metering. In the embodiment of the present invention, the parameters associated with the pulse error that the electric energy meter can collect by itself are recorded as basic parameters.

[0104] For example, the load current of the electric energy meter, the temperature of the electric energy meter, etc. can be used as basic parameters. Furthermore, the temperature of the electric energy meter can be specifically refined into the temperature of the single chip computer inside the electric energy meter and the temperature of the terminal seat of the electric energy meter as the basic parameters of the electric energy meter, and the embodiment of the present invention does not specifically limit this.

[0105] In order to quickly and conveniently realize error compensation for the electric energy meter, an error estimation model can be deployed locally in the electric energy meter. Here, the error estimation model is a neural network model for error estimation. It can be understood that the error estimation model is deployed locally in the electric energy meter here, specifically storing the parameters of the error estimation model locally in the electric energy meter, so that when error compensation is required for the electric energy meter, the electric energy meter can perform error estimation on the electric energy meter based on the basic parameters of the electric energy meter obtained by the collection and the parameters of the error estimation model stored locally, thereby obtaining the local estimated error of the electric energy meter.

[0106] During specific operation, the electric energy meter can use the basic parameters of the electric energy meter as input parameters of the locally deployed error estimation model, that is, input the basic parameters of the electric energy meter into the error estimation model, and implement error estimation based on the basic parameters of the electric energy meter by running the error estimation model locally, thereby using the output result of the error estimation model as the local estimation error. Here, the local estimation error is the positive active pulse error of the electric energy meter estimated locally in the electric energy meter.

[0107] On this basis, in order to ensure the reliability and accuracy of the error estimation model itself, it is necessary to monitor the model error of the error estimation model. The model error referred to here can be understood as the difference between the local estimated error based on the output of the error estimation model and the true error. The smaller the difference between the two, the smaller the model error, and the higher the reliability and accuracy of the error estimation model itself; conversely, the larger the difference between the two, the larger the model error, and the lower the reliability and accuracy of the error estimation model itself.

[0108] Considering that the computing power of the electric energy meter itself is limited, and updating and optimizing the parameters of the error estimation model usually requires a lot of computing power support, this type of complex calculation is obviously not suitable for running on the electric energy meter. In view of this situation, in an embodiment of the present invention, the parameters of the error estimation model are updated based on the data master station, and the data master station remotely configures the updated parameters to the electric energy meter.

[0109] Here, the data master station is a device with powerful computing power, and the data master station has the function of communicating with the electric energy meter. For example, at the equipment site, a portable computer equipped with a data transmission device can be used as a data master station. The data master station can monitor the model error of the error estimation model deployed locally at the electric energy meter, and when the model error is greater than a preset threshold, the parameters of the error estimation model are updated and optimized, and then the updated parameters are remotely configured to the electric energy meter, thereby updating the parameters of the error estimation model at the electric energy meter, and ensuring the reliability and accuracy of the error estimation performed locally at the electric energy meter.

[0110] Here, the preset threshold is a pre-set threshold, and the preset threshold can be understood as the maximum value of the tolerable model error.

[0111] It can be understood that, therefore, at the electric energy meter, the updated model parameters can be stored, and the updated model parameters can be applied to achieve error estimation for the electric energy meter, and the estimated local estimation error can be applied to achieve electric energy meter compensation.

[0112] The parameter updating method of the error estimation model provided by the present invention first uses the error estimation model to perform error compensation on the electric energy meter, which can effectively ensure the reliability of the error estimation. And on this basis, when the model error of the error estimation model is greater than the preset threshold, the model parameters are remotely updated through the data master station, that is, the electric energy meter on site is used for error estimation, and the remote data master station is used for parameter update. By combining the on-site operation with the master station operation, under the premise that the computing power resources of the electric energy meter itself are limited, the reliability and accuracy of the error compensation are further guaranteed.

[0113] The electric energy meter error compensation device provided by the present invention is described below. The electric energy meter error compensation device described below and the electric energy meter error compensation method described above can be referred to each other.

[0114] Figure 4 Schematic diagram of the structure of the electric energy meter error compensation device provided by the present invention. Figure 4 As shown, the device comprises: A parameter determination unit 410, configured to determine basic parameters of the electric energy meter; An error estimation unit 420 is used to input the basic parameters into an error estimation model deployed locally at the electric energy meter to obtain a local estimated error output by the error estimation model, wherein the parameters of the error estimation model are remotely updated by a data master station, and the data master station updates the parameters of the error estimation model when the model error of the error estimation model is greater than a preset threshold; The error compensation unit 430 is used to perform error compensation on the measurement result of the electric energy meter based on the local estimation error.

[0115] The electric energy meter error compensation device provided by the present invention first adopts an error estimation model to perform error compensation on the electric energy meter, which can effectively ensure the reliability of error estimation. And on this basis, when the model error of the error estimation model is greater than a preset threshold, the model parameters are remotely updated through the data master station, that is, the electric energy meter on site is used for error estimation, and the remote data master station is used for parameter update. By combining on-site calculation with master station calculation, under the premise that the computing power resources of the electric energy meter itself are limited, the reliability and accuracy of error compensation are further guaranteed.

[0116] Based on any of the above embodiments, the basic parameters include the load current, the temperature of the single chip computer and the temperature of the terminal seat of the electric energy meter; The error estimation unit is specifically used for: The load current, the temperature of the single-chip computer and the temperature of the terminal seat of the electric energy meter are input into an error estimation model deployed locally on the electric energy meter to obtain a local estimation error output by the error estimation model.

[0117] Based on any of the above embodiments, the model error is the difference between the local estimation error and the master station estimation error; The master station estimation error is obtained by the data master station performing error estimation on the electric energy meter based on a master station estimation model. The master station estimation model is deployed at the data master station, and the scale of the master station estimation model is greater than or equal to the scale of the error estimation model.

[0118] Based on any of the above embodiments, the input parameters of the master station estimation model include the basic parameters and the operating time, power factor and grid frequency of the electric energy meter.

[0119] Based on any of the above embodiments, the power factor includes an instantaneous value of the power factor and an average value of the power factor; The grid frequency includes an instantaneous grid frequency value and an average grid frequency value.

[0120] Based on any of the above embodiments, the error estimation model is a back propagation neural network.

[0121] The parameter updating device of the error estimation model provided by the present invention is described below. The parameter updating device of the error estimation model described below and the parameter updating method of the error estimation model described above can correspond to each other.

[0122] Figure 5 Schematic diagram of the structure of the parameter updating device of the error estimation model provided by the present invention. Figure 5 As shown, the device comprises: The error acquisition unit 510 is used to acquire a model error of an error estimation model, wherein the error estimation model is deployed locally at the electric energy meter, wherein the error estimation model is used to output a local estimation error based on basic parameters of the electric energy meter, wherein the local estimation error is used to perform error compensation on a metering result of the electric energy meter; The parameter updating unit 520 is used to update the parameters of the error estimation model when the model error is greater than a preset threshold, and configure the updated parameters locally to the electric energy meter.

[0123] The parameter updating device of the error estimation model provided by the present invention firstly adopts the error estimation model to perform error compensation on the electric energy meter, which can effectively ensure the reliability of the error estimation. And on this basis, when the model error of the error estimation model is greater than the preset threshold, the model parameters are remotely updated through the data master station, that is, the electric energy meter on site is used for error estimation, and the remote data master station is used for parameter update. By combining the on-site operation with the master station operation, under the premise that the computing power resources of the electric energy meter itself are limited, the reliability and accuracy of the error compensation are further ensured.

[0124] Figure 6 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 6 As shown, the electronic device may include: a processor 610, a communication interface 620, a memory 630 and a communication bus 640, wherein the processor 610, the communication interface 620 and the memory 630 communicate with each other through the communication bus 640. The processor 610 may call the logic instructions in the memory 630 to execute the electric energy meter error compensation method, which includes: Determining basic parameters of the electric energy meter; The basic parameters are input into an error estimation model deployed locally at the electric energy meter to obtain a local estimated error output by the error estimation model, wherein the parameters of the error estimation model are remotely updated by a data master station, and the data master station updates the parameters of the error estimation model when the model error of the error estimation model is greater than a preset threshold value; Based on the local estimated error, error compensation is performed on the measurement result of the electric energy meter.

[0125] Alternatively, the processor 610 may call the logic instructions in the memory 630 to execute the parameter updating method of the error estimation model, the method comprising: Obtaining a model error of an error estimation model, where the error estimation model is deployed locally at the electric energy meter, the error estimation model is used to output a local estimation error based on basic parameters of the electric energy meter, and the local estimation error is used to perform error compensation on a metering result of the electric energy meter; When the model error is greater than a preset threshold, parameters of the error estimation model are updated, and the updated parameters are configured locally in the electric energy meter.

[0126] In addition, the logic instructions in the above-mentioned memory 630 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the relevant technology. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0127] On the other hand, the present invention further provides a computer program product, the computer program product includes a computer program, the computer program can be stored in a non-transitory computer-readable storage medium, when the computer program is executed by a processor, the computer can execute the electric energy meter error compensation method provided by the above methods, the method includes: Determining basic parameters of the electric energy meter; The basic parameters are input into an error estimation model deployed locally at the electric energy meter to obtain a local estimated error output by the error estimation model, wherein the parameters of the error estimation model are remotely updated by a data master station, and the data master station updates the parameters of the error estimation model when the model error of the error estimation model is greater than a preset threshold value; Based on the local estimated error, error compensation is performed on the measurement result of the electric energy meter.

[0128] Alternatively, when the computer program is executed by a processor, the computer can execute the parameter updating method of the error estimation model provided by the above methods, the method comprising: Obtaining a model error of an error estimation model, where the error estimation model is deployed locally at the electric energy meter, the error estimation model is used to output a local estimation error based on basic parameters of the electric energy meter, and the local estimation error is used to perform error compensation on a metering result of the electric energy meter; When the model error is greater than a preset threshold, parameters of the error estimation model are updated, and the updated parameters are configured locally in the electric energy meter.

[0129] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for compensating an electric energy meter error provided by the above methods is implemented, and the method comprises: Determining basic parameters of the electric energy meter; The basic parameters are input into an error estimation model deployed locally at the electric energy meter to obtain a local estimated error output by the error estimation model, wherein the parameters of the error estimation model are remotely updated by a data master station, and the data master station updates the parameters of the error estimation model when the model error of the error estimation model is greater than a preset threshold value; Based on the local estimated error, error compensation is performed on the measurement result of the electric energy meter.

[0130] Alternatively, when the computer program is executed by a processor, the method for updating parameters of the error estimation model provided by the above methods is implemented, and the method includes: Obtaining a model error of an error estimation model, where the error estimation model is deployed locally at the electric energy meter, the error estimation model is used to output a local estimation error based on basic parameters of the electric energy meter, and the local estimation error is used to perform error compensation on a metering result of the electric energy meter; When the model error is greater than a preset threshold, parameters of the error estimation model are updated, and the updated parameters are configured locally in the electric energy meter.

[0131] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0132] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the relevant technology can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiment.

[0133] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for compensating an electric energy meter error, characterized in that: Applied to an electric energy meter, the electric energy meter error compensation method includes: Determining basic parameters of the electric energy meter; The basic parameters are input into an error estimation model deployed locally at the electric energy meter to obtain a local estimated error output by the error estimation model, wherein the parameters of the error estimation model are remotely updated by a data master station, and the data master station updates the parameters of the error estimation model when the model error of the error estimation model is greater than a preset threshold value; Based on the local estimated error, error compensation is performed on the measurement result of the electric energy meter.

2. The method for compensating an electric energy meter error according to claim 1, characterized in that: The basic parameters include the load current, the temperature of the single chip microcomputer and the temperature of the terminal block of the electric energy meter; The step of inputting the basic parameters into an error estimation model deployed locally at the electric energy meter to obtain a local estimation error output by the error estimation model includes: The load current, the temperature of the single-chip computer and the temperature of the terminal seat of the electric energy meter are input into an error estimation model deployed locally on the electric energy meter to obtain a local estimation error output by the error estimation model.

3. The electric energy meter error compensation method according to claim 1 or 2, characterized in that: The model error is the difference between the local estimation error and the master station estimation error; The master station estimation error is obtained by the data master station performing error estimation on the electric energy meter based on a master station estimation model. The master station estimation model is deployed at the data master station, and the scale of the master station estimation model is greater than or equal to the scale of the error estimation model.

4. The method for compensating an electric energy meter error according to claim 3, characterized in that: The input parameters of the master station estimation model include the basic parameters and the operation time, power factor and grid frequency of the electric energy meter.

5. The method for compensating an electric energy meter error according to claim 4, characterized in that: The power factor includes the instantaneous value of the power factor and the average value of the power factor; The grid frequency includes an instantaneous grid frequency value and an average grid frequency value.

6. The electric energy meter error compensation method according to claim 1 or 2, characterized in that: The error estimation model is a back propagation neural network.

7. A parameter updating method for an error estimation model, characterized in that: Applied to the data master station, the parameter updating method of the error estimation model includes: Obtaining a model error of an error estimation model, where the error estimation model is deployed locally at the electric energy meter, the error estimation model is used to output a local estimation error based on basic parameters of the electric energy meter, and the local estimation error is used to perform error compensation on a metering result of the electric energy meter; When the model error is greater than a preset threshold, parameters of the error estimation model are updated, and the updated parameters are configured locally in the electric energy meter.

8. 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, it implements the electric energy meter error compensation method as described in any one of claims 1 to 6, or implements the parameter updating method of the error estimation model as described in claim 7.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the electric energy meter error compensation method as described in any one of claims 1 to 6, or implements the parameter updating method of the error estimation model as described in claim 7.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, it implements the electric energy meter error compensation method as described in any one of claims 1 to 6, or implements the parameter updating method of the error estimation model as described in claim 7.

Images