Intelligent electric energy meter data acquisition method and system based on Internet of Things

Abstract

The invention relates to the technical field of data processing, in particular to an intelligent electric energy meter data acquisition method and system based on the Internet of Things, and the method comprises the steps: obtaining a predicted electric energy increment at a current moment according to a prediction model; the real-time error is compared with an error threshold value, the running state is obtained, and the real-time error is the absolute value of the difference value between the predicted electric energy increment and the actually-measured electric energy increment; when the running state is normal, the reading of the electric energy meter is stored, and otherwise, early warning is given out. According to the technical scheme, the running state of the electric energy meter can be accurately judged, and it is guaranteed that the electric energy meter can collect accurate electricity consumption.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method and system for collecting data of a smart electric energy meter based on the Internet of Things. Background Art

[0002] The electric energy meter is mainly used to record the amount of electricity consumed by the user's electrical equipment. The amount of electricity recorded by the electric energy meter is an important basis for the power management department to charge the user for electricity. Therefore, it is necessary to collect the readings of the electric energy meter to accurately record the amount of electricity used by the user.

[0003] At present, the patent application document with publication number CN116029621A discloses a method, device, electronic device and storage medium for evaluating the operation of electric energy meters in a substation, wherein the method includes: obtaining the electricity consumption data corresponding to at least two electric energy meters in a target substation; inputting the electricity consumption data into a mathematical model, and outputting a target error coefficient corresponding to at least one target electric energy meter among at least two electric energy meters, wherein the target electric energy meter is determined based on the judgment conditions in the mathematical model, and the judgment conditions include an indicator interval corresponding to a verification indicator determined by the mathematical model under a target hit rate, and the difference between the verification hit rate corresponding to the verification indicator and the training hit rate is less than a preset threshold; respectively comparing each target error coefficient with the error threshold to determine the operation evaluation result of each target electric energy meter in the target substation.

[0004] The above method can judge the operating status of the target electric energy meter by comparing each target error coefficient with the error threshold. However, since the electricity consumption patterns of users in different periods are different, which leads to errors in the output results of the mathematical model, the above method only judges the operating status of the target electric energy meter based on a fixed error threshold, and cannot achieve accurate judgment of the operating status of the electric energy meter, resulting in the electric energy meter being unable to collect accurate electricity consumption. Summary of the invention

[0005] In order to solve the technical problem that the electric energy meter cannot collect accurate electricity consumption, the present application provides a smart electric energy meter data collection method and system based on the Internet of Things, which can accurately determine the operating status of the electric energy meter and ensure that the electric energy meter can collect accurate electricity consumption.

[0006] In a first aspect, the present application provides a method for collecting data of a smart electric energy meter based on the Internet of Things, the method comprising: obtaining a predicted electric energy increment at the current moment according to a prediction model; comparing a real-time error with an error threshold to obtain an operating state, wherein the real-time error is the absolute value of the difference between the predicted electric energy increment and the measured electric energy increment; in response to the operating state being normal, storing the indication of the electric energy meter, otherwise, issuing an early warning; the training process of the prediction model is as follows: inputting the time series of the operating parameters at any moment in the electric energy increment sample sequence into the prediction model to obtain a prediction result at the said moment; performing an initial training on the prediction model according to a first loss, wherein the first loss is a mean square error function, and the operating parameters include voltage and current; after the initial training is iterated for a first preset number of times, calculating the electric energy volatility at any moment, and performing a secondary training on the prediction model according to a second loss, until the secondary training is iterated for a second preset number of times, and the training is completed; the second loss for: , is the batch size, and For the moment The predicted results and actual increments, and For the moment The predicted results and actual increments, and For the moment and time The fluctuation of electric energy, is the adjustment coefficient; the error threshold is: , To measure the energy fluctuation of the energy increment, The number of moments in the energy increment sample sequence.

[0007] The predicted electric energy increment at the current moment is obtained according to the prediction model; the real-time error is compared with the error threshold. If the real-time error is greater than the error threshold, it means that the operation status is normal and the reading of the electric energy meter is stored. Otherwise, an early warning is issued and the electric energy meter is repaired in time. Considering the differences in electricity consumption patterns at different times, the current moment is The variance or sample entropy of the electric energy increments at multiple previous moments is taken as the electric energy volatility. The greater the electric energy volatility, the more difficult it is for the prediction model to accurately output the predicted electric energy increment. Therefore, determining the error threshold based on the electric energy volatility can accurately judge the operating status of the electric energy meter and ensure that the electric energy meter can collect accurate electricity consumption.

[0008] Preferably, the operating parameters also include power factor and power of load equipment.

[0009] The load device is directly connected to the electric energy meter, and its power can reflect the load device's ability to consume electric energy. The power factor is used to reflect the reactive power consumption of the power system. Using the power factor and the power of the load device as inputs to the prediction model can obtain accurate predictions of electric energy increments.

[0010] Preferably, the time The power fluctuation is The variance or sample entropy of the energy increments at multiple previous moments.

[0011] Variance or sample entropy can quantify the time The degree of confusion of the power increment at multiple moments before reflects the user's The rules of electricity usage.

[0012] Preferably, the first loss for: , and For the moment The predicted results and actual increments, is the batch size.

[0013] Preferably, the first loss for: , and For the moment The predicted results and actual increments, is the batch size, For the moment The fluctuation of electric energy.

[0014] The electric energy fluctuation at any moment in the electric energy increment sample sequence can reflect the electricity consumption pattern at that moment. The greater the electric energy fluctuation, the greater the change in electricity consumption between adjacent moments. At this time, it is more difficult for the prediction model to output accurate prediction results. Therefore, in order to enable the prediction model to output accurate predicted electric energy increments under any electricity consumption pattern, when calculating the first loss, weights can be assigned to the prediction errors at different moments according to the electric energy fluctuation, so that more attention can be paid to the moments with larger electric energy fluctuations during the training process of the prediction model.

[0015] Preferably, the process of obtaining the first preset number of times is: during the initial training process, obtain the first loss reduction amount of M iterations before the current iteration; if the first loss reduction amounts in the M iterations are all less than the minimum threshold, use the current iteration number as the first preset number.

[0016] When the reduction of the first loss in M ​​iterations is less than the minimum threshold, it means that the value of the first loss tends to be stable, and the training effect of the initial training has declined. Continuing the initial training will cause the prediction model to overfit. At this time, it is necessary to reduce the training difficulty and perform secondary training on the prediction model based on the second loss. Therefore, the current number of iterations is used as the first preset number, and the initial training can be stopped in time to avoid overfitting.

[0017] Preferably, the comparing the real-time error and the error threshold to obtain the operating state includes: in response to the real-time error being greater than the error threshold, the operating state is normal; otherwise, the operating state is abnormal.

[0018] In the second aspect of the present application, there is also provided a smart energy meter data collection system based on the Internet of Things, comprising a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a smart energy meter data collection method based on the Internet of Things described in the first aspect of the present application is implemented.

[0019] The technical solution of this application has the following beneficial technical effects: The predicted electric energy increment at the current moment is obtained according to the prediction model; the real-time error is compared with the error threshold. If the real-time error is greater than the error threshold, it means that the operation status is normal and the reading of the electric energy meter is stored. Otherwise, an early warning is issued and the electric energy meter is repaired in time. Considering the differences in electricity consumption patterns at different times, the current moment is The variance or sample entropy of the electric energy increments at multiple previous moments is taken as the electric energy volatility. The greater the electric energy volatility, the more difficult it is for the prediction model to accurately output the predicted electric energy increment. Therefore, determining the error threshold based on the electric energy volatility can accurately judge the operating status of the electric energy meter and ensure that the electric energy meter can collect accurate electricity consumption.

[0020] Furthermore, in order to ensure the accuracy of the predicted electric energy increment, during the training process of the prediction model, the time series of the operating parameters at any moment in the electric energy increment sample sequence is input into the prediction model to obtain the prediction result at the said moment; the prediction model is initially trained according to the mean square error function; after the first preset number of iterations of the initial training, the value of the first loss tends to be stable, indicating that the training effect of the initial training has declined, and continuing the initial training will cause the prediction model to overfit. At this time, it is necessary to reduce the training difficulty, and the prediction model is trained for the second time according to the second loss, so that the prediction model can better achieve the expected training effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present application will become easy to understand. In the accompanying drawings, several embodiments of the present application are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein: Figure 1 It is a flow chart of a smart electric energy meter data collection method based on the Internet of Things according to an embodiment of the present application; Figure 2 is a flowchart of a prediction model training process according to an embodiment of the present application; Figure 3 It is a structural block diagram of a smart electricity meter data acquisition system based on the Internet of Things according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0023] It should be understood that when the terms "first", "second", etc. are used in the claims, specification and drawings of the present application, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the specification and claims of the present application indicate the presence of the described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or their collections.

[0024] According to the first aspect of the present application, the present application provides a smart electricity meter data collection method based on the Internet of Things, which is used to determine whether there is an error in the indication of the smart electricity meter. If there is no error in the indication of the smart electricity meter, the indication of the electricity meter is stored to realize data collection of the smart electricity meter; if there is an error in the indication of the smart electricity meter, an early warning is issued and the electricity meter is inspected and repaired in time to ensure that the smart electricity meter can collect accurate electricity consumption.

[0025] Figure 1 FIG. 1 is a flow chart of a method for collecting data of a smart electric energy meter based on the Internet of Things according to an embodiment of the present application. Figure 1 As shown, the smart energy meter data collection method based on the Internet of Things includes steps S101 to S103, which are described in detail below.

[0026] S101, obtaining the predicted electric energy increment at the current moment according to the prediction model.

[0027] In one embodiment, the electric energy meter corresponds to an indication at each moment, and the indication at the current moment is subtracted from the indication at the previous adjacent moment to obtain the measured electric energy increment at the current moment. The predicted electric energy increment at the current moment is obtained through the prediction model, and the predicted electric energy increment is the electric energy increment of the electric energy meter under normal conditions. By comparing the predicted electric energy increment with the measured electric energy increment, it can be determined whether the operating state is normal or abnormal.

[0028] Among them, the prediction model is a recurrent neural network, the input of the prediction model is the time series of operating parameters at any time, and the output is the predicted electric energy increment at the time, the operating parameters include voltage and current; the prediction model extracts the time series features of each operating parameter to obtain the time series features of each operating parameter, and transforms the dimension of the multidimensional data composed of the time series features of each operating parameter, and outputs the predicted electric energy increment.

[0029] It can be understood that, taking the operating parameter voltage as an example, at the current moment The voltage time series includes the current moment Voltage values ​​at multiple previous moments. In other embodiments, the operating parameters also include power factor and power of the load device.

[0030] In one embodiment, in order for the prediction model to accurately output the predicted electric energy increment at the current moment, the prediction model needs to be trained. Figure 2 : is a flowchart of a prediction model training process according to an embodiment of the present application, wherein the prediction model training process includes: S201, inputting the time series of the operating parameters at any moment in the electric energy increment sample sequence into a prediction model to obtain a prediction result at the said moment.

[0031] The electric energy increment sample sequence includes the electric energy increments at multiple moments collected when the smart electric energy meter is in normal operation, and the electric energy increments in the electric energy increment sample sequence are the real increments at the corresponding moments. The time series of each operating parameter can be collected at any moment in the electric energy increment sample sequence.

[0032] By inputting the time series of the operating parameters at any moment in the electric energy increment sample sequence into the prediction model, the prediction result at that moment can be output.

[0033] S202, performing initial training on the prediction model according to a first loss, where the first loss is a mean square error function.

[0034] The first loss is used to constrain the prediction result at this moment to be equal to the actual increment at this moment in the electric energy increment sample sequence. for: , and For the moment The predicted results and actual increments, is the batch size.

[0035] Understandably, For the moment The difference between the predicted result and the actual increment, used to reflect the moment prediction error.

[0036] The batch size is set to 6. That is, after the prediction model outputs the prediction results for 6 moments, the first loss can be calculated and the prediction model can be updated using the gradient descent method.

[0037] In another embodiment, the power fluctuation at any moment in the power increment sample sequence can reflect the power consumption pattern at that moment. The greater the power fluctuation, the greater the change in power consumption between adjacent moments. At this time, the more difficult it is for the prediction model to output accurate prediction results. Therefore, in order to enable the prediction model to output accurate predicted power increments under any power consumption pattern, when calculating the first loss, weights can be assigned to the prediction errors at different moments according to the power fluctuation, so that more attention is paid to moments with greater power fluctuations during the training process of the prediction model. Specifically, the first loss for: , and For the moment The predicted results and actual increments, is the batch size, For the moment The fluctuation of electric energy.

[0038] Among them, the time The power fluctuation is The variance or sample entropy of the energy increments at multiple previous moments.

[0039] In this way, the prediction model is initially trained using the first loss, so that the prediction result approaches the true increment, even if the error between the prediction result and the true increment approaches 0.

[0040] S203, after the initial training is iterated for a first preset number of times, the power fluctuation at any time is calculated, and the prediction model is trained for a second time according to the second loss, until the second training is iterated for a second preset number of times, and the training is completed.

[0041] Among them, the purpose of the initial iteration is to make the error between the predicted result and the actual increment approach 0. The first preset number of times is 50. After the first preset number of initial training iterations, the value of the first loss will tend to be stable. At this time, the initial training is stopped.

[0042] In one embodiment, the process of obtaining the first preset number of times is: during the initial training process, obtain the first loss reduction amount of M iterations before the current iteration; if the first loss reduction amounts in the M iterations are all less than the minimum threshold, use the current iteration number as the first preset number.

[0043] Among them, the value of M is 3, and the minimum threshold is 0.01; the initial training has high requirements for the prediction model, strictly constraining the error between the prediction result and the actual increment to approach 0, and the training difficulty is relatively high; when the reduction of the first loss in M ​​iterations is less than the minimum threshold, it means that the value of the first loss tends to be stable, and the training effect of the initial training decreases. Continuing the initial training will cause the prediction model to overfit. At this time, it is necessary to reduce the training difficulty and perform secondary training on the prediction model based on the second loss; that is to say, the training difficulty of the secondary training is less than that of the initial training, so that the prediction model can better achieve the expected training effect.

[0044] Specifically, the second training of the prediction model according to the second loss includes: obtaining the prediction results of the prediction model at N moments and the power fluctuation at the N moments; the second loss for: , is the batch size, and For the moment The predicted results and actual increments, and For the moment The predicted results and actual increments, and For the moment and time The fluctuation of electric energy, The prediction model is updated by using the gradient descent method, and the prediction model is trained twice until the second training iteration reaches a second preset number of times and the training is completed. The second preset number of times is 100.

[0045] in, It is used to constrain the error between the predicted result and the actual increment to approach 0. For the moment Error Power fluctuation The ratio between is used to reflect the time Correlation between errors and power fluctuations; It is used to constrain the correlation between errors and power fluctuations at different times to be stable, and the adjustment coefficient The value of is 0.5.

[0046] It is understandable that the greater the power fluctuation, the more difficult it is to obtain accurate prediction results. Therefore, errors in the prediction results of the prediction model are allowed to exist during the secondary training process, and the correlation between the prediction results and the power fluctuation is constrained, so that the ratio of the error of the prediction results to the power fluctuation tends to a stable value.

[0047] In this way, the training process of the prediction model is completed, and a trained prediction model is obtained; the time series of each operating parameter at the current moment is input into the prediction model to obtain the predicted electric energy increment at the current moment.

[0048] S102, comparing the real-time error with the error threshold to obtain the operating status, where the real-time error is the absolute value of the difference between the predicted electric energy increment and the actually measured electric energy increment.

[0049] In one embodiment, the error threshold is 0.5. The comparing the real-time error and the error threshold to obtain the operation state includes: in response to the real-time error being greater than the error threshold, the operation state is normal; otherwise, the operation state is abnormal.

[0050] The error threshold is positively correlated with the power fluctuation. The error threshold is: , To measure the energy fluctuation of the energy increment, is the number of moments in the energy increment sample sequence, and For the moment The predicted results and actual increments, For the moment The fluctuation of electric energy.

[0051] S103, in response to the operating status being normal, storing the reading of the electric energy meter, otherwise, issuing an early warning.

[0052] In one embodiment, when the operating status is normal, it means that the indication of the electric energy meter can accurately reflect the user's electricity consumption, and the indication of the electric energy meter is stored. Otherwise, it means that the indication of the electric energy meter cannot accurately reflect the user's electricity consumption, an early warning is issued, and the electric energy meter is repaired in time.

[0053] The storing of the indication of the electric energy meter comprises: compressing and storing the indication of the electric energy meter by using a preset compression algorithm, wherein the preset compression algorithm is revolving door compression, Huffman coding or run-length coding.

[0054] The above specific embodiments introduce the technical principle and implementation details of the data collection method of smart energy meters based on the Internet of Things. The predicted energy increment at the current moment is obtained based on the prediction model; the real-time error and the error threshold are compared. If the real-time error is greater than the error threshold, it means that the operation status is normal and the reading of the energy meter is stored. Otherwise, an early warning is issued and the energy meter is repaired in time; considering the differences in electricity consumption patterns at different times, the current moment is recorded. The variance or sample entropy of the electric energy increments at multiple previous moments is taken as the electric energy volatility. The greater the electric energy volatility, the more difficult it is for the prediction model to accurately output the predicted electric energy increment. Therefore, determining the error threshold based on the electric energy volatility can accurately judge the operating status of the electric energy meter and ensure that the electric energy meter can collect accurate electricity consumption.

[0055] Furthermore, in order to ensure the accuracy of the predicted electric energy increment, during the training process of the prediction model, the time series of the operating parameters at any moment in the electric energy increment sample sequence is input into the prediction model to obtain the prediction result at the said moment; the prediction model is initially trained according to the mean square error function; after the first preset number of iterations of the initial training, the value of the first loss tends to be stable, indicating that the training effect of the initial training has declined, and continuing the initial training will cause the prediction model to overfit. At this time, it is necessary to reduce the training difficulty, and the prediction model is trained for the second time according to the second loss, so that the prediction model can better achieve the expected training effect.

[0056] According to the second aspect of the present application, the present application also provides a smart electricity meter data acquisition system based on the Internet of Things. Figure 3 1 is a block diagram of a smart electric energy meter data acquisition system based on the Internet of Things according to an embodiment of the present application. Figure 3 As shown, the system 50 includes a processor and a memory, the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a smart electric energy meter data acquisition method based on the Internet of Things according to the first aspect of the present application is implemented. The system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, whose settings and functions are known in the art, so they are not described here.

[0057] In the present application, the aforementioned memory may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. For example, a computer-readable storage medium may be any suitable magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module, or both. Any such computer storage medium may be part of a device or accessible or connectable to a device. Any application or module described in this application may be implemented using computer-readable / executable instructions that may be stored or otherwise maintained by such a computer-readable medium.

[0058] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0059] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the patent application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent application shall be subject to the attached claims.

Claims

1. A smart energy meter data collection method based on the Internet of Things, characterized in that: The method comprises: The predicted electric energy increment at the current moment is obtained according to the prediction model; the real-time error is compared with the error threshold to obtain the operating status, and the real-time error is the absolute value of the difference between the predicted electric energy increment and the measured electric energy increment; In response to the operating state being normal, the reading of the electric energy meter is stored, otherwise, an early warning is issued; The training process of the prediction model is as follows: inputting the time series of the operating parameters at any time in the electric energy increment sample sequence into the prediction model to obtain the prediction result at the said time; performing initial training on the prediction model according to the first loss, wherein the first loss is a mean square error function, and the operating parameters include voltage and current; after the initial training is iterated for a first preset number of times, calculating the electric energy volatility at any time, and performing secondary training on the prediction model according to the second loss, until the secondary training is iterated for a second preset number of times and the training is completed; the second loss for: , is the batch size, and For the moment The predicted results and actual increments, and For the moment The predicted results and actual increments, and For the moment and time The fluctuation of electric energy, is the adjustment coefficient; The error threshold is: , To measure the energy fluctuation of the energy increment, The number of moments in the energy increment sample sequence.

2. The method for collecting data of a smart electric energy meter based on the Internet of Things according to claim 1 is characterized in that: The operating parameters also include power factor and power of the load device.

3. The method for collecting data of a smart electric energy meter based on the Internet of Things according to claim 1 is characterized in that: time The power fluctuation is The variance or sample entropy of the energy increments at multiple previous moments.

4. The method for collecting data of a smart electric energy meter based on the Internet of Things according to claim 1 is characterized in that: The first loss for: , and For the moment The predicted results and actual increments, is the batch size.

5. The method for collecting data of a smart electric energy meter based on the Internet of Things according to claim 1 is characterized in that: The first loss for: , and For the moment The predicted results and actual increments, is the batch size, For the moment The fluctuation of electric energy.

6. The method for collecting data of a smart electric energy meter based on the Internet of Things according to claim 1 is characterized in that: The process of obtaining the first preset number of times is as follows: During the initial training process, the first loss reduction of M iterations before the current iteration is obtained; If the first loss reduction amount in M ​​iterations is less than the minimum threshold, the current number of iterations is used as the first preset number.

7. The method for collecting data of a smart electric energy meter based on the Internet of Things according to claim 1 is characterized in that: The comparing the real-time error and the error threshold to obtain the operating status includes: In response to the real-time error being greater than the error threshold, the operating state is normal; otherwise, the operating state is abnormal.

8. A smart electric energy meter data acquisition system based on the Internet of Things, characterized in that: It comprises a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a smart electric energy meter data collection method based on the Internet of Things according to any one of claims 1 to 7 is implemented.

Images