Method and System for Detecting the State of an Electric Energy Meter Communication Module Based on Data Analysis

Abstract

The present invention relates to the field of data processing. The present invention relates to a method and system for detecting the state of an electric energy meter communication module based on data analysis. The method includes: collecting historical data packets of the electric energy meter communication module with various lengths, and environmental data corresponding to each historical data packet at the corresponding historical moment; there are multiple historical data packets of each length; for each length of historical data packet, obtain the degree of interference of the environment on the determination of the communication module state for each type of communication data, and the importance of various communication data for determining the communication module state; then construct a corresponding classification model, train it to obtain a state detection model, and collect in real time the communication data packets sent by the electric energy meter communication module, and obtain the state of the electric energy meter communication module by combining the state detection model corresponding to the historical data packet with the closest length. Using the method of the present invention can greatly improve the accuracy of detecting the state of the communication module.

Description

Technical Field

[0001] The present invention relates to the field of data processing. More specifically, the present invention relates to a method and system for detecting the state of an electricity meter communication module based on data analysis. Background Art

[0002] With the rapid development of the economy, the demand for electricity in all walks of life is increasing, and the role of electricity meters is becoming more and more important. An electricity meter is an instrument used to measure electric energy, also known as a watt-hour meter, a kilowatt-hour meter, or a meter for measuring various electrical quantities. When using an electricity meter, it should be noted that in the case of low voltage (below 500 volts) and small current (tens of amperes), the electricity meter can be directly connected to the circuit for measurement. In the case of high voltage or large current, the electricity meter cannot be directly connected to the line and needs to be used in conjunction with a voltage transformer or a current transformer.

[0003] An electricity meter generally includes a power module, a metering module, a display module, a communication module, a security module, a clock module, a storage module, and a power on / off module. Among them, the communication module is used for data transmission and communicates with a host or other devices.

[0004] After the electricity meter collects the electricity consumption data of the power grid, it will transmit the data to the monitoring center or the power supply company through its communication module. If the communication module of the electricity meter fails, the collected data cannot be transmitted in time. Therefore, the fault detection of the communication module of the electricity meter is crucial. The existing detection methods mainly rely on simple signal transmission states (such as whether the connection is successful, whether the data is sent successfully, etc.) to detect the state of the electricity meter communication module. Since this method does not consider the interference of environmental factors on the state detection of the electricity meter communication module, misjudgment may occur when detecting the state of the electricity meter communication module, resulting in poor accuracy of the state detection results. Summary of the Invention

[0005] To solve the technical problem of poor accuracy of the detection results of the existing electricity meter communication module detection methods, the present invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides a method for detecting the state of an electricity meter communication module based on data analysis, including: collecting historical data packets of the electricity meter communication module with various lengths, and environmental data corresponding to each historical data packet at the corresponding historical moment; there are multiple historical data packets of each length;

[0007] For each length of historical data packet, obtain the degree of environmental interference when each communication data is used to judge the status of the communication module, and the importance of various communication data for judging the status of the communication module; construct a classification model and train it using a loss function to obtain a status detection model. The calculation expression of the loss function used for training the classification model corresponding to the i-th length data packet is:

[0008]

[0009] In the formula, is the classification loss function of the n-th communication data in the classification model corresponding to the -th length data packet; is the label of the status of the communication module corresponding to the -th communication data; is the predicted probability of the n-th communication data output by the classification model; ; is the degree of environmental interference when the m-th communication data corresponding to the -th length data packet is used to judge the status of the communication module; is the importance of the -th communication data corresponding to the -th length data packet for judging the status of the communication module;

[0010] Collect communication data packets sent by the power meter communication module in real time, and obtain the status of the power meter communication module by combining the status detection model corresponding to the historical data packet with the closest length. Assume that there are two types of lengths of historical data packets closest to the length of the communication data packet collected in real time, then select the status detection model corresponding to the historical data packet with a length smaller than the length of the communication data packet collected in real time to detect the power meter communication module.

[0011] The beneficial effects are as follows: When constructing the classification model, the method for detecting the status of the power meter communication module of the present invention takes into account that when using data packets to detect the communication module, data packets of different lengths are affected by the environment to different degrees. Different classification models (i.e., clustering modeling) are constructed for historical data packets of different lengths. After collecting real-time communication data packets, select the classification model of the historical data packet with the closest length to detect the communication module, which greatly improves the accuracy of communication module status detection.

[0012] In addition, during the process of training the classification model, considering that different types of communication data may have different degrees of importance for judging the status of the communication module, and also considering the interference of environmental data on the judgment of the communication module status by communication data, the loss function is corrected based on the degree of interference of environmental data and the degree of importance of various communication data for judging the status of the communication module, thereby avoiding the limitations of a single data source, and also enabling the classification model to suppress real-time environmental interference (such as sudden temperature changes), and improving the classification accuracy of the model in complex environments.

[0013] Preferably, the calculation method of the degree of interference of the m-th type of communication data corresponding to the -th length data packet on the judgment of the status of the communication module includes: calculating the degree of interference of temperature, the degree of interference of humidity, and the degree of interference of voltage fluctuation on the judgment of the status of the communication module by the m-th type of communication data corresponding to the -th length data packet;

[0014] Performing weighted summation on the degree of interference of temperature, the degree of interference of humidity, and the degree of interference of voltage fluctuation, so as to obtain the degree of interference of the m-th type of communication data corresponding to the -th length data packet on the judgment of the status of the communication module by the environment.

[0015] Its beneficial effect is that when calculating the degree of interference of environmental data on the judgment of the status of the communication module of the electricity meter by communication data, the degree of interference of temperature, humidity, and voltage fluctuation on the judgment of the status of the communication module of the electricity meter is comprehensively considered, so that the calculation result of the degree of interference of environmental data on the judgment of the status of the communication module of the electricity meter by communication data is more accurate.

[0016] Preferably, the calculation expression of the degree of interference of the m-th type of communication data corresponding to the -th length data packet on the judgment of the status of the communication module by the environment is:

[0017] ;

[0018] In the formula, is the information carrying capacity of the -th type of environmental data sequence; are respectively the information carrying capacities of the temperature sequence, humidity sequence, and voltage fluctuation sequence under the -th length data packet, , and respectively represent the degree of interference of temperature, the degree of interference of humidity, and the degree of interference of voltage fluctuation on the judgment of the status of the communication module by the m-th type of communication data corresponding to the -th length data packet.

[0019] The beneficial effects are as follows: When calculating the interference degree of environmental data on the communication data for judging the state of the communication module of the electric energy meter, the relationship between the information carrying capacity of the environmental data and the interference degree on the state of the communication module is further considered. When performing weighted summation on the interference degrees corresponding to various environmental data, a larger weight is assigned to the environmental data with a larger information carrying capacity, and a smaller weight is assigned to the environmental data with a smaller information carrying capacity, further improving the accuracy of the interference degree of environmental data on the communication data for judging the state of the communication module of the electric energy meter.

[0020] Preferably, the calculation expression for the information carrying capacity of the nth type of environmental data is:

[0021] ;

[0022] In the formula, is the standard deviation of the nth type of environmental data sequence corresponding to the kth type of length data packet, is the Pearson correlation coefficient between the nth type of environmental data sequence and the nth type of environmental data sequence corresponding to the kth type of length data packet, respectively represent the temperature sequence, humidity sequence, and voltage fluctuation sequence.

[0023] Preferably, the calculation expression for the interference degree of the nth type of length data packet corresponding to the mth type of communication data on judging the state of the communication module by temperature is:

[0024] ;

[0025] In the formula, and are respectively the maximum value and minimum value in the mth type of communication data sequence corresponding to the nth type of length data packet, and are respectively the maximum value and minimum value of the temperature in the temperature sequence under the nth type of length data packet, is the hyperbolic tangent function, is the absolute value of the Pearson correlation coefficient between the temperature sequence and the

[0026] Preferably, the nth type of length data packet corresponding to the The importance level of the communication data and the information gain of the m-th type of communication data corresponding to the i-th length data packet for judging the status of the electric energy meter communication module are positively correlated, and the calculation expression of the information gain is as follows:

[0027] ;

[0028] In the formula, is the information entropy of the status y of the electric energy meter communication module under the i-th length data packet; is the conditional entropy of the status y of the electric energy meter communication module after knowing the m-th type of communication data under the i-th length data packet.

[0029] Preferably, the communication data includes communication time, signal strength, and packet loss rate.

[0030] Preferably, it further includes: performing linear interpolation processing on the collected historical data packets and environmental data, and denoising them using wavelet transform.

[0031] Preferably, the classification model is a neural network model or a machine learning model.

[0032] In a second aspect, the present invention further provides a system for detecting the status of an electric energy meter communication module based on data analysis, including a processor and a memory. The memory stores computer program instructions for detecting the status of the electric energy meter communication module. When the computer program instructions are executed by the processor, the method for detecting the status of the electric energy meter communication module based on data analysis of the present invention is implemented.

[0033] In summary, the beneficial effects of the present invention are as follows: By using the method of the present invention, the accuracy of the detection result of the status of the electric energy meter communication module can be greatly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is a flowchart schematically showing a method for detecting the status of an electric energy meter communication module based on data analysis according to an embodiment of the present invention;

[0035] Figure 2 is a schematic structural diagram showing a system for detecting the status of an electric energy meter communication module based on data analysis according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0036] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0037] The specific embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0038] Embodiment of the method for detecting the state of the communication module of the electric energy meter based on data analysis:

[0039] As Figure 1 shown, the method for detecting the state of the communication module of the electric energy meter based on data analysis of the present invention includes:

[0040] S101. Collect historical data packets of the communication module of the electric energy meter with various lengths and corresponding environmental data. Specifically: collect historical data packets of the communication module of the electric energy meter with various lengths, and environmental data at the historical moments corresponding to each historical data packet; there are multiple historical data packets of each length.

[0041] After collecting the historical data packets, it is necessary to obtain the state of the communication module of the electric energy meter corresponding to the historical data packet, so as to label the historical data packet. The states of the communication module include two states: normal and abnormal. The label for the normal state is 0, and the label for the abnormal state is 1. Indicates that the state of the communication module of the electric energy meter is normal, Indicates that the state of the communication module of the electric energy meter is abnormal.

[0042] The environmental data at the historical moment corresponding to the historical data packet refers to the environmental data collected at the historical moment corresponding to the historical data packet. The state labels of the communication module of the electric energy meter corresponding to each historical data packet are collected while collecting the historical data packets.

[0043] The historical data packets of the communication module of the electric energy meter can be collected through the sensor network and the communication interface device.

[0044] S102. Obtain the interference degree of the environmental data on the judgment of the communication module state and the importance degree of the communication data on the judgment of the communication module state. Specifically: for each length of historical data packet, obtain the interference degree of the environment on the judgment of the communication module state by each communication data, and the importance degree of various communication data for judging the communication module state;

[0045] In this embodiment, judging the state of the communication module by a certain kind of communication data means judging the state of the communication module based on this kind of communication data.

[0046] For the communication data and environmental data of historical data packets with the same length, calculate the interference degree of the environmental data on the judgment of the communication module state by the communication data and the importance degree of the communication data for judging the communication module state, providing a basis for subsequent sub-scenario modeling and avoiding the insufficient adaptability of the global model to different data packet lengths.

[0047] S103. Obtain the state detection models corresponding to historical data packets of various lengths. Specifically, for historical data packets of each length, construct a classification model respectively, and train it using a loss function to obtain the state detection model. The calculation expression of the loss function used for training the classification model corresponding to the data packet of the i-th length is:

[0048]

[0049] In the formula, is the classification loss function of the n-th communication data in the classification model corresponding to the data packet of the -th length; is the label of the communication module status corresponding to the -th communication data; is the predicted probability of the n-th communication data output by the classification model; ; is the degree of interference of the m-th communication data corresponding to the data packet of the -th length on judging the communication module status by the environment; is the importance degree of the -th communication data corresponding to the data packet of the -th length for judging the communication module status.

[0050] Under the condition that the label of the communication module status corresponding to the -th communication data is 1 (i.e., the communication module is abnormal), , the larger they are, the smaller it is, that is, the greater the difficulty of classifying this communication data, the greater the penalty factor of the loss function for judging this communication data, and the greater the corresponding loss function value; conversely, , the smaller they are, the larger it is, that is, the smaller the difficulty of classifying this communication data, the smaller the penalty factor of the loss function for judging this communication data, and the smaller the corresponding loss function value.

[0051] Similarly, under the condition that the label of the communication module status corresponding to the -th communication data is 0 (i.e., the communication module is normal), , the larger they are, the larger it is, that is, the greater the penalty factor of the loss function when the communication data is judged incorrectly, and the greater the corresponding loss function value; conversely, , the smaller they are, the smaller it is, that is, the smaller the penalty factor of the loss function when the communication data is judged incorrectly, and the smaller the corresponding loss function value.

[0052] Therefore, when calculating the loss function using the loss function calculation expression of this embodiment, it can ensure that a smaller loss function value is set when the difficulty of classifying communication data is relatively large, and a larger loss function value is set when the difficulty of classifying communication data is relatively small, thereby making the training effect of the classification model better.

[0053] In this embodiment, when calculating the loss function, the interference degree of the environmental data under the i-th length data packet on the communication data for judging the state of the electric energy meter communication module and the importance degree of the communication data for judging the state of the electric energy meter communication module are introduced to correct the loss function of the classification model. Compared with only relying on the original binary cross-entropy loss function, the loss function can more accurately and effectively judge whether the state of the electric energy meter communication module is abnormal during the training process. The method for calculating the loss function in this embodiment can more accurately distinguish the credibility of each communication data under different length data packets, provide more refined discriminant information, thereby improving the model's ability to distinguish different states (normal and abnormal) of the electric energy meter communication module, and finally making the loss function more discriminative and capable of better optimizing the classification performance during training, especially when the electric energy meter communication module is greatly interfered by environmental characteristics.

[0054] In this embodiment, the communication data includes communication time, signal strength, and packet loss rate. In other embodiments, the communication data may also include other types of data.

[0055] In this embodiment, the classification model is a neural network model or a machine learning model. In other embodiments, the classification model may also be other suitable classification models.

[0056] If the classification model is a neural network model, the training process of the classification model corresponding to the i-th length historical data packet is as follows:

[0057] A. Obtain the working state of the electric energy meter communication module corresponding to each collected i-th length historical data packet, and then label each i-th length historical data packet to obtain a training set;

[0058] For a certain historical data packet, if the working state of the corresponding electric energy meter communication module is normal, the label of this historical data packet is 0; otherwise, the label of this historical data packet is 1.

[0059] B. Select the network architecture of the classification model and initialize the parameters;

[0060] Convolutional Neural Network (CNN) or Recurrent Neural Network (RNN) can be selected.

[0061] Initializing the parameters includes initializing the weights and biases in the network, and the initialization method can adopt the random initialization method.

[0062] C. Forward Propagation: Compute the input data through the network to obtain the predicted output.

[0063] D. Calculate Loss: Calculate the value of the loss function based on the predicted output and the true label.

[0064] E. Backward Propagation: Calculate the gradient through the chain rule and update the network parameters to reduce the loss.

[0065] F. Parameter Update: Use the optimizer to update the network parameters according to the calculated gradient.

[0066] G. Set Hyperparameters: Determine hyperparameters such as the learning rate, batch size, and number of training epochs.

[0067] H. Iterative Training: Perform multiple iterations on the training set, and each time use batch data for training.

[0068] S104. Real-time collect the communication data packets sent by the electric energy meter communication module, and obtain the status of the electric energy meter communication module by combining with the status detection model corresponding to the historical data packet with the closest length; assume that there are two types of lengths of the historical data packets closest to the length of the real-time collected communication data packet, then select the status detection model corresponding to the historical data packet with a length smaller than the length of the real-time collected communication data packet to detect the electric energy meter communication module.

[0069] For example: Assume that the length of the communication data packet sent by the electric energy meter communication module collected in real time is 5, and the historical data packets collected in step S101 include historical data packets with lengths of 4, 5, and 6. Then select the status detection model corresponding to the historical data packet with a length of 5 as the current status detection model, and input the communication data packet sent by the electric energy meter communication module collected in real time into the current status detection model to obtain the status of the electric energy meter communication module.

[0070] Assume that there are two types of lengths of the historical data packets closest to the length of the real-time collected communication data packet, then select the status detection model corresponding to the historical data packet with a length smaller than the length of the real-time collected communication data packet to detect the electric energy meter communication module.

[0071] For example: Assume that the length of the communication data packet sent by the real-time electricity meter communication module collected is 3, and the historical data packets collected in step S1 include historical data packets with lengths of 2 and 4. Then, the classification model corresponding to the communication data packet with a length of 2 is preferentially selected to detect the state of the electricity meter communication module. Since the length of the real-time data packet is 3, which is closer to the data packet with a length of 2 in terms of the degree of environmental interference, the model may be more likely to adapt to this slight change, thus maintaining a better classification effect. While the data packet with a length of 4 may have a relatively large difference in the degree of environmental interference from the real-time data packet, and the model may require more adjustments to handle this change, resulting in unstable classification results.

[0072] Since when using data packets to detect the communication module, data packets of different lengths are affected by the environment to different degrees. If the same classification model is used to detect data packets of different lengths, the accuracy of the detection results will be poor. In the method of this embodiment, when constructing the classification model, different classification models are constructed for historical data packets of different lengths (i.e., clustering modeling). After collecting real-time communication data packets, the classification model corresponding to the historical data packet with the length closest to the real-time communication data packet is selected to detect the state of the electricity meter communication module (i.e., select the most suitable classification model), thereby greatly improving the accuracy of the state detection of the electricity meter communication module.

[0073] In addition, when training the classification model, the importance of different types of communication data for judging the state of the communication module is comprehensively considered, avoiding the limitations of a single data source. Moreover, by correcting the loss function based on the degree of interference of environmental data and the importance of communication data for judging the state of the communication module, real-time environmental interference (such as sudden temperature changes) can be suppressed, and the classification accuracy of the model in a complex environment can be improved.

[0074] In one embodiment, the calculation method for the degree of interference of the m-th type of communication data corresponding to the data packet of the -th length on the state of the communication module includes:

[0075] S201. Calculate the degree of interference of temperature, the degree of interference of humidity, and the degree of interference of voltage fluctuation of the m-th type of communication data corresponding to the data packet of the -th length on the state of the communication module;

[0076] S202. Perform weighted summation on the degree of interference of temperature, the degree of interference of humidity, and the degree of interference of voltage fluctuation, so as to obtain the degree of interference of the m-th type of communication data corresponding to the data packet of the -th length on the state of the communication module by the environment.

[0077] In this embodiment, when calculating the interference degree of environmental data on the judgment of the communication module status of the electricity meter by communication data, instead of considering the interference degree of a single environmental factor on the judgment of the communication module status by communication data, the interference degrees of temperature on the judgment of the communication module status by communication data, humidity on the judgment of the communication module status by communication data, and voltage fluctuation on the judgment of the communication module status by communication data are comprehensively considered, so that the calculation result of the interference degree of environmental data on the judgment of the communication module status by communication data is more accurate.

[0078] In one embodiment, the calculation expression for the interference degree of the m-th type of communication data corresponding to the n-th type of length data packet on the judgment of the communication module status by the environment is: ;

[0079] ;

[0080] In the formula, is the information carrying capacity of the k-th type of environmental data sequence; are the information carrying capacities of the temperature sequence, humidity sequence, and voltage fluctuation sequence under the n-th type of length data packet respectively, respectively represent the interference degree of temperature, the interference degree of humidity, and the interference degree of voltage fluctuation on the judgment of the communication module status by the m-th type of communication data corresponding to the n-th type of length data packet. ; , and respectively represent the interference degree of temperature, the interference degree of humidity, and the interference degree of voltage fluctuation on the judgment of the communication module status by the m-th type of communication data corresponding to the n-th type of length data packet. ;

[0081] In this embodiment, when calculating the interference degree of environmental data on the judgment of the m-th type of communication data of the electricity meter communication module status, not only the interference degrees of various environmental data on the judgment of the electricity meter communication module status by communication data are considered, but also the information carrying capacities of various environmental data are further considered. When the environmental data with a larger information carrying capacity changes, the interference degree on the communication module status is larger. When performing weighted summation on the interference degrees corresponding to various environmental data, the weight corresponding to the environmental data with a larger information carrying capacity is made larger, and the weight corresponding to the environmental data with a smaller information carrying capacity is made smaller, thereby further improving the interference degree of environmental data on the judgment of the m-th type of communication data of the electricity meter communication module status.

[0082] In one embodiment, the calculation expression for the information carrying capacity of the k-th type of environmental data is: ;

[0083] ;

[0084] In the formula, is the The standard deviation of the th environmental data sequence corresponding to the th length data packet, and the Pearson correlation coefficient between the th environmental data sequence and the th environmental data sequence corresponding to the th length data packet, respectively represent the temperature sequence, humidity sequence, and voltage fluctuation sequence.

[0085] The larger the standard deviation of the th environmental data sequence, the greater the data fluctuation degree of the th environmental data, and the greater the information carrying capacity of the th environmental data; conversely, the smaller the standard deviation of the th environmental data sequence, the smaller the data fluctuation degree of the th environmental data, and the smaller the information carrying capacity of the th environmental data. The larger the Pearson correlation coefficient between the th environmental data sequence and the th environmental data sequence, the stronger the correlation between the th environmental data sequence and the th environmental data sequence, and the smaller the conflict between the th environmental data and the th environmental data (i.e., is smaller), indicating more identical information, and the smaller the information carrying capacity of the th environmental data; conversely, the smaller the Pearson correlation coefficient between the th environmental data sequence and the th environmental data sequence, the weaker the correlation between the

[0086] In one embodiment, the calculation expression for the degree of interference of temperature on the judgment of the communication module state by the th communication data corresponding to the

[0087] th length data packet is:

[0088] In the formula, and are respectively the maximum and minimum values in the th communication data sequence corresponding to the th length data packet, and are respectively the maximum and minimum values of the temperature in the temperature sequence under the -th length data packet. is the hyperbolic tangent function. is the absolute value of the Pearson correlation coefficient between the temperature sequence under the -th length data packet and the

[0089] is the ratio of the change amount of the -th communication data to the temperature change range. The larger this value is, the greater the average fluctuation amplitude of the -th communication data when the temperature changes by 1°C under the -th length data packet, and the greater the interference degree of the temperature on the judgment of the communication module status of the electricity meter by the

[0090] -th communication data. The smaller this value is, the smaller the average fluctuation amplitude of the -th communication data when the temperature changes by 1°C under the -th length data packet, and the smaller the interference degree of the temperature on the judgment of the communication module status of the electricity meter by the -th communication data.

[0091] The larger the is, the more relevant the temperature sequence and the -th communication data sequence are under the -th length data packet, and the greater the interference degree of the temperature on the judgment of the communication module status of the electricity meter by the

[0092] -th communication data. The smaller this value is, the less relevant the temperature sequence and the -th communication data sequence are under the -th length data packet, and the smaller the interference degree of the temperature on the judgment of the communication module status of the electricity meter by the -th communication data. In this embodiment, when calculating the interference degree of the temperature on the judgment of the communication module status of the electricity meter by the -th communication data, the relationship between the ratio of the change amount of the -th communication data to the temperature change range and the interference degree corresponding to the temperature, as well as the relationship between the correlation between the temperature sequence and the -th communication data sequence and the interference degree corresponding to the temperature, are considered, so that the calculated interference degree of the temperature on the judgment of the communication module status of the electricity meter by the -th communication data is more accurate.

[0090] The interference degree of the humidity on the judgment of the communication module status of the electricity meter by the -th communication data , and the interference degree of the voltage fluctuation on the judgment of the communication module status of the electricity meter by the -th communication data are calculated in the same way as the above.

[0091] In one embodiment, the importance degree of the -th communication data corresponding to the -th length data packet is positively correlated with the information gain of the -th communication data corresponding to the -th length data packet for judging the communication module status of the electricity meter. The calculation expression of the information gain is:

[0092] ;

[0093] Wherein, is the information entropy of the state 𝑦 of the electricity meter communication module under the 𝑖-th length data packet; is the conditional entropy of the state 𝑦 of the electricity meter communication module after the 𝑚-th communication data is known under the 𝑖-th length data packet.

[0094] In this embodiment, it is possible to directly make the equal to the , or make the proportional to the , and the proportionality coefficient is positive.

[0095] The greater the information gain, the higher the importance of the 𝑚-th communication data for judging the state of the electricity meter communication module under the -th length data packet. The smaller the information gain, the lower the importance of the 𝑚-th communication data for judging the state of the electricity meter communication module under the -th length data packet. Therefore, by using the method of this embodiment, the accuracy of calculating the importance degree of the -th communication data for judging the state of the electricity meter communication module can be improved.

[0096] In one embodiment, it further includes: performing linear interpolation processing on the collected historical data packets and environmental data, and denoising them by using wavelet transform.

[0097] By performing linear interpolation processing on the collected historical data packets and environmental data, missing values can be filled. By using wavelet transform to denoise the collected data, the accuracy of the collected data can be improved, thereby further improving the accuracy of the detection result of the state of the electricity meter communication module.

[0098] Embodiment of the electricity meter communication module state detection system based on data analysis:

[0099] The present invention also provides an electricity meter communication module state detection system based on data analysis. As Figure 2 shown, the electricity meter communication module state detection system based on data analysis includes a processor and a memory. The memory stores computer program instructions for detecting the state of the electricity meter communication module. When the computer program instructions are executed by the processor, it implements an electricity meter communication module state detection method according to the above embodiments.

[0100] The electricity meter communication module state detection system based on data analysis further includes a communication bus and a communication interface and other components well known to those skilled in the art. Their settings and functions are known in the art, so they will not be described in detail here.

[0101] In the description of this specification, the meanings of "a plurality of" and "several" are at least two, such as two, three or more, etc., unless otherwise specifically defined.

[0102] Although this specification has shown and described multiple embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will think of many changes, alterations and alternative ways without departing from the spirit and concept of the present invention. It should be understood that various alternatives to the embodiments of the present invention described herein may be employed in the practice of the present invention.

Claims

1. A method for detecting the state of an electric energy meter communication module based on data analysis, characterized in that: include: Collect historical data packets of various lengths from the communication modules of electric energy meters, as well as environmental data at historical moments corresponding to each historical data packet; There are multiple historical data packets of each length; For each length of historical data packets, obtain the degree of interference from the environment when judging the state of the communication module for each communication data, as well as the importance of various communication data for judging the state of the communication module; build a classification model, train it using the loss function to obtain a state detection model, and the loss function calculation expression used when training the classification model corresponding to the i-th length data packet is: In the formula, For the The classification loss function of the nth communication data in the classification model corresponding to the data packet of this length; For the The label of the communication module status corresponding to the secondary communication data; is the output of the classification model Prediction of communication data probability; For the The mth type of communication data corresponding to the data packet of the type length determines the interference degree of the environment when the state of the communication module is judged; For the The length of the data packet corresponding to The importance of the communication data for determining the state of the communication module; The communication data packets sent by the electric energy meter communication module are collected in real time, and the state of the electric energy meter communication module is obtained by combining the state detection model corresponding to the historical data packet with the closest length to it; assuming that there are two types of historical data packet lengths closest to the length of the communication data packet collected in real time, the state detection model corresponding to the historical data packet with a smaller length than the communication data packet collected in real time is selected to detect the electric energy meter communication module.

2. The method for detecting the state of an electric energy meter communication module based on data analysis according to claim 1, characterized in that: No. The method for calculating the degree of interference of the environment when judging the state of the communication module by the mth type of communication data corresponding to the data packet of the type length comprises: calculating the mth type of communication data corresponding to the data packet of the type length The mth type of communication data corresponding to the data packet of the type length is used to judge the interference degree of the communication module state caused by temperature, humidity and voltage fluctuation; The interference degree due to temperature, the interference degree due to humidity and the interference degree due to voltage fluctuation are weighted and summed to obtain the first The mth type of communication data corresponding to the data packet of the type length is used to judge the interference degree of the environment when the state of the communication module is judged.

3. The method for detecting the state of an electric energy meter communication module based on data analysis according to claim 2, characterized in that: No. The calculation expression for the degree of interference from the environment when the mth type of communication data corresponding to the data packet of the type length is used to judge the state of the communication module is: ； In the formula, For the The information carrying capacity of the environmental data sequence; Respectively The information carrying capacity of temperature sequence, humidity sequence and voltage fluctuation sequence under the data packets of different lengths, , and Respectively represent The mth type of communication data corresponding to the data packet of the type length is used to judge the interference degree of the communication module state caused by temperature, humidity and voltage fluctuation.

4. The method for detecting the state of an electric energy meter communication module based on data analysis according to claim 3, characterized in that: No. The information carrying capacity of environmental data The calculation expression is: ； In the formula, For the The length of the data packet corresponding to The standard deviation of the environmental data series, For the The length of the data packet corresponding to Environmental data series and Pearson correlation coefficient of environmental data series, They represent temperature series, humidity series and voltage fluctuation series respectively.

5. The method for detecting the state of an electric energy meter communication module based on data analysis according to claim 3, characterized in that: No. The calculation expression for the degree of interference of temperature when judging the state of the communication module by the mth type of communication data corresponding to the data packet of the type length is: ； In the formula, and Respectively The maximum and minimum values ​​in the mth communication data sequence corresponding to the data packet of the length, and Respectively The maximum and minimum values ​​of the temperature in the temperature sequence under the data packet of the same length. is the hyperbolic tangent function, For the The absolute value of the Pearson correlation coefficient between the temperature sequence under the data packet of the mth length and the mth communication data sequence.

6. The method for detecting the state of an electric energy meter communication module based on data analysis according to claim 1, characterized in that: No. The length of the data packet corresponding to The importance of the mth communication data and the information gain of the mth communication data corresponding to the i-th length data packet for judging the state of the electric energy meter communication module is positively correlated, and the calculation expression of the information gain is: ； In the formula, is the information entropy of the state 𝑦 of the electric energy meter communication module under the 𝑖th length data packet; is the conditional entropy of the communication module state y of the energy meter after the mth communication data is known under the i-th length data packet.

7. The method for detecting the state of an electric energy meter communication module based on data analysis according to claim 1, characterized in that: The communication data includes communication time, signal strength and packet loss rate.

8. The method for detecting the state of an electric energy meter communication module based on data analysis according to claim 1, characterized in that: The method also includes: performing linear interpolation processing on the collected historical data packets and environmental data, and performing denoising on the data by using wavelet transform.

9. The method for detecting the state of an electric energy meter communication module based on data analysis according to any one of claims 1 to 8, characterized in that: The classification model is a neural network model or a machine learning model.

10. A power meter communication module status detection system based on data analysis, characterized in that: It comprises a processor and a memory, wherein the memory stores computer program instructions for detecting the state of an electric energy meter communication module, and when the computer program instructions are executed by the processor, the electric energy meter communication module state detection method based on data analysis as described in any one of claims 1 to 9 is implemented.

Images