Electricity meter metering method and device based on electric pattern fluctuation

Abstract

The invention discloses an electricity meter metering method and device based on electric ripple fluctuation, and the method comprises the steps: directly capturing and processing time domain signals of current, voltage and power at an electricity meter end, converting the time domain signals into frequency domain signals, and recognizing the electric power fingerprint features of different electric appliances through a neural network model, thereby achieving the local, real-time and high-precision electric appliance state recognition. And generating an ammeter metering report based on the actual state information, and finally uploading the ammeter metering report to the server. Compared with a traditional cloud data acquisition mode, the method can significantly reduce errors and time delay in the data transmission process, improves the accuracy of power fingerprint identification, breaks through the limitation of hardware data transmission bandwidth, enables the electric fingerprint data analysis to be more efficient and accurate, can monitor the operation state of an electric appliance in real time, and improves the safety of the electric appliance. And finer and more reliable data support can be provided for power consumption management and energy-saving optimization.

Description

Technical Field

[0001] The present invention relates to the field of electric meter measurement, and in particular to an electric meter measurement method and device based on electric ripple fluctuation. Background Art

[0002] Power fingerprint recognition technology is developed from power load recognition technology and is an important part of power fingerprint technology. Power fingerprint recognition technology collects electrical data from electrical equipment and uses artificial intelligence technology and big data technology to mine the "power fingerprint" features that can characterize the characteristics of the equipment.

[0003] However, the existing electric fingerprint data is collected and analyzed through the cloud, which may lead to errors such as missing, offset, and delay in the collection of electric fingerprint data, resulting in large errors in the feature recognition of electric fingerprints. At the same time, the data transmission of electric fingerprints is limited by the data transmission bandwidth of the hardware, which limits the application of electric fingerprint data. Summary of the invention

[0004] In view of this, an object of the present invention is to provide an electric meter measurement method and device based on electric ripple fluctuation.

[0005] In order to achieve the above technical objectives, the technical solution adopted by the present invention is:

[0006] In a first aspect, the present invention provides an electric meter measurement method based on electric pattern fluctuation, comprising:

[0007] Obtaining metering information from the electric meter at a preset frequency, the metering information including current value and power value;

[0008] Performing time-domain-frequency-domain transformation on the measurement information to obtain multiple frequency information;

[0009] Input multiple frequency information into the neural network model to obtain the electrical appliance information corresponding to the frequency information;

[0010] Mapping and storing the frequency information and the electrical appliance information, obtaining the electrical pattern characteristic information of each electrical appliance;

[0011] Identify the actual status information of the current electrical appliance based on the electrical pattern feature information;

[0012] Generate an electric meter measurement report based on the actual status information and upload the electric meter measurement report to the server.

[0013] In some embodiments, the measurement information is transformed from the time domain to the frequency domain to obtain a plurality of frequency information including:

[0014] Arrange the data points in the measurement information in chronological order to form a first time series;

[0015] Performing noise filtering on the first time series to obtain a first time domain signal;

[0016] Performing a fast Fourier transform on the first time domain signal to obtain frequency information, each frequency information including an amplitude value and a phase value corresponding to each data point.

[0017] In some embodiments, inputting the plurality of frequency information into the neural network model comprises:

[0018] For each frequency information, perform the following steps:

[0019] Draw a discrete point distribution graph according to the frequency information, the abscissa of the discrete point distribution graph is the acquisition timestamp of each data point of the current frequency information, and the ordinate of the discrete point distribution graph is the amplitude value of each data point of the current frequency information;

[0020] Connect the data points in the discrete point distribution diagram in sequence to obtain the actual waveform diagram of the current frequency information;

[0021] Input the actual waveform graph into the neural network model.

[0022] In some embodiments, the neural network is trained by:

[0023] Construct a basic neural network model, the input parameter of which is the sample frequency information in the sample database, and the output result is the sample electrical appliance information corresponding to the sample frequency information, the sample frequency information includes the sample amplitude value and the sample phase value, and the sample electrical appliance information includes the electrical appliance model and the electrical appliance name;

[0024] The sample frequency information in the sample database is read to repeatedly train the basic neural network until the accuracy of the basic neural network model is within the range of a preset accuracy threshold, thereby obtaining a trained neural network model.

[0025] In some embodiments, the frequency information is mapped and stored with the appliance information to obtain the electric signature feature information of each appliance, including:

[0026] The actual waveform diagram and the electrical appliance information are mapped and stored.

[0027] In some embodiments, identifying the actual state information of the current electrical appliance according to the electric pattern feature information includes:

[0028] Retrieving reference state information and a reference waveform diagram associated with the reference state information from a database according to the electrical appliance information, wherein the reference state information includes at least one of a closed state, an open state, and a working state, and the reference waveform diagram includes at least one of a closed waveform diagram, an open waveform diagram, and a working waveform diagram;

[0029] Compare the actual waveform with the reference waveform to obtain a similarity value between the actual waveform and each reference waveform;

[0030] The multiple similarity values ​​are arranged according to numerical values, and the reference state information associated with the reference waveform corresponding to the similarity value arranged in the first position is selected as the actual state information of the current electrical appliance.

[0031] In some embodiments, generating an electric meter measurement report according to actual status information includes:

[0032] The actual state information corresponding to the current electrical appliances within the preset time period is sorted into an actual power consumption list, wherein the actual power consumption list includes the actual state information and the actual total power consumption corresponding to each actual state information;

[0033] Obtain the historical electricity consumption list of the current appliance, which includes historical status information and the total historical electricity consumption corresponding to each historical status information;

[0034] Draw the actual electricity consumption list and the historical electricity consumption list into an electricity consumption comparison chart;

[0035] Compare the actual total electricity consumption and the historical total electricity consumption at the same time point in the electricity consumption comparison diagram to obtain the comparison result;

[0036] Generate an electricity meter measurement report based on the comparison results and electricity usage comparison chart.

[0037] In some embodiments, the actual total power consumption and the historical total power consumption at the same time node are compared in the power consumption comparison graph, and the comparison results include:

[0038] Calculate whether the actual status information at the same time node is consistent with the historical status information;

[0039] If they are consistent, the difference between the actual total electricity consumption and the historical total electricity consumption at the same time node is calculated and recorded as the electricity consumption difference;

[0040] Determine whether the power consumption difference is within a preset threshold value;

[0041] If yes, the output comparison result is normal;

[0042] If not, obtain the reference power consumption of the current electrical appliance under the current actual state information;

[0043] Calculating the ratio of the actual total power consumption to the reference power consumption, and obtaining the number of electrical appliances in the current electrical appliance information according to the ratio, which is recorded as the first quantity information;

[0044] Obtain the number of electrical appliances of the user's current electrical appliance information, and record it as the second quantity information;

[0045] Determine whether the first quantity information is consistent with the second quantity information. If not, output the comparison result as abnormal.

[0046] In a second aspect, the present invention provides an electric meter measuring device based on electric ripple fluctuations, comprising an electric meter body and a meter module, the electric meter body comprising a first incoming terminal, a first outgoing terminal and a circuit breaker, the electric meter body being used to collect metering information, the metering information comprising at least one of a voltage value, a current value and a power value; the meter module comprising a second incoming terminal and a second outgoing terminal, the second incoming terminal being connected in parallel with the first incoming terminal, the second outgoing terminal being connected in parallel with the first outgoing terminal, the meter module being used to execute the method described in the first aspect.

[0047] By adopting the above technical solution, the present invention has the following beneficial effects compared with the prior art:

[0048] Different from the existing technology, this technical solution directly captures and processes the time domain signals of current, voltage and power at the meter end, converts them into frequency domain signals, and uses a neural network model to identify the power fingerprint characteristics of different electrical appliances, thereby realizing localized, real-time and high-precision electrical appliance status recognition, and generating a meter measurement report based on the actual status information, and finally uploading the meter measurement report to the server. Compared with the traditional cloud data collection method, this method can significantly reduce the error and delay in the data transmission process, improve the accuracy of power fingerprint recognition, and break through the limitation of hardware data transmission bandwidth, making power fingerprint data analysis more efficient and accurate, not only to monitor the operating status of electrical appliances in real time, but also to provide more refined and reliable data support for power management and energy-saving optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only 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.

[0050] Figure 1 It is a step diagram of the electric meter measurement method;

[0051] Figure 2 It is a schematic diagram of the distribution of discrete points;

[0052] Figure 3 It is a schematic diagram of the actual waveform diagram and the reference waveform diagram;

[0053] Figure 4 It is a schematic diagram of the electricity comparison diagram. DETAILED DESCRIPTION

[0054] The present invention will be further described in detail below in conjunction with the accompanying drawings and examples. It is particularly noted that the following examples are only used to illustrate the present invention, but are not intended to limit the scope of the present invention. Similarly, the following examples are only partial embodiments of the present invention rather than all embodiments, and all other embodiments obtained by those of ordinary skill in the art without creative work are within the scope of protection of the present invention.

[0055] See also Figures 1 to 4 In a first aspect, this embodiment provides an electric meter measurement method based on electric pattern fluctuation, comprising:

[0056] S101, obtaining metering information on an electric meter according to a preset frequency, where the metering information includes a current value and a power value;

[0057] S102, performing time domain-frequency domain transformation on the measurement information to obtain multiple frequency information;

[0058] S103, inputting multiple frequency information into a neural network model to obtain electrical appliance information corresponding to the frequency information;

[0059] S104, mapping and storing the frequency information and the electrical appliance information to obtain the electrical signature feature information of each electrical appliance;

[0060] S105, identifying the actual state information of the current electrical appliance according to the electric pattern feature information;

[0061] S106: Generate an electric meter measurement report according to the actual status information, and upload the electric meter measurement report to the server.

[0062] In this embodiment, the preset frequency can be 1000Hz, and the preset frequency can also be set according to actual needs. In step S101, the metering information on the electric meter is obtained according to the preset frequency, which can be understood as performing data collection on the electric meter 1000 times per second. The metering information includes current value and power value, and optionally, it can also include voltage value. The voltage value is the total voltage measured on the electric meter, the current value is the total current measured on the electric meter, and the power value is the total power measured on the electric meter. It can be understood that this step corresponds to the structure of the electric meter metering device shown in the second aspect. The structure of the electric meter itself is not modified, but based on the detection principle of the existing electric meter, the metering data of the electric meter is shared. In this embodiment, different electric pattern information can be obtained by comparing the integral of the power value with the current value.

[0063] In step S102, the metering information is transformed from time domain to frequency domain. The time domain signal is a signal that changes with time, such as the change of current with time, and the frequency domain signal is a representation of the time domain signal decomposed into different frequency information. The time domain to frequency domain transformation is to convert the time domain information into frequency information. The frequency information is continuous amplitude change data. Taking the current value as an example, the total current value can obtain multiple different current frequencies after the time domain to frequency domain transformation. Each current frequency has a continuous amplitude change to characterize different electrical signals.

[0064] In step S103, the frequency information is identified and classified using a neural network model to obtain corresponding appliance information. For example, a certain brand of air conditioner corresponds to frequency information A, and a certain brand of washing machine corresponds to frequency information B. Through the identification and classification of the neural network model, the frequency information can be matched to different appliance information.

[0065] In step S104, the frequency information and the electrical appliance information are mapped and stored to form electrical pattern feature information, which can also be understood as electrical pattern data, electrical pattern features, etc. of the electrical appliance.

[0066] In step S105, the actual state information of the current appliance is identified based on the electric pattern characteristic information. Specifically, it can be understood in combination with the above-mentioned current example as follows: different appliances have different current signals in different states. For example, when the washing machine is turned on, the current signal will jump from a lower amplitude to a higher amplitude. When the washing machine is turned off, the current signal will jump from a higher amplitude to a lower amplitude. When the washing machine is on standby, the current signal will remain at a higher amplitude stage. The electric pattern characteristic information contains the current amplitude data of the washing machine. By identifying the frequency information in the electric pattern characteristic information, the current state of the appliance associated with this frequency information, that is, the actual state information, can be obtained.

[0067] In some optional embodiments, the actual state information of the current electrical appliance can be obtained by comprehensive judgment based on the electric pattern characteristic information and the power value in the metering information. The specific steps may be: when a jump in the current value is detected, the timestamp of the current value jump is recorded in real time, recorded as the first jump timestamp, and the current value is recorded as the first jump current value; the power value corresponding to the first jump timestamp is synchronously obtained, recorded as the first jump power value. The first jump power value is integrated to obtain the first power integral, the first jump current value is transformed from the time domain to the frequency domain to obtain the first electric pattern characteristic information, and the actual state information of the electrical appliance is determined based on the first electric pattern characteristic information and the first power integral. Specifically, when the electrical appliance is started, the first power integral is a positive value, that is, the first jump power value suddenly increases, and when the electrical appliance is turned off, the first power integral is a negative value, that is, the first jump power value suddenly decreases. It can be understood that the first electric pattern characteristic information plays a role in distinguishing between the working state and the on-off state in this judgment process, that is, the amplitude change contained in the first electric pattern characteristic information can be used to distinguish between the working state and the on-off state. Generally speaking, when the appliance is in a working state for a long time, the amplitude of the first electric pattern characteristic information will stabilize in a certain range. When the appliance is switched on or off, the amplitude of the first electric pattern characteristic information will suddenly change from a certain range to another range. On this basis, the specific judgment of the on-off state is achieved by judging the positive and negative values ​​of the first power integral.

[0068] In step S106, an electric meter measurement report is generated according to the actual status information, such as Figure 4 As shown, the meter measurement report can be presented in the form of a chart, or the meter measurement report can cover the user's electricity consumption data in the current period, and be sent to the value server in the form of a list or data set. The server updates this data to the historical data of the current user, and when it is necessary to generate the current user's electricity consumption distribution map, it is retrieved from the database accordingly, thereby realizing the user's electricity consumption overview, which is convenient for subsequent data analysis, data processing, etc.

[0069] It should be noted that the metering information collected in this embodiment is collected at a high frequency, and steps S101 to S106 are all performed locally instead of uploading to the server. This can reduce the feature loss during data transmission and reduce the amount of data collected between the server and the device collection end, thereby avoiding the risk of data loss caused by network instability, etc., and can maintain the feature integrity of local data and improve the accuracy of generating meter metering reports.

[0070] This embodiment directly captures and processes the time domain signals of current, voltage and power at the meter end, converts them into frequency domain signals, and uses a neural network model to identify the power fingerprint characteristics of different electrical appliances, thereby achieving localized, real-time and high-precision electrical appliance status recognition, and generates a meter measurement report based on the actual status information, and finally uploads the meter measurement report to the server. Compared with traditional cloud data collection methods, this method can significantly reduce errors and delays in the data transmission process, improve the accuracy of power fingerprint recognition, and break through the limitations of hardware data transmission bandwidth, making power fingerprint data analysis more efficient and accurate, not only to monitor the operating status of electrical appliances in real time, but also to provide more refined and reliable data support for power management and energy-saving optimization.

[0071] In the second aspect, the present embodiment provides an electric meter measuring device based on electric ripple fluctuations, including an electric meter body and a meter module, the electric meter body including a first incoming terminal, a first outgoing terminal and a circuit breaker, the electric meter body is used to collect metering information, the metering information includes at least one item of voltage value, current value and power value; the meter module includes a second incoming terminal and a second outgoing terminal, the second incoming terminal is connected in parallel with the first incoming terminal, the second outgoing terminal is connected in parallel with the first outgoing terminal, and the meter module is used to execute the method described in the first aspect.

[0072] The electric meter body is an existing electric meter, and the circuit breaker is also an equipped structure in the existing electric meter. The circuit breaker has an electric quantity monitoring function. On this basis, the incoming terminal and the outgoing terminal of the circuit breaker are recorded as a first incoming terminal and a first outgoing terminal; the meter module is a hardware structure newly added on the basis of the electric meter body, which may specifically include a processing chip and a storage chip, etc. The processing chip is used to execute the method described in the first aspect, and the storage chip is used to store the collected data. The incoming terminal and the outgoing terminal of the meter module are recorded as a second incoming terminal and a second outgoing terminal, then the second incoming terminal is connected in parallel with the first incoming terminal, and the second outgoing terminal is connected in parallel with the first outgoing terminal to synchronize the electric quantity data, that is, to obtain metering information.

[0073] Corresponding to the aforementioned method, the electric meter measuring device shown in this embodiment connects the metering information of the electric meter in parallel on the basis of the existing electric meter structure, and can execute the method described in the first aspect.

[0074] In some embodiments, the measurement information is transformed from the time domain to the frequency domain to obtain a plurality of frequency information including:

[0075] Arrange the data points in the measurement information in chronological order to form a first time series;

[0076] Performing noise filtering on the first time series to obtain a first time domain signal;

[0077] Performing a fast Fourier transform on the first time domain signal to obtain frequency information, each frequency information including an amplitude value and a phase value corresponding to each data point.

[0078] In this embodiment, metering information is collected from the electric meter, and the metering information generally includes voltage value, current value, power value, etc. For subsequent analysis, it is necessary to organize these data into a one-dimensional data array in chronological order, called the first time series, which can be specifically expressed by formula (1), which is as follows:

[0079] D=[x[0],x[1],x[2],…,x[n]];

[0080] In formula (1), D is the first time series, x[n] represents the measurement information collected at the nth moment, n is a natural number, n∈N-1, N is the total number of sampling points, and when 1000 Hz is used, N=1000.

[0081] During the data acquisition process, it may be affected by external interference or noise from the device itself, so it is necessary to filter the noise of the first time series to improve the accuracy of subsequent analysis. You can choose a suitable filter, such as a low-pass filter, a band-pass filter, or an adaptive filter, to filter according to the signal characteristics and noise frequency characteristics.

[0082] Perform a fast Fourier transform (FFT) on the filtered first time domain signal to convert the time domain signal into a frequency domain signal, thereby extracting frequency information. This is expressed by formula (2), which is as follows:

[0083]

[0084] In formula (2), X[k] is the complex number representation of the frequency domain signal, and k is the serial number of the frequency information. It can be understood that k is a natural number, and the first frequency information can be recorded as the first frequency signal. Correspondingly, k=1.

[0085] The amplitude value of each frequency information is extracted from the complex array output by FFT. The amplitude value can be expressed using formula (3), which is as follows:

[0086]

[0087] In formula (3), |X[k]| is the amplitude value of the kth frequency information, Re(X[k]) is the real part of X[k], and Im(X[k]) is the imaginary part of X[k];

[0088] At the same time, the phase value of each frequency information is calculated. The phase value can be expressed using formula (4). Formula (4) is as follows:

[0089]

[0090] In formula (4), φ[k] is the phase value of the kth frequency information.

[0091] Through the above steps, this embodiment realizes the time domain-frequency domain transformation of the metering information and successfully extracts multiple frequency information, including the amplitude value and phase value corresponding to each frequency information. These frequency information provides an important basis for subsequent electrical appliance identification and status analysis, and can effectively improve the accuracy and reliability of power fingerprint identification.

[0092] See also Figure 2 In some embodiments, inputting the plurality of frequency information into the neural network model comprises:

[0093] For each frequency information, perform the following steps:

[0094] Draw a discrete point distribution graph according to the frequency information, the abscissa of the discrete point distribution graph is the acquisition timestamp of each data point of the current frequency information, and the ordinate of the discrete point distribution graph is the amplitude value of each data point of the current frequency information;

[0095] Connect the data points in the discrete point distribution diagram in sequence to obtain the actual waveform diagram of the current frequency information;

[0096] Input the actual waveform graph into the neural network model.

[0097] In this embodiment, the specific process of inputting multiple frequency information into the neural network model is described in detail, including the steps of drawing a discrete point distribution diagram, generating an actual waveform diagram, and inputting it into the neural network model. In this embodiment, multiple frequency information has been obtained through time domain-frequency domain transformation, and each frequency information includes a series of data points, wherein each data point contains a timestamp and an amplitude value, the timestamp indicates the specific time when the data point was collected; the amplitude value indicates the amplitude value corresponding to the time point. For each set of frequency information, a discrete point distribution diagram is drawn, with the timestamp as the horizontal coordinate and the amplitude value as the vertical coordinate, such as Figure 2 and Figure 3 As shown, a discrete point distribution diagram is formed. Further, the data points in the discrete point distribution diagram are connected in sequence to form an actual waveform diagram. Specifically, linear interpolation or other interpolation methods can be used to connect every two adjacent data points to obtain a continuous actual waveform diagram. From the above steps, multiple actual waveform diagrams can be obtained, each of which corresponds to a single frequency information.

[0098] The actual waveform diagram is converted into a format suitable for neural network model input, for example, it can be input into the neural network model in the form of a matrix, and the neural network model obtains output results, such as appliance type, status information or other characteristics.

[0099] Through the above steps, this embodiment realizes the drawing of multiple frequency information into discrete point distribution diagrams and the generation of actual waveform diagrams, which are further input into the neural network model. This process not only effectively utilizes the frequency information, but also enhances the intuitiveness of data expression through graphical methods, providing stronger support for power fingerprint recognition. Ultimately, the neural network model can accurately identify the state of the electrical appliance based on the input waveform diagram, improving the accuracy and reliability of recognition.

[0100] In some embodiments, the neural network is trained by:

[0101] Construct a basic neural network model, the input parameter of which is the sample frequency information in the sample database, and the output result is the sample electrical appliance information corresponding to the sample frequency information, the sample frequency information includes the sample amplitude value and the sample phase value, and the sample electrical appliance information includes the electrical appliance model and the electrical appliance name;

[0102] The sample frequency information in the sample database is read to repeatedly train the basic neural network until the accuracy of the basic neural network model is within the range of a preset accuracy threshold, thereby obtaining a trained neural network model.

[0103] In this embodiment, a basic neural network model is constructed. The design of the model can be determined according to the needs of the specific task, and usually includes an input layer, a hidden layer, and an output layer. The input layer receives the sample frequency information, and the output layer generates the corresponding electrical appliance information. The input parameters of the basic neural network are the sample frequency information in the sample database, which includes the sample amplitude value, the original phase value, the output result, the electrical appliance model, and the electrical appliance name. The sample amplitude value represents the strength of each frequency information. The sample phase value represents the phase information of each frequency information. The output result of the output result network is the electrical appliance information corresponding to the sample frequency information, including: the electrical appliance model is the specific model of the identified electrical appliance, and the electrical appliance name is the name of the identified electrical appliance.

[0104] Prepare a sample database containing multiple sets of labeled frequency information and their corresponding appliance information. These samples will be used to train the basic neural network. By reading the sample frequency information in the sample database, the neural network model will gradually adjust its internal parameters. The training process usually includes two stages: forward propagation and backward propagation. Forward propagation passes the input sample frequency information through the network and calculates the output appliance information. Backward propagation calculates the gradient and updates the model parameters based on the error between the output result of the model and the true label to reduce the prediction error. The basic neural network is trained for multiple iterations, using different sample frequency information to enhance the learning ability and generalization ability of the model. After each iteration, the accuracy of the model is evaluated. During the training process, the accuracy of the model is evaluated regularly, usually using cross-validation or validation sets to monitor the performance of the model. The training process continues until the accuracy of the basic neural network model reaches or exceeds the preset accuracy threshold. This threshold can be set according to application requirements and actual conditions to ensure the effectiveness and reliability of the model in actual applications. Once the accuracy of the basic neural network model reaches the preset threshold, a trained neural network model is obtained. This model can be used in actual electrical appliance recognition tasks and can quickly and accurately identify the model and name of the appliance based on the input frequency information.

[0105] Through the above steps, the neural network training process in this embodiment includes building a basic model, reading sample data, performing multiple repeated trainings, and evaluating accuracy, and finally obtaining a trained neural network model that can effectively identify electrical appliance information. This process ensures that the model has good performance and reliability in practical applications.

[0106] In some embodiments, the frequency information is mapped and stored with the appliance information to obtain the electric signature feature information of each appliance, including:

[0107] The actual waveform diagram and the electrical appliance information are mapped and stored.

[0108] The mapping stored procedure specifically includes:

[0109] Uniquely identify and encode the actual waveform of each appliance;

[0110] Establish association between characteristic data of the actual waveform and corresponding electrical appliance information (such as model and name);

[0111] Pair and store waveform features and electrical appliance information according to a predefined data structure (such as a dictionary or a database table);

[0112] Generate a unique electric signature information record for each appliance;

[0113] During the storage process, ensure that each record contains the actual waveform data, the corresponding basic information of the electrical appliance, and the frequency information (amplitude value, phase value);

[0114] A retrieval and matching mechanism is established so that the corresponding electrical appliance information can be quickly located through the waveform diagram later.

[0115] This embodiment can construct an electrical signature feature information library of electrical appliances through this mapping storage method, providing a data basis for subsequent electrical appliance identification.

[0116] In some embodiments, identifying the actual state information of the current electrical appliance according to the electric pattern feature information includes:

[0117] Retrieving reference state information and a reference waveform diagram associated with the reference state information from a database according to the electrical appliance information, wherein the reference state information includes at least one of a closed state, an open state, and a working state, and the reference waveform diagram includes at least one of a closed waveform diagram, an open waveform diagram, and a working waveform diagram;

[0118] Compare the actual waveform with the reference waveform to obtain a similarity value between the actual waveform and each reference waveform;

[0119] The multiple similarity values ​​are arranged according to numerical values, and the reference state information associated with the reference waveform corresponding to the similarity value arranged in the first position is selected as the actual state information of the current electrical appliance.

[0120] In this embodiment, according to the appliance information (such as model, name, etc.) of the current appliance, the database is queried and reference state information related to the appliance and its corresponding reference waveform are retrieved. This step ensures that reference data matching the characteristics of the current appliance can be obtained.

[0121] The actual waveform of the current appliance is compared with each reference waveform for similarity. A variety of algorithms can be used to calculate the similarity, such as Euclidean distance, cosine similarity or DTW similarity algorithm. Optionally, the DTW similarity algorithm is an algorithm specifically used to compare time series data, which can effectively handle the time transformation of the waveform.

[0122] In some preferred embodiments, a dynamic time warping (DTW) algorithm is used to calculate the similarity between the actual waveform of the current appliance and each reference waveform. A two-dimensional DTW matrix is ​​constructed, the starting point is initialized, and an infinite value is set to fill the matrix. The matrix is ​​traversed to calculate the difference between the corresponding points of the current actual waveform and the current reference waveform, and the values ​​in the matrix are updated, considering different alignment paths. After the calculation is completed, the DTW distance is read from the lower right corner of the matrix and converted into a similarity value. Finally, the above steps are repeated for all reference waveforms to obtain the similarity value of each reference waveform, and the reference waveform with the highest similarity is selected to determine the actual state information of the current appliance.

[0123] For each reference waveform, calculate the similarity value with the current actual waveform. Arrange all the calculated similarity values ​​according to the numerical value, and identify which reference waveform is most similar to the current actual waveform. Select the similarity value ranked first to find the reference waveform that is most similar to the current actual waveform.

[0124] The extracted reference actual state information is returned as the actual state information of the current electrical appliance. For example, if the state corresponding to the selected reference waveform is "working state", the actual state information of the current electrical appliance is determined to be "working state".

[0125] Through the above steps, this embodiment realizes the process of identifying the actual state information of the current electrical appliance based on the electrical pattern feature information. The process includes retrieving relevant reference actual state information and reference waveform from the database, calculating the similarity between the current actual waveform and the reference waveform, and finally determining the actual state of the electrical appliance by selecting the best matching reference state information. This process can effectively improve the accuracy and reliability of electrical appliance state recognition.

[0126] See also Figure 4 In some embodiments, generating an electric meter measurement report according to actual status information includes:

[0127] The actual state information corresponding to the current electrical appliances within the preset time period is sorted into an actual power consumption list, wherein the actual power consumption list includes the actual state information and the actual total power consumption corresponding to each actual state information;

[0128] Obtain the historical electricity consumption list of the current appliance, which includes historical status information and the total historical electricity consumption corresponding to each historical status information;

[0129] Draw the actual electricity consumption list and the historical electricity consumption list into an electricity consumption comparison chart;

[0130] Compare the actual total electricity consumption and the historical total electricity consumption at the same time point in the electricity consumption comparison diagram to obtain the comparison result;

[0131] Generate an electricity meter measurement report based on the comparison results and electricity usage comparison chart.

[0132] In this embodiment, a preset time segment (such as one day, one week or one month) can be selected to collect all actual state information of the current electrical appliance within the time segment, and for each actual state information, the corresponding actual total power consumption is calculated to generate a detailed list containing the state information and the total power consumption; extract historical power consumption records from the electrical appliance power consumption history database, obtain the same or similar historical time segment as the current time segment, extract the historical total power consumption corresponding to each historical state, and organize the historical state information and the corresponding total power consumption; convert the actual power consumption list and the historical power consumption list into a chart, with the horizontal axis being the time node and the vertical axis being the total power consumption, and specifically, a line chart, a bar chart or the like that is convenient for comparison can be selected; in the power consumption comparison chart, the actual total power consumption and the historical total power consumption of the corresponding time node are compared one by one, and the absolute difference between the actual total power consumption and the historical total power consumption is calculated, and the relative difference percentage is calculated to mark the significantly different power consumption nodes; the finally generated meter measurement report may include an overview of the actual total power consumption, a comparison of the historical total power consumption, a description of the power consumption trend, a prompt of abnormal power consumption, and may also include power efficiency evaluation, energy-saving suggestions, power consumption behavior analysis, etc.

[0133] This embodiment uses a systematic method to compare the actual power consumption of electrical appliances with historical data, and generates a detailed and intuitive meter measurement report to help users understand power consumption and improve power efficiency.

[0134] In some embodiments, the actual total power consumption and the historical total power consumption at the same time node are compared in the power consumption comparison graph, and the comparison results include:

[0135] Calculate whether the actual status information at the same time node is consistent with the historical status information;

[0136] If they are consistent, the difference between the actual total electricity consumption and the historical total electricity consumption at the same time node is calculated and recorded as the electricity consumption difference;

[0137] Determine whether the power consumption difference is within a preset threshold value;

[0138] If yes, the output comparison result is normal;

[0139] If not, obtain the reference power consumption of the current electrical appliance under the current actual state information;

[0140] Calculating the ratio of the actual total power consumption to the reference power consumption, and obtaining the number of electrical appliances in the current electrical appliance information according to the ratio, which is recorded as the first quantity information;

[0141] Obtain the number of electrical appliances of the user's current electrical appliance information, and record it as the second quantity information;

[0142] Determine whether the first quantity information is consistent with the second quantity information. If not, output the comparison result as abnormal.

[0143] In this embodiment, a specific time node in the power usage comparison diagram is selected, and the actual state information and historical state information of the time node are extracted. The actual state information (such as on, off, working, etc.) of the current appliance at the time node is compared with the historical state information to see if they are consistent. If the actual state information is consistent with the historical state information, proceed to the next step. If not, the comparison result is output as "inconsistent state" and the comparison of the time node is terminated.

[0144] Furthermore, a preset threshold value (e.g., ±10%) is set according to user needs or historical data analysis to determine whether the power difference is within a reasonable range. It is determined whether the calculated power difference is within the preset threshold range. If the power difference is within the threshold range, the comparison result is output as "normal". If the power difference exceeds the threshold, the next step is performed.

[0145] Specifically, the next step of processing includes: when the power consumption difference exceeds the threshold, obtain the reference power consumption of the current appliance under the current actual status information. This reference power consumption is usually based on the historical data of the appliance or the standard provided by the manufacturer. Calculate the ratio of the actual total power consumption to the reference power consumption, and according to the calculated ratio, deduce the number of electrical appliances of the current appliance information, which is recorded as the first quantity information. For example, if the ratio is 1.2, it means that the number of appliances is 1.2. It can be further judged whether this value is reasonable. In principle, the first quantity information is a positive integer. If there is a decimal, it means that the current appliance is abnormal in its usage status. For example, the air conditioner has a long service life, resulting in increased energy consumption, which increases the actual total power consumption, resulting in the generation of the first quantity information with a ratio of 1.2.

[0146] Compare the first quantity information with the second quantity information: If the two are consistent, the comparison result is output as "normal". If they are inconsistent, the comparison result is output as "abnormal", and the user may be prompted to check the status of the appliance or the power consumption. The above-mentioned 1.2 units are also abnormal, which can be understood as the abnormal number of air conditioners due to disrepair, that is, the power and power consumption of the air conditioners are too large. Record the comparison results of each time node to form a complete power consumption comparison report, which is convenient for users to conduct subsequent analysis and decision-making.

[0147] Through the above steps, this embodiment systematically compares the actual total power consumption with the historical total power consumption, analyzes the consistency of the power consumption status and quantity information of the electrical appliances, and provides users with a detailed power consumption monitoring and abnormality identification mechanism. This process not only helps users understand the power consumption of electrical appliances, but also can timely discover potential problems and improve power efficiency.

[0148] Different from the existing technology, this technical solution directly captures and processes the time domain signals of current, voltage and power at the meter end, converts them into frequency domain signals, and uses a neural network model to identify the power fingerprint characteristics of different electrical appliances, thereby realizing localized, real-time and high-precision electrical appliance status recognition, and generating a meter measurement report based on the actual status information, and finally uploading the meter measurement report to the server. Compared with the traditional cloud data collection method, this method can significantly reduce the error and delay in the data transmission process, improve the accuracy of power fingerprint recognition, and break through the limitation of hardware data transmission bandwidth, making power fingerprint data analysis more efficient and accurate, not only to monitor the operating status of electrical appliances in real time, but also to provide more refined and reliable data support for power management and energy-saving optimization.

[0149] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0150] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) or a processor (processor) to perform all or part of the steps of 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 and other media that can store program codes.

[0151] The above descriptions are only some embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Any equivalent device or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A meter measurement method based on electric ripple fluctuation, characterized in that: include: Acquire metering information on the electric meter according to a preset frequency, wherein the metering information includes a current value and a power value; Performing a time domain-frequency domain transformation on the measurement information to obtain a plurality of frequency information; Inputting a plurality of the frequency information into a neural network model to obtain electrical appliance information corresponding to the frequency information; Mapping and storing the frequency information and the electrical appliance information to obtain electrical signature information of each electrical appliance; Identifying actual status information of the current electrical appliance according to the electric pattern characteristic information; Generate an electric meter measurement report according to the actual status information, and upload the electric meter measurement report to the server.

2. The electric meter measurement method based on electric ripple fluctuation according to claim 1 is characterized in that: The measurement information is transformed from time domain to frequency domain to obtain multiple frequency information including: Arranging the data points in the measurement information in chronological order into a first time series; Performing noise filtering on the first time series to obtain a first time domain signal; Perform a fast Fourier transform on the first time domain signal to obtain frequency information, where each frequency information includes an amplitude value and a phase value corresponding to each data point.

3. The electric meter measurement method based on electric ripple fluctuation according to claim 1 is characterized in that: Inputting the plurality of frequency information into the neural network model comprises: For each frequency information, the following steps are performed: Draw a discrete point distribution graph according to the frequency information, wherein the abscissa of the discrete point distribution graph is the acquisition timestamp of each data point of the current frequency information, and the ordinate of the discrete point distribution graph is the amplitude value of each data point of the current frequency information; Connect the data points in the discrete point distribution diagram in sequence to obtain an actual waveform diagram of the current frequency information; The actual waveform diagram is input into the neural network model.

4. The electric meter measurement method based on electric ripple fluctuation according to claim 1 is characterized in that: The neural network is trained by the following steps: Constructing a basic neural network model, the input parameter of which is the sample frequency information in the sample database, and the output result is the sample electrical appliance information corresponding to the sample frequency information, the sample frequency information includes the sample amplitude value and the sample phase value, and the sample electrical appliance information includes the electrical appliance model and the electrical appliance name; The sample frequency information in the sample database is read to repeatedly train the basic neural network until the accuracy of the basic neural network model is within a preset accuracy threshold, thereby obtaining a trained neural network model.

5. The electric meter measurement method based on electric ripple fluctuation according to claim 4 is characterized in that: The frequency information is mapped and stored with the electrical appliance information to obtain the electrical signature feature information of each electrical appliance, including: The actual waveform diagram and the electrical appliance information are mapped and stored.

6. The electric meter measurement method based on electric ripple fluctuation according to claim 4 is characterized in that: Identifying the actual status information of the current electrical appliance according to the electric pattern characteristic information includes: Retrieving reference state information and a reference waveform diagram associated with the reference state information in a database according to the electrical appliance information, wherein the reference state information includes at least one of a closed state, an open state, and a working state, and the reference waveform diagram includes at least one of a closed waveform diagram, an open waveform diagram, and a working waveform diagram; Comparing the actual waveform with the reference waveform to obtain a similarity value between the actual waveform and each reference waveform; The plurality of similarity values ​​are arranged according to numerical values, and the reference state information associated with the reference waveform corresponding to the similarity value arranged in the first position is selected as the actual state information of the current electrical appliance.

7. The electric meter measurement method based on electric ripple fluctuation according to claim 6 is characterized in that: Generating an electric meter measurement report according to the actual status information includes: Arrange the actual status information corresponding to the current electrical appliances within the preset time period into an actual power consumption list, wherein the actual power consumption list includes the actual status information and the actual total power consumption corresponding to each of the actual status information; Obtaining a historical electricity consumption list of the current electrical appliance, the historical electricity consumption list including historical status information and a total historical electricity consumption corresponding to each piece of historical status information; Draw the actual electricity consumption list and the historical electricity consumption list into an electricity consumption comparison chart; Compare the actual total electricity consumption and the historical total electricity consumption at the same time node in the electricity consumption comparison graph to obtain a comparison result; The electricity meter measurement report is generated according to the comparison result and the electricity usage comparison diagram.

8. The electric meter measurement method based on electric ripple fluctuation according to claim 7 is characterized in that: The actual total power consumption and the historical total power consumption at the same time node are compared in the power consumption comparison graph to obtain a comparison result including: Calculate whether the actual status information at the same time node is consistent with the historical status information; If they are consistent, the difference between the actual total power consumption and the historical total power consumption at the same time node is calculated and recorded as the power consumption difference; Determining whether the power consumption difference is within a preset threshold value; If yes, the output comparison result is normal; If not, obtain the reference power consumption of the current electrical appliance under the current actual state information; Calculating a ratio of the actual total power consumption to the reference power consumption, and obtaining the number of electrical appliances in current electrical appliance information according to the ratio, which is recorded as first quantity information; Obtain the number of electrical appliances of the user's current electrical appliance information, and record it as the second quantity information; Determine whether the first quantity information is consistent with the second quantity information, and if not, output the comparison result as abnormal.

9. An electric meter measuring device based on electric wave fluctuation, characterized in that: include: An electric meter body, comprising a first inlet terminal, a first outlet terminal and a circuit breaker, wherein the electric meter body is used to collect metering information, wherein the metering information comprises at least one of a voltage value, a current value and a power value; A meter module, comprising a second inlet terminal and a second outlet terminal, wherein the second inlet terminal is connected in parallel with the first inlet terminal, and the second outlet terminal is connected in parallel with the first outlet terminal, and the meter module is used to execute the method described in any one of claims 1-8.

Images