A method and device for classifying the measurement performance of electric energy meters

Abstract

The present invention discloses a method and device for classifying the metering performance of an electric energy meter. The method comprises: after obtaining historical normalized data, actual electric energy data, and a first electric energy loss value of the electric energy meter, using the historical normalized data, the actual electric energy data, and the first electric energy loss value to construct an energy conservation model for the electric energy meter; using the energy conservation model and a preset topological neural network to calculate an energy error value related to metering performance; and classifying the metering performance of the electric energy meter according to the energy error value according to a preset risk decision tree to obtain a classification result related to metering performance. The present invention can construct an energy conservation model using the historical normalized data and actual electric energy data of the electric energy meter, calculate the energy error value of the electric energy meter according to the energy conservation model and the preset topological neural network, and classify the metering performance of the electric energy meter based on the energy error value to determine the metering performance, thereby improving the accuracy of detection.

Description

Technical Field

[0001] The present invention relates to the technical field of electric energy meters, and in particular to a method and device for classifying the metering performance of electric energy meters. Background Art

[0002] An energy meter (also known as an energy meter) is a device used to measure and record power generation, supplied (inter-supplied), plant power consumption, line loss, and consumer power consumption. It consists of an energy meter (active and reactive energy meters, maximum demand meters, multi-rate meters, etc.), metering transformers (including voltage and current transformers), and secondary connecting wires.

[0003] To accurately analyze the performance of various energy meters, detailed analysis of the metering performance is required to improve the accuracy of the performance analysis. Currently, a common method involves detecting the instantaneous values ​​of the energy meter and the metering transformer in the energy metering device, comparing each instantaneous value with a corresponding threshold, and then determining the performance of the energy metering device based on the comparison results.

[0004] However, the currently commonly used methods have the following technical problems: since the detected values ​​are all instantaneous values, the detection errors are large. For example, a certain electricity meter performs well at a certain moment, and the instantaneous value detected at this time meets the corresponding threshold requirements, but it cannot continue to work well after the detection moment. At this time, the deviation between the detection result and the actual result is small, which reduces the accuracy of the detection and is prone to misjudgment, which in turn causes risks. Summary of the Invention

[0005] The present invention proposes a logistics warehouse screening method and device based on order information. The method can use the historical normalized data and actual electric energy data of the electric energy meter to construct an energy conservation model, calculate the energy error value of the electric energy meter according to the energy conservation model and a preset topological neural network, and classify the metering performance of the electric energy meter based on the energy error value to determine the metering performance, thereby improving the detection accuracy.

[0006] A first aspect of an embodiment of the present invention provides a method for classifying metering performance of an electric energy meter, the method comprising:

[0007] After acquiring historical normalized data, actual electric energy data, and a first electric energy loss value of the electric energy meter, constructing an energy conservation model of the electric energy meter using the historical normalized data, the actual electric energy data, and the first electric energy loss value;

[0008] Calculating energy error values ​​related to metrological performance using the energy conservation model and a preset topological neural network;

[0009] The metering performance of the electric energy meter is classified according to the energy error value according to a preset risk decision tree to obtain a classification result on the metering performance.

[0010] In a possible implementation of the first aspect, a method for constructing the preset topological neural network includes:

[0011] A preset electric energy meter is used as a meter graph node, and a metering performance corresponding to the preset electric energy meter is used as a performance graph node;

[0012] Connecting the meter graph node with the performance graph node to obtain a node edge;

[0013] A metering deviation value of the metering performance corresponding to a preset electric energy meter is obtained, and the metering deviation value is used as the weight of the node edge to obtain a topological neural network about the metering performance of the electric energy meter.

[0014] In a possible implementation of the first aspect, a method for acquiring the historical normalized data includes:

[0015] Acquiring historical power data of a plurality of power meters with different functions, wherein the historical power data is power data of the power meters in several different time periods in the past;

[0016] Converting the historical data into a historical power matrix in a matrix format;

[0017] The historical electric energy matrix is ​​normalized to obtain historical normalized data.

[0018] In a possible implementation of the first aspect, the normalization formula is shown as follows:

[0019]

[0020]

[0021] in, Represents the new historical power matrix parameters of the normalized historical normalized data, Represents the normalized old historical electric energy matrix parameters, N represents the total amount of parameters, i represents the serial number of the parameters, Represents the i-th parameter in the new historical electric energy matrix parameter, Represents the i-th parameter in the old historical electric energy matrix parameters, Indicates the minimum value of the new parameter in the new historical power matrix parameters, Indicates the minimum value of the old parameters in the old historical power matrix parameters.

[0022] In a possible implementation of the first aspect, a method for acquiring the first power loss value includes:

[0023] Measure the actual pulse number of the energy meter using a preset oscilloscope;

[0024] Calculating a first power loss value using the historical normalized data, the actual power data, and the measured number of pulses;

[0025] The calculation of the first power loss value is shown in the following formula:

[0026]

[0027] Wherein, γ is the first power loss value, m0 is the calculated pulse number, m is the measured pulse number, C b is the pulse constant of the electric energy meter, P is the actual electric energy data, K J K is the connection coefficient of the electric energy meter. U This is historical normalized data.

[0028] In a possible implementation of the first aspect, the energy conservation model is constructed according to a preset energy conservation formula, which is shown as follows:

[0029]

[0030] Among them, the is the historical normalized data within the jth measurement period within the lth preset time period, is the actual electric energy data in the jth metering time period, L j is the first electric energy loss value in the jth measurement time period, n is the measurement sum, i is the time period count, and j is the time period count different from i.

[0031] In a possible implementation of the first aspect, calculating an energy error value related to metrology performance using the energy conservation model and a preset topological neural network includes:

[0032] Obtaining a second electric energy loss value using the energy conservation model, and obtaining a calculation weight value related to metering performance using a preset topological neural network;

[0033] An energy error value related to metering performance is calculated using the calculated weight value and the second electric energy loss value.

[0034] In a possible implementation manner of the first aspect, the energy error value is calculated as shown in the following formula:

[0035]

[0036] Where a is the energy error value, M is the total number of metrological performance, i is the count of metrological performance, Si is the second power loss value, To calculate the weight value.

[0037] In a possible implementation of the first aspect, the preset risk decision tree is shown as follows:

[0038]

[0039] Among them, g(x) is the output value of the decision tree function, x is the parameter of the decision tree function, f(y) is the input value of the decision model, and α, β, and θ are the classification parameters output by the decision tree function.

[0040] A second aspect of an embodiment of the present invention provides a device for classifying metering performance of an electric energy meter, the device comprising:

[0041] a model building module configured to, after acquiring historical normalized data, actual electric energy data, and a first electric energy loss value of the electric energy meter, build an energy conservation model of the electric energy meter using the historical normalized data, the actual electric energy data, and the first electric energy loss value;

[0042] An energy error value calculation module, configured to calculate an energy error value related to metering performance using the energy conservation model and a preset topological neural network;

[0043] The performance classification module is used to classify the metering performance of the electric energy meter according to the energy error value according to a preset risk decision tree to obtain a classification result on the metering performance.

[0044] Compared with the prior art, the embodiment of the present invention provides a method and device for classifying the metering performance of an electric energy meter, which has the beneficial effect that: the present invention can use the historical normalized data and actual electric energy data of the electric energy meter to construct an energy conservation model, calculate the energy error value of the electric energy meter according to the energy conservation model and a preset topological neural network, and classify the metering performance of the electric energy meter based on the energy error value to determine the metering performance, thereby improving the detection accuracy to avoid misjudgment and reducing the risk of using the electric energy meter. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 This is a flow chart of a method for classifying metering performance of an electric energy meter provided by one embodiment of the present invention;

[0046] Figure 2 This is an operational flow chart of a method for classifying metering performance of an electric energy meter provided by one embodiment of the present invention;

[0047] Figure 3The present invention is a schematic structural diagram of a device for classifying metering performance of an electric energy meter according to an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0049] In order to solve the above problems, a method for classifying the metering performance of an electric energy meter provided in an embodiment of the present application will be introduced and explained in detail through the following specific embodiments.

[0050] Reference Figure 1 , which shows a flow chart of a method for classifying metering performance of an electric energy meter provided by an embodiment of the present invention.

[0051] As an example, the method for classifying the metering performance of an electric energy meter may include:

[0052] S11. After acquiring historical normalized data, actual power data, and a first power loss value of the power meter, construct an energy conservation model of the power meter using the historical normalized data, the actual power data, and the first power loss value.

[0053] In one embodiment, historical normalized data, actual power data, and a first power loss value may be obtained respectively, and then an energy conservation model for the power meter may be constructed using the historical normalized data, actual power data, and the first power loss value.

[0054] Since energy in nature neither appears nor disappears, it simply transforms from one form to another or transfers from one object to another, while the total amount of energy remains constant. This theorem allows for the development of energy conservation models, making it easier to observe the various energy transformation pathways surrounding electrical energy meters.

[0055] The historical normalized data may be data obtained by normalizing the historical electric energy data of the electric energy meter.

[0056] The actual power data can be obtained from actual power consumption records obtained by the power supply bureau. The actual power data can include data measured by a power meter and power consumption of the power meter.

[0057] The first power loss value may be the power loss of the power meter calculated according to the actually measured pulse number of the power meter.

[0058] As an example, the operation of obtaining the historical normalized data includes the following sub-steps:

[0059] S21. Acquire historical power data of multiple power meters with different functions, where the historical power data is power data of the power meters in several different time periods in the past.

[0060] In the embodiment of the present invention, historical electric energy data within M time periods can be directly obtained through an electric energy meter, such as parameters displayed by an electric energy meter, parameters displayed by a current transformer, and the like.

[0061] In an optional embodiment, to improve calculation accuracy, historical energy data from multiple energy meters with different functions may be obtained. For example, historical energy data from N energy meters with different functions over the past M time periods may be obtained, where both N and M are positive integers greater than 1.

[0062] S22. Convert the historical data into a historical power matrix according to a matrix format.

[0063] S23. Normalize the historical electric energy matrix to obtain historical normalized data.

[0064] Specifically, since the obtained parameter value ranges and units are different, in order to facilitate unified processing of the historical power data, the historical power data needs to be converted into a unified evaluation standard, so the historical power data needs to be normalized.

[0065] In one embodiment, the normalization formula is as follows:

[0066]

[0067] in, Represents the new historical power matrix parameters of the normalized historical normalized data, Represents the normalized old historical electric energy matrix parameters, N represents the total amount of parameters, i represents the serial number of the parameters, Represents the i-th parameter in the new historical electric energy matrix parameter, Represents the i-th parameter in the old historical electric energy matrix parameters, Indicates the minimum value of the new parameter in the new historical power matrix parameters, Indicates the minimum value of the old parameters in the old historical power matrix parameters.

[0068] In this embodiment, the historical electric energy data after normalization can improve the accuracy of the calculation results, so that data of different units and magnitudes can be calculated uniformly, which facilitates the subsequent analysis of the metering performance corresponding to the electric energy meter, improves the accuracy of the analysis, and speeds up the analysis efficiency.

[0069] As an example, the operation of the first power loss value includes the following sub-steps:

[0070] S31. Use a preset oscilloscope to measure the actual pulse number of the electric energy meter.

[0071] S32. Calculate a first power loss value using the historical normalized data, the actual power data, and the measured number of pulses.

[0072] The calculation of the first power loss value is shown in the following formula:

[0073]

[0074] Wherein, γ is the first power loss value, m0 is the calculated pulse number, m is the measured pulse number, C b is the pulse constant of the electric energy meter, P is the actual electric energy data, K J K is the connection coefficient of the electric energy meter. U This is historical normalized data.

[0075] In one embodiment, mathematical formulas and equations provide a universal language and tool for describing and solving a variety of problems. An energy meter is a device used to measure energy consumption. Its principle is to calculate energy using three parameters: current, voltage, and time. According to the law of conservation of energy, energy consumption must equal energy generation. Therefore, when using an energy meter, an energy conservation model for the meter can be established based on the energy conservation formula.

[0076] As an example, the energy conservation model is constructed according to a preset energy conservation formula, which is shown as follows:

[0077]

[0078] Among them, the is the historical normalized data within the jth measurement period within the i-th preset time period, is the actual electric energy data in the jth metering time period, L j is the first electric energy loss value in the jth measurement time period, n is the measurement sum, i is the time period count, and j is the time period count different from i.

[0079] Specifically, establishing an energy conservation model can make the energy conversion between electric energy meters clearer, facilitate further analysis of the metering performance, and improve the accuracy and efficiency of the analysis.

[0080] S12. Calculate an energy error value related to metering performance using the energy conservation model and a preset topological neural network.

[0081] After obtaining the energy conservation model, the energy conservation model and a preset topological neural network can be used to calculate an energy error value related to metering performance. The energy error value can be an error value of metered energy related to metering performance of the electric energy meter.

[0082] As an example, the construction method of the preset topological neural network includes:

[0083] S41: Using a preset electric energy meter as a meter graph node, and using the metering performance corresponding to the preset electric energy meter as a performance graph node.

[0084] S42: Connect the meter graph node and the performance graph node to obtain a node edge.

[0085] S43: Obtain a metering deviation value of the metering performance corresponding to a preset electric energy meter, use the metering deviation value as the weight of the node edge, and obtain a topological neural network regarding the metering performance of the electric energy meter.

[0086] In the embodiments of the present invention, electric energy metering refers to the accurate measurement of consumed electric energy, which is an important part of power production, marketing, and safe operation of the power grid. The electric energy meter includes many devices, such as electric energy meters, current transformers, voltage transformers, etc.

[0087] Every measuring instrument has its own unique metrological performance. For example, the metrological performance of a balance includes stability, accuracy, sensitivity, and invariance. The metrological performance of different electric energy meters also varies.

[0088] Specifically, a preset electric energy meter may be used as a graph node to obtain a meter graph node; similarly, a metering performance corresponding to the electric energy meter may be used as a corresponding graph node to obtain a performance graph node.

[0089] By connecting the meter graph nodes with the performance graph nodes to obtain node edges, a simple directed graph can be formed. In this graph, the energy meter can be represented as an entity node, and the corresponding metering performance can be represented as an attribute node or a relationship node.

[0090] Specifically, mapping can help better monitor the performance of energy meters and quickly identify any potential problems. By forming such a map, it is easier to understand how the entire system is operating, making forecasting and planning easier.

[0091] Since the metrological deviation value of the metrological performance is obtained, wherein the metrological deviation value is the difference between the preset metrological performance and the actual metrological performance value, for example, if the metrological performance is accuracy, the metrological deviation value is the accuracy difference between the preset standard accuracy and the actual accuracy of the measurement.

[0092] Furthermore, the topological neural network for metrological performance is formed by a weighted directed graph, in which nodes represent neurons and edges represent connections between neurons. Each connection has a weight that determines the degree of influence of the signal transmitted from one neuron to another.

[0093] Furthermore, in a metering performance topology neural network, nodes represent devices or system components in the network, and edges represent the connections and communications between them. Each edge is assigned a weight, representing the metering deviation of metering performance metrics such as latency and bandwidth for the corresponding connection. These weights can be used to reflect key performance indicators such as the transmission time of a packet from one node to another and the packet loss rate.

[0094] In the embodiment of the present invention, establishing the metering performance topology neural network can mathematically express the metering performance corresponding to each electric energy meter, which is more intuitive and convenient for subsequent calls.

[0095] In an optional embodiment, step S12 may include the following sub-steps:

[0096] S121: Obtain a second electric energy loss value using the energy conservation model, and obtain a calculation weight value related to metering performance using a preset topological neural network.

[0097] S122: Calculate an energy error value related to metering performance using the calculated weight value and the second electric energy loss value.

[0098] In this embodiment of the present invention, a second power loss value can be obtained based on an energy conservation model. Since the energy conservation model for each power meter requires the use of historical normalized data and actual power data to calculate the power loss value, the second power loss value can be the same as the first power loss value described above.

[0099] This process is similar to training a model, using known data to "teach" the model to make predictions and calculations. Once the energy conservation model for each electricity meter is established, the model can be used to make more accurate estimates of electricity loss values.

[0100] Specifically, the energy error value is calculated as follows:

[0101]

[0102] Where a is the energy error value, M is the total number of metrological performance, i is the count of metrological performance, S i is the second power loss value, To calculate the weight value.

[0103] Specifically, since the calculated energy error value is an important analytical basis for the metrological performance, the calculation step is indispensable. The energy error value calculated according to the formula is accurate, the result is easy to analyze, and the calculation process is fast, which is an indispensable step in analyzing metrological performance.

[0104] S13. Classify the metering performance of the electric energy meter according to the energy error value according to a preset risk decision tree to obtain a classification result on the metering performance.

[0105] Since the energy meter will inevitably have loss values ​​within the preset range when collecting current values, it will not affect the metering performance. When the error between historical energy data and actual energy data is too large, it will affect the metering performance of the energy meter.

[0106] In one embodiment, the preset risk decision tree is shown as follows:

[0107]

[0108] Among them, g(x) is the output value of the decision tree function, x is the parameter of the decision tree function, f(y) is the input value of the decision model, and α, β, and θ are the classification parameters output by the decision tree function.

[0109] The energy error value can be used as the input value of the decision tree function. The classification of the metrological performance corresponding to the output energy error value is calculated by the decision tree function to obtain three classification parameters α, β, and θ.

[0110] Specifically, the metering performance of the electric energy meter can be classified according to the specific output classification parameters, and the classification operation can be as follows:

[0111] When the classification of the output metrological performance is α, that is, when the input value is less than the parameter of the decision tree function, α is used to classify the metrological performance;

[0112] When the classification of the output metrological performance is β, that is, when the input value is greater than the parameter of the decision tree function, β is used to classify the metrological performance;

[0113] When the classification of the output metrological performance is θ, that is, when the input value is equal to the parameter of the decision tree function, γ is used to classify the metrological performance.

[0114] In an optional embodiment, α, β, and θ are classification levels of the metering performance in question, for example, α corresponds to normal performance, β corresponds to low-risk performance, and θ corresponds to high-risk performance.

[0115] According to the decision tree function, the metering performance can be simply and quickly analyzed and accurately classified, thereby improving the analysis efficiency of the metering performance corresponding to the electric energy meter, thereby achieving accurate analysis.

[0116] Reference Figure 2 , shows an operational flow chart of a method for classifying metering performance of an electric energy meter provided by an embodiment of the present invention.

[0117] Specifically, the operation of the method for classifying the metering performance of the electric energy meter may be as follows:

[0118] The first step is to establish a metering performance topology neural network based on the metering performance of the preset electric energy meter.

[0119] The second step is to obtain historical power data of N power meters with different functions in the past M time periods, and normalize the historical power data to obtain historical normalized data.

[0120] The third step is to obtain preset actual power data, calculate the power loss value based on the historical normalized data and the actual power data, and establish an energy conservation model for each power meter based on the historical normalized data, the actual power data and the power loss value.

[0121] The fourth step is to calculate the energy error value of the metrological performance according to the energy conservation model and the metrological performance topological neural network.

[0122] The fifth step is to classify the metering performance of the electric energy meter according to the energy error values ​​of different computer meters using a preset risk decision tree, and complete the metering performance analysis of the electric energy meter.

[0123] The embodiment of the present invention obtains preset actual electric energy data, calculates the electric energy loss value based on the historical normalized data and the actual electric energy data, and establishes an energy conservation model for each electric energy meter based on the historical normalized data, the actual electric energy data and the electric energy loss value; establishing the energy conservation model can make the energy conversion between the electric energy meters clearer, facilitate further analysis of the metering performance, and improve the accuracy and efficiency of the analysis; calculate the energy error value of the metering performance based on the energy conservation model, the result is easy to analyze, the calculation process is fast, and it is an indispensable step in analyzing the metering performance; use the preset risk decision tree to classify the metering performance of the electric energy meter according to the energy error values ​​of different computer meters, which can simply and quickly analyze the metering performance and accurately classify it, thereby improving the analysis efficiency of the metering performance corresponding to the electric energy meter, thereby achieving accurate analysis.

[0124] In this embodiment, an embodiment of the present invention provides a method for classifying the metering performance of an electric energy meter, and its beneficial effects are: the present invention can use the historical normalized data and actual electric energy data of the electric energy meter to construct an energy conservation model, calculate the energy error value of the electric energy meter according to the energy conservation model and a preset topological neural network, and classify the metering performance of the electric energy meter based on the energy error value to determine the metering performance, thereby improving the detection accuracy to avoid misjudgment and reducing the risk of using the electric energy meter.

[0125] The embodiment of the present invention also provides a device for classifying the metering performance of an electric energy meter, see Figure 3 , which shows a structural diagram of a metering performance classification device for an electric energy meter provided by an embodiment of the present invention.

[0126] As an example, the device for classifying the metering performance of the electric energy meter may include:

[0127] A model building module 301 is configured to, after acquiring historical normalized data, actual power data, and a first power loss value of the power meter, build an energy conservation model for the power meter using the historical normalized data, the actual power data, and the first power loss value;

[0128] An energy error value calculation module 302 is configured to calculate an energy error value related to metering performance using the energy conservation model and a preset topological neural network;

[0129] The performance classification module 303 is configured to classify the metering performance of the electric energy meter according to the energy error value in accordance with a preset risk decision tree, and obtain a classification result on the metering performance.

[0130] Optionally, the preset topological neural network is constructed by:

[0131] A preset electric energy meter is used as a meter graph node, and a metering performance corresponding to the preset electric energy meter is used as a performance graph node;

[0132] Connecting the meter graph node with the performance graph node to obtain a node edge;

[0133] A metering deviation value of the metering performance corresponding to a preset electric energy meter is obtained, and the metering deviation value is used as the weight of the node edge to obtain a topological neural network about the metering performance of the electric energy meter.

[0134] Optionally, the method for obtaining the historical normalized data includes:

[0135] Acquiring historical power data of a plurality of power meters with different functions, wherein the historical power data is power data of the power meters in several different time periods in the past;

[0136] Converting the historical data into a historical power matrix in a matrix format;

[0137] The historical electric energy matrix is ​​normalized to obtain historical normalized data.

[0138] Optionally, the normalization formula is as follows:

[0139]

[0140] in, Represents the new historical power matrix parameters of the normalized historical normalized data, Represents the normalized old historical electric energy matrix parameters, N represents the total amount of parameters, i represents the serial number of the parameters, Represents the i-th parameter in the new historical electric energy matrix parameter, Represents the i-th parameter in the old historical electric energy matrix parameters, Indicates the minimum value of the new parameter in the new historical power matrix parameters, Indicates the minimum value of the old parameters in the old historical power matrix parameters.

[0141] Optionally, a method for obtaining the first power loss value includes:

[0142] Measure the actual pulse number of the energy meter using a preset oscilloscope;

[0143] Calculating a first power loss value using the historical normalized data, the actual power data, and the measured number of pulses;

[0144] The calculation of the first power loss value is shown in the following formula:

[0145]

[0146] Wherein, γ is the first power loss value, m0 is the calculated pulse number, m is the measured pulse number, C b is the pulse constant of the electric energy meter, P is the actual electric energy data, K J K is the connection coefficient of the electric energy meter. U This is historical normalized data.

[0147] Optionally, the energy conservation model is constructed according to a preset energy conservation formula, which is shown as follows:

[0148]

[0149] Among them, the is the historical normalized data within the jth measurement period within the i-th preset time period, is the actual electric energy data in the jth metering time period, L j is the first electric energy loss value in the jth measurement time period, n is the measurement sum, i is the time period count, and j is the time period count different from i.

[0150] Optionally, the calculating of the energy error value related to the metrology performance by using the energy conservation model and a preset topological neural network includes:

[0151] Obtaining a second electric energy loss value using the energy conservation model, and obtaining a calculation weight value related to metering performance using a preset topological neural network;

[0152] An energy error value related to metering performance is calculated using the calculated weight value and the second electric energy loss value.

[0153] Optionally, the energy error value is calculated as shown in the following formula:

[0154]

[0155] Where a is the energy error value, M is the total number of metrological performance, i is the count of metrological performance, S i is the second power loss value, To calculate the weight value.

[0156] Optionally, the preset risk decision tree is shown as follows:

[0157]

[0158] Among them, g(x) is the output value of the decision tree function, x is the parameter of the decision tree function, and f(y) is the input value of the decision model;

[0159] The energy error value is used as an input value of a decision tree function, and the classification of the metrological performance corresponding to the energy error value is calculated and output by the decision tree function, where α, β, and θ are classification parameters output by the decision tree function respectively.

[0160] Those skilled in the art can clearly understand that, for the sake of convenience and brevity, the specific working process of the device described above can refer to the corresponding process in the aforementioned method embodiment and will not be repeated here.

[0161] Furthermore, an embodiment of the present application also provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for classifying the metering performance of the electric energy meter as described in the above embodiment is implemented.

[0162] Furthermore, an embodiment of the present application also provides a computer-readable storage medium, which stores a computer-executable program, and the computer-executable program is used to enable a computer to execute the metering performance classification method for electric energy meters as described in the above embodiment.

[0163] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for classifying the metering performance of electric energy meters, characterized in that: The method comprises: After acquiring historical normalized data, actual electric energy data, and a first electric energy loss value of the electric energy meter, constructing an energy conservation model of the electric energy meter using the historical normalized data, the actual electric energy data, and the first electric energy loss value; Calculating energy error values ​​related to metrological performance using the energy conservation model and a preset topological neural network; Classifying the metering performance of the electric energy meter according to the energy error value according to a preset risk decision tree to obtain a classification result on the metering performance; The construction method of the preset topological neural network includes: A preset electric energy meter is used as a meter graph node, and a metering performance corresponding to the preset electric energy meter is used as a performance graph node; Connecting the meter graph node with the performance graph node to obtain a node edge; Obtaining a metering deviation value of metering performance corresponding to a preset electric energy meter, using the metering deviation value as a weight of the node edge, and obtaining a topological neural network regarding the metering performance of the electric energy meter; The energy conservation model and the preset topological neural network are used to calculate the energy error value related to the metering performance, including: Obtaining a second electric energy loss value using the energy conservation model, and obtaining a calculation weight value related to metering performance using a preset topological neural network; An energy error value related to metering performance is calculated using the calculated weight value and the second electric energy loss value.

2. The method for classifying the metering performance of electric energy meters according to claim 1, wherein: The method for obtaining the historical normalized data includes: Acquiring historical power data of a plurality of power meters with different functions, wherein the historical power data is power data of the power meters in several different time periods in the past; Converting the historical power data into a historical power matrix in a matrix format; The historical electric energy matrix is ​​normalized to obtain historical normalized data.

3. The method for classifying the metering performance of electric energy meters according to claim 2, wherein: The normalization formula is as follows: in, Represents the new historical power matrix parameters of the normalized historical normalized data, Represents the normalized old historical electric energy matrix parameters, N represents the total amount of parameters, i represents the serial number of the parameters, Represents the i-th parameter in the new historical electric energy matrix parameter, Represents the i-th parameter in the old historical electric energy matrix parameters, Indicates the minimum value of the new parameter in the new historical power matrix parameters, Indicates the minimum value of the old parameters in the old historical power matrix parameters.

4. The method for classifying the metering performance of electric energy meters according to claim 1, wherein: The method for obtaining the first power loss value includes: Measure the actual pulse number of the energy meter using a preset oscilloscope; Calculating a first power loss value using the historical normalized data, the actual power data, and the measured number of pulses; The calculation of the first power loss value is shown in the following formula: Wherein, γ is the first power loss value, m0 is the calculated pulse number, m is the measured pulse number, C b is the pulse constant of the electric energy meter, P is the actual electric energy data, K J K is the connection coefficient of the electric energy meter. U This is historical normalized data.

5. The method for classifying the metering performance of electric energy meters according to claim 1, wherein: The energy conservation model is constructed according to a preset energy conservation formula, which is shown as follows: Among them, the is the historical normalized data within the jth measurement period within the i-th preset time period, is the actual electric energy data in the jth metering time period, L j is the first electric energy loss value in the jth measurement time period, n is the measurement sum, i is the time period count, and j is the time period count different from i.

6. The method for classifying the metering performance of electric energy meters according to claim 1, wherein: The energy error value is calculated as follows: Where a is the energy error value, M is the total number of metrological performance, i is the count of metrological performance, S i is the second power loss value, To calculate the weight value.

7. The method for classifying the metering performance of electric energy meters according to claim 1, wherein: The preset risk decision tree is shown as follows: Among them, g(x) is the output value of the decision tree function, x is the parameter of the decision tree function, f(y) is the input value of the decision model, and α, β, and θ are the classification parameters output by the decision tree function.

8. A device for classifying the measurement performance of an electric energy meter, characterized in that: The device comprises: a model building module configured to, after acquiring historical normalized data, actual electric energy data, and a first electric energy loss value of the electric energy meter, build an energy conservation model of the electric energy meter using the historical normalized data, the actual electric energy data, and the first electric energy loss value; An energy error value calculation module, configured to calculate an energy error value related to metering performance using the energy conservation model and a preset topological neural network; a performance classification module, configured to classify the metering performance of the electric energy meter according to the energy error value according to a preset risk decision tree, and obtain a classification result on the metering performance; The construction method of the preset topological neural network includes: A preset electric energy meter is used as a meter graph node, and a metering performance corresponding to the preset electric energy meter is used as a performance graph node; Connecting the meter graph node with the performance graph node to obtain a node edge; Obtaining a metering deviation value of metering performance corresponding to a preset electric energy meter, using the metering deviation value as a weight of the node edge, and obtaining a topological neural network regarding the metering performance of the electric energy meter; The energy conservation model and the preset topological neural network are used to calculate the energy error value related to the metering performance, including: Obtaining a second electric energy loss value using the energy conservation model, and obtaining a calculation weight value related to metering performance using a preset topological neural network; An energy error value related to metering performance is calculated using the calculated weight value and the second electric energy loss value.

Images