Electric energy metering abnormal data identification system and method based on electric power gap system

Abstract

The invention discloses an electric energy metering abnormal data identification system and method based on an electric power gap system, and relates to the technical field of electric energy metering abnormal data identification. The electric energy metering abnormal data identification method based on the electric power wide-gap system comprises the following specific steps: step 1, collecting electric energy metering related data of an electric power system in real time based on the electric power wide-gap system, and extracting data characteristics of original electric energy metering data; the method comprises the following steps: step 1, constructing a multi-scale prediction model oriented to electric energy metering data characteristics, and step 2, comparing a prediction value output by the prediction model with an actual observation value, and identifying an abnormal part in the electric energy metering data by combining a preset abnormality judgment rule. Abnormal data in electric energy metering can be accurately identified and corrected.

Description

Technical Field

[0001] The present invention belongs to the technical field of abnormal power metering data identification, and particularly relates to an abnormal power metering data identification system and method based on the Power HarmonyOS system. Background Art

[0002] With the rapid development of technology, the power system is undergoing a profound transformation from the traditional mode towards intelligence and digitization. This transformation not only improves the operation efficiency of the power system but also enhances its stability and reliability. Against this background, the accuracy and reliability of power metering data have become key indicators for measuring the performance of the power system.

[0003] Power metering data plays multiple roles in the power system. Firstly, it is the basis for power user billing and directly relates to the economic interests of power enterprises and users. Secondly, these data are also important bases for advanced applications such as power system load forecasting, fault diagnosis, and energy efficiency management in the power system. By deeply analyzing power metering data, potential problems in the power system can be discovered in a timely manner, resource allocation can be optimized, and overall energy efficiency can be improved.

[0004] To address this issue, the industry has developed various identification technologies for abnormal power metering data. These technologies mainly detect and identify abnormal data by analyzing the statistical characteristics, change trends of power metering data, and comparison with historical data. However, with the increasing complexity of the power system and the dynamic changes in data characteristics, existing technologies may not be able to meet the massive data processing requirements of large-scale power systems or accurately identify abnormal data caused by complex faults. Existing identification technologies for abnormal power metering data cannot adapt to complex power system environments and dynamic data characteristics. Summary of the Invention

[0005] The purpose of the present invention is to provide an abnormal power metering data identification system and method based on the Power HarmonyOS system to solve the technical problem that existing technologies cannot adapt to complex power system environments and dynamic data characteristics.

[0006] To achieve the above purpose, the present invention adopts the following technical solutions:

[0007] An abnormal power metering data identification method based on the Power HarmonyOS system includes:

[0008] Step 1: Collect power metering-related data of the power system in real time based on the Power HarmonyOS system, and extract the data characteristics of the original power metering data;

[0009] The power metering related data includes the instantaneous values of grid voltage, current and power factor. Feature extraction is performed on the power metering related data within each time window, and key feature data of the power metering status is extracted from the instantaneous values of voltage, current and power factor;

[0010] Step 2: Based on the operating characteristics of the power system and the historical data pattern, construct a multi-scale prediction model for the characteristics of power metering data;

[0011] Step 3: Compare the predicted values output by the prediction model with the actual observed values, and combine the preset anomaly judgment rules to identify the abnormal parts in the power metering data.

[0012] Further, key feature data of the power metering status is extracted from the instantaneous value of voltage. The specific method is as follows:

[0013] Using the formula represents the voltage fluctuation feature, where i represents the i-th time window, t represents time t, uV(i) represents the voltage fluctuation feature in the i-th time window, t0 represents the start time of the time window, T represents the length of the time window, and V(i, t) represents the instantaneous voltage value at time t in the i-th time window. represents the rated voltage value of the i-th time window, and ΔmaxV(i) represents the maximum voltage fluctuation in the i-th time window, that is, the difference between the maximum instantaneous voltage value and the minimum instantaneous voltage value in the i-th time window, and a is a constant value.

[0014] Further, key feature data of the power metering status is extracted from the instantaneous value of current. The specific method is as follows:

[0015] Using the formula represents the current fluctuation feature, where i represents the i-th time window, t represents time t, uI(i) represents the current fluctuation feature in the i-th time window, t0 represents the start time of the time window, T represents the length of the time window, and I(i, t) represents the instantaneous current value at time t in the i-th time window. represents the average current value of the i-th time window, and ΔmaxI(i) represents the maximum current fluctuation in the i-th time window, that is, the difference between the maximum instantaneous current value and the minimum instantaneous current value in the i-th time window, and a is a constant value.

[0016] Further, key feature data of the power metering status is extracted from the instantaneous value of power factor. The specific method is as follows:

[0017] Using the formula represents the power fluctuation feature, where i represents the i-th time window, t represents time t, and uC(i) represents the power fluctuation feature in the i-th time window. The average power factor of the \(i\)th time window is denoted as \(\overline{C}(i)\), the maximum power factor of the \(i\)th time window is denoted as \(maxC(i)\), the minimum power factor of the \(i\)th time window is denoted as \(minC(i)\), the time interval length between the maximum power factor and the minimum power factor within the \(i\)th time window is denoted as \(h(maxC(i), minC(i))\), and \(e\) represents the base of the natural logarithm.

[0018] Furthermore, a multi-scale prediction model for power metering data characteristics is constructed. The specific method is as follows:

[0019] Based on the active power change of the power system in historical time windows, the load disturbance factor \(f_h\) of the power system is defined;

[0020] Based on the fault-free working time of the power system in historical time windows, the fault disturbance factor \(g_z\) of the power system is defined;

[0021] Based on the key feature data of the power metering status in historical time windows, combined with the fault disturbance factor and load disturbance factor of the power system, using the formula represents the prediction model of the key features of the power metering status. Among them, \(i\) represents the \(i\)th time window, \(W\) represents the key feature data of the power metering status, that is, \(W(i)=\{I(i), U(i), C(i)\}\), represents the predicted value of the key feature data of the power metering status in the \((i + 1)\)th time window, \(k\) w represents the influence coefficient of the prediction of the key feature data of the power metering status, \(\{a_0, a_1, \cdots, a\) j \} represents the prediction model coefficients, \(m\) represents the order of the prediction model, \(g_z\) represents the fault disturbance factor of the power system, and \(f_h\) represents the load disturbance factor of the power system;

[0022] Set the prediction error threshold. When the error between the predicted value obtained from the prediction model and the actual value is greater than or equal to the prediction error threshold, it is necessary to adjust the prediction model coefficients, order, and weight coefficients until the error between the predicted value and the actual value is less than the error threshold.

[0023] Furthermore, based on the active power change of the power system in historical time windows, the load disturbance factor \(f_h\) of the power system is defined. The specific method is as follows:

[0024] Using the formula represents the load disturbance factor of the power system, where \(T\) represents the duration of a time window, \(n\) represents the total number of historical time windows for collecting the active power of the power system, \(i\) represents the \(i\)th time window, \(p_b\) represents the reference power value of the power system, \(p(i)\) represents the average active power value of the power system in the \(i\)th time window, and \(\eta\) represents the time decay coefficient.

[0025] Further, based on the fault-free operation time of the power system in the historical time window, a fault disturbance factor gz of the power system is defined. The specific method is as follows:

[0026] Using the formula represents the fault disturbance factor of the power system, where T represents the duration of a time window, n represents the total number of historical time windows for collecting the active power of the power system, i represents the i-th time window, sj(i) represents the fault operation time of the power system in time window i, cs(i) represents the number of fault emissions of the power system in time window i, k1 represents the fault degree weight coefficient, and k2 represents the fault frequency weight coefficient.

[0027] Further, the present invention also provides a power metering abnormal data identification system based on the Power HarmonyOS, which is applied to the power metering abnormal data identification method based on the Power HarmonyOS, and includes:

[0028] A power metering feature extraction module, configured to collect power metering-related data of the power system in real time through the Power HarmonyOS and extract the data features of the original power metering data;

[0029] A power metering data prediction module, configured to construct a multi-scale prediction model for power metering data features based on the operation characteristics of the power system and the laws of historical data;

[0030] A power metering abnormal identification module, configured to compare the predicted value output by the prediction model with the actual observed value, and combine the preset abnormal judgment rules to identify the abnormal part in the power metering data.

[0031] In summary, due to the adoption of the above technical solutions, the beneficial effects of the present invention are as follows:

[0032] 1. The present invention provides an important basis for the abnormal detection and fault warning of the power system through key data such as the extracted voltage fluctuation characteristics, current fluctuation characteristics, and power fluctuation characteristics. These characteristic data can reflect the operation state and change trend of the power system. By applying technologies such as real-time data acquisition and processing, feature extraction, and abnormal detection, the present invention not only improves the accuracy and reliability of power metering data but also provides strong support for the intelligent and digital transformation of the power system;

[0033] 2. By comprehensively considering the operating characteristics of the power system and the laws of historical data, the multi-scale prediction model constructed by the present invention can more accurately capture the changing trend of electric energy metering data. This model not only considers the load disturbance factor, but also incorporates the fault disturbance factor, thereby improving the comprehensiveness and accuracy of the prediction. By comparing the predicted value output by the prediction model with the actual observed value and combining the preset abnormal judgment rules, the abnormal part in the electric energy metering data can be efficiently identified. This method not only helps to improve the efficiency of abnormal detection, but also enhances the accuracy of detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0035] Figure 1 Shows the flowchart of the method for identifying abnormal electric energy metering data based on the Power HarmonyOS system;

[0036] Figure 2 Shows the module diagram of the system for identifying abnormal electric energy metering data based on the Power HarmonyOS system. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0037] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0038] Embodiment 1. As Figure 1 shown, the method for identifying abnormal electric energy metering data based on the Power HarmonyOS system specifically includes the following steps:

[0039] Step 1. Based on the Power HarmonyOS system, collect the data related to electric energy metering of the power system in real time, and extract the data characteristics of the original electric energy metering data.

[0040] Deploy an intelligent acquisition unit at the edge computing node of the Power Harmony system. According to the real-time data stream of the Power Harmony system, dynamically divide time windows. The length T of the time window is set according to actual requirements. Within each time window, collect the power metering-related data of the power system in real time through the distributed data bus. The power metering-related data includes the instantaneous values of grid voltage, current, and power factor. Denote the power metering-related data within one time window as W(t) = {V(t), I(t), C(t)}, where t ∈ [t0, t0 + T]. Here, V(t) represents the instantaneous voltage value at time t, I(t) represents the instantaneous current value at time t, C(t) represents the power factor at time t, t0 represents the start time of the time window, and T represents the length of the time window. To ensure the accuracy and consistency of the data, perform preprocessing on the collected raw data. The preprocessing steps include data cleaning, denoising, and normalization processing;

[0041] Extract features from the power metering-related data within each time window. From the instantaneous values of voltage, current, and power factor, extract the key feature data of the power metering state, including voltage fluctuation features, current fluctuation features, and power fluctuation features;

[0042] Voltage fluctuation features:

[0043]

[0044] Among them, i represents the i-th time window, t represents time t, uV(i) represents the voltage fluctuation feature in the i-th time window, t0 represents the start time of the time window, T represents the length of the time window, and V(i, t) represents the instantaneous voltage value at time t in the i-th time window. represents the rated voltage value of the i-th time window, ΔmaxV(i) represents the maximum voltage fluctuation in the i-th time window, that is, the difference between the maximum instantaneous voltage value and the minimum instantaneous voltage value within the i-th time window, and a is a constant value.

[0045] Current fluctuation features:

[0046]

[0047] Among them, i represents the i-th time window, t represents time t, uI(i) represents the current fluctuation feature in the i-th time window, t0 represents the start time of the time window, T represents the length of the time window, and I(i, t) represents the instantaneous current value at time t in the i-th time window. represents the average current value of the i-th time window, ΔmaxI(i) represents the maximum current fluctuation in the i-th time window, that is, the difference between the maximum instantaneous current value and the minimum instantaneous current value within the i-th time window, and a is a constant value.

[0048] Power fluctuation characteristics:

[0049]

[0050] Among them, i represents the i-th time window, t represents time t, and uC(i) represents the power fluctuation characteristics in the i-th time window. represents the average power factor in the i-th time window, maxC(i) represents the maximum power factor in the i-th time window, minC(i) represents the minimum power factor in the i-th time window, h(maxC(i), minC(i)) represents the time interval length between the maximum power factor and the minimum power factor in the i-th time window, and e represents the base of the natural logarithm.

[0051] Step 2: Based on the operating characteristics of the power system and the historical data law, construct a multi-scale prediction model for the characteristics of electric energy metering data.

[0052] Based on the key factors affecting power fluctuations in the power system, define the disturbance factor affecting power fluctuations. Based on the active power change of the power system in the historical time window, define the load disturbance factor of the power system. The specific formula is as follows:

[0053]

[0054] Among them, fh represents the load disturbance factor of the power system, T represents the duration of a time window, n represents the total number of historical time windows for collecting the active power of the power system, i represents the i-th time window, pb represents the reference power value of the power system, p(i) represents the average active power value of the power system in the i-th time window, η represents the time decay coefficient, and in this embodiment, η is set to 0.08.

[0055] Based on the fault-free working time of the power system in the historical time window, define the fault disturbance factor of the power system. The specific formula is as follows:

[0056]

[0057] Among them, gz represents the fault disturbance factor of the power system, T represents the duration of a time window, n represents the total number of historical time windows for collecting the active power of the power system, i represents the i-th time window, sj(i) represents the fault operation time of the power system in time window i, cs(i) represents the number of fault emissions of the power system in time window i, k1 represents the fault degree weight coefficient, and k2 represents the fault frequency weight coefficient. In this embodiment, k1 is set to 0.7 and k2 is set to 0.3.

[0058] Based on the key feature data of the power metering status in the historical time window, combined with the fault disturbance factor and load disturbance factor of the power system, a prediction model for the key features of the power metering status is established. The specific formula is as follows:

[0059]

[0060] Among them, i represents the i-th time window, W represents the key feature data of the power metering status, that is, W(i) = {I(i), U(i), C(i)}, represents the predicted value of the key feature data of the power metering status in the (i + 1)-th time window, k w represents the influence coefficient of the prediction of the key feature data of the power metering status, {a0, a1,..., a j} represents the prediction model coefficients, m represents the order of the prediction model, gz represents the fault disturbance factor of the power system, and fh represents the load disturbance factor of the power system;

[0061] Set the prediction error threshold. When the error between the predicted value obtained from the prediction model and the actual value is greater than or equal to the prediction error threshold, it is necessary to adjust the prediction model coefficients, order, and weight coefficients until the error between the predicted value and the actual value is less than the error threshold.

[0062] Step 3: Compare the predicted value output by the prediction model with the actual observed value, and combine the preset anomaly judgment rules to identify the abnormal part in the power metering data.

[0063] For each time window i, compare the predicted values of the key feature data of the power metering status output by the prediction model, including the predicted values of voltage fluctuation, current fluctuation, and power fluctuation characteristics, with the actual observed values, calculate the absolute error between the predicted value and the actual value, and set different absolute error thresholds for the voltage fluctuation, current fluctuation, and power fluctuation characteristic data according to the actual situation of the power system and the requirements of anomaly detection. When the error between the predicted value and the actual value of the key feature data of the power metering status is greater than or equal to the corresponding threshold, it is determined that the power metering data in this time window is abnormal, record the time, location, and anomaly type information of the abnormal data, construct an abnormal event graph based on the intelligent diagnosis of the Power Harmony knowledge graph, and use the graph structure to represent the equipment, anomaly type, environmental parameters in the power system and their relationships. The node set V includes equipment, anomaly type, environmental parameters, etc., and the edge set E includes the inducing relationship and the time sequence relationship. Among them, the anomaly type refers to the cause of the anomaly, such as equipment failure, data transmission error, external interference. Based on the constructed abnormal event graph, use graph neural network or other graph processing technologies for intelligent diagnosis to further analyze the cause of the anomaly and identify potential faulty equipment or system problems.

[0064] Example 2, such asFigure 2 The power metering abnormal data identification system based on the Power HarmonyOS shown in the figure specifically includes the following:

[0065] Power metering feature extraction module: Deploy intelligent acquisition units at the edge computing nodes of the Power HarmonyOS. According to the real-time data stream of the Power HarmonyOS, dynamically divide time windows. The length T of the time window is set according to actual requirements. In each time window, collect power metering-related data of the power system in real time through the distributed data bus. The power metering-related data includes the instantaneous values of grid voltage, current, and power factor. Denote the power metering-related data in one time window as W(t) = {V(t), I(t), C(t)}, t ∈ [t0, t0 + T], where V(t) represents the instantaneous voltage value at time t, I(t) represents the instantaneous current value at time t, C(t) represents the power factor at time t, t0 represents the start time of the time window, and T represents the length of the time window. To ensure the accuracy and consistency of the data, preprocess the collected raw data. The preprocessing steps include data cleaning, denoising, and normalization processing;

[0066] Extract features from the power metering-related data in each time window, and extract key feature data of the power metering state from the instantaneous values of voltage, current, and power factor, including voltage fluctuation features, current fluctuation features, and power fluctuation features;

[0067] Voltage fluctuation features:

[0068]

[0069] Among them, i represents the i-th time window, t represents time t, uV(i) represents the voltage fluctuation feature in the i-th time window, t0 represents the start time of the time window, T represents the length of the time window, and V(i, t) represents the instantaneous voltage value at time t in the i-th time window, represents the rated voltage value of the i-th time window, ΔmaxV(i) represents the maximum voltage fluctuation in the i-th time window, that is, the difference between the maximum instantaneous voltage value and the minimum instantaneous voltage value in the i-th time window, and a is a constant value.

[0070] Current fluctuation features:

[0071]

[0072] Among them, i represents the i-th time window, t represents time t, uI(i) represents the current fluctuation feature in the i-th time window, t0 represents the start time of the time window, T represents the length of the time window, and I(i, t) represents the instantaneous current value at time t in the i-th time window, represents the average current value of the i-th time window, and ΔmaxI(i) represents the maximum current fluctuation value of the i-th time window, that is, the difference between the maximum instantaneous current value and the minimum instantaneous current value within the i-th time window, and a is a constant value.

[0073] Power fluctuation characteristics:

[0074]

[0075] Among them, i represents the i-th time window, t represents time t, and uC(i) represents the power fluctuation characteristics in the i-th time window. represents the average power factor value of the i-th time window, maxC(i) represents the maximum power factor value of the i-th time window, minC(i) represents the minimum power factor value of the i-th time window, h(maxC(i), minC(i)) represents the time interval length between the maximum power factor value and the minimum power factor value within the i-th time window, and e represents the base of the natural logarithm.

[0076] The electric energy metering data prediction module, based on the key factors affecting power fluctuations in the power system, defines the disturbance factors affecting power fluctuations, and based on the active power changes of the power system in historical time windows, defines the load disturbance factor of the power system. The specific formula is as follows:

[0077]

[0078] Among them, fh represents the load disturbance factor of the power system, T represents the duration of a time window, n represents the total number of historical time windows for collecting the active power of the power system, i represents the i-th time window, pb represents the reference power value of the power system, p(i) represents the average active power value of the power system in the i-th time window, η represents the time decay coefficient, and in this embodiment, η is set to 0.08.

[0079] Based on the fault-free working time of the power system in historical time windows, the fault disturbance factor of the power system is defined. The specific formula is as follows:

[0080]

[0081] Among them, gz represents the fault disturbance factor of the power system, T represents the duration of a time window, n represents the total number of historical time windows for collecting the active power of the power system, i represents the i-th time window, sj(i) represents the fault operation time of the power system in time window i, cs(i) represents the number of fault emissions of the power system in time window i, k1 represents the fault degree weight coefficient, and k2 represents the fault frequency weight coefficient. In this embodiment, k1 is set to 0.7 and k2 is set to 0.3.

[0082] Based on the key feature data of the power metering status in the historical time window, combined with the fault disturbance factor and load disturbance factor of the power system, a prediction model for the key features of the power metering status is established. The specific formula is as follows:

[0083]

[0084] Among them, i represents the i-th time window, W represents the key feature data of the power metering status, that is, W(i) = {I(i), U(i), C(i)}, represents the predicted value of the key feature data of the power metering status in the (i + 1)-th time window, k w represents the influence coefficient of the prediction of the key feature data of the power metering status, {a0, a1,..., a j} represents the prediction model coefficients, m represents the order of the prediction model, gz represents the fault disturbance factor of the power system, and fh represents the load disturbance factor of the power system;

[0085] Set the prediction error threshold. When the error between the predicted value and the actual value obtained from the prediction model is greater than or equal to the prediction error threshold, it is necessary to adjust the prediction model coefficients, order, and weight coefficients until the error between the predicted value and the actual value is less than the error threshold.

[0086] Power metering anomaly identification module. For each time window i, compare the predicted values of the key feature data of the power metering status output by the prediction model, including the predicted values of voltage fluctuation, current fluctuation, and power fluctuation characteristics, with the actual observed values, calculate the absolute error between the predicted value and the actual value, and set different absolute error thresholds for the voltage fluctuation, current fluctuation, and power fluctuation characteristic data according to the actual situation of the power system and the requirements of anomaly detection. When the error between the predicted value and the actual value of the key feature data of the power metering status is greater than or equal to the corresponding threshold, it is determined that the power metering data in this time window is abnormal, record the time, location, and anomaly type information of the abnormal data, construct an abnormal event graph based on the intelligent diagnosis of the Power Harmony knowledge graph, and use the graph structure to represent the equipment, anomaly types, environmental parameters, and their relationships in the power system. The node set V includes equipment, anomaly types, environmental parameters, etc., and the edge set E includes the inducing relationship and the time sequence relationship. Among them, the anomaly type refers to the cause of the anomaly, such as equipment failure, data transmission error, external interference. Based on the constructed abnormal event graph, use graph neural networks or other graph processing technologies for intelligent diagnosis, further analyze the cause of the anomaly, and identify potential faulty equipment or system problems.

[0087] As described above, it is only the preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, making equivalent replacements or changes, shall be covered by the protection scope of the present invention.

[0088] The preferred embodiments of the present invention disclosed above are only used to help illustrate the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the present invention to only the specific embodiments. Obviously, according to the content of this specification, many modifications and changes can be made. This specification selects and specifically describes these embodiments in order to better explain the principle and practical application of the present invention, so that those skilled in the art can well understand and utilize the present invention. The present invention is only limited by the claims and their full scope and equivalents.

Claims

1. The method for identifying abnormal data of electric energy metering based on the electric power Hongmeng system is characterized by: include: Step 1: Based on the Hongmeng Electric Power System, the power system's electric energy metering-related data is collected in real time, and the data features of the original electric energy metering data are extracted; The data related to electric energy metering include the instantaneous values ​​of grid voltage, current and power factor. The features of the data related to electric energy metering in each time window are extracted, and the key feature data of the electric energy metering state are extracted from the instantaneous values ​​of voltage, current and power factor. Step 2: Based on the operation characteristics of the power system and the rules of historical data, a multi-scale prediction model for the characteristics of electric energy metering data is constructed; Step 3: Compare the predicted value output by the prediction model with the actual observed value, and identify the abnormal part in the electric energy metering data in combination with the preset abnormal judgment rules.

2. According to the method for identifying abnormal data of electric energy metering based on the electric power Hongmeng system according to claim 1, it is characterized in that: From the instantaneous value of voltage, extract the key characteristic data of the electric energy metering state. The specific method is as follows: Using the formula represents the voltage fluctuation characteristics, where i represents the i-th time window, t represents time t, uV(i) represents the voltage fluctuation characteristics in the i-th time window, t0 represents the start time of the time window, T represents the length of the time window, V(i, t) represents the instantaneous value of the voltage in the i-th time window at time t, represents the rated voltage value in the i-th time window, ΔmaxV(i) represents the maximum voltage fluctuation in the i-th time window, that is, the difference between the maximum instantaneous voltage value and the minimum instantaneous voltage value in the i-th time window, and a is a constant value.

3. The method for identifying abnormal data of electric energy metering based on the electric power Hongmeng system according to claim 1 is characterized in that: From the instantaneous value of the current, extract the key characteristic data of the electric energy metering state. The specific method is as follows: Using the formula represents the current fluctuation characteristics, where i represents the i-th time window, t represents time t, uI(i) represents the current fluctuation characteristics in the i-th time window, t0 represents the start time of the time window, T represents the length of the time window, and I(i, t) represents the instantaneous value of the current in the i-th time window at time t. represents the average current value in the i-th time window, ΔmaxI(i) represents the maximum current fluctuation in the i-th time window, that is, the difference between the maximum instantaneous current value and the minimum instantaneous current value in the i-th time window, and a is a constant value.

4. The method for identifying abnormal data of electric energy metering based on the electric power Hongmeng system according to claim 1 is characterized in that: From the instantaneous value of the power factor, the key characteristic data of the electric energy metering state is extracted. The specific method is as follows: Using the formula represents the power fluctuation characteristics, where i represents the i-th time window, t represents time t, and uC(i) represents the power fluctuation characteristics in the i-th time window. represents the average value of the power factor in the i-th time window, maxC(i) represents the maximum value of the power factor in the i-th time window, minC(i) represents the minimum value of the power factor in the i-th time window, h(maxC(i), minC(i)) represents the time interval between the maximum value of the power factor and the minimum value of the power factor in the i-th time window, and e represents the base of the logarithm of the natural number.

5. The method for identifying abnormal data of electric energy metering based on the electric power Hongmeng system according to claim 1 is characterized in that: Construct a multi-scale prediction model for electric energy metering data features. The specific method is as follows: Based on the active power change of the power system in the historical time window, the load disturbance factor fh of the power system is defined; Based on the fault-free working time of the power system in the historical time window, the fault disturbance factor gz of the power system is defined; Based on the key characteristic data of the electric energy metering state in the historical time window, combined with the fault disturbance factor and load disturbance factor of the power system, the formula is used A prediction model for the key features of the electric energy metering state, wherein i represents the i-th time window, W represents the key feature data of the electric energy metering state, that is, W(i) = {I(i), U(i), C(i)}, The predicted value of the key characteristic data representing the energy metering status in the i+1th time window, k w The predicted influence coefficient of the key characteristic data representing the state of electric energy metering, {a0, a1, …, a j } represents the prediction model coefficient, m represents the prediction model order, gz represents the fault disturbance factor of the power system, and fh represents the load disturbance factor of the power system; Set a prediction error threshold. When the error between the predicted value and the actual value obtained by the prediction model is greater than or equal to the prediction error threshold, it is necessary to adjust the prediction model coefficient, order and weight coefficient until the error between the predicted value and the actual value is less than the error threshold.

6. The method for identifying abnormal data of electric energy metering based on the electric power Hongmeng system according to claim 5 is characterized in that: Based on the active power change of the power system in the historical time window, the load disturbance factor fh of the power system is defined. The specific method is: Using the formula represents the load disturbance factor of the power system, where T represents the duration of a time window, n represents the total number of historical time windows for collecting the active power of the power system, i represents the i-th time window, pb represents the benchmark power value of the power system, p(i) represents the average active power value of the power system in the i-th time window, and η represents the time attenuation coefficient.

7. The method for identifying abnormal data of electric energy metering based on the electric power Hongmeng system according to claim 1 is characterized in that: Based on the fault-free working time of the power system in the historical time window, the fault disturbance factor gz of the power system is defined. The specific method is: Using the formula represents the fault disturbance factor of the power system, where T represents the length of a time window, n represents the total number of historical time windows for collecting the active power of the power system, i represents the i-th time window, sj(i) represents the fault operation time of the power system in time window i, cs(i) represents the number of fault transmissions of the power system in time window i, k1 represents the fault degree weight coefficient, and k2 represents the fault frequency weight coefficient.

8. The electric energy metering abnormal data identification system based on the electric power Hongmeng system is applied to the electric energy metering abnormal data identification method based on the electric power Hongmeng system according to any one of claims 1 to 7, characterized in that: include: The electric energy metering feature extraction module is used to collect the electric energy metering related data of the power system in real time through the power Hongmeng system and extract the data features of the original electric energy metering data; The electric energy metering data prediction module is used to build a multi-scale prediction model for electric energy metering data characteristics based on the power system operation characteristics and historical data rules; The electric energy metering anomaly identification module is used to compare the predicted value output by the prediction model with the actual observed value, and identify the abnormal part in the electric energy metering data in combination with the preset anomaly judgment rules.

Images