An electric quantity data automatic acquisition system

Abstract

This invention provides an automatic power data acquisition system, relating to the field of data acquisition technology. It includes a data acquisition module for acquiring data and converting the acquired data signals; a data processing module for preprocessing the data; an anomaly identification module for identifying anomalies in the target real-time data using a power prediction model trained on historical target data, and marking the anomaly data with a first label; and a data transmission module for transmitting the target real-time data and the first-labeled data to a control terminal for correction and fault diagnosis of the first-labeled data. By converting the acquired real-time data to obtain first data, and using the preprocessed historical data to train a model to obtain a power prediction model, the system predicts the preprocessed first data, thereby achieving online identification of anomaly data and correction based on actual power data. This ensures intelligent data acquisition while reducing losses caused by anomalies.

Description

Technical Field

[0001] This invention relates to the field of data acquisition technology, and in particular to an automatic power data acquisition system. Background Technology

[0002] Electricity consumption data is an important parameter that can be used to determine the working status and perform statistical analysis. Currently, electricity consumption data collection systems have basically achieved full coverage and can provide data support for various businesses.

[0003] However, while the electricity market is developing rapidly, it also faces the problem of frequently collecting abnormal electricity data due to the huge amount of data required, external interference, or unstable communication channels. This results in huge economic losses and hinders the safety of the power system. Therefore, it is particularly important to realize real-time intelligent collection of electricity data while accurately identifying abnormal data and taking corresponding measures.

[0004] Therefore, the present invention provides an automatic power data acquisition system. Summary of the Invention

[0005] This invention provides an automatic power data acquisition system, which converts real-time acquired data to obtain first data, trains a model using preprocessed historical data to obtain a power prediction model, and predicts the preprocessed first data, thereby realizing online identification of abnormal data and correcting it based on actual power data, ensuring intelligent data acquisition while reducing losses caused by abnormal data.

[0006] This invention provides an automatic power data acquisition system, comprising:

[0007] Data acquisition module: used to acquire target parameter data in real time using several power calculation devices, convert the target parameter data into signals, obtain first data, and store it;

[0008] Data processing module: used to preprocess the historical electricity data of the preset amount extracted from the electricity database and the first data respectively to obtain the target historical data and the target real-time data;

[0009] Anomaly detection module: Used to detect anomalies in the target real-time data by using the power prediction model obtained by training the model with the target historical data. If there are anomalies in the target real-time data, the abnormal data is marked as the first anomaly.

[0010] Data transmission module: used to transmit the target real-time data and the first marker data to the control terminal to correct and troubleshoot the first marker data.

[0011] Preferably, the first data is data in which the target parameter data is converted from analog signals to digital signals.

[0012] Preferably, the data processing module includes:

[0013] Data cleaning unit: used to sort the first data and historical electricity data in time sequence, and then calculate and judge the attribute similarity between two adjacent data in the first data and the historical electricity data. If the similarity exceeds the preset value, it is determined that the two adjacent data are duplicates, and at this time, one duplicate data is deleted.

[0014] Data denoising unit: It is used to smooth the first data and historical power data after cleaning to achieve denoising, and then normalize the denoised data to finally obtain the target real-time data and target historical data.

[0015] Preferably, the anomaly detection module includes:

[0016] Model training unit: used to perform preliminary point anomaly judgment by taking any two-thirds of the target's historical data as the first key data and obtaining the first anomalous data point;

[0017] The average value of the power consumption data before and after the first abnormal data point is obtained to correct the first abnormal data point and obtain the first corrected data.

[0018] By combining the first set of corrected data, a training dataset is obtained;

[0019] A power consumption prediction model is obtained by training the power consumption trend component, seasonal component, and the influence of external factors on power consumption, combined with the training dataset.

[0020] Anomaly threshold determination unit: used to manually annotate the target historical data other than the first key data to obtain the second key data, and to determine the anomaly threshold based on the anomaly identification accuracy and false negative rate of the second key data;

[0021] The formula for calculating the anomaly threshold is as follows:

[0022]

[0023]

[0024]

[0025] Where Y represents the anomaly threshold; P1 represents the accuracy of identifying anomaly data; P2 represents the false negative rate of identifying anomaly data; ω1 represents the contribution factor of accuracy to the calculation of the anomaly threshold; ω2 represents the contribution factor of false negative rate to the calculation of the anomaly threshold; MTP This represents the number of data points that were identified as outliers; M FP This represents the number of normal data points that were identified as outliers; M FN This represents the number of data points that were initially considered abnormal but were later determined to be normal.

[0026] Anomaly identification unit: Used to identify anomalies in the results obtained from real-time data of the target predicted by the power prediction model based on anomaly threshold.

[0027] Preferably, the anomaly detection unit includes:

[0028] Prediction block: Based on the target real-time data before time t, it uses the power prediction model to predict the probability distribution of power consumption at time t and obtain the first prediction result;

[0029] Identification block: Used to compare the electricity consumption collected at time t with the first prediction result to obtain the abnormal deviation value. The formula for calculating the abnormal deviation value is as follows:

[0030]

[0031] Among them, Q t The abnormal deviation value at time t is represented by d. t This represents the actual electricity consumption collected at time t; Let F represent the mean of the predicted power probability distribution at time t; let F represent the variance; and let δ represent the calculated loss coefficient in the process of obtaining the abnormal deviation value.

[0032] If the abnormal deviation value is greater than the abnormal threshold, the electricity data is determined to be abnormal data and marked as abnormal, thus obtaining the first marked data.

[0033] Preferably, the data transmission module includes:

[0034] Data transmission unit: used to transmit real-time target data and the first marker data to the control terminal via the communication transmission channel;

[0035] Anomaly correction unit: used to correct the first marked data based on the daily electricity consumption curve corresponding to the target historical data in the same season as the first marked data;

[0036] Response unit: Used to judge the correction effect of the first correction data. If the correction effect value is less than the preset threshold, the first correction data is marked to obtain the second marked data, and the power calculation device corresponding to the second marked data is set to disabled for manual emergency fault repair.

[0037] Preferably, the anomaly correction unit includes:

[0038] Decision block: Used to determine the user type and seasonal attribute of the first labeled data;

[0039] Calculation block: Used to calculate the first seasonal daily electricity consumption curve L=l using the extracted target historical data of the same period based on the user type and seasonal attributes of the first labeled data. (i) , i = 1, 2, ..., n, where l (i) This is represented as the electricity consumption value at the i-th moment corresponding to the daily electricity consumption curve;

[0040] Comparison block: Used to set the remaining normal time C=c for the day containing the first marked data. (j) j = 1, 2, ..., m, where c (j) This is represented by the electricity consumption value corresponding to the j-th remaining normal time, and compared with the electricity consumption curve L=l of the first seasonal day. (i) The electricity consumption values ​​of i = 1, 2, ..., n are obtained by unquantizing the values ​​and obtaining C′ = c′. (j) j = 1, 2, ... m and L′ = l′ (i) For i = 1, 2, ... m, perform correlation analysis to obtain several correlation degrees;

[0041] The first similarity is obtained by averaging all the correlation scores.

[0042] Correction block: Used to weight the electricity consumption value corresponding to the time when the first marker data appears in the first seasonal daily electricity consumption curve using the first similarity to obtain the correction value and replace the corresponding first marker data as the first correction data output.

[0043] Preferably, the reaction unit comprises:

[0044] Effect judgment block: used to judge the effect of the first correction data by calculating the correction effect value of the first correction data;

[0045] The formula for calculating the correction effect value is as follows:

[0046]

[0047] Among them, X (k) Let x represent the correction effect value of the k-th first correction data; let a represent the total number of first correction data; x k This is represented by the kth actual electricity consumption data value; z k Let θ1 represent the k-th correction value; θ2 represent the weighting coefficient of the influence of the root mean square error on the correction effect; and θ3 represent the weighting coefficient of the influence of the average error on the correction effect, where θ1 + θ2 = 1 and θ1 > θ2.

[0048] Troubleshooting block: Used to analyze the obtained correction effect value. If it is less than the preset threshold, the corresponding first correction data is marked to obtain the second marked data. The corresponding power calculation device of the second marked data is obtained and disabled, and then emergency manual repair is performed.

[0049] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings.

[0050] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0051] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0052] Figure 1 This is a structural diagram of an automatic power data acquisition system according to an embodiment of the present invention. Detailed Implementation

[0053] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0054] This invention provides an automatic power data acquisition system, such as... Figure 1 As shown, it includes:

[0055] Data acquisition module: used to acquire target parameter data in real time using several power calculation devices, convert the target parameter data into signals, obtain first data, and store it;

[0056] Data processing module: used to preprocess the historical electricity data of the preset amount extracted from the electricity database and the first data respectively to obtain the target historical data and the target real-time data;

[0057] Anomaly detection module: Used to detect anomalies in the target real-time data by using the power prediction model obtained by training the model with the target historical data. If there are anomalies in the target real-time data, the abnormal data is marked as the first anomaly.

[0058] Data transmission module: used to transmit the target real-time data and the first marker data to the control terminal to correct and troubleshoot the first marker data.

[0059] In this embodiment, the power calculation device is a device consisting of an energy meter, a voltage transformer, a current transformer, and secondary connecting wires, used to measure power generation and user power consumption; the target parameter data refers to the user's real-time power consumption data; the first data is the data obtained by converting the target parameter data from analog signals to digital signals.

[0060] In this embodiment, the power database mainly consists of historical power data, current power data, user information, and user behavior information; the preset amount is pre-set based on the amount of power data collected in real time.

[0061] In this embodiment, the target historical data and target real-time data are obtained by preprocessing the historical power consumption data and the first data, wherein the data preprocessing includes data cleaning, data denoising and data normalization.

[0062] In this embodiment, the power consumption prediction model is a model trained using historical target data, which takes into account the trend component, seasonal component and external influencing factors of power consumption. It is used to predict the probability distribution of power consumption data in the future in real time. Anomaly identification refers to comparing the actual collected power consumption with the results predicted by the power consumption prediction model to identify abnormal data. The first labeled data refers to the target real-time data that has been identified as abnormal data and labeled.

[0063] In this embodiment, the purpose of correcting the first marker data is to reduce the losses caused by abnormal data to a certain extent, which is conducive to maintaining the safety of the power system; the premise of fault repair is that after correcting the abnormal data, the power data still has a huge error compared with the normal data.

[0064] The beneficial effects of the above technical solution are: by converting the collected real-time data to obtain the first data, and using the preprocessed historical data to train the model to obtain the power prediction model, the preprocessed first data is used to predict, thereby realizing the online identification of abnormal data and correcting it based on the actual power data, ensuring intelligent data collection while reducing losses caused by abnormal data.

[0065] This invention provides an automatic power data acquisition system, wherein the data processing module includes:

[0066] Data cleaning unit: used to sort the first data and historical electricity data in time sequence, and then calculate and judge the attribute similarity between two adjacent data in the first data and the historical electricity data. If the similarity exceeds the preset value, it is determined that the two adjacent data are duplicates, and at this time, one duplicate data is deleted.

[0067] Data denoising unit: It is used to smooth the first data and historical power data after cleaning to achieve denoising, and then normalize the denoised data to finally obtain the target real-time data and target historical data.

[0068] In this embodiment, chronological arrangement refers to arranging data in chronological order; the attributes of the data include numerical attributes and label attributes, such as time and user name.

[0069] In this embodiment, the attribute similarity is obtained by first calculating the similarity between the numerical attribute values ​​and the label attribute values ​​between adjacent data, and then calculating the weighted average value according to the weight of the attributes; the preset similarity is set in advance, generally 0.85.

[0070] In this embodiment, for example, there are two pairs of adjacent data, data 1 and data 2, and data 3 and data 4; the corresponding attribute similarities are 0.9 and 0.65, respectively. At this time, it is determined that the adjacent data, data 1 and data 2, are duplicates, and one data is deleted arbitrarily.

[0071] In this embodiment, data denoising is used to reduce the negative impact of noisy data on subsequent data analysis and modeling; smoothing mainly refers to averaging the data and using the calculated average value as the value of all data; normalization is used to eliminate the influence of units; target real-time data refers to the data obtained after preprocessing the first data; target historical data is the data obtained after preprocessing historical electricity data, which is used to train the prediction model and to help correct the data.

[0072] The beneficial effects of the above technical solution are as follows: by judging the similarity of the attributes of two adjacent data in the first data and the historical power data, duplicate data is deleted to perform data cleaning; the cleaned first data and historical power data are denoised and normalized to obtain the target real-time data and the target historical data, which lays the data support for subsequent model training, abnormal data identification and data correction.

[0073] This invention provides an automatic power data acquisition system, wherein the anomaly identification module includes:

[0074] Model training unit: used to perform preliminary point anomaly judgment by taking any two-thirds of the target's historical data as the first key data and obtaining the first anomalous data point;

[0075] The average value of the power consumption data before and after the first abnormal data point is obtained to correct the first abnormal data point and obtain the first corrected data.

[0076] By combining the first set of corrected data, a training dataset is obtained;

[0077] A power consumption prediction model is obtained by training the power consumption trend component, seasonal component, and the influence of external factors on power consumption, combined with the training dataset.

[0078] Anomaly threshold determination unit: used to manually annotate the target historical data other than the first key data to obtain the second key data, and to determine the anomaly threshold based on the anomaly identification accuracy and false negative rate of the second key data;

[0079] The formula for calculating the anomaly threshold is as follows:

[0080]

[0081]

[0082]

[0083] Where Y represents the anomaly threshold; P1 represents the accuracy of identifying anomaly data; P2 represents the false negative rate of identifying anomaly data; ω1 represents the contribution factor of accuracy to the calculation of the anomaly threshold; ω2 represents the contribution factor of false negative rate to the calculation of the anomaly threshold; M TP This represents the number of data points that were identified as outliers; M FP This represents the number of normal data points that were identified as outliers; M FN This represents the number of data points that were initially considered abnormal but were later determined to be normal.

[0084] Anomaly identification unit: Used to identify anomalies in the results obtained from real-time data of the target predicted by the power prediction model based on anomaly threshold.

[0085] In this embodiment, the first key data refers to the target historical data that accounts for two-thirds; the point anomaly refers to a point anomaly value, which is the first abnormal data point, when there is a power value that is significantly different from the normal power value based on the global trend of the first key data.

[0086] In this embodiment, for example, there exists a first abnormal data point A with a corresponding power value of a1, and the power values ​​before and after data point A are a0 and a2, respectively. In this case, the first abnormal data point A is corrected to obtain the power value of the first corrected data A′.

[0087] In this embodiment, the training dataset consists of all the obtained first correction data; the trend component mainly refers to the upward slope and downward slope; the seasonal component includes the four seasons of spring, summer, autumn and winter; external influencing factors include factors such as temperature and holidays; the power prediction model is a model trained based on the target historical data, used to predict the probability distribution of power data at future moments in real time.

[0088] In this embodiment, the second key data refers to manually labeled abnormal data in the target historical data other than the first key data; the abnormal threshold is used to identify abnormal data; the abnormal identification accuracy rate refers to the proportion of correctly identified data among the data that is identified as abnormal; the false negative rate is the proportion of data that is identified as abnormal out of all abnormal data.

[0089] The beneficial effects of the above technical solution are: by using two-thirds of the target's historical data to train the model to obtain the power prediction model, and then using the abnormal data in the remaining one-third of the target's historical data to obtain the abnormal threshold, the combination of the two can realize the abnormal identification of the target's real-time data, improve the data quality and enhance the data identification performance.

[0090] This invention provides an automatic power data acquisition system, wherein the anomaly identification unit includes:

[0091] Prediction block: Based on the target real-time data before time t, it uses the power prediction model to predict the probability distribution of power consumption at time t and obtain the first prediction result;

[0092] Identification block: Used to compare the electricity consumption collected at time t with the first prediction result to obtain the abnormal deviation value. The formula for calculating the abnormal deviation value is as follows:

[0093]

[0094] Among them, Q t The abnormal deviation value at time t is represented by d. t This represents the actual electricity consumption collected at time t; Let F represent the mean of the predicted power probability distribution at time t; let F represent the variance; and let δ represent the calculated loss coefficient in the process of obtaining the abnormal deviation value.

[0095] If the abnormal deviation value is greater than the abnormal threshold, the electricity data is determined to be abnormal data and marked as abnormal, thus obtaining the first marked data.

[0096] In this embodiment, the first prediction result is obtained by using an electricity prediction model and is compared with the actual electricity consumption to determine whether there is an anomaly.

[0097] In this embodiment, for example, if the abnormal deviation between the power data B collected at time t1 and the prediction result is greater than the abnormal threshold, then the power data B is marked and output as the first marked data.

[0098] The beneficial effects of the above technical solution are: by comparing the prediction result of the current time using the power consumption prediction model with the actual power consumption data at the current time, the abnormal deviation value is obtained, and the abnormal threshold is used to determine whether there is an anomaly in the data collected at the current time, thus effectively realizing the accurate identification of abnormal data.

[0099] This invention provides an automatic power data acquisition system, wherein the data transmission module includes:

[0100] Data transmission unit: used to transmit real-time target data and the first marker data to the control terminal via the communication transmission channel;

[0101] Anomaly correction unit: used to correct the first marked data based on the daily electricity consumption curve corresponding to the target historical data in the same season as the first marked data;

[0102] Response unit: Used to judge the correction effect of the first correction data. If the correction effect value is less than the preset threshold, the first correction data is marked to obtain the second marked data, and the power calculation device corresponding to the second marked data is set to disabled for manual emergency fault repair.

[0103] In this embodiment, the communication transmission channel is used to transmit the collected data to the control terminal; the control terminal is used for; the daily electricity consumption curve refers to a curve that shows the trend of electricity consumption changes on a given day.

[0104] In this embodiment, the first corrected data is obtained by averaging the similarity values ​​obtained by correlation analysis between the seasonal daily electricity consumption curve of the target historical data and the electricity consumption values ​​corresponding to the remaining normal time of the day in which the first marked data is located, and then performing weighted processing to obtain the corresponding corrected value to replace the first marked data.

[0105] In this embodiment, the correction effect refers to the degree of closeness between the first corrected data and the normal real data; the preset threshold is set in advance.

[0106] In this embodiment, for example, there are first correction data C1 and C2, and the corresponding correction effect values ​​are all less than a preset threshold. At this time, the first correction data C1 and C2 are marked and output as second marked data. The power calculation device corresponding to the first correction data C1 and C2 is disabled, and manual fault repair is arranged.

[0107] The beneficial effects of the above technical solution are as follows: by transmitting the real-time data of the target and the first marker to the control terminal through the communication transmission channel, and then correcting the first marker data based on the historical data of the target, and taking targeted measures according to the correction effect, the real-time intelligent data acquisition is guaranteed, while accurately identifying abnormal data and responding to it, thus ensuring data quality, improving the safety of the power system to a certain extent, and minimizing economic losses.

[0108] This invention provides an automatic power data acquisition system, wherein the anomaly correction unit includes:

[0109] Decision block: Used to determine the user type and seasonal attribute of the first labeled data;

[0110] Calculation block: Used to calculate the first seasonal daily electricity consumption curve L=l using the extracted target historical data of the same period based on the user type and seasonal attributes of the first labeled data. (i) , i = 1, 2, ..., n, where l Ii) This is represented as the electricity consumption value at the i-th moment corresponding to the daily electricity consumption curve;

[0111] Comparison block: Used to set the remaining normal time C=c for the day containing the first marked data. (j) j = 1, 2, ..., m, where c (j) This is represented by the electricity consumption value corresponding to the j-th remaining normal time, and compared with the electricity consumption curve L=l of the first seasonal day. (i) The electricity consumption values ​​of i = 1, 2, ..., n are obtained by unquantizing the values ​​and obtaining C′ = c′. (j) j = 1, 2, ... m and L′ = l′ (i) For i = 1, 2, ... m, perform correlation analysis to obtain several correlation degrees;

[0112] The first similarity is obtained by averaging all the correlation scores.

[0113] Correction block: Used to weight the electricity consumption value corresponding to the time when the first marker data appears in the first seasonal daily electricity consumption curve using the first similarity to obtain the correction value and replace the corresponding first marker data as the first correction data output.

[0114] In this embodiment, user types include residents and non-residents, such as individual business owners in shopping malls; seasonal attributes mainly refer to the four seasons of spring, summer, autumn and winter; the first seasonal daily electricity consumption curve is mainly used to describe the trend of daily electricity consumption in different seasons, such as the typical electricity consumption curve trend in autumn and winter is more different, and electricity consumption is the greatest in winter, with the daily electricity consumption curve having a bimodal characteristic; non-quantification processing refers to eliminating the negative impact on subsequent data analysis caused by different units and magnitudes; correlation refers to the degree of closeness at each time point.

[0115] The beneficial effects of the above technical solution are as follows: by performing correlation analysis on the seasonal daily electricity consumption curve of the target historical data and the electricity consumption value of the remaining normal time of the day where the first marked data is located without quantification, a first similarity is obtained; by using the first similarity to weight the electricity consumption value to obtain a correction value to replace the corresponding first marked data, the abnormal data can be effectively corrected and the data quality can be improved.

[0116] This invention provides an automatic power data acquisition system, wherein the reaction unit includes:

[0117] Effect judgment block: used to judge the effect of the first correction data by calculating the correction effect value of the first correction data;

[0118] The formula for calculating the correction effect value is as follows:

[0119]

[0120] Among them, X (k) Let x represent the correction effect value of the k-th first correction data; let a represent the total number of first correction data; x k This is represented by the kth actual electricity consumption data value; z k Let θ1 represent the k-th correction value; θ2 represent the weighting coefficient of the influence of the root mean square error on the correction effect; and θ3 represent the weighting coefficient of the influence of the average error on the correction effect, where θ1 + θ2 = 1 and θ1 > θ2.

[0121] Troubleshooting block: Used to analyze the obtained correction effect value. If it is less than the preset threshold, the corresponding first correction data is marked to obtain the second marked data. The corresponding power calculation device of the second marked data is obtained and disabled, and then emergency manual repair is performed.

[0122] In this embodiment, for example, there are first correction data C1 and C2, and the corresponding correction effect values ​​are all less than a preset threshold. At this time, the first correction data C1 and C2 are marked and output as second marked data. The power calculation device corresponding to the first correction data C1 and C2 is disabled, and manual fault repair is arranged.

[0123] The beneficial effects of the above technical solution are: by comparing the correction effect value obtained by the formula with the preset threshold, the first correction data that does not meet the preset effect is obtained and the corresponding power calculation device is disabled. Then, manual maintenance is arranged, which effectively ensures the data quality, maintains the safety of the power system, and reduces the possibility of economic losses to a certain extent.

[0124] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. An automatic power data collection system, characterized by, The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. wherein Y represents an abnormality threshold value; P1 represents an accuracy rate of identifying abnormal data; P2 represents a missed detection rate of identifying abnormal data; ω1 represents a contribution factor of the accuracy rate to the abnormality threshold value; ω2 represents a contribution factor of the missed detection rate to the abnormality threshold value; M TP represents the number of data points determined as abnormal points; M FP represents the number of data points determined as normal points; M FN represents the number of data points determined as normal points; The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring.

2. The automatic electric quantity data acquisition system according to claim 1, wherein The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring.

3. The automatic electric quantity data acquisition system of claim 1, wherein, The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring.

4. The automatic electric quantity data acquisition system of claim 1, wherein The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. wherein Q t represents the abnormal deviation value at time t; d t represents the actual power consumption collected at time t; represents the mean value of the predicted power consumption probability distribution at time t; F represents the variance; and δ represents the calculation loss coefficient in the process of obtaining the abnormal deviation value. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring.

5. The automatic electric quantity data acquisition system of claim 1, wherein, The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs to the field of electric quantity data real-time monitoring. The utility model relates to an electric quantity data real-time monitoring method and device, and belongs The data transmission unit is configured to transmit the target real-time data and the first marked data to the control terminal based on a communication transmission channel. The anomaly correction unit is configured to correct the first marked data based on a daily power consumption curve corresponding to the target historical data in the same season as the first marked data. The reaction unit is configured to judge the correction effect of the first corrected data, and if the correction effect value is less than a preset threshold, mark the first corrected data to obtain second marked data, set the power consumption calculation device state corresponding to the second marked data as disabled, and perform manual emergency fault maintenance.

6. The automatic electric quantity data acquisition system of claim 5, wherein, The anomaly correction unit comprises: The judgment block is configured to judge the user type and the seasonal attribute of the first marked data. The computing block is configured to calculate a first seasonal daily power consumption curve L = l by using the extracted user type based on the first mark data and the target historical data of the corresponding time period of the seasonal attribute. (i) l = l (i) represents the power consumption value corresponding to the i th time point of the daily power consumption curve. Contrast block: for the remaining normal time C=c of the day on which the first mark data is located (j) , j=1, 2, …m, wherein c (j) represents the power consumption value corresponding to the jth remaining normal time, and the power consumption value of the first seasonal day power curve L=l (i) , i=1, 2, …n after the non-quantitative treatment of the power consumption value of the first seasonal day power curve L=l (j) , j=1, 2, …m and L'=l' (i) , i=1, 2, …m respectively, and a plurality of correlation degrees are obtained through correlation analysis. The first similarity is obtained by averaging the evaluation of all the correlation degrees. The correction block is configured to obtain a correction value by weighting the power consumption value corresponding to the time when the first marked data appears in the first seasonal daily power consumption curve using the first similarity, and replace the corresponding first marked data as the first corrected data output.

7. The automatic electric quantity data acquisition system of claim 5, wherein the electric quantity data acquisition device is connected to the electric quantity data acquisition device through the electric quantity data acquisition device connection unit. The reaction unit comprises: The effect judgment block is configured to judge the correction effect of the first corrected data by calculating the correction effect value of the first corrected data. The correction effect value is calculated according to the following formula: wherein X (k) represents the correction effect value of the kth first correction data; a represents the total number of the first correction data; x k represents the kth real power consumption data value; z k represents the kth correction value corresponding thereto; θ1represents the influence weight coefficient of the root mean square error on the correction effect; θ2represents the influence weight coefficient of the average error on the correction effect, wherein θ1+ θ2= 1, and θ1> θ2; The fault troubleshooting block is configured to analyze the obtained correction effect value, and if the correction effect value is less than a preset threshold, mark the corresponding first corrected data to obtain second marked data, obtain the corresponding power consumption calculation device of the second marked data and disable it, and then perform emergency manual maintenance.

Images