CN-121980322-A - Electrical energy metering equipment abnormality detection method and system based on time sequence analysis

CN121980322ACN 121980322 ACN121980322 ACN 121980322ACN-121980322-A

Abstract

The application relates to the technical field of data processing, in particular to an electric energy metering equipment abnormality detection method and system based on time sequence analysis. The method comprises the steps of collecting electricity consumption data by using the electric energy metering equipment, calculating the change degree of the electricity consumption data according to the electricity consumption data, taking the change degree of the electricity consumption data as a weighted autocorrelation coefficient of a trending sequence of the weight calculation electricity consumption data, obtaining a maximum weighted autocorrelation coefficient, calculating a correction threshold of the trending sequence by combining the maximum weighted autocorrelation coefficient and the variation coefficient, and judging the abnormal condition of the electric energy metering equipment according to the correction threshold. The application can obviously reduce the false alarm rate and the missing report rate of the electric energy metering equipment.

Inventors

WU BEI
LIU JINSONG
XU YIFANG
JU PENG
WU CHENGRUI
SUN HUANLIN
Shao Yuexin
LUO HENGJUAN

Assignees

无锡市恒通电器有限公司

Dates

Publication Date: 20260505
Application Date: 20260407

Claims (9)

1. The method for detecting the abnormality of the electric energy metering equipment based on the time sequence analysis is characterized by comprising the following steps: collecting electricity consumption data by using the electric energy metering equipment, calculating the change degree of the electricity consumption data according to the electricity consumption data, taking the change degree of the electricity consumption data as a weighted autocorrelation coefficient of a trending sequence of the weight calculation electricity consumption data, obtaining a maximum weighted autocorrelation coefficient, calculating a correction threshold of the trending sequence by combining the variation coefficient, and judging the abnormal condition of the electric energy metering equipment according to the correction threshold; the power consumption data is an original time sequence data sequence, and comprises a time stamp and corresponding power consumption; The method for calculating the change degree of the electricity consumption data comprises the following steps: Obtaining a trending sequence according to an original time sequence, presetting a hysteresis order range, calculating autocorrelation coefficients of the trending sequence in the hysteresis order range under a plurality of hysteresis orders, obtaining a significant peak hysteresis set according to the hysteresis orders and the corresponding autocorrelation coefficients, calculating electricity utilization change rate of any hysteresis order in the significant peak hysteresis set, and further calculating a periodicity value of the trending sequence according to standard deviation and average value of the trending sequence and the autocorrelation coefficients and the electricity utilization change rates corresponding to the hysteresis orders; dividing the trending sequence continuously and non-overlapping according to the size of a preset noise window to obtain a plurality of trending subsequences, and obtaining the noise level of the power consumption data according to the autocorrelation coefficient of each trending subsequence when the hysteresis order is 1 and the standard deviation of each trending subsequence; the product of the difference between 1 and the periodicity value and the noise level is taken as the degree of change in the electricity consumption data.
2. The method for detecting abnormality of electric energy metering equipment based on time sequence analysis according to claim 1, wherein the obtaining the detrending sequence according to the original time sequence data sequence comprises fitting the original time sequence data sequence by using a least square method to obtain a fitting value at each moment, calculating a difference between an actual value at any moment and the fitting value as the detrending value at the moment, and forming the detrending values at all moments into the detrending sequence; And the actual value at any time is the power consumption data at the corresponding time in the original time sequence data sequence.
3. The method for detecting abnormality of electrical energy metering equipment based on time sequence analysis according to claim 1, wherein the obtaining the significant peak lag set according to the lag order and the corresponding autocorrelation coefficient comprises the steps of forming an autocorrelation curve by the continuous lag order and the corresponding autocorrelation coefficient, obtaining the lag order corresponding to the maximum point of the autocorrelation curve, and taking the lag order corresponding to the maximum point of each correlation curve as the significant peak lag set.
4. The method for detecting abnormality of an electric energy metering device based on time series analysis according to claim 1, wherein the calculating the electricity utilization rate at any hysteresis order in the set of significant peak hysteresis includes: for any hysteresis order, calculating the ratio of the absolute value of the difference value of two data points of any interval hysteresis order in the detrending sequence to the data point before, and calculating the average value of all the ratios under the corresponding hysteresis order as the corresponding electricity utilization change rate of the hysteresis order.
5. The abnormality detection method for an electric energy meter based on time series analysis according to claim 1, wherein the periodic value calculation method for the detrending sequence comprises: calculating the ratio of the sum of products of the autocorrelation coefficients corresponding to each hysteresis order and the power utilization change rate to the sum of the power utilization change rates in the trending sequence; and calculating the negative number of the ratio of the standard deviation of the detrending sequence to the mean value of the detrending sequence, multiplying the negative number with the super parameter, and then calculating the exponential function value as the periodicity value corresponding to the detrending sequence.
6. The abnormality detection method for an electric energy meter device based on time series analysis according to claim 1, wherein calculating a weighted autocorrelation coefficient of a trending sequence of electric energy data using a degree of change of the electric energy data as a weight includes: for any hysteresis order, calculating the product of the difference value between any data point in the detrending sequence and the average value of the detrending sequence and the difference value between the data point of the data point interval response hysteresis order and the average value of the detrending sequence, and multiplying the product with the change degree of corresponding power utilization data to obtain the weighted covariance of the data point; Similarly, acquiring weighted covariance and weighted variance of other data points in the trending sequence, and further acquiring weighted covariance sum and weighted variance sum; and calculating the ratio of the weighted covariance sum to obtain a weighted autocorrelation coefficient of the trending sequence under the corresponding hysteresis order.
7. The electrical energy metering device abnormality detection method based on time series analysis according to claim 1, wherein the calculation method of the correction threshold value is as follows: Presetting an initial threshold; Calculating the ratio of the absolute value of the maximum value of the weighted autocorrelation coefficients in all hysteresis orders to the variation coefficient of the trending sequence as an adjustment factor, avoiding the excessive or insufficient adjustment factor, and respectively adding super-parameters to the molecular denominator of the adjustment factor; the product of the initial threshold and the adjustment factor is calculated as a correction threshold.
8. The method for detecting abnormal condition of electric energy metering equipment based on time sequence analysis according to claim 1, wherein the judging of abnormal condition of the electric energy metering equipment according to the correction threshold comprises the steps of calculating standard scores of data points at any moment in a trend sequence by utilizing z-score, and if the standard scores are larger than the correction threshold, marking the data points at corresponding moments as abnormal points and outputting all the abnormal points as an abnormal point list.
9. An electrical energy metering device anomaly detection system based on time series analysis, comprising a processor and a memory, the memory storing computer program instructions which, when executed by the processor, implement the electrical energy metering device anomaly detection method based on time series analysis of any one of claims 1-8.

Description

Electrical energy metering equipment abnormality detection method and system based on time sequence analysis Technical Field The application relates to the technical field of data processing, in particular to an electric energy metering equipment abnormality detection method and system based on time sequence analysis. Background The electric energy metering device is used as key equipment for operation monitoring and electric charge settlement of an electric power system, and the accurate detection and abnormal detection of the operation state of the electric energy metering device are particularly important. At present, the abnormality detection technology of the electric energy metering device is mainly divided into two main types, namely an off-line method, namely a method for detecting and identifying on site by using special hardware equipment, wherein the method has high cost and low efficiency, and an on-line method for identifying abnormal states mainly by analyzing metering data. In the prior art, anomaly detection of an electric energy metering device mainly depends on a time series data analysis method. For example, a sliding window averaging method or a simple threshold comparison is employed to monitor fluctuations in the amount of electricity used. Specifically, the existing system collects time sequence data (such as a power consumption sequence) of the electric energy metering equipment, calculates a short-term average value, compares the short-term average value with a preset threshold value, and judges that the electric energy metering equipment is abnormal if the short-term average value exceeds the threshold value. This method is based on statistical principles and aims to capture sudden changes in data. However, the prior art ignores the long-term correlation of time series data and the influence of external environmental factors, resulting in high false alarm rate or high false alarm rate in complex scenes. For example, during seasonal peak hours, the natural fluctuation of electricity consumption is large, and normal fluctuation and abnormality are difficult to distinguish by a simple threshold method. Disclosure of Invention In order to solve the problem that the prior art ignores long-term correlation of time sequence data and influence of external environmental factors to cause high false alarm rate or high missing report rate of electric energy metering equipment in a complex scene, so that the abnormal detection accuracy of the electric energy metering equipment is influenced, the application provides the electric energy metering equipment abnormal detection method and system based on time sequence analysis. In a first aspect, the present application provides a method for detecting an abnormality of an electric energy metering device based on time sequence analysis, which adopts the following technical scheme: The method for detecting the abnormality of the electric energy metering equipment based on time sequence analysis comprises the following steps: collecting electricity consumption data by using the electric energy metering equipment, calculating the change degree of the electricity consumption data according to the electricity consumption data, taking the change degree of the electricity consumption data as a weighted autocorrelation coefficient of a trending sequence of the weight calculation electricity consumption data, obtaining a maximum weighted autocorrelation coefficient, calculating a correction threshold of the trending sequence by combining the variation coefficient, and judging the abnormal condition of the electric energy metering equipment according to the correction threshold; the power consumption data is an original time sequence data sequence, and comprises a time stamp and corresponding power consumption; The method for calculating the change degree of the electricity consumption data comprises the following steps: Obtaining a trending sequence according to an original time sequence, presetting a hysteresis order range, calculating autocorrelation coefficients of the trending sequence in the hysteresis order range under a plurality of hysteresis orders, obtaining a significant peak hysteresis set according to the hysteresis orders and the corresponding autocorrelation coefficients, calculating electricity utilization change rate of any hysteresis order in the significant peak hysteresis set, and further calculating a periodicity value of the trending sequence according to standard deviation and average value of the trending sequence and the autocorrelation coefficients and the electricity utilization change rates corresponding to the hysteresis orders; dividing the trending sequence continuously and non-overlapping according to the size of a preset noise window to obtain a plurality of trending subsequences, and obtaining the noise level of the power consumption data according to the autocorrelation coefficient of each trending subsequence when the hyste