A method for collecting power data for a photovoltaic energy storage system

Abstract

The present invention relates to the field of data processing. The present invention relates to a method for collecting power data for a photovoltaic energy storage system, including: installing smart electricity meters at the connection points between the AC output terminals of the photovoltaic energy storage system and the power grid respectively to collect power data and transmit it. The smart electricity meters are communicatively connected to each other. The transmission method includes: obtaining the historical power consumption data sequences of the smart electricity meters and the electricity meters connected thereto and calculating the periods of each power consumption data sequence; obtaining the subsequences under each period and obtaining the clustering clusters under each period; for each electricity meter, calculating its comprehensive anomaly degree; and transmitting the power consumption data sequence of the electricity meter with the highest comprehensive anomaly degree through the network node with the highest communication quality. By using the method of the present invention, it is possible to avoid transmission errors in the power consumption data of abnormal electricity meters and ensure timely detection of anomalies at each connection point between the photovoltaic energy storage system and the power grid.

Description

Technical Field

[0001] The present invention relates to the field of data processing. More specifically, the present invention relates to a method for collecting power data for a photovoltaic energy storage system. Background Art

[0002] A photovoltaic energy storage system is an integrated energy solution that combines photovoltaic power generation and electrical energy storage technologies. Its core goal is to achieve efficient utilization and stable supply of clean energy. A photovoltaic energy storage system generally consists of a photovoltaic array, an inverter, an energy storage device, an energy storage converter, and an energy management system. The photovoltaic array is composed of multiple solar panels, which convert solar energy into direct current (DC). The inverter is used to convert the DC power generated by the photovoltaic into alternating current (AC) for use by household or industrial equipment. The energy storage device is used to store the electrical energy generated by the photovoltaic array. The energy storage device uses lithium batteries, lead-acid batteries, or flow batteries. The energy storage converter is used to control the charging and discharging process of the battery and optimize the energy flow; the energy management system is used to monitor the photovoltaic power generation, energy storage status, load demand, and grid electricity price in real time.

[0003] In order to ensure the reliable operation of the power network connected to the photovoltaic energy storage system, it is necessary to collect the power data at the connection point between the photovoltaic energy storage system and the grid in real time to monitor whether each connection point is operating normally. In the prior art, intelligent electricity meters are usually used to collect the power data of the photovoltaic energy storage system.

[0004] An intelligent electricity meter refers to an electronic electricity meter that not only has accurate electricity metering but also has multiple functions such as data storage, information transmission, remote power on / off, and abnormal alarm. The intelligent electricity meter can accurately measure forward, reverse, active, and reactive electric energy, and record key parameters such as maximum demand, voltage and current harmonics, and power factor. It has high accuracy, wide range, wide power factor, sensitive start, and long-term unchanged accuracy, and does not require cycle calibration. The intelligent electricity meter is built with a data processing unit, which can perform real-time processing and analysis on the collected electricity consumption data, generate electricity consumption reports, trend charts, energy consumption comparisons, etc., providing more data support for power supply companies and users, and helping to optimize electricity consumption management and the formulation of energy-saving measures.

[0005] With the continuous development of smart grid and Internet of Things technologies, intelligent electricity meters, as an important part of the power system, have been applied in various places such as homes, businesses, and industries. The intelligent electricity meter has a two-way communication function and can exchange data with other intelligent substations in the smart grid. Through the built-in communication module, it can remotely read meters and transmit electricity consumption data to the management system of power supply companies or users, eliminating the cumbersome process of traditional manual meter reading and improving the accuracy and efficiency of data.

[0006] Such as Figure 1As shown in the figure, the method of collecting and transmitting power data of a photovoltaic energy storage system using smart electricity meters in the prior art is as follows: Smart electricity meters are installed at the connection points between the AC output terminals of the photovoltaic energy storage system and the power grid respectively to collect power data and transmit it. When transmitting, the power consumption data of multiple smart electricity meters are uniformly transmitted through a main smart electricity meter. However, in this data transmission method, since the power consumption data of each electricity meter is transmitted without discrimination during transmission and the amount of transmitted data is large, if a certain electricity meter is abnormal, the power consumption data collected by this abnormal electricity meter may not be transmitted in time, resulting in the inability to detect the abnormality of the grid connection point corresponding to the abnormal electricity meter in time. Summary of the Invention

[0007] To solve the technical problem that the existing electricity meter data transmission method cannot detect the abnormalities of each connection point between the photovoltaic energy storage system and the power grid in time, the present invention provides solutions in the following aspects.

[0008] In the first aspect, the present invention provides a method for collecting power data for a photovoltaic energy storage system, including: installing smart electricity meters at the connection points between the AC output terminals of the photovoltaic energy storage system and the power grid respectively to collect power data and transmit it. The smart electricity meters are communicatively connected to each other. The transmission method includes: obtaining the historical power consumption data sequences of the smart electricity meters and the electricity meters connected thereto and calculating the periods of each power consumption data sequence;

[0009] Intercepting a period length from multiple power consumption data sequences according to multiple periods respectively, thereby obtaining multiple subsequences, clustering each of the multiple subsequences corresponding to each period respectively, thereby obtaining clustering clusters under each period. The length of the subsequence is equal to the corresponding period; for each electricity meter, calculating its first abnormal degree under each period and performing weighted summation to obtain a comprehensive abnormal degree; and transmitting the power consumption data sequence of the electricity meter with the highest comprehensive abnormal degree through the network node with the highest communication quality; The expression of the first abnormal degree is:

[0010] ;

[0011] In the formula, represents the abnormal degree of the s-th electricity meter in the t-th period, represents the hyperbolic tangent function, represents the current power consumption data of the s-th electricity meter, represents the probability density function of the subsequence of the s-th electricity meter in the t-th period in the clustering cluster.

[0012] The beneficial effects are as follows: Before transmitting the power consumption data sequences of multiple electricity meters, the power data acquisition method for the photovoltaic energy storage system of the present invention first calculates the abnormality degree of each electricity meter according to the power consumption data sequences, and uses the network node with the highest communication quality to transmit the power consumption data sequence of the electricity meter with the highest abnormality degree, so as to obtain the abnormal situation of the electricity meter faster, avoid transmission errors of the power consumption data of the abnormal electricity meter, and ensure timely discovery of the abnormal situation at each connection point between the photovoltaic energy storage system and the power grid. In addition, when calculating the abnormality degree of the electricity meter, considering the periodicity of the power consumption data sequence, the power consumption data sequence is intercepted according to the length of each period of the power consumption data sequence to obtain subsequences, and the abnormality degree of the electricity meter in each period is calculated by using the subsequences in each period. Compared with calculating the abnormality degree of the electricity meter by using the entire power consumption data sequence, the data processing amount is greatly reduced, and the calculation efficiency of the abnormality degree of the electricity meter is improved.

[0013] Furthermore, when calculating the abnormality degree in a certain period, first cluster the subsequence in this period to obtain each cluster. Considering that the smaller the difference between the current power consumption data of the electricity meter and the center of this distribution, the greater the proximity between the two, the greater the possibility that the current power consumption data belongs to the corresponding probability distribution, and the smaller the possibility that the power consumption data sequence of the electricity meter is abnormal, let the abnormality degree of the electricity meter in a certain period be positively correlated with the difference, so as to calculate the abnormality degree of the electricity meter in a certain period more accurately.

[0014] Preferably, the calculation expression of the comprehensive abnormality degree is:

[0015] ;

[0016] In the formula, represents the comprehensive abnormality degree of the s-th electricity meter, represents the probability that the power consumption data of the cluster where the subsequence of the s-th electricity meter in the t-th period is located follows a normal distribution, represents the mean value of the within-class distance of the cluster where the subsequence of the s-th electricity meter in the t-th period is located, and T represents the total number of the periods.

[0017] The beneficial effects are as follows: The present invention calculates the comprehensive anomaly degree by weighted summation of the first anomaly degrees in different periods. Generally, the greater the probability that the electricity consumption data within the clustering cluster corresponding to the electricity meter follows a normal distribution, the greater the credibility of the calculated first anomaly degree. When performing weighted summation, the weight of this first anomaly degree should also be greater; the smaller the within-class distance of the clustering cluster where the electricity meter is located in different periods, the greater the similarity of this clustering cluster, and the greater the weight of the first anomaly degree calculated through this clustering cluster. Therefore, when performing weighted summation of the first anomaly degrees in different periods, the corresponding weight is made positively correlated with the probability that the electricity consumption data within the clustering cluster follows a normal distribution and inversely proportional to the within-class distance of the clustering cluster where the electricity meter is located in the corresponding period, so as to calculate the comprehensive anomaly degree of the electricity meter more accurately.

[0018] Preferably, for a certain clustering cluster, the method for obtaining the probability that its electricity consumption data follows a normal distribution includes: calculating the mean and variance of the electricity meter readings within the clustering cluster, and then using the A-D test to obtain the probability that the electricity consumption data within the clustering cluster follows a normal distribution.

[0019] The beneficial effects are as follows: Using the A-D test can better achieve normality testing, can sensitively reveal the potential asymmetry of the data, and make the finally obtained probability of following a normal distribution more accurate.

[0020] Preferably, for a certain electricity meter, its comprehensive anomaly degree is equal to the mean of the anomaly degrees in its respective periods.

[0021] Preferably, the method used to calculate the period of the electricity consumption data sequence is the autocorrelation function.

[0022] The beneficial effects are as follows: The autocorrelation function can detect the periodic components in the signal. When there are periodic characteristics in the signal (such as the electricity consumption data sequence), the autocorrelation function will show peaks at the corresponding lag times. The positions of these peaks directly correspond to the periods of the signal, so as to accurately identify and analyze the periodic characteristics of the electricity consumption data sequence. In addition, the electricity consumption data sequence is often affected by various noises, which may mask the true periodic characteristics of the signal. Through the autocorrelation function, these random noises can be filtered out and the true periodic information of the signal can be extracted. Therefore, using the autocorrelation function to calculate the period of the electricity consumption data sequence can make the period calculation result more accurate.

[0023] Preferably, when clustering the subsequences in a certain period, the calculation expression of the clustering distance is:

[0024] ;

[0025] In the formula, represents the clustering distance between the i-th subsequence and the j-th subsequence, represents the cross-correlation between the $i$-th subsequence and the $j$-th subsequence.

[0026] Its beneficial effect is that: the greater the cross-correlation between two subsequences, the more similar the two subsequences are, and the smaller the clustering distance between the two subsequences. Let the clustering distance be negatively correlated with the cross-correlation between two subsequences, and the clustering distance corresponding to the two subsequences can be calculated more accurately.

[0027] Preferably, when clustering the subsequences in a certain period, the calculation expression of the clustering distance is:

[0028] ;

[0029] In the formula, represents the clustering distance between the $i$-th subsequence and the $j$-th subsequence, represents the $i$-th subsequence, represents the $j$-th subsequence, represents the norm corresponding to the $i$-th subsequence and the $j$-th subsequence.

[0030] Its beneficial effect is that: in a multi-dimensional space, the L2 norm can accurately reflect the absolute distance between vectors. In addition, when calculating the distance, the L2 norm magnifies the influence of larger differences through the sum of squares, but at the same time smooths the smaller differences to a certain extent. This characteristic makes the L2 norm have a certain robustness to noise when calculating the distance. Therefore, in the present invention, by using the L2 norm to calculate the distance between two subsequences, the calculation result of the clustering distance is more accurate.

[0031] Preferably, when intercepting the power consumption data sequence, it is intercepted from the tail of the power consumption data sequence.

[0032] Its beneficial effect is that: since the abnormal conditions of the electricity meter may change over time. Therefore, detecting the abnormal conditions of the electricity meter based on the power consumption data collected by the electricity meter in a time period closer to the current moment can ensure a higher accuracy. The present invention further improves the accuracy of calculating the abnormal degree of the electricity meter by intercepting from the tail of the power consumption data sequence to obtain the subsequence.

[0033] Preferably, the k-means clustering algorithm is used when clustering the subsequences.

[0034] Preferably, if the transmission volume of the network node with the highest communication quality reaches saturation, the power consumption data sequence of the electricity meter with the highest comprehensive abnormal degree is transmitted through the network node with the second highest communication quality.

[0035] In summary, the beneficial effects of the present invention are as follows: By using the method of the present invention, it is possible to avoid transmission errors in the power consumption data of abnormal electricity meters, obtain the abnormal conditions of the electricity meters in a timely manner, and have a relatively high transmission efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features, and advantages of the exemplary embodiments of the present invention will become readily understood. In the drawings, several embodiments of the present invention are shown by way of illustration and not limitation, and like or corresponding reference numerals denote like or corresponding parts, wherein:

[0037] Figure 1 schematically shows a method for transmitting data of an intelligent electricity meter in the prior art;

[0038] Figure 2 schematically shows a structural diagram of a method for collecting power data for a photovoltaic energy storage system according to an embodiment of the present invention;

[0039] Figure 3 schematically shows a flowchart of a method for transmitting power consumption data according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0040] Hereinafter, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0041] The following will describe in detail the specific embodiments of the present invention in conjunction with the accompanying drawings.

[0042] Embodiment of the method for collecting power data for a photovoltaic energy storage system:

[0043] The method for collecting power data for a photovoltaic energy storage system of the present invention includes: installing intelligent electricity meters at the connections between the respective AC output terminals of the photovoltaic energy storage system and the power grid to collect power data and transmit it, and the intelligent electricity meters are communicatively connected to each other.

[0044] As Figure 2 shown, the intelligent electricity meter includes: a power consumption data collection unit, a communication unit, and a controller. The communication unit is connected to a plurality of electricity meters and a plurality of network nodes; the controller is used to transmit the power consumption data of the intelligent electricity meter and the electricity meters connected thereto through the communication unit.

[0045] The power consumption data collection unit includes a metering unit, a voltage sampling unit, and a current sampling unit.

[0046] In this embodiment, the structure of the controller of the smart electricity meter is a conventional prior art, and the structure of the controller is the same as that of the controller in the Chinese patent with the authorization announcement number CN113130235B.

[0047] The communication unit is connected to the user management system of the monitoring center through multiple network nodes.

[0048] As Figure 3 shown, the transmission method includes:

[0049] S101. Obtain the historical power consumption data sequences of the smart electricity meter and the electricity meters connected thereto, and calculate the periods of each power consumption data sequence;

[0050] In this embodiment, the historical power consumption data sequence of the electricity meter refers to the data sequence composed of the power consumption data collected by the electricity meter at historical moments. Since there are peak power consumption periods and off-peak power consumption periods in the distribution network, and the power consumption conditions of the loads connected to the distribution network are different at different times of the day, the power consumption data collected by the electricity meter has a certain periodicity.

[0051] S102. Obtain the subsequences under each period, specifically: respectively intercept the lengths of one period from multiple power consumption data sequences according to multiple periods, so as to obtain multiple subsequences, and perform clustering on the multiple subsequences corresponding to each period respectively, so as to obtain the clustering clusters under each period, and the length of the subsequence is equal to the corresponding period;

[0052] For example: assume that the periods calculated in step S101 are 2, 3, and 5, and there are three power consumption data sequences in total. Then, the lengths of 2, 3, and 5 are used as the data lengths to intercept the three power consumption data sequences respectively. After intercepting, the lengths of the three subsequences corresponding to the period 2 are all 2; the lengths of the three subsequences corresponding to the period 3 are all 3; the lengths of the three subsequences corresponding to the period 5 are all 5.

[0053] In this embodiment, the k-means clustering algorithm is used for clustering the subsequences. In other embodiments, appropriate clustering algorithms such as hierarchical clustering, DBSCAN clustering algorithm, and K-medoids clustering algorithm can also be used.

[0054] S103. Obtain the comprehensive anomaly degree of each electricity meter, specifically: for each electricity meter, calculate its first anomaly degree under each period, and perform weighted summation to obtain the comprehensive anomaly degree; the calculation expression of the first anomaly degree is:

[0055] ;

[0056] In the formula, represents the anomaly degree of the s-th electricity meter under the t-th period, represents the hyperbolic tangent function, represents the current power consumption data of the s-th electricity meter, represents the probability density function of the cluster where the subsequence of the s-th electricity meter is located in the t-th cycle.

[0057] The current power consumption data of the electricity meter refers to the power consumption data collected by the electricity meter at the current moment.

[0058] In this embodiment, there are two methods for obtaining the probability density function. The first one is the histogram method, which is a non-parametric method and does not require assuming the distribution type of X. Divide the value range of X into several small intervals, then calculate the frequency or frequency density in each interval according to the statistical results of X, and finally represent it with a histogram. The shape of the histogram can reflect the distribution characteristics of X, such as whether it is symmetric, whether it is unimodal, whether it has skewness, etc. The other method is the parameter estimation method, which is a parametric method and requires assuming the distribution type of X, such as normal distribution, exponential distribution, uniform distribution, etc. The basic idea of the parameter estimation method is to estimate the distribution parameters of X, such as mean, variance, shape parameter, etc., according to the statistical results of X by using the maximum likelihood method, moment method or Bayesian method, etc. Then substitute the estimated parameters into the distribution function of X to obtain the probability density distribution function f(x) of X.

[0059] represents the position of the current power consumption data of the s-th electricity meter in the probability density function. The closer this position is to the center, the greater the possibility that it belongs to the probability distribution corresponding to the probability density function. represents the difference between the current power consumption data of the s-th electricity meter and the central power consumption data of this distribution. The smaller the difference, the closer the current power consumption data of the s-th electricity meter is to the center of this distribution, the greater the possibility that the current power consumption data belongs to this distribution, and the smaller the possibility of its abnormality. Therefore, the calculation expression of the first degree of abnormality in this embodiment can accurately calculate the degree of abnormality of the electricity meter in a certain cycle.

[0060] Since the greater the degree of abnormality of the electricity meter in each cycle, the greater the possibility that the electricity meter has an abnormality. Therefore, when calculating the comprehensive degree of abnormality of the electricity meter, it is necessary to make the comprehensive degree of abnormality of the electricity meter positively correlated with the degree of abnormality in each cycle. For example: The average value of the degree of abnormality of the electricity meter in each cycle can be used as the comprehensive degree of abnormality of the electricity meter.

[0061] S104. Transmit the power consumption data sequence of the electricity meter with the highest comprehensive degree of abnormality through the network node with the highest communication quality.

[0062] Before transmitting the power consumption data sequences of multiple electricity meters in the power data acquisition method for the photovoltaic energy storage system of this embodiment, the abnormality degree of each electricity meter is first calculated based on the power consumption data sequence. For the power consumption data sequence of the electricity meter with the highest abnormality degree, it is transmitted using the network node with the highest communication quality, so as to obtain the abnormality of the electricity meter faster, avoid transmission errors in the power consumption data of the abnormal electricity meter, and ensure timely detection of the abnormalities at each connection point between the photovoltaic energy storage system and the power grid. In addition, when calculating the abnormality degree of the electricity meter, considering the periodicity of the power consumption data sequence, the power consumption data sequence is intercepted according to the length of each period of the power consumption data sequence to obtain subsequences, and the abnormality degree of the electricity meter in each period is calculated using the subsequences under each period. Compared with calculating the abnormality degree of the electricity meter using the entire power consumption data sequence, the data processing amount is greatly reduced, the calculation efficiency of the abnormality degree of the electricity meter is improved, and thus the transmission efficiency of the power consumption data is improved. Furthermore, when calculating the abnormality degree in a certain period, the subsequence in this period is first clustered to obtain each clustering cluster. Considering that the smaller the difference between the current power consumption data of the electricity meter and the center of this distribution, the greater the closeness between the two, the greater the possibility that the current power consumption data belongs to the corresponding probability distribution, and the smaller the possibility that the power consumption data sequence of the electricity meter is abnormal. Let the abnormality degree of the electricity meter in a certain period be positively correlated with the difference, so as to calculate the abnormality degree of the electricity meter in a certain period more accurately.

[0063] In the above embodiment, the abnormality degree of the s-th electricity meter in the t-th period is calculated based on the probability density function of the clustering cluster it belongs to. In another embodiment, the calculation expression of the abnormality degree of the s-th electricity meter in the t-th period is: ;

[0064] In the formula, represents the abnormality degree of the s-th electricity meter in the t-th period, represents the current power consumption data of the s-th electricity meter, represents the mean value of the power consumption data in the clustering cluster where the subsequence of the s-th electricity meter in the t-th period is located, represents the standard deviation of the power consumption data in the clustering cluster where the subsequence of the s-th electricity meter in the t-th period is located.

[0065] The greater the difference between the reading of an electricity meter and the mean of the electricity meter readings in the clustering cluster corresponding to the electricity meter, the greater the degree of abnormality of the electricity consumption data collected by the electricity meter; in the case where the current electricity consumption data of two electricity meters are the same but the clustering clusters corresponding to the two electricity meters are different, considering that the dispersion degrees of the electricity consumption data in different clustering clusters are different (that is, the standard deviations of the electricity consumption data in different clustering clusters are different), by taking the ratio of the difference between the reading of the electricity meter and the mean of the electricity meter readings in the clustering cluster corresponding to the electricity meter to the standard deviation of the electricity consumption data in the corresponding clustering cluster as the degree of abnormality of the electricity meter in a certain period, it is ensured that in the case where the current electricity consumption data of two electricity meters are the same but the clustering clusters corresponding to the two electricity meters are different, the calculated first degrees of abnormality of the two electricity meters are different, making the calculated first degree of abnormality more accurate.

[0066] In one embodiment, the calculation expression of the comprehensive degree of abnormality is:

[0067] ;

[0068] In the formula, represents the comprehensive degree of abnormality of the s-th electricity meter, represents the probability that the electricity consumption data of the clustering cluster where the subsequence of the s-th electricity meter in the t-th period is located follows a normal distribution, represents the mean of the within-class distances of the clustering cluster where the subsequence of the s-th electricity meter in the t-th period is located, T represents the total number of the periods, represents the degree of abnormality of the s-th electricity meter in the t-th period.

[0069] In this embodiment, the comprehensive degree of abnormality is calculated by weighted summation of the first degrees of abnormality in different periods. Generally, the greater the probability that the electricity consumption data within the clustering cluster corresponding to the electricity meter follows a normal distribution, the greater the credibility of the calculated first degree of abnormality, and when performing weighted summation, the weight of this first degree of abnormality should also be greater; the smaller the within-class distance of the clustering cluster where the electricity meter is located in different periods, the greater the similarity of the clustering cluster, and the greater the weight of the first degree of abnormality calculated through this clustering cluster should also be. Therefore, when performing weighted summation of the first degrees of abnormality in different periods, the corresponding weights are made positively correlated with the probability that the electricity consumption data within the clustering cluster follows a normal distribution and inversely proportional to the within-class distance of the clustering cluster where the electricity meter is located in the corresponding period, so as to calculate the comprehensive degree of abnormality of the electricity meter more accurately.

[0070] In one embodiment, for a certain clustering cluster, the method for obtaining the probability that its electricity consumption data follows a normal distribution includes: calculating the mean and variance of the electricity meter readings within the clustering cluster, and then using the A-D test to obtain the probability that the electricity consumption data within the clustering cluster follows a normal distribution.

[0071] The A-D test can be used to better achieve normality testing, which can sensitively reveal the potential asymmetry of the data, making the probability of finally obtaining a normal distribution more accurate.

[0072] In one embodiment, for a certain electricity meter, its comprehensive degree of abnormality is equal to the mean value of the degrees of abnormality in each cycle.

[0073] In one embodiment, the method used to calculate the period of the electricity consumption data sequence is the autocorrelation function.

[0074] The autocorrelation function can detect the periodic components in a signal. When there are periodic characteristics in a signal (such as an electricity consumption data sequence), the autocorrelation function will show peaks at the corresponding lag times. The positions of these peaks directly correspond to the period of the signal, thus accurately identifying and analyzing the periodic characteristics of the electricity consumption data sequence. In addition, the electricity consumption data sequence is often affected by various noises, which may mask the true periodic characteristics of the signal. Through the autocorrelation function, these random noises can be filtered out, and the true periodic information of the signal can be extracted. Therefore, using the autocorrelation function to calculate the period of the electricity consumption data sequence can make the period calculation result more accurate.

[0075] In one embodiment, when clustering the subsequences in a certain cycle, the calculation expression of the clustering distance is:

[0076] ;

[0077] In the formula, represents the clustering distance between the i-th subsequence and the j-th subsequence, represents the cross-correlation between the i-th subsequence and the j-th subsequence.

[0078] The greater the cross-correlation between two subsequences, the more similar the two subsequences are, and the smaller the clustering distance between the two subsequences. Letting the clustering distance be negatively correlated with the cross-correlation between two subsequences can more accurately calculate the clustering distance corresponding to the two subsequences.

[0079] In one embodiment, when clustering the subsequences in a certain cycle, the calculation expression of the clustering distance is:

[0080] ;

[0081] In the formula, represents the clustering distance between the i-th subsequence and the j-th subsequence, represents the i-th subsequence, represents the j-th subsequence, represents the norm corresponding to the i-th subsequence and the j-th subsequence.

[0082] The L2 norm, also known as the Euclidean norm, represents the straight-line distance from the origin to the end point of the vector, that is, the length of the vector in Euclidean space. This property makes the L2 norm intuitive and has a clear geometric meaning when measuring the distance between vectors. In a multi-dimensional space, the L2 norm can accurately reflect the absolute distance between vectors. In addition, when calculating the distance, the L2 norm amplifies the influence of larger differences through the sum of squares, but at the same time smooths smaller differences to a certain extent. This characteristic makes the L2 norm have a certain robustness to noise when calculating the distance. In this embodiment, the L2 norm is used to calculate the distance between two subsequences, making the calculation result of the clustering distance more accurate.

[0083] In one embodiment, when intercepting the power consumption data sequence, it starts from the tail of the power consumption data sequence.

[0084] For example: Suppose there are three periods, 2, 3, and 4, calculated in step S101 of the above embodiment, and there are three power consumption data sequences in total. The first power consumption data sequence is [a, b, c, d, e], the second power consumption data sequence is [f, g, h, i, j], and the third power consumption data sequence is [i, j, k, l, m]. In the case of a period of 2, the subsequences obtained after interception are: [d, e], [i, j], [l, m]. In the case of a period of 3, the subsequences obtained after interception are: [c, d, e], [h, i, j], [k, l, m]. In the case of a period of 4, the subsequences obtained after interception are: [b, c, d, e], [g, h, i, j], [j, k, l, m].

[0085] Since the abnormal situation of the electricity meter may change over time, for example: at the beginning, the electricity meter is in a normal state and an abnormality occurs after a period of time. Therefore, detecting the abnormal situation of the electricity meter based on the power consumption data collected in the time period closer to the current moment of the electricity meter can ensure a high degree of accuracy. In this embodiment, the subsequences are obtained by starting from the tail of the power consumption data sequence, thereby further improving the accuracy of calculating the abnormal degree of the electricity meter.

[0086] In one embodiment, if the transmission volume of the network node with the highest communication quality reaches saturation, the power consumption data sequence of the electricity meter with the highest comprehensive abnormal degree is transmitted through the network node with the second-highest communication quality.

[0087] If the transmission volume of the network node with the second-highest communication quality also reaches the saturation state, then the network node with the highest communication quality and unsaturated transmission volume is selected from the remaining network nodes for transmission.

[0088] By using the method for transmitting the electricity consumption data sequence according to this embodiment, it is possible to effectively and timely transmit the electricity consumption data sequence of the electricity meter with the highest comprehensive abnormality degree under the state where the transmission volume of the network node with the highest communication quality reaches saturation.

[0089] Although this specification has shown and described multiple embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will think of many changes, alterations, and alternative ways without departing from the spirit and scope of the present invention. It should be understood that various alternatives to the embodiments of the present invention described herein may be employed in practicing the present invention.

Claims

1. A method for collecting power data for a photovoltaic energy storage system, characterized in that: include: Smart energy meters are installed at the connection points between each AC output terminal of the photovoltaic energy storage system and the power grid to collect power data and transmit it. Each smart energy meter is connected to each other in communication. The transmission method includes: obtaining the historical power consumption data sequence of the smart energy meter and the power meter connected to it and calculating the period of each power consumption data sequence; According to multiple cycles, multiple power consumption data sequences are cut off by the length of one cycle to obtain multiple subsequences, and multiple subsequences corresponding to each cycle are clustered to obtain cluster clusters under each cycle, and the length of the subsequence is equal to the corresponding cycle; for each electric energy meter, the first abnormality degree under each cycle is calculated, and a weighted sum is performed to obtain the comprehensive abnormality degree; and the power consumption data sequence of the electric energy meter with the highest comprehensive abnormality degree is transmitted through the network node with the highest communication quality; the expression of the first abnormality degree is: ； In the formula, Indicates the abnormality of the sth electric energy meter in the tth cycle, represents the hyperbolic tangent function, represents the current power consumption data of the sth electric energy meter, Represents the probability density function of the cluster to which the subsequence of the s-th electric energy meter in the t-th period belongs.

2. The method for collecting power data for a photovoltaic energy storage system according to claim 1, characterized in that: The calculation expression of the comprehensive abnormality degree is: ； In the formula, Indicates the comprehensive abnormality degree of the sth electric energy meter, It represents the probability that the power consumption data of the cluster where the subsequence of the sth electric energy meter in the tth period belongs to follows the normal distribution, represents the mean of the intra-class distances of the subsequences of the s-th electric energy meter in the t-th period, and T represents the total number of the periods.

3. The method for collecting power data for a photovoltaic energy storage system according to claim 2, characterized in that: For a certain cluster, the method for obtaining the probability that its electricity consumption data obeys the normal distribution includes: calculating the mean and variance of the electric energy meter readings in the cluster, and then using AD test to obtain the probability that the electricity consumption data in the cluster obeys the normal distribution.

4. The method for collecting power data for a photovoltaic energy storage system according to claim 1, characterized in that: For a certain electric energy meter, the comprehensive abnormality degree is equal to the average value of the abnormality degrees in each cycle of the meter.

5. The method for collecting power data for a photovoltaic energy storage system according to claim 1, characterized in that: The method used to calculate the period of the electricity consumption data series is the autocorrelation function.

6. The method for collecting power data for a photovoltaic energy storage system according to claim 1, characterized in that: When clustering subsequences under a certain period, the calculation expression of clustering distance is: ； In the formula, represents the clustering distance between the ith subsequence and the jth subsequence, represents the mutual correlation between the i-th subsequence and the j-th subsequence.

7. The method for collecting power data for a photovoltaic energy storage system according to claim 1, characterized in that: When clustering subsequences under a certain period, the calculation expression of clustering distance is: ； In the formula, represents the clustering distance between the ith subsequence and the jth subsequence, represents the i-th subsequence, represents the jth subsequence, Indicates the correspondence between the i-th subsequence and the j-th subsequence Norm.

8. The method for collecting power data for a photovoltaic energy storage system according to claim 1, characterized in that: When intercepting the power consumption data sequence, interception starts from the end of the power consumption data sequence.

9. The method for collecting power data for a photovoltaic energy storage system according to any one of claims 1 to 8, characterized in that: The k-means clustering algorithm is used to cluster subsequences.

10. The method for collecting power data for a photovoltaic energy storage system according to any one of claims 1 to 8, characterized in that: If the transmission volume of the network node with the highest communication quality reaches saturation, the electricity consumption data sequence of the electric energy meter with the highest comprehensive abnormality will be transmitted through the network node with the second highest communication quality.

Images