Abnormal state assessment method used in solar cell operation process

Abstract

The invention relates to the field of data processing, in particular to an abnormal state evaluation method used in the operation process of a solar cell, and the method comprises the steps: carrying out the collection of all monitoring data in the operation process of a solar cell panel of a power station, and carrying out the window division; determining the short-time abnormal degree of the window according to the importance degree of the data of each dimension in the window, and carrying out short-time abnormal judgment; determining a trend analysis factor according to the frequency change of the short-time abnormal window; according to the trend analysis factor and the short-time abnormal degree of the solar cell panel multi-dimensional operation monitoring time sequence data window, determining a distance measurement evaluation factor of the data window; distance measurement between the multi-dimensional operation monitoring time sequence data windows is determined according to the distance measurement evaluation factors, a clustering process is completed based on the distance measurement, and abnormal state evaluation is performed according to a result of the clustering process, so that the accuracy and effect of abnormal state evaluation are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electrical digital data processing, and in particular to a method for evaluating abnormal conditions during the operation of a solar cell. Background Art

[0002] The long-term stability and efficiency of solar panels directly impact the performance and economic benefits of the entire solar power generation system. Over time, panels may be affected by factors such as environmental changes, equipment aging, and external interference, leading to decreased efficiency or failure, and even potential safety hazards. Real-time monitoring and timely assessment of panel abnormalities can identify potential problems early and enable effective maintenance measures, avoiding equipment downtime or performance degradation, minimizing financial losses, extending the panel's lifespan, and improving system safety and reliability. Therefore, abnormality assessment is crucial to the operation of solar cells and is fundamental to ensuring stable system operation and achieving efficient power generation.

[0003] During the operation of solar panels, real-time monitoring and abnormal condition assessment are critical to ensuring their efficient and stable operation. Traditional anomaly detection methods primarily rely on cluster analysis based on Euclidean distance. This method clusters the multi-dimensional operational monitoring data of solar panels to identify normal and abnormal conditions. However, these methods often fail to fully account for the temporal characteristics of the data and are susceptible to short-term anomalies caused by external environmental factors (such as weather changes and cloud cover), leading to misidentification of abnormal conditions. In particular, when solar panel operational data fluctuates due to short-term weather changes, traditional methods may misinterpret these short-term anomalies as equipment failures, thereby affecting subsequent fault detection and maintenance decisions. Furthermore, existing methods lack sensitivity to long-term trends (such as equipment aging or efficiency decline), and can easily overlook the gradual degradation of solar panels.

[0004] Therefore, how to ensure the accuracy of abnormal pattern recognition in solar cell operation monitoring data has become an urgent problem to be solved. Summary of the Invention

[0005] In view of this, an embodiment of the present invention provides a method for evaluating abnormal conditions during operation of a solar cell, so as to solve the problem of how to ensure the accuracy of abnormal pattern recognition in solar cell operation monitoring data.

[0006] An embodiment of the present invention provides a method for evaluating an abnormal state during operation of a solar cell. The method includes the following steps:

[0007] Collect all monitoring data during the operation of the solar panels of the power station to obtain multi-dimensional operation monitoring time series data of each solar panel, divide the multi-dimensional operation monitoring time series data of all solar panels into windows, and obtain multi-dimensional operation monitoring time series data windows of all solar panels;

[0008] For any solar panel, a multi-dimensional operation monitoring time series data window of the solar panel is obtained, and the importance of each dimension is determined according to the change information of each dimension data in the multi-dimensional operation monitoring time series data window of the solar panel; the short-term abnormality degree of the window is determined according to the importance of each dimension in the multi-dimensional operation monitoring time series data window of the solar panel; a threshold value is determined according to the short-term abnormality degree of the window of the multi-dimensional operation monitoring time series data window of the solar panel, a short-term abnormality window is obtained, and a trend analysis factor is determined according to the frequency change of the short-term abnormality window;

[0009] Determining a distance metric evaluation factor of the multidimensional operation monitoring time series data window according to a trend analysis factor and a short-term abnormality degree of the multidimensional operation monitoring time series data window of the solar panel;

[0010] According to the distance metric evaluation factor of the multidimensional abnormality monitoring time series data window, the timing optimization distance metric between the multidimensional operation monitoring time series data windows is determined, and the clustering process is completed based on the timing optimization distance metric, and the abnormal state evaluation is performed according to the result of the clustering process.

[0011] Preferably, determining the importance of a dimension based on change information of data of each dimension in the multi-dimensional operation monitoring time series data window of the solar panel includes:

[0012] According to the distance between the monitoring data points in each dimension in the multidimensional operation monitoring time series data window of the solar panel, the connectivity path distance sequence of each monitoring data point in the multidimensional monitoring time series data window in each monitoring dimension is determined; the absolute value of the difference between the connectivity path distance sequence of the multidimensional monitoring data point corresponding to each timestamp in the window in each dimension and the connectivity path distance sequences of other timestamps in the dimension at the corresponding position of the path distance sequence is taken as the first difference, the reciprocal of the path number in the path distance sequence is used as the weight to perform weighted summation on the first difference and divide the sum by the number of paths in the path distance sequence to obtain a first path distance difference mean; normalization processing is performed according to the mean of the first path distance difference means between the connectivity path distance sequence of the multidimensional monitoring data point corresponding to each timestamp in the window in each dimension and the connectivity path distance sequences of all other timestamps in the dimension to obtain a normalization processing result, and the normalization processing result is used as the importance of the multidimensional monitoring data point corresponding to each timestamp in the window in each dimension.

[0013] Preferably, determining the connectivity path distance sequence of each monitoring data point in each monitoring dimension in the multidimensional monitoring time series data window according to the distance between monitoring data points in each dimension in the multidimensional operation monitoring time series data window of the solar panel comprises:

[0014] For any monitoring data point in any dimension of the window, the monitoring data point is taken as the starting point, and the path traversal is performed in the window according to the minimum Euclidean distance, and the path corresponding to the traversal result is taken as the connected path; the Euclidean distance between the two data points corresponding to each sub-path in the connected path is taken as the path distance of the sub-path, and the connected path distance sequence is obtained according to the path distance of each sub-path in the connected path.

[0015] Preferably, determining the short-term abnormality degree of the window according to the importance of each dimension in the solar panel multi-dimensional operation monitoring time series data window includes:

[0016] Obtain the importance of the multidimensional monitoring data point corresponding to each timestamp in the window in each dimension, and take the average of the importance of all multidimensional monitoring data points in any dimension in any window as the window dimension importance of the dimension; take the average of the window dimension importance of any dimension in all windows as the overall dimension importance of the dimension; divide the overall dimension importance of any dimension by the sum of the overall importance of all dimensions as the fusion importance of the dimension; take the average of the monitoring data of any dimension in the window as the first mean of the window;

[0017] For any dimension in the window, starting from the second data point, the importance of the data point is multiplied by the difference in the monitoring value between the data point and the previous data point as the first difference value; based on the importance as the weight, the result of weighted mean calculation of all the first difference values ​​of any dimension in the window is used as the second difference value of the dimension; the square of the result of subtracting the monitoring value corresponding to any timestamp of any dimension in the window from the first mean of the window is used as the first mean difference; the result of weighted mean of all timestamps in any dimension in the window raised to the power of half of the importance corresponding to each timestamp is used as the first fluctuation degree of the dimension;

[0018] The result of calculating the first fluctuation degree of all dimensions in the window to the power of one-half after weighted mean according to the fusion importance is used as the second fluctuation degree of the window; the result of calculating the second difference value of all dimensions in the window to the power of one-half after weighted mean according to the fusion importance is used as the third difference value of the window;

[0019] A second fluctuation degree of the window and a third difference value of the window are obtained, and the second fluctuation degree of the window is divided by the third difference value of the window and normalized to obtain a result as the short-term abnormality degree of the window.

[0020] Preferably, the method of performing threshold judgment based on the degree of short-term abnormality of the multi-dimensional operation monitoring time series data window of the solar panel, obtaining the short-term abnormal window and determining the trend analysis factor based on the frequency change of the short-term abnormal window includes:

[0021] Obtain a short-term anomaly threshold, judge the short-term anomaly degree of the window according to the short-term anomaly threshold, and obtain all short-term anomaly windows; obtain a second window, and obtain the short-term anomaly frequency change factor in the second window according to the short-term anomaly change information of the first window in the second window; obtain the least squares fitting result of the multidimensional monitoring time series data in the second window through the least squares fitting result according to the multidimensional monitoring time series data in the second window, and use the average slope of the fitting result corresponding to each dimension in the monitoring time series data in the second window according to the fusion importance of the dimension as the window trend feature factor in the second window Sub-process; obtaining a trend item decomposition result of all multidimensional monitoring time series data in the second window through trend item decomposition according to all multidimensional monitoring time series data of the solar panel, and performing weighted mean calculation on the average slope of the trend item decomposition result of each dimension in the second window by the fusion importance of the dimension as the overall window trend characteristic factor of the trend item; obtaining the window trend characteristic factor in the second window and the overall window trend characteristic factor of the trend item in the second window, and calculating the absolute value of the subtraction between the window trend characteristic factor in the second window and the overall window trend characteristic factor of the trend item in the second window as the first trend characteristic factor;

[0022] Obtain the short-term abnormal frequency change factor in the second window and the first trend characteristic factor of the second window, take the product of the short-term abnormal frequency change factor and the first trend characteristic factor as the second trend characteristic factor, and take the result of subtracting the constant 1 from the second trend characteristic factor as the trend analysis factor of the second window.

[0023] Preferably, the trend analysis factor and the short-term abnormality degree of the solar panel multi-dimensional operation monitoring time series data window are used to determine the distance metric evaluation factor of the multi-dimensional operation monitoring time series data window, including:

[0024] Obtain the trend analysis factor of the second window and the short-term abnormality degree of the first window, and use the trend analysis factor of the second window as the trend analysis factor of all first windows in the second window; use the calculation result of subtracting the constant 1 from the trend analysis factor of the first window as the first trend characteristic item of the first window; use the calculation result of multiplying the short-term abnormality degree of the first window by the first trend characteristic item of the first window as the distance measurement evaluation factor of the multidimensional operation monitoring time series data window.

[0025] Preferably, determining the time series optimization distance metric between the multidimensional operation monitoring time series data windows according to the distance metric evaluation factor of the multidimensional abnormality monitoring time series data window, and completing the clustering process based on the distance metric, includes:

[0026] Obtain two multidimensional monitoring data that need to be distance-measured and the first windows corresponding to the two multidimensional monitoring data, use the average of the distance measurement evaluation factors of the first windows corresponding to the two multidimensional monitoring data as the timing difference weight, multiply the timing difference weight by the DTW distance between the first windows corresponding to the two multidimensional monitoring data as the timing difference metric; subtract a constant 1 from the timing difference weight as the Euclidean distance difference weight, and multiply the Euclidean distance difference weight by the Euclidean distance between the two multidimensional monitoring data as the Euclidean distance difference metric;

[0027] Obtain the time series difference metric and the Euclidean distance difference metric of the two multidimensional monitoring data and the two multidimensional monitoring time series data windows corresponding to the two multidimensional monitoring data, add the time series difference metric and the Euclidean distance difference metric as the time series optimization distance metric; and perform the optimization according to the time series optimization distance metric between the two multidimensional monitoring data. Clustering obtains clustering results.

[0028] Preferably, performing abnormal state assessment according to the result of the clustering process includes:

[0029] Obtain the result of the clustering process, and use the average Euclidean distance between the cluster center point of any cluster in the clustering result and the center points of all other clusters as the first Euclidean distance of the cluster; and use the average Euclidean distance between the cluster data points in any cluster in the clustering result and the cluster center point of the cluster as the second Euclidean distance of the cluster;

[0030] Obtaining a first Euclidean distance of the cluster and a second Euclidean distance of the cluster, multiplying the first Euclidean distance by the second Euclidean distance and performing normalization processing on the result as the abnormality degree of the cluster;

[0031] The cluster abnormality degree threshold is obtained, and the abnormality degree of the cluster is compared with the cluster abnormality degree threshold. If the abnormality degree of the cluster is greater than or equal to the cluster abnormality degree threshold, the cluster is determined to be an abnormal state cluster; if the abnormality degree of the cluster is less than the cluster abnormality degree threshold, the cluster is determined to be a normal state cluster, thereby completing the abnormal state assessment during the operation of the solar cell.

[0032] Preferably, the obtaining of the second window, obtaining a short-term abnormal frequency change factor in the second window according to the short-term abnormal change information of the first window in the second window; and obtaining a window trend characteristic factor in the second window according to the short-term abnormal frequency change factor in the second window through a least squares fitting result, includes:

[0033] The second window contains at least three first windows, the number of short-time abnormal first windows in the second window is used as the short-time abnormal frequency change assessment factor of the second window, a short-time abnormal frequency factor sequence is obtained according to the short-time abnormal frequency change assessment factors of all second windows in the multi-dimensional operation monitoring time series data of the solar panel, the short-time abnormal frequency factor sequence is fitted with the least squares method to obtain the short-time abnormal frequency change assessment factor fitting result, and the normalized calculation result of the slope corresponding to the second window in the short-time abnormal frequency factor fitting result is used as the short-time abnormal frequency change factor in the second window.

[0034] Preferably, the step of dividing the multi-dimensional operation monitoring time series data of all solar panels into windows to obtain the multi-dimensional operation monitoring time series data windows of all solar panels includes:

[0035] The multidimensional operation monitoring time series data of the solar panel and a preset window length are respectively obtained; the multidimensional operation monitoring time series data of the solar panel are traversed continuously with a constant step length of 1 according to the preset window length, and a subsequence contained in each movement of the window of the preset window length is used as the multidimensional operation monitoring time series data window corresponding to the center point of the window.

[0036] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0037] The present invention collects all monitoring data during the operation of solar panels in a power station, collects all monitoring data during the operation of solar panels in a power station, obtains multi-dimensional operation monitoring time series data of each solar panel, divides the multi-dimensional operation monitoring time series data of all solar panels into windows, and obtains multi-dimensional operation monitoring time series data windows of all solar panels; for any solar panel, obtains the multi-dimensional operation monitoring time series data window of the solar panel, determines the importance of the dimension according to the change information of the data of each dimension in the multi-dimensional operation monitoring time series data window of the solar panel; determines the importance of the dimension according to the change information of the data of each dimension in the multi-dimensional operation monitoring time series data window of the solar panel The method comprises the following steps: determining the degree of short-term abnormality of the window according to the importance of the window; performing a threshold judgment according to the short-term abnormality degree of the solar panel multi-dimensional operation monitoring time series data window, obtaining the short-term abnormality window, and determining a trend analysis factor according to the frequency change of the short-term abnormality window; determining a distance metric evaluation factor of the multi-dimensional operation monitoring time series data window according to the trend analysis factor and the short-term abnormality degree of the solar panel multi-dimensional operation monitoring time series data window; determining a time series optimized distance metric between the multi-dimensional operation monitoring time series data windows according to the distance metric evaluation factor of the multi-dimensional abnormality monitoring time series data window, completing a clustering process based on the time series optimized distance metric, and performing an abnormal state evaluation according to the result of the clustering process. Among them, by introducing the time series difference measurement of multi-dimensional monitoring time series data in the clustering process, the interference of fluctuations caused by short-term anomalies on clustering is eliminated, thereby reducing misjudgment and improving the accuracy of abnormal state identification. In the clustering process for abnormal state assessment, the Euclidean distance and time series difference measurement are dynamically evaluated through the distance measurement evaluation factor, which can effectively deal with environmental fluctuations caused by seasonal changes or weather factors and enhance the adaptability to these environmental factors. By dynamically adjusting the weight of the distance measurement evaluation factor, short-term environmental fluctuations are avoided from being misjudged as abnormal states, thereby improving the adaptability of the abnormal assessment method to environmental changes. In the introduction of the trend analysis factor, the distance measurement evaluation factor is adaptively adjusted according to different operating data and abnormal patterns, thereby ensuring that the abnormal detection method can dynamically adjust the detection strategy, distinguish short-term weather fluctuations from long-term recession trends, and improve the accuracy of abnormal state assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The accompanying drawings, which constitute part of the present invention, are provided to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are provided to explain the present invention and do not constitute an undue limitation of the present invention. In the accompanying drawings:

[0039] Figure 1 This is a flow chart of a method for evaluating abnormal conditions during operation of a solar cell, provided in Example 1 of the present invention; DETAILED DESCRIPTION

[0040] The following describes in detail embodiments of the present disclosure, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present disclosure, and should not be construed as limiting the present disclosure.

[0041] It should be noted that the terms "first," "second," and the like in the specification of the present disclosure and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate so that the embodiments of the present disclosure described herein can be implemented in an unexpected order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure.

[0042] In order to illustrate the technical solution of the present invention, specific embodiments are provided below.

[0043] The specific scenarios targeted by this invention are:

[0044] See also Figure 1 , is a flow chart of a method for evaluating abnormal conditions during operation of a solar cell provided by the first embodiment of the present invention, such as Figure 1 As shown, a method for evaluating an abnormal state during operation of a solar cell may include:

[0045] Step S101: All monitoring data during the operation of solar panels in a power station are collected to obtain multi-dimensional operation monitoring time series data of each solar panel, and the multi-dimensional operation monitoring time series data of all solar panels are divided into windows to obtain multi-dimensional operation monitoring time series data windows of all solar panels.

[0046] In a solar power station, each solar panel is equipped with multiple monitoring sensors for collecting operational monitoring data. These sensors include temperature sensors, voltage sensors, current sensors, power sensors, and irradiance sensors. Therefore, in the embodiments of the present invention, all multi-dimensional monitoring time series data for each solar panel in the power station is collected. It should be noted that the multi-dimensional operational monitoring time series data for the solar panels is all known historical monitoring data, and is collected once every minute. This is not a limitation and can be set based on real-time scenarios.

[0047] After obtaining the multi-dimensional operation monitoring time series data of each solar cell panel, it is necessary to divide the window of the multi-dimensional operation monitoring time series data of the solar cell panel according to the preset time series window length. Preferably, the window length is set in the embodiment of the present invention. The window traverses all the multi-dimensional operation monitoring time series data of the solar panel with a step length of 1. Each time the window moves, the data in the window is used as the multi-dimensional operation monitoring time series data window corresponding to the center point of the current window.

[0048] Step S102: For any solar panel, a multi-dimensional operation monitoring time series data window of the solar panel is obtained, and the importance of each dimension is determined based on the change information of the data of each dimension in the window; the short-term abnormality degree of the window is determined based on the importance of each dimension in the window; a threshold value is determined based on the short-term abnormality degree of the window to obtain a short-term abnormal window; and a trend analysis factor is determined based on the frequency change of the short-term abnormal window.

[0049] After obtaining the multi-dimensional operation monitoring time series data of each solar panel and the multi-dimensional operation monitoring time series data window, the operation mode of the multi-dimensional monitoring data points of the solar panel can be identified and divided using a clustering algorithm based on the differences between the multi-dimensional monitoring data points of the solar panel. Traditionally, when using a clustering algorithm to identify and divide the operation mode of the solar panel operation monitoring data, it is achieved based on the Euclidean distance between the multi-dimensional operation monitoring data points of the solar panel. However, when the solar panel experiences a sudden change in light or a change in the ambient wind speed, the multi-dimensional monitoring data points may undergo a sudden change, resulting in short-term anomalies being identified as abnormal modes in the clustering-based solar panel operation mode identification process. Therefore, when measuring the distance between the multi-dimensional operation monitoring data points of the solar panel in the clustering process, it is necessary to comprehensively measure the Euclidean distance between the multi-dimensional operation monitoring data points of the solar panel in combination with the time series change differences between the multi-dimensional operation monitoring time series data windows corresponding to the multi-dimensional operation monitoring time series data points of the solar panel, so as to achieve the purpose of optimizing the Euclidean distance in the traditional clustering process.

[0050] In the process of optimizing the time series change difference of the Euclidean distance between the multidimensional monitoring data points of the solar panel, because different degrees of time series change differences need to be optimized under different operating conditions, it is necessary to evaluate the distance measurement weight between the Euclidean distance between the multidimensional monitoring data points and the time series difference between the multidimensional operation monitoring time series data windows corresponding to the multidimensional monitoring data points, that is, the distance measurement evaluation factor, which is used for the optimized distance measurement between the multidimensional operation monitoring data of the solar panel in the clustering process. In the process of evaluating the distance metric evaluation factor, since the monitoring data of the solar panel is multidimensional data, each timestamp in the multidimensional operation monitoring time series data corresponds to the data of multiple monitoring dimensions of the solar panel at the sampling time. Therefore, in the process of analysis, it is necessary to evaluate the dimensional importance of the monitoring data points in each multidimensional operation monitoring time series data window, so as to perform multidimensional information fusion through the dimensional importance, so as to ensure the accuracy of the distance metric evaluation factor in the distance optimization process. After obtaining the importance of the data points in each dimension in the time series window, because the performance degradation caused by long-term use of the solar panel causes abnormal changes in the multidimensional monitoring data points similar to short-term abnormalities, it will appear as a continuous increase in the frequency of short-term abnormal windows. Therefore, after obtaining the multidimensional operation monitoring time series data of each solar panel, the short-term abnormality degree of the window is further determined according to the importance of each dimension in the window, and a threshold judgment is performed according to the short-term abnormality degree to obtain the short-term abnormality time series data window in all the multidimensional operation monitoring time series data windows of the solar panel, and the short-term abnormality frequency of the short-term abnormality time series data window is evaluated through the second window, so as to obtain the trend analysis factor of the short-term abnormality window.

[0051] The specific optimization process is as follows: determine the importance of the dimension based on the change information of the data of each dimension in the window; determine the short-term abnormality of the window based on the importance of each dimension in the window; perform threshold judgment based on the short-term abnormality of the window to obtain the short-term abnormal window; determine the trend analysis factor based on the frequency change of the short-term abnormal window.

[0052] The importance of a dimension is determined according to the change information of each dimension data in the window, including:

[0053] According to the distance between the monitoring data points in each dimension in the window, determine the connectivity path distance sequence of each monitoring data point in the multidimensional monitoring time series data window in each monitoring dimension; take the absolute value of the difference between the connectivity path distance sequence of the multidimensional monitoring data point corresponding to each timestamp in the window in each dimension and the connectivity path distance sequence of other timestamps in the dimension at the corresponding position of the path distance sequence as the first difference, use the reciprocal of the path number in the path distance sequence as the weight to perform weighted summation on the first difference and divide it by the number of paths in the path distance sequence to obtain the first path distance difference mean; perform normalization processing on the mean of the first path distance difference between the connectivity path distance sequence of the multidimensional monitoring data point corresponding to each timestamp in the window in each dimension and the connectivity path distance sequences of all other timestamps in the dimension to obtain the normalization processing result, and use the normalization processing result as the importance of the multidimensional monitoring data point corresponding to each timestamp in the window in each dimension.

[0054] In one embodiment, the Take the solar panel multi-dimensional operation monitoring time series data window as an example, record the The window is , No. In the window In the dimension The connected path distance sequence of the data point corresponding to the timestamp The path distance is , then the solar panel In the window In the dimension The calculation expression for the importance of the data point corresponding to the timestamp is:

[0055] ;

[0056] in, Indicates the solar panel In the window In the dimension The importance of the data point corresponding to the timestamp, represents the normalization function, represents the number of data points in the window, represents the path number in the path distance sequence, Indicates the In the window In the dimension The path distance sequence corresponding to the timestamp The path distance and In the window In the dimension The path distance sequence corresponding to the timestamp The first difference between the path distances.

[0057] The step of determining the connected path distance sequence of each monitoring data point in each monitoring dimension in the multi-dimensional monitoring time series data window according to the distance between the monitoring data points in each dimension in the window includes:

[0058] For any monitoring data point in any dimension of the window, the monitoring data point is taken as the starting point, and the path traversal is performed in the window according to the minimum Euclidean distance, and the path corresponding to the traversal result is taken as the connected path; the Euclidean distance between the two data points corresponding to each sub-path in the connected path is taken as the path distance of the sub-path, and the connected path distance sequence is obtained according to the path distance of each sub-path in the connected path.

[0059] After obtaining the importance of the multidimensional monitoring data point corresponding to each timestamp in the window in each dimension, it is necessary to determine the short-term abnormality degree of the window according to the importance of each dimension in the window, so as to judge the short-term abnormality window through the short-term abnormality degree of the window, and thus evaluate the trend analysis factor through the frequency changes presented by the short-term abnormality window.

[0060] The step of determining the short-term abnormality of a window according to the importance of each dimension in the window includes:

[0061] Obtain the importance of the multidimensional monitoring data point corresponding to each timestamp in the window in each dimension, and take the average of the importance of all multidimensional monitoring data points in any dimension in any window as the window dimension importance of that dimension;

[0062] In one embodiment, the In the window The calculation expression of the window dimension importance of the dimension is:

[0063] ;

[0064] in, Indicates the In the window The importance of the window dimension of each dimension, represents the number of timestamps in the window, i.e. the window length, Indicates the In the time series data window The multidimensional data point corresponding to the timestamp The importance of a dimension.

[0065] The average of the window dimension importance of any dimension in all windows is taken as the overall dimension importance of the dimension;

[0066] In one embodiment, the The calculation expression of the overall dimension importance of a dimension is:

[0067] ;

[0068] in, Indicates the The overall dimension importance of each dimension, Indicates the In the window The importance of the window dimension of each dimension, Represents the total number of windows in the multi-dimensional operation monitoring time series data of all solar panels.

[0069] The overall importance of any dimension is divided by the sum of the overall importance of all dimensions as the fusion importance of the dimension;

[0070] In one embodiment, the The calculation expression of the fusion importance of each dimension is:

[0071] ;

[0072] in, Indicates the The importance of integration of the dimensions, Indicates the The overall dimension importance of each dimension, Indicates the The overall dimension importance of each dimension, Indicates the total number of dimensions.

[0073] The mean of the monitoring data in any dimension of the window is taken as the first mean of the window;

[0074] In one embodiment, the In the window In the dimension The monitoring data corresponding to the timestamp is , then The window in The calculation expression of the first mean of the window in dimension is:

[0075] ;

[0076] in, Indicates the The window in The window-first mean in dimension, Indicates the In the window In the dimension The monitoring data corresponding to the timestamp, Indicates the number of timestamps in the window.

[0077] For any dimension in the window, starting from the second data point, the importance of the data point is multiplied by the difference in monitoring values ​​between the data point and the previous data point as a first difference value;

[0078] In one embodiment, the In the window In the dimension The monitoring data corresponding to the timestamp is , then In the time series data window The dimension in The calculation expression for the first difference value of the data point corresponding to the timestamp is:

[0079] ;

[0080] in, Indicates the In the time series data window The dimension in The first difference value of the data point corresponding to the timestamp, Indicates the In the window In the dimension The monitoring data corresponding to the timestamp, Indicates the In the window In the dimension The monitoring data corresponding to the timestamp, Indicates the In the time series data window The multidimensional data point corresponding to the timestamp The importance of the dimension, Indicates the absolute value calculation symbol.

[0081] It should be noted that for In the time series data window The dimension in The first difference value of the data point corresponding to the timestamp is only Start acquiring data points.

[0082] According to the importance, all the first difference values ​​of any dimension in the window are weighted and the result of weighted mean calculation is used as the second difference value of the dimension;

[0083] In one embodiment, the In the window The calculation expression of the second difference value of the dimension is:

[0084] ;

[0085] in, Indicates the In the window The second difference value of the dimension, Indicates the In the time series data window The multidimensional data point corresponding to the timestamp The importance of the dimension, Indicates the In the time series data window The dimension in The first difference value of the data point corresponding to the timestamp, represents the number of timestamps in the window, Indicates the absolute value calculation symbol.

[0086] The square of the result of subtracting the monitoring value corresponding to any timestamp of any dimension in the window from the first mean of the window is used as the first mean difference; the result of calculating the weighted mean of the first mean difference of all timestamps in any dimension in the window by raising the weighted mean of the importance corresponding to each timestamp to the power of one-half is used as the first fluctuation degree of the dimension;

[0087] In one embodiment, the In the window In the dimension The monitoring data corresponding to the timestamp is , then In the window The calculation expression of the first fluctuation degree of each dimension is:

[0088] ;

[0089] in, Indicates the In the window The first degree of fluctuation of the dimension, Indicates the In the time series data window The multidimensional data point corresponding to the timestamp The importance of the dimension, Indicates the In the window In the dimension The monitoring data corresponding to the timestamp, Indicates the The window in The window-first mean in dimension, Indicates the number of timestamps in the window.

[0090] The result of calculating the second fluctuation degree of the window by raising the square of the first fluctuation degree of all dimensions in the window to the power of one-half after performing weighted mean according to the fusion importance is used as the second fluctuation degree of the window;

[0091] In one embodiment, the The calculation expression of the second fluctuation degree of a window is:

[0092] ;

[0093] in, Indicates the The second degree of fluctuation of the window, Indicates the The importance of integration of the dimensions, Indicates the In the window The square of the first fluctuation degree of the dimension, Indicates the total number of dimensions.

[0094] The result of calculating the second difference value of all dimensions in the window to the power of half after performing weighted mean according to the fusion importance is used as the third difference value of the window;

[0095] In one embodiment, the The calculation expression of the third difference value of a window is:

[0096] ;

[0097] in, Indicates the The third difference value of the window, Indicates the The importance of integration of the dimensions, Indicates the In the window The second difference value of the dimension, Indicates the total number of dimensions.

[0098] A second fluctuation degree of the window and a third difference value of the window are obtained, and the second fluctuation degree of the window is divided by the third difference value of the window and normalized to obtain a result as the short-term abnormality degree of the window.

[0099] In one embodiment, the The calculation expression of the short-term abnormality degree of a window is:

[0100] ;

[0101] in, Indicates the The short-term abnormality of the window, Indicates the The second degree of fluctuation of the window, Indicates the The third difference value of the window. Represents the normalization function.

[0102] It should be noted that the short-term abnormality of the time series data window is evaluated by using the fluctuation degree of the monitored value in the time series data window and the change information of the monitored data in the window. The higher the value, the higher the short-term abnormality. The third difference value of the window based on the short-term abnormality of the window The larger it is, the lower the short-term abnormality of the window.

[0103] After obtaining the short-term abnormality degree of the solar panel multi-dimensional operation time series data window, a short-term abnormality window is judged according to the short-term abnormality degree of the time series data window using a short-term abnormality threshold, thereby evaluating the trend analysis factor based on the change in the number of short-term windows and the overall trend of the solar panel multi-dimensional operation monitoring time series data.

[0104] The step of performing threshold judgment based on the short-term abnormality degree of the window, obtaining the short-term abnormal window, and determining the trend analysis factor based on the frequency change of the short-term abnormal window includes:

[0105] Obtain a short-term anomaly threshold, judge the short-term anomaly degree of the window according to the short-term anomaly threshold, and obtain all short-term anomaly windows; obtain a second window, and obtain the short-term anomaly frequency change factor in the second window according to the short-term anomaly change information of the first window in the second window; obtain the least squares fitting result of the multidimensional monitoring time series data in the second window through the least squares fitting result according to the multidimensional monitoring time series data in the second window, and use the average slope of the fitting result corresponding to each dimension in the monitoring time series data in the second window according to the fusion importance of the dimension as the window trend feature factor in the second window Sub-process; obtaining a trend item decomposition result of all multidimensional monitoring time series data in the second window through trend item decomposition according to all multidimensional monitoring time series data of the solar panel, and performing weighted mean calculation on the average slope of the trend item decomposition result of each dimension in the second window by the fusion importance of the dimension as the overall window trend characteristic factor of the trend item; obtaining the window trend characteristic factor in the second window and the overall window trend characteristic factor of the trend item in the second window, and calculating the absolute value of the subtraction between the window trend characteristic factor in the second window and the overall window trend characteristic factor of the trend item in the second window as the first trend characteristic factor;

[0106] In one embodiment, a short-term abnormality threshold of a window is set to perform short-term abnormal window judgment. Preferably, the short-term abnormality threshold of the window is set to , the windows whose short-term abnormality degree is higher than the short-term abnormality degree threshold are identified as short-term abnormal windows, thereby completing the short-term abnormal window judgment.

[0107] Obtain the short-term abnormal frequency change factor in the second window and the first trend characteristic factor of the second window, take the product of the short-term abnormal frequency change factor and the first trend characteristic factor as the second trend characteristic factor, and take the result of subtracting the constant 1 from the second trend characteristic factor as the trend analysis factor of the second window.

[0108] The acquiring of the second window, acquiring the short-term abnormal frequency change factor in the second window according to the short-term abnormal change information of the first window in the second window, includes:

[0109] The second window contains at least three first windows, the number of short-time abnormal first windows in the second window is used as the short-time abnormal frequency change assessment factor of the second window, a short-time abnormal frequency factor sequence is obtained according to the short-time abnormal frequency change assessment factors of all second windows in the multi-dimensional operation monitoring time series data of the solar panel, the short-time abnormal frequency factor sequence is fitted with the least squares method to obtain the short-time abnormal frequency change assessment factor fitting result, and the normalized calculation result of the slope corresponding to the second window in the short-time abnormal frequency factor fitting result is used as the short-time abnormal frequency change factor in the second window.

[0110] In one embodiment, the The short-term abnormal frequency change factor in the second time series data window is , No. The first trend characteristic factor in the second time series data window is , then The calculation expression of the trend analysis factor of the second time series data window is:

[0111] ;

[0112] in, Indicates the The trend analysis factor of the second time series data window, Indicates the The short-term abnormal frequency change factor in the second time series data window, Indicates the The first trend characteristic factor in the second time series data window, represents the normalization function, and 1 represents the constant 1.

[0113] Step S103 : determining a distance metric evaluation factor of the multi-dimensional operation monitoring time series data window according to the trend analysis factor and the short-term abnormality degree of the multi-dimensional operation monitoring time series data window of the solar panel.

[0114] After obtaining the trend analysis factor of the solar panel multi-dimensional operation monitoring time series data window, the trend analysis factor can be used to optimize and evaluate the long-term trend change information appearing in the window during the evaluation process of the distance metric evaluation factor of the distance optimization metric, thereby avoiding the long-term attenuation trend appearing in the operation process of the solar panel from being eliminated in the distance metric based on the time series optimization, thereby ensuring accurate abnormal state evaluation of the solar panel multi-dimensional operation monitoring data.

[0115] Determining the distance metric evaluation factor of the multi-dimensional operation monitoring time series data window according to the trend analysis factor and the short-term abnormality degree of the solar panel multi-dimensional operation monitoring time series data window includes:

[0116] Obtain the trend analysis factor of the second window and the short-term abnormality degree of the first window, and use the trend analysis factor of the second window as the trend analysis factor of all first windows in the second window; use the calculation result of subtracting the constant 1 from the trend analysis factor of the first window as the first trend characteristic item of the first window; use the calculation result of multiplying the short-term abnormality degree of the first window by the first trend characteristic item of the first window as the distance measurement evaluation factor of the multidimensional operation monitoring time series data window.

[0117] In one embodiment, the The calculation expression of the distance metric evaluation factor of the multi-dimensional operation monitoring time series data window of a solar panel is:

[0118] ;

[0119] in, Indicates the The distance metric evaluation factor of the multi-dimensional operation monitoring time series data window of a solar panel, Indicates the The short-term abnormality of the window, Indicates the Trend analysis factor for a time series data window.

[0120] It should be noted that the distance metric evaluation factor is evaluated using the short-term anomaly degree and trend analysis factor of the solar panel multi-dimensional operation monitoring time series data window. The higher the short-term anomaly degree, the more likely short-term anomalies are to occur in the window, such as sudden changes in monitoring data due to weather changes. The higher the distance metric evaluation factor, the higher the trend analysis factor of the time series data window, the higher the trend change in the current window, and the smaller the distance metric evaluation factor.

[0121] Step S104, determining the distance metric between the multidimensional operation monitoring time series data windows according to the distance metric evaluation factor of the multidimensional operation monitoring time series data windows, completing the clustering process based on the distance metric, and performing abnormal state evaluation according to the results of the clustering process.

[0122] After obtaining the distance metric evaluation factor of the multidimensional operation monitoring time series data window, the distance metric evaluation factor of the window is used to optimize the distance metric in the clustering process of the solar panel multidimensional monitoring data, thereby obtaining accurate clustering results.

[0123] The step of determining the time series optimization distance metric between the multidimensional operation monitoring time series data windows according to the distance metric evaluation factor of the multidimensional abnormality monitoring time series data window and completing the clustering process based on the distance metric includes:

[0124] Obtain two multidimensional monitoring data that need to be distance-measured and the first windows corresponding to the two multidimensional monitoring data, use the average of the distance measurement evaluation factors of the first windows corresponding to the two multidimensional monitoring data as the timing difference weight, multiply the timing difference weight by the DTW distance between the first windows corresponding to the two multidimensional monitoring data as the timing difference metric; subtract a constant 1 from the timing difference weight as the Euclidean distance difference weight, and multiply the Euclidean distance difference weight by the Euclidean distance between the two multidimensional monitoring data as the Euclidean distance difference metric;

[0125] Obtaining a time series difference metric and a Euclidean distance difference metric between the two multidimensional monitoring data and two multidimensional monitoring time series data windows corresponding to the two multidimensional monitoring data, and adding the time series difference metric and the Euclidean distance difference metric as the time series optimized distance metric;

[0126] In one embodiment, assuming that The multi-dimensional operation monitoring data of solar panels is , No. The multi-dimensional operation monitoring data of solar panels is , then the calculation expression of the time series optimization distance metric between the two multi-dimensional operation monitoring data points is:

[0127] ;

[0128] in, Indicates the Multi-dimensional operation monitoring data of solar panels and The time series optimization distance measurement between the multi-dimensional operation monitoring data of solar panels, represents the mean value of the distance measurement evaluation factor between the two multi-dimensional operation monitoring data points, that is, the time series difference weight, represents the Euclidean distance between two multidimensional monitoring data points, Represents the difference between the multidimensional operation monitoring time series data windows corresponding to two multidimensional monitoring data points distance, 1 means constant 1.

[0129] After obtaining the time series optimization distance metric between the two multi-dimensional operation monitoring data points, the time series optimization distance metric between the two multi-dimensional monitoring data points is used to perform Clustering obtains clustering results.

[0130] It should be noted that the Clustered The value was determined by the elbow method. Clustered The value is a known technology, so it will not be described here in detail.

[0131] After obtaining the clustering results of the solar panel operation monitoring data, an abnormal state assessment can be performed based on the results of the clustering process.

[0132] The abnormal state assessment according to the result of the clustering process includes:

[0133] Obtain the result of the clustering process, and use the average Euclidean distance between the cluster center point of any cluster in the clustering result and the center points of all other clusters as the first Euclidean distance of the cluster; and use the average Euclidean distance between the cluster data points in any cluster in the clustering result and the cluster center point of the cluster as the second Euclidean distance of the cluster;

[0134] Obtaining a first Euclidean distance of the cluster and a second Euclidean distance of the cluster, multiplying the first Euclidean distance by the second Euclidean distance and performing normalization processing on the result as the abnormality degree of the cluster;

[0135] In one embodiment, the The calculation expression of the abnormality degree of a cluster is:

[0136] ;

[0137] in, Indicates the The abnormality of each cluster, Indicates the The first Euclidean distance of the clusters, Indicates the The second Euclidean distance of the clusters, Represents the normalization function.

[0138] It should be noted that the first Euclidean distance and the second Euclidean distance of the cluster are used to evaluate the abnormality of the cluster. The larger the The farther the distance between a cluster and other clusters, the higher the degree of abnormality. The larger the The more dispersed the data points in a cluster are, the higher the degree of anomaly.

[0139] The cluster abnormality degree threshold is obtained, and the abnormality degree of the cluster is compared with the cluster abnormality degree threshold. If the abnormality degree of the cluster is greater than or equal to the cluster abnormality degree threshold, the cluster is determined to be an abnormal state cluster; if the abnormality degree of the cluster is less than the cluster abnormality degree threshold, the cluster is determined to be a normal state cluster, thereby completing the abnormal state assessment during the operation of the solar cell.

[0140] In one embodiment, a cluster abnormality threshold is set to evaluate the abnormal state of the clusters in the clustering result. Preferably, the cluster abnormality threshold is set to , the clusters with a degree of abnormality greater than the cluster abnormality threshold are identified as abnormal states in the operation process of the solar cell, thereby completing the abnormal state evaluation in the operation process of the solar cell.

[0141] In summary, the embodiment of the present invention collects all monitoring data during the operation of the solar panels of the power station, obtains the multi-dimensional operation monitoring time series data of each solar panel, divides the multi-dimensional operation monitoring time series data of all solar panels into windows, and obtains the multi-dimensional operation monitoring time series data windows of all solar panels; for any solar panel, obtains the multi-dimensional operation monitoring time series data window of the solar panel, determines the importance of the dimension according to the change information of each dimension data in the multi-dimensional operation monitoring time series data window of the solar panel; determines the window according to the importance of each dimension in the multi-dimensional operation monitoring time series data window of the solar panel. short-term abnormality degree; performing threshold judgment according to the short-term abnormality degree of the multi-dimensional operation monitoring time series data window of the solar panel, obtaining the short-term abnormality window and determining the trend analysis factor according to the frequency change of the short-term abnormality window; determining the distance metric evaluation factor of the multi-dimensional operation monitoring time series data window according to the trend analysis factor and the short-term abnormality degree of the multi-dimensional operation monitoring time series data window of the solar panel; determining the time series optimized distance metric between the multi-dimensional operation monitoring time series data windows according to the distance metric evaluation factor of the multi-dimensional abnormality monitoring time series data window, completing a clustering process based on the time series optimized distance metric, and performing abnormal state evaluation according to the result of the clustering process. Among them, by introducing the time series difference measurement of multi-dimensional monitoring time series data in the clustering process, the interference of fluctuations caused by short-term anomalies on clustering is eliminated, thereby reducing misjudgment and improving the accuracy of abnormal state identification. In the clustering process for abnormal state assessment, the Euclidean distance and the time series difference measurement are dynamically compared through the distance measurement evaluation factor, which can effectively handle environmental fluctuations caused by seasonal changes or weather factors and enhance the adaptability to these environmental factors. By dynamically adjusting the weight of the distance measurement evaluation factor, short-term environmental fluctuations are avoided from being misjudged as abnormal states, thereby improving the adaptability of the abnormal assessment method to environmental changes. In the introduction of the trend analysis factor, the distance measurement evaluation factor is adaptively adjusted according to different operating data and abnormal patterns, thereby ensuring that the abnormal detection method can dynamically adjust the detection strategy, distinguish short-term weather fluctuations from long-term recession trends, and improve the accuracy of abnormal state assessment.

[0142] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. A method for evaluating abnormal conditions during operation of a solar cell, characterized in that: The abnormal state evaluation method for a solar cell during operation includes: Collect all monitoring data during the operation of the solar panels of the power station, obtain multi-dimensional operation monitoring time series data of each solar panel, divide the multi-dimensional operation monitoring time series data of all solar panels into windows, and obtain multi-dimensional operation monitoring time series data windows of all solar panels; For any solar panel, a multi-dimensional operation monitoring time series data window of the solar panel is obtained, and the importance of the dimension is determined according to the change information of the data of each dimension in the multi-dimensional operation monitoring time series data window of the solar panel; the short-term abnormality of the window is determined according to the importance of each dimension in the multi-dimensional operation monitoring time series data window of the solar panel; a threshold value is judged according to the short-term abnormality of the window of the multi-dimensional operation monitoring time series data window of the solar panel, a short-term abnormal window is obtained, and a trend analysis factor is determined according to the frequency change of the short-term abnormal window; Determining a distance metric evaluation factor of the multidimensional operation monitoring time series data window according to a trend analysis factor and a short-term abnormality degree of the multidimensional operation monitoring time series data window of the solar panel; According to the distance metric evaluation factor of the multidimensional abnormal monitoring time series data window, the timing optimization distance metric between the multidimensional operation monitoring time series data windows is determined, and the clustering process is completed based on the timing optimization distance metric, and the abnormal state evaluation is performed according to the result of the clustering process.

2. The method for evaluating abnormal state during operation of a solar cell according to claim 1, characterized in that: The step of determining the importance of a dimension according to the change information of each dimension data in the multi-dimensional operation monitoring time series data window of the solar panel includes: According to the distances between the monitoring data points in each dimension in the multidimensional operation monitoring time series data window of the solar panel, the connected path distance sequence of each monitoring data point in the multidimensional monitoring time series data window in each monitoring dimension is determined; the absolute value of the difference between the connected path distance sequence of the multidimensional monitoring data point corresponding to each timestamp in the window in each dimension and the connected path distance sequence of other timestamps in the dimension at the corresponding position of the path distance sequence is taken as the first difference, the reciprocal of the path number in the path distance sequence is used as the weight to perform weighted summation on the first difference and divide it by the number of paths in the path distance sequence to obtain the first path distance difference mean; according to the mean of the first path distance difference means between the connected path distance sequence of the multidimensional monitoring data point corresponding to each timestamp in the window in each dimension and the connected path distance sequences of all other timestamps in the dimension, normalization processing is performed to obtain a normalization processing result, and the normalization processing result is used as the importance of the multidimensional monitoring data point corresponding to each timestamp in the window in each dimension.

3. The method for evaluating abnormal state during operation of a solar cell according to claim 2, characterized in that: Determining the connected path distance sequence of each monitoring data point in each monitoring dimension in the multi-dimensional monitoring time series data window according to the distance between the monitoring data points in each dimension in the multi-dimensional operation monitoring time series data window of the solar panel comprises: For any monitoring data point in any dimension of the window, the monitoring data point is taken as the starting point, and the path traversal is performed in the window according to the minimum Euclidean distance, and the path corresponding to the traversal result is taken as the connected path; the Euclidean distance between the two data points corresponding to each sub-path in the connected path is taken as the path distance of the sub-path, and the connected path distance sequence is obtained according to the path distance of each sub-path in the connected path.

4. The method for evaluating abnormal state during operation of a solar cell according to claim 1, characterized in that: Determining the short-term abnormality degree of the window according to the importance of each dimension in the solar panel multi-dimensional operation monitoring time series data window includes: Obtain the importance of the multidimensional monitoring data point corresponding to each timestamp in the window in each dimension, and take the average of the importance of all the multidimensional monitoring data points in any dimension in any window as the window dimension importance of the dimension; take the average of the window dimension importance of any dimension in all windows as the overall dimension importance of the dimension; take the overall dimension importance of any dimension divided by the sum of the overall importance of all dimensions as the fusion importance of the dimension; take the mean of the monitoring data in any dimension in the window as the first mean of the window; For any dimension in the window, starting from the second data point, the importance of the data point is multiplied by the difference in the monitored values ​​between the data point and the previous data point as the first difference value; according to the importance as the weight, the result of weighted mean calculation of all the first difference values ​​of any dimension in the window is taken as the second difference value of the dimension; the square of the result of subtracting the monitored value corresponding to any timestamp of any dimension in the window from the first mean of the window is taken as the first mean difference; the first mean difference of all timestamps in any dimension in the window is calculated by raising the weighted mean of the importance corresponding to each timestamp to the power of one-half as the first fluctuation degree of the dimension; The result of calculating the square of the first fluctuation degree of all dimensions in the window to the power of one-half after weighted mean according to the fusion importance is used as the second fluctuation degree of the window; the result of calculating the second difference value of all dimensions in the window to the power of one-half after weighted mean according to the fusion importance is used as the third difference value of the window; The second fluctuation degree of the window and the third difference value of the window are obtained, and the second fluctuation degree of the window is divided by the third difference value of the window and the result of normalization is taken as the short-term abnormality degree of the window.

5. The method for evaluating abnormal state during operation of a solar cell according to claim 1, characterized in that: The method of performing threshold judgment according to the short-term abnormality degree of the window of the multi-dimensional operation monitoring time series data window of the solar panel, obtaining the short-term abnormal window and determining the trend analysis factor according to the frequency change of the short-term abnormal window includes: Obtain a short-term anomaly threshold, judge the short-term anomaly degree of the window according to the short-term anomaly threshold, and obtain all short-term anomaly windows; obtain a second window, and obtain the short-term anomaly frequency change factor in the second window according to the short-term anomaly change information of the first window in the second window; obtain the least squares fitting result of the multidimensional monitoring time series data in the second window according to the least squares fitting result of the multidimensional monitoring time series data in the second window, and use the average slope of the fitting result corresponding to each dimension in the monitoring time series data in the second window according to the fusion importance of the dimension as the window trend feature factor in the second window. Sub; obtaining the trend item decomposition result of all the multidimensional monitoring time series data in the second window through trend item decomposition according to all the multidimensional monitoring time series data of the solar panel, and calculating the weighted average of the average slope of the trend item decomposition result of each dimension in the second window by the fusion importance of the dimension as the overall window trend characteristic factor of the trend item; obtaining the window trend characteristic factor in the second window and the overall window trend characteristic factor of the trend item in the second window, and calculating the absolute value of the subtraction between the window trend characteristic factor in the second window and the overall window trend characteristic factor of the trend item in the second window as the first trend characteristic factor; Obtain the short-time abnormal frequency change factor in the second window and the first trend characteristic factor of the second window, take the product of the short-time abnormal frequency change factor and the first trend characteristic factor as the second trend characteristic factor, and take the result of subtracting the constant 1 from the second trend characteristic factor as the trend analysis factor of the second window.

6. The method for evaluating abnormal state during operation of a solar cell according to claim 5, characterized in that: According to the trend analysis factor and the short-term abnormality degree of the solar panel multi-dimensional operation monitoring time series data window, determining the distance metric evaluation factor of the multi-dimensional operation monitoring time series data window includes: Obtain the trend analysis factor of the second window and the short-term abnormality degree of the first window, and use the trend analysis factor of the second window as the trend analysis factor of all first windows in the second window; use the result of subtracting the constant 1 from the trend analysis factor of the first window as the first trend characteristic item of the first window; use the result of multiplying the short-term abnormality degree of the first window by the first trend characteristic item of the first window as the distance measurement evaluation factor of the multidimensional operation monitoring time series data window.

7. The method for evaluating abnormal state during operation of a solar cell according to claim 1, characterized in that: Determining the time series optimization distance metric between the multidimensional operation monitoring time series data windows according to the distance metric evaluation factor of the multidimensional abnormal monitoring time series data window, and completing the clustering process based on the distance metric, includes: Obtain two multidimensional monitoring data that need to be distance measured and the first windows corresponding to the two multidimensional monitoring data, use the average of the distance measurement evaluation factors of the first windows corresponding to the two multidimensional monitoring data as the timing difference weight, multiply the timing difference weight by the DTW distance between the first windows corresponding to the two multidimensional monitoring data as the timing difference measurement; subtract a constant 1 from the timing difference weight as the Euclidean distance difference weight, and multiply the Euclidean distance difference weight by the Euclidean distance between the two multidimensional monitoring data as the Euclidean distance difference measurement; Obtain the timing difference metric and the Euclidean distance difference metric of the two multidimensional monitoring data and the two multidimensional monitoring time series data windows corresponding to the two multidimensional monitoring data, add the timing difference metric and the Euclidean distance difference metric as the timing optimization distance metric; and perform the optimization according to the timing optimization distance metric between the two multidimensional monitoring data. Clustering obtains clustering results.

8. The method for evaluating abnormal state during operation of a solar cell according to claim 1, characterized in that: The performing abnormal state assessment according to the result of the clustering process includes: Obtain the result of the clustering process, and use the average Euclidean distance between the cluster center point of any cluster in the clustering result and the center points of all other clusters as the first Euclidean distance of the cluster; use the average Euclidean distance between the cluster data points in any cluster in the clustering result and the cluster center point of the cluster as the second Euclidean distance of the cluster; Obtaining a first Euclidean distance of the cluster and a second Euclidean distance of the cluster, and multiplying the first Euclidean distance by the second Euclidean distance and performing normalization processing on the result as the abnormality degree of the cluster; The cluster abnormality degree threshold is obtained, and the abnormality degree of the cluster is compared with the cluster abnormality degree threshold. If the abnormality degree of the cluster is greater than or equal to the cluster abnormality degree threshold, the cluster is determined to be an abnormal state cluster; if the abnormality degree of the cluster is less than the cluster abnormality degree threshold, the cluster is determined to be a normal state cluster, thereby completing the abnormal state evaluation during the operation of the solar cell.

9. The method for evaluating abnormal state during operation of a solar cell according to claim 5, characterized in that: The acquiring of the second window, acquiring the short-term abnormal frequency change factor in the second window according to the short-term abnormal change information of the first window in the second window, comprises: The second window contains at least three first windows, the number of short-time abnormal first windows in the second window is used as the short-time abnormal frequency change assessment factor of the second window, a short-time abnormal frequency factor sequence is obtained according to the short-time abnormal frequency change assessment factors of all second windows in the solar panel multi-dimensional operation monitoring time series data, the short-time abnormal frequency factor sequence is fitted by least squares to obtain the short-time abnormal frequency change assessment factor fitting result, and the normalized calculation result of the slope corresponding to the second window in the short-time abnormal frequency factor fitting result is used as the short-time abnormal frequency change factor in the second window.

10. The method for evaluating abnormal state during operation of a solar cell according to claim 5, characterized in that: The multi-dimensional operation monitoring time series data of all solar panels are divided into windows to obtain the multi-dimensional operation monitoring time series data windows of all solar panels, including: The multidimensional operation monitoring time series data of the solar panel and the preset window length are respectively obtained; the multidimensional operation monitoring time series data of the solar panel are traversed continuously with a constant step length of 1 according to the preset window length, and the subsequence contained in each movement of the window of the preset window length is used as the multidimensional operation monitoring time series data window corresponding to the center point of the window.

Images