A fault diagnosis system for a photovoltaic power station

Abstract

This application relates to the field of photovoltaic system fault identification technology, specifically to a fault diagnosis system for photovoltaic power plants. The system includes: a fault diagnosis data acquisition module for real-time acquisition of various fault diagnosis data for each photovoltaic panel in the photovoltaic power plant to be analyzed; a fault diagnosis data analysis module for determining the response difference coefficient between each photovoltaic panel and its nearest neighbor photovoltaic panels; determining the fault response sensitivity coefficient of various fault diagnosis data for each photovoltaic panel based on the dispersion of various fault diagnosis data of each photovoltaic panel and its multiple nearest neighbor photovoltaic panels, and the consistency of the asynchronous response monitoring matrix between each photovoltaic panel and its multiple nearest neighbor photovoltaic panels; obtaining the monitoring response vector of the photovoltaic power plant to be analyzed; and a fault diagnosis module for obtaining the fault diagnosis results of the photovoltaic power plant to be analyzed. This application aims to improve the accuracy of fault diagnosis in photovoltaic power plants.

Description

Technical Field

[0001] This application relates to the field of photovoltaic system fault identification technology, specifically to a fault diagnosis system for photovoltaic power plants. Background Technology

[0002] A photovoltaic (PV) system is a system that converts solar energy into electrical energy. A PV power station is a PV power generation system that is based on a PV system, connected to the power grid, and transmits electricity to the grid. However, PV arrays, as a key component of PV power stations, are prone to various faults due to their large footprint and wide distribution, such as cracked PV modules, aging wiring, and hot spots. Therefore, timely fault diagnosis is necessary for PV power stations.

[0003] Currently, this goal is often achieved through online monitoring of the operational status of photovoltaic power plant equipment. In a photovoltaic power plant, solar energy conversion occurs on a unit basis, with photovoltaic panels interconnected in parallel or series. This connection method results in data exhibiting correlated changes during actual operation. However, traditional methods do not consider the differences in fault diagnosis data and fault responses during monitoring, leading to low accuracy in diagnosing faults in photovoltaic power plants. Summary of the Invention

[0004] In view of the above, it is necessary to provide a fault diagnosis system for photovoltaic power plants, which improves the accuracy of fault diagnosis for photovoltaic power plants compared with traditional fault diagnosis systems for photovoltaic power plants.

[0005] The fault diagnosis system for photovoltaic power plants proposed in this application adopts the following technical solution: One embodiment of this application provides a fault diagnosis system for photovoltaic power plants, the system comprising: The fault diagnosis data acquisition module is used to acquire multiple fault diagnosis data for each photovoltaic panel in the photovoltaic power station to be analyzed in real time. The fault diagnosis data analysis module is used to determine the response difference coefficient between each photovoltaic panel and its neighboring photovoltaic panels based on the similarity of the changing trends of various fault diagnosis data between each photovoltaic panel and its neighboring photovoltaic panels, and the differences of various fault diagnosis data between each photovoltaic panel and its neighboring photovoltaic panels at the current moment. Based on the response difference coefficient between each photovoltaic panel and its multiple neighboring photovoltaic panels, the asynchronous response monitoring matrix of each photovoltaic panel is obtained; based on the dispersion of various fault diagnosis data of each photovoltaic panel and its multiple neighboring photovoltaic panels, and the consistency of the asynchronous response monitoring matrix between each photovoltaic panel and its multiple neighboring photovoltaic panels, the fault response sensitivity coefficient of various fault diagnosis data of each photovoltaic panel is determined. Based on the distribution of the fault response sensitivity coefficient, the monitoring response vector of the photovoltaic power station to be analyzed is obtained; The fault diagnosis module is used to obtain fault diagnosis results of the photovoltaic power station to be analyzed based on the monitoring response vector using a neural network.

[0006] In one embodiment, the fault diagnosis data includes: total irradiance, direct irradiance, diffuse irradiance, temperature, humidity, and output active power data.

[0007] In one embodiment, the process of obtaining the response difference coefficient is as follows: The various fault diagnosis data of each photovoltaic panel at all sampling times are used as the input of the trend verification algorithm, and the trend statistics of various fault diagnosis data of each photovoltaic panel are output. The trend statistics of all kinds of fault diagnosis data of each photovoltaic panel are combined into a dynamic response vector. The similarity of the dynamic response vectors between each photovoltaic panel and its nearest neighbor photovoltaic panels is denoted as response similarity. All fault diagnosis data of each photovoltaic panel at the current moment are combined into a fault diagnosis data group; the difference between the fault diagnosis data group of each photovoltaic panel and its neighboring photovoltaic panels is recorded as the data difference. The response difference coefficient is positively correlated with the data difference and negatively correlated with the response similarity.

[0008] In one embodiment, the formula for calculating the response difference coefficient is: ;in, This represents the response difference coefficient between the x-th photovoltaic panel and its y-th nearest neighbor photovoltaic panel; and Let x and y represent the fault diagnosis data sets of the x-th photovoltaic panel and its y-th nearest neighbor photovoltaic panel at the current moment, respectively; s() represents the difference measurement function. and Let X and Y represent the dynamic response vectors of the x-th photovoltaic panel and its y-th nearest neighbor photovoltaic panel, respectively; Y() represents the similarity measurement function; exp[] represents the exponential function with the natural constant as the base.

[0009] In one embodiment, the elements in the asynchronous response monitoring matrix are the response difference coefficients between each photovoltaic panel and its multiple neighboring photovoltaic panels. The arrangement of each element is the same as the arrangement of the corresponding neighboring photovoltaic panels in the photovoltaic array, and the central element of the asynchronous response monitoring matrix is ​​0.

[0010] In one embodiment, the process for determining the fault response sensitivity coefficient is as follows: Using all types of fault diagnosis data of any photovoltaic panel as input, the response weight of each type of fault diagnosis data of any photovoltaic panel is obtained by using an objective weighting method. The response weights of various fault diagnosis data of each photovoltaic panel and its neighboring photovoltaic panels are integrated; Calculate the consistency ratio of the asynchronous response monitoring matrix between each photovoltaic panel and its nearest neighbor photovoltaic panels; The fault response sensitivity coefficient is positively correlated with the fusion result and negatively correlated with the consistency ratio.

[0011] In one embodiment, the fault response sensitivity coefficient is determined by the ratio of the fusion result to the consistency ratio.

[0012] In one embodiment, the fault response sensitivity coefficient is calculated as follows: ;in, This indicates the x-th photovoltaic panel. Fault response sensitivity coefficient of various fault diagnosis data; and These represent the x-th photovoltaic panel and its y-th nearest neighbor photovoltaic panel, respectively. The response weights of the fault diagnosis data; τ represents the consistency ratio of the asynchronous response monitoring matrix between the x-th photovoltaic panel and its y-th nearest neighbor photovoltaic panel; n represents the number of nearest neighbor photovoltaic panels of the x-th photovoltaic panel; τ represents a preset constant greater than 0.

[0013] In one embodiment, the method for obtaining the monitoring response vector is as follows: calculate the mean value of the fault response sensitivity coefficient of any fault diagnosis data of all photovoltaic panels, and form the monitoring response vector by normalizing the mean value of all fault diagnosis data.

[0014] In one embodiment, the fault diagnosis result includes the fault type and fault location.

[0015] This application has at least the following beneficial effects: This application analyzes the similarity of the changing trends of various parameters between each photovoltaic panel and its neighboring photovoltaic panels in a photovoltaic power station, as well as the degree of difference in the corresponding parameters between each photovoltaic panel and its neighboring photovoltaic panels at the current moment, to obtain a response difference coefficient, reflecting the trend response difference of parameter changes in different directions within the local area where each photovoltaic panel is located. Furthermore, by combining the asynchronous response differences between photovoltaic panels and the sensitivity of each photovoltaic panel to fault response, a fault response sensitivity coefficient is obtained, taking into account the correlation characteristics of photovoltaic panel fault response caused by changes in the connection method between photovoltaic panels. Finally, by synthesizing the fault response sensitivity coefficients of all photovoltaic panels in the photovoltaic power station, a monitoring response vector of the photovoltaic power station is obtained, which can accurately reflect the sensitivity characteristics of the photovoltaic power station's fault response. Based on the monitoring response vector, a neural network is used to obtain the fault diagnosis results of the photovoltaic power station, improving the accuracy of fault diagnosis of the photovoltaic power station. Attached Figure Description

[0016] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A block diagram of a fault diagnosis system for a photovoltaic power plant provided in this application; Figure 2 A schematic diagram illustrating the process of obtaining the difference coefficient in response; Figure 3 A schematic diagram illustrating the process for determining the fault response sensitivity coefficient; Figure 4 This is a schematic diagram of the process for obtaining fault diagnosis results. Detailed Implementation

[0018] In the description of the embodiments in this application, the words "exemplary," "or," and "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the words "exemplary," "or," and "for example" is intended to present the relevant concepts in a specific manner.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. It should be understood that, unless otherwise stated, " / " in this application means "or".

[0020] It should also be noted that the terms "first" and "second" in this application are used to distinguish similar objects, rather than to describe a specific order or sequence.

[0021] The following description, in conjunction with the accompanying drawings, details a specific solution for a fault diagnosis system for photovoltaic power plants provided in this application.

[0022] Please see Figure 1 The diagram illustrates a fault diagnosis system for a photovoltaic power plant according to an embodiment of the present invention. The system includes: a fault diagnosis data acquisition module 101, a fault diagnosis data analysis module 102, and a fault diagnosis module 103.

[0023] The fault diagnosis data acquisition module 101 is used to acquire multiple fault diagnosis data of each photovoltaic panel in the photovoltaic power station to be analyzed in real time.

[0024] Photovoltaic power plants are generally built in outdoor environments with abundant sunlight. The photovoltaic panels are mainly composed of photovoltaic arrays formed by connecting lines in series and parallel. They are connected to the power grid in parallel through combiner boxes, inverters, and transformers. In addition, photovoltaic power plants are equipped with corresponding data acquisition equipment to collect relevant data on photovoltaic power generation.

[0025] Furthermore, since the photovoltaic array in a photovoltaic power station is composed of connected photovoltaic panels, various fault diagnosis data of each photovoltaic panel in the photovoltaic array are collected in real time according to the acquisition equipment set in the photovoltaic power station to be analyzed. The fault diagnosis data includes total irradiance, direct irradiance, diffuse irradiance, temperature, humidity and output active power data. Among them, total irradiance, direct irradiance and diffuse irradiance are obtained through light sensors, temperature is obtained through temperature sensors, humidity is obtained through humidity sensors, and output active power data is obtained through power sensors.

[0026] In this embodiment, during the collection of fault diagnosis data, the collection time interval is 1 second. The value of the collection time interval is preset by the user and can be set by the implementer. This application does not impose any special restrictions.

[0027] Because environmental interference may occur during the data acquisition process, resulting in noise interference in the acquired fault diagnosis data, noise reduction processing is performed on the acquired fault diagnosis data.

[0028] In this embodiment, a Wiener filter is used to denoise the fault diagnosis data. As another implementation, based on the ability to denoise the fault diagnosis data, the implementer may use other existing technologies to denoise the fault diagnosis data, such as Kalman filters, etc. This application does not impose any special restrictions.

[0029] The fault diagnosis data analysis module 102 is used to determine the response difference coefficient between each photovoltaic panel and its neighboring photovoltaic panels; determine the fault response sensitivity coefficient of various fault diagnosis data of each photovoltaic panel; and then obtain the monitoring response vector of the photovoltaic power station to be analyzed based on the distribution of the fault response sensitivity coefficient.

[0030] Typically, photovoltaic (PV) power plants operate in complex environments. During normal operation, environmental interference can disrupt the collected fault diagnosis data. If the parameters of any PV panel in the PV array change, the parameters of other panels in the array will show corresponding changes. However, due to different connection methods between PV panels, the connection method and the distance between PV panels will affect the response to parameter changes. Moreover, in the PV array of a PV power plant, damage to the surface of PV panels, the formation of local hot spots, and differences in the movement angle caused by loose PV panel supports can all cause a chain reaction in the parameters of PV panels, and different parameters respond differently to faults. For example, in the actual operation of a PV power plant, since PV panels are connected in series and in parallel, if one PV panel has a problem, it will not only affect its own output but also the output of other PV panels, resulting in a combined loss of the entire PV array and a reduction in the overall output power of the PV array.

[0031] Based on the above analysis, if a photovoltaic power station experiences an operational fault, the parameter changes of photovoltaic panels at different locations will exhibit varying responses. Therefore, during fault diagnosis of a photovoltaic power station, the asynchronous differences in the parameters of the photovoltaic panels within the photovoltaic array over time should be analyzed in conjunction with the parameter response status. Based on the analysis results, the comparative characteristics of the response changes of different parameters in the photovoltaic power station to a fault are obtained. The specific calculation and analysis process is as follows: (1) Based on the above analysis, when diagnosing the faults of photovoltaic power plants, in addition to considering the differences in various fault diagnosis data at different sampling times, it is also necessary to consider the differences in the response changes of different types of fault diagnosis data when a fault occurs in the photovoltaic power plant. Therefore, the local response status of the parameters of the photovoltaic panel is analyzed by combining the fault diagnosis data at different sampling times. Specifically, the various fault diagnosis data of each photovoltaic panel at all sampling times are used as the input of the trend test algorithm, and the trend statistics of various fault diagnosis data of each photovoltaic panel are output. The trend statistics of all types of fault diagnosis data of each photovoltaic panel are used to form the dynamic response vector of each photovoltaic panel. The dynamic response vector reflects the response status of different parameters of the photovoltaic panel during the fault monitoring process.

[0032] In this embodiment, the trend test algorithm is the Mann-Kendall trend test method, and the trend statistic is the statistic Z in the Mann-Kendall trend test method. As other implementation methods, based on the ability to measure the changing trend of various fault diagnosis data, the implementer can use other existing technologies for measurement, such as the slope method, Cox-Stuart test, etc. This application does not impose any special restrictions.

[0033] To further analyze the differences in parameter response changes of photovoltaic panels in a local area at the current moment, and thus accurately reflect the correlation characteristics of photovoltaic panel fault response caused by changes in the connection method between photovoltaic panels during the operation of the photovoltaic power station; based on the similarity of the dynamic response vectors between each photovoltaic panel and its nearest neighboring photovoltaic panels, and the differences in various fault diagnosis data between each photovoltaic panel and its nearest neighboring photovoltaic panels at the current moment, the response difference coefficient between each photovoltaic panel and its nearest neighboring photovoltaic panels is determined, and the calculation formula is as follows: ;in, This represents the response difference coefficient between the x-th photovoltaic panel and its y-th nearest neighbor photovoltaic panel; all fault diagnosis data for each photovoltaic panel at the current moment are combined to form a fault diagnosis data set for each photovoltaic panel at the current moment. and Let x and y represent the fault diagnosis data sets of the x-th photovoltaic panel and its y-th nearest neighbor photovoltaic panel at the current moment, respectively; s() represents the difference measurement function. and Let X and Y represent the dynamic response vectors of the x-th photovoltaic panel and its y-th nearest neighbor photovoltaic panel, respectively; Y() represents the similarity measurement function; exp[] represents the exponential function with the natural constant as the base, in order to avoid the denominator being 0.

[0034] In this embodiment, each photovoltaic panel in the photovoltaic array is taken as the center, and the 8-neighbor areas of each photovoltaic panel are considered as the 8 nearest neighbor photovoltaic panels. These 8 neighbor areas are the photovoltaic panels adjacent to the central photovoltaic panel in the directions of 0 degrees, 45 degrees, 90 degrees, 135 degrees, 180 degrees, 225 degrees, 270 degrees, and 315 degrees, with the horizontal rightward direction in the photovoltaic array considered as the 0-degree direction. The number of nearest neighbor photovoltaic panels for each photovoltaic panel can be set by the implementer, but they must be adjacent to each photovoltaic panel in different directions.

[0035] In this embodiment, the difference measurement function is the SAD (Sum of Absolute Difference) value. As other implementation methods, based on the ability to measure the difference between elements at the same location in the fault diagnosis data set, the implementer may use other existing technologies for measurement, such as Euclidean distance, mean square error, etc. This application does not impose any special restrictions.

[0036] In this embodiment, the similarity measurement function can be cosine similarity. As other implementation methods, based on the ability to measure the similarity between dynamic response vectors, the implementer can use other calculation methods for measurement, such as the reciprocal of Hamming distance, the reciprocal of Euclidean distance, etc. This application does not impose any special restrictions.

[0037] It should be noted that: the larger the calculated response difference coefficient, the stronger the difference between the x-th photovoltaic panel and its i-th panel. The greater the difference in the correlation response between neighboring photovoltaic panels, the more likely the photovoltaic panels are to experience localized correlated failures. A schematic diagram illustrating the process of obtaining the response difference coefficient is shown below. Figure 2 As shown.

[0038] (2) Based on the above analysis, due to the differences in the connection methods between photovoltaic panels and the orientation of photovoltaic panels, the correlation response in a local area may exhibit asynchronous response characteristics in different directions. Therefore, the trend response differences of parameter changes in different directions within the local area of ​​each photovoltaic panel are analyzed to obtain the trend response coefficients between each photovoltaic panel and other photovoltaic panels in the local area. In this embodiment, the trend response differences of parameter changes in the 8-neighborhood directions within the local area of ​​each photovoltaic panel are analyzed.

[0039] Furthermore, if the trend response differences in different local areas of the photovoltaic array change, it indicates that during the operation of the photovoltaic power station, due to faults such as damage to the surface of the photovoltaic panels or differences in the angle of movement, the local areas exhibit asynchronous trend response characteristics. Therefore, for each photovoltaic panel in the photovoltaic array, the response difference coefficient between the photovoltaic panel and its nearest neighbor photovoltaic panels is used as each element in the asynchronous response monitoring matrix. The arrangement position of each element in the asynchronous response monitoring matrix is ​​the same as the arrangement position of the nearest neighbor photovoltaic panels corresponding to each element in the photovoltaic array.

[0040] In this embodiment, the asynchronous response monitoring matrix is ​​a 3×3 matrix. If the positions of the x-th photovoltaic panel and its 8 nearest neighbor photovoltaic panels are arranged as follows: Then, the asynchronous response monitoring matrix formed by the response difference coefficients between the x-th photovoltaic panel and its 8 nearest neighbor photovoltaic panels is: ,in, This represents the x-th photovoltaic panel; These represent the 1st, 2nd, 3rd, 4th, 5th, 6th, 7th, and 8th nearest neighbor photovoltaic panels of the xth photovoltaic panel, respectively. These represent the response difference coefficients between the x-th photovoltaic panel and its 1st, 2nd, 3rd, 4th, 5th, 6th, 7th, and 8th nearest neighbor photovoltaic panels, respectively.

[0041] Furthermore, if a photovoltaic power station malfunctions, the difference in parameter response between different photovoltaic panels in a local area will lead to asynchronous characteristics of local trend response, affecting the operational stability of the photovoltaic power station. Therefore, for each photovoltaic panel in the photovoltaic power station to be analyzed, the consistency ratio of the asynchronous response monitoring matrix between the photovoltaic panel and its nearest neighbor photovoltaic panels is calculated. The larger the consistency ratio, the smaller the asynchronous difference of trend response in the local area.

[0042] Using all types of fault diagnosis data collected from any photovoltaic panel as input, an objective weighting method is used to obtain the response weight of each type of fault diagnosis data for any photovoltaic panel. The larger the response weight of any type of fault diagnosis data, the faster the response of that type of fault diagnosis data to state changes during the monitoring of the photovoltaic panel.

[0043] In this embodiment, the entropy weight method is used to obtain the response weight of each fault diagnosis data. As another implementation method, based on the ability to obtain the response weight of each fault diagnosis data, the implementer may use other existing technologies to obtain the response weight of each fault diagnosis data, such as the standard deviation method, the CRITIC method, etc. This application does not impose any special restrictions.

[0044] Furthermore, based on the response weights of various fault diagnosis data of each photovoltaic panel and its multiple neighboring photovoltaic panels, and the consistency of the asynchronous response monitoring matrix between each photovoltaic panel and its multiple neighboring photovoltaic panels, the fault response sensitivity coefficients of various fault diagnosis data of each photovoltaic panel are determined, and the calculation formula is as follows: ;in, This indicates the x-th photovoltaic panel. Fault response sensitivity coefficient of various fault diagnosis data; and These represent the x-th photovoltaic panel and its y-th nearest neighbor photovoltaic panel, respectively. The response weights of the fault diagnosis data; The consistency ratio of the asynchronous response monitoring matrix between the x-th photovoltaic panel and its y-th nearest neighbor photovoltaic panel is represented; n represents the number of nearest neighbor photovoltaic panels of the x-th photovoltaic panel; τ represents a preset constant greater than 0, the value of τ is preset by the user and can be set by the implementer. In this embodiment, the value of τ is 0.01.

[0045] It should be noted that a larger fault response sensitivity coefficient indicates a faster response of the i-th type of fault diagnosis data within a local range during the fault monitoring process of the photovoltaic power station under analysis. A schematic diagram illustrating the process for determining the fault response sensitivity coefficient is shown below. Figure 3 As shown.

[0046] (3) Based on the above analysis, in the fault monitoring process of the photovoltaic power station to be analyzed, the asynchronous response differences between photovoltaic panels and the sensitivity of each photovoltaic panel to fault response were combined when different types of faults occurred in the local area of ​​the photovoltaic array, and the fault response sensitivity coefficient of each photovoltaic panel was obtained for each type of fault diagnosis data of each photovoltaic panel.

[0047] To further determine the sensitivity characteristics of each type of fault diagnosis data to the fault response during the fault monitoring process of the photovoltaic power station to be analyzed, the mean value of the fault response sensitivity coefficient of each type of fault diagnosis data of all photovoltaic panels is calculated, and the normalized value of the mean value of all types of fault diagnosis data is used to form the monitoring response vector of the photovoltaic power station to be analyzed. The monitoring response vector integrates the asynchronous response differences between photovoltaic panels in a local area of ​​the photovoltaic power station to be analyzed and the response characteristics of the fault diagnosis data of each photovoltaic panel, accurately reflecting the sensitivity characteristics of the fault response of the photovoltaic power station to be analyzed.

[0048] In this embodiment, the Min-Max normalization method is used to normalize the mean. As another implementation, based on the ability to normalize the mean, the implementer may use other existing technologies to normalize the mean, such as the Z-Score normalization method, the decimal scaling normalization method, etc. This application does not impose any special restrictions.

[0049] The fault diagnosis module 103 is used to obtain the fault diagnosis results of the photovoltaic power station to be analyzed based on the monitoring response vector using a neural network.

[0050] Using the same acquisition method as for the monitoring response vectors of the photovoltaic power station to be analyzed, a preset number of monitoring response vectors from photovoltaic power stations are obtained. These preset number of monitoring response vectors are then divided into a training set and a test set in a 7:3 ratio. Labels are added to the monitoring response vectors of the preset number of photovoltaic power stations, with the labels indicating fault type and fault location. A gated recurrent neural network is trained based on the divided training set, and the gated recurrent neural network is validated using the test set. The specific training process of the gated recurrent neural network is a well-known technique and will not be described in detail in this application. Implementers can choose other existing feasible neural networks, and this embodiment does not impose any restrictions on this.

[0051] In this embodiment, the preset quantity is 1000. The preset quantity is preset by the user and can be set by the implementer. This application does not impose any special restrictions.

[0052] Based on the monitoring response vector of the photovoltaic power station to be analyzed, a pre-trained gated recurrent neural network is used to obtain the fault diagnosis results of the photovoltaic power station to be analyzed. The fault diagnosis results include the fault type and fault location. The optimization algorithm of the gated recurrent neural network is stochastic gradient descent, and the loss function is cross-entropy loss. A schematic diagram of the fault diagnosis result acquisition process is shown below. Figure 4 As shown.

[0053] In summary, this application analyzes the similarity of the changing trends of various parameters between each photovoltaic panel and its neighboring photovoltaic panels in a photovoltaic power station, as well as the degree of difference in the corresponding parameters between each photovoltaic panel and its neighboring photovoltaic panels at the current moment, to obtain a response difference coefficient. This coefficient reflects the trend response difference of parameter changes in different directions within the local area where each photovoltaic panel is located. Furthermore, by combining the asynchronous response differences between photovoltaic panels and the sensitivity of each photovoltaic panel to fault response, a fault response sensitivity coefficient is obtained, taking into account the correlation characteristics of photovoltaic panel fault response caused by changes in the connection method between photovoltaic panels. Finally, by synthesizing the fault response sensitivity coefficients of all photovoltaic panels in the photovoltaic power station, a monitoring response vector of the photovoltaic power station is obtained, which can accurately reflect the sensitivity characteristics of the photovoltaic power station's fault response. Based on the monitoring response vector, a neural network is used to obtain the fault diagnosis results of the photovoltaic power station, thereby improving the accuracy of fault diagnosis of the photovoltaic power station.

[0054] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than that shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. In the descriptions corresponding to the flowcharts and block diagrams in the accompanying drawings, the operations or steps corresponding to different blocks may also occur in a different order than disclosed in the description, and sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. Each block in a block diagram and / or flowchart, and combinations of blocks in a block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0055] It will be apparent to those skilled in the art that this application is not limited to the details of the exemplary embodiments described above, and that this application can be implemented in other specific forms without departing from its essential characteristics. Therefore, the embodiments described above should be considered exemplary and non-limiting in all respects.

Claims

1. A fault diagnosis system for photovoltaic power plants, characterized in that, The system includes: The fault diagnosis data acquisition module is used to acquire multiple fault diagnosis data for each photovoltaic panel in the photovoltaic power station to be analyzed in real time. The fault diagnosis data analysis module is used to determine the response difference coefficient between each photovoltaic panel and its neighboring photovoltaic panels based on the similarity of the changing trends of various fault diagnosis data between each photovoltaic panel and its neighboring photovoltaic panels, and the differences of various fault diagnosis data between each photovoltaic panel and its neighboring photovoltaic panels at the current moment. Based on the response difference coefficient between each photovoltaic panel and its multiple neighboring photovoltaic panels, the asynchronous response monitoring matrix of each photovoltaic panel is obtained; based on the dispersion of various fault diagnosis data of each photovoltaic panel and its multiple neighboring photovoltaic panels, and the consistency of the asynchronous response monitoring matrix between each photovoltaic panel and its multiple neighboring photovoltaic panels, the fault response sensitivity coefficient of various fault diagnosis data of each photovoltaic panel is determined. Based on the distribution of the fault response sensitivity coefficient, the monitoring response vector of the photovoltaic power station to be analyzed is obtained; The fault diagnosis module is used to obtain fault diagnosis results of the photovoltaic power station to be analyzed based on the monitoring response vector using a neural network.

2. The fault diagnosis system for photovoltaic power plants as described in claim 1, characterized in that, The fault diagnosis data includes: total irradiance, direct irradiance, diffuse irradiance, temperature, humidity, and output active power data.

3. The fault diagnosis system for photovoltaic power plants as described in claim 1, characterized in that, The process for obtaining the response difference coefficient is as follows: The various fault diagnosis data of each photovoltaic panel at all sampling times are used as the input of the trend verification algorithm, and the trend statistics of various fault diagnosis data of each photovoltaic panel are output. The trend statistics of all kinds of fault diagnosis data of each photovoltaic panel are combined into a dynamic response vector. The similarity of the dynamic response vectors between each photovoltaic panel and its nearest neighbor photovoltaic panels is denoted as response similarity. All fault diagnosis data of each photovoltaic panel at the current moment are combined into a fault diagnosis data group; the difference between the fault diagnosis data group of each photovoltaic panel and its neighboring photovoltaic panels is recorded as the data difference. The response difference coefficient is positively correlated with the data difference and negatively correlated with the response similarity.

4. The fault diagnosis system for photovoltaic power plants as described in claim 3, characterized in that, The formula for calculating the response difference coefficient is as follows: ;in, This represents the response difference coefficient between the x-th photovoltaic panel and its y-th nearest neighbor photovoltaic panel; and Let x and y represent the fault diagnosis data sets of the x-th photovoltaic panel and its y-th nearest neighbor photovoltaic panel at the current moment, respectively; s() represents the difference measurement function. and Let X and Y represent the dynamic response vectors of the x-th photovoltaic panel and its y-th nearest neighbor photovoltaic panel, respectively; Y() represents the similarity measurement function; exp[] represents the exponential function with the natural constant as the base.

5. A fault diagnosis system for photovoltaic power plants as described in claim 1, characterized in that, The elements in the asynchronous response monitoring matrix are the response difference coefficients between each photovoltaic panel and its multiple neighboring photovoltaic panels. The arrangement of each element is the same as the arrangement of the corresponding neighboring photovoltaic panels in the photovoltaic array. The central element of the asynchronous response monitoring matrix is ​​0.

6. A fault diagnosis system for photovoltaic power plants as described in claim 1, characterized in that, The process for determining the fault response sensitivity coefficient is as follows: Using all types of fault diagnosis data of any photovoltaic panel as input, the response weight of each type of fault diagnosis data of any photovoltaic panel is obtained by using an objective weighting method. The response weights of various fault diagnosis data of each photovoltaic panel and its neighboring photovoltaic panels are integrated; Calculate the consistency ratio of the asynchronous response monitoring matrix between each photovoltaic panel and its nearest neighbor photovoltaic panels; The fault response sensitivity coefficient is positively correlated with the fusion result and negatively correlated with the consistency ratio.

7. A fault diagnosis system for photovoltaic power plants as described in claim 6, characterized in that, The fault response sensitivity coefficient is determined by the ratio of the fusion result to the consistency ratio.

8. A fault diagnosis system for photovoltaic power plants as described in claim 7, characterized in that, The calculation relationship for the fault response sensitivity coefficient is as follows: ;in, This indicates the x-th photovoltaic panel. Fault response sensitivity coefficient of various fault diagnosis data; and These represent the x-th photovoltaic panel and its y-th nearest neighbor photovoltaic panel, respectively. The response weights of the fault diagnosis data; τ represents the consistency ratio of the asynchronous response monitoring matrix between the x-th photovoltaic panel and its y-th nearest neighbor photovoltaic panel; n represents the number of nearest neighbor photovoltaic panels of the x-th photovoltaic panel; τ represents a preset constant greater than 0.

9. A fault diagnosis system for photovoltaic power plants as described in claim 1, characterized in that, The method for obtaining the monitoring response vector is as follows: calculate the mean value of the fault response sensitivity coefficient of any fault diagnosis data of all photovoltaic panels, and form the monitoring response vector by normalizing the mean value of all fault diagnosis data.

10. A fault diagnosis system for a photovoltaic power station as described in claim 1, characterized in that, The fault diagnosis results include the fault type and fault location.

Images