Intelligent fault operation and maintenance system and method for photovoltaic new energy power generation system

Abstract

The invention relates to the technical field of photovoltaic operation and maintenance, in particular to an intelligent fault operation and maintenance system and method for a photovoltaic new energy power generation system, and the method comprises the steps: enabling an inspection cleaning robot to move on a photovoltaic panel, and collecting the first image information and second image information of the photovoltaic panel; the operation and maintenance central control module analyzes the state of the photovoltaic panel according to the first image information, generates first state information and generates a photovoltaic panel cleaning instruction according to the first state information; the inspection cleaning robot cleans the photovoltaic panel according to the photovoltaic panel cleaning instruction; the operation and maintenance central control module analyzes the state of the photovoltaic panel according to the second image information to generate second state information; according to the first state information and the second state information of the same photovoltaic panel, whether the photovoltaic panel is damaged is analyzed, and if yes, damage information is generated; according to the scheme, the photovoltaic panel can be cleaned, whether the photovoltaic panel is damaged or not is detected, the accuracy of photovoltaic panel damage judgment is improved, extra photovoltaic panel inspection is not needed, the time cost is saved, and the inspection efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of photovoltaic operation and maintenance technology, and specifically to an intelligent fault operation and maintenance system and method for a photovoltaic new energy power generation system. Background Art

[0002] With the adjustment of the global energy structure, photovoltaic new energy power generation systems have been widely used due to their environmental protection and sustainability.

[0003] Photovoltaic new energy power generation systems collect solar energy and convert it into electricity through photovoltaic panels. Photovoltaic panels are widely distributed and the environment is complex and changeable. The traditional manual inspection method is costly and has low inspection accuracy. Moreover, photovoltaic panels will become dirty as they are used, which affects the collection of solar energy. Large-area photovoltaic panels are spliced ​​together, and it is difficult for inspection personnel to detect the dirtiness of the internal photovoltaic panels. Therefore, there is an urgent need for an intelligent fault operation and maintenance system and method for photovoltaic new energy power generation systems to solve the above problems. Summary of the invention

[0004] One of the purposes of the present invention is to provide an intelligent fault operation and maintenance system for a photovoltaic new energy power generation system, which can clean the photovoltaic panels and detect whether the photovoltaic panels are damaged at the same time, thereby improving the accuracy of judging whether the photovoltaic panels are damaged. There is no need for additional inspections of the photovoltaic panels, saving time and cost, and improving inspection efficiency.

[0005] The basic solution 1 provided by the present invention is: an intelligent fault operation and maintenance system for a photovoltaic new energy power generation system, comprising: a patrol cleaning robot and an operation and maintenance central control module;

[0006] The inspection and cleaning robot is used to move on the photovoltaic panel and collect image information of the photovoltaic panel; wherein the image information includes first image information and second image information, the first image information includes: the image of the next photovoltaic panel that the inspection and cleaning robot will drive on; the second image information includes: the image of the previous photovoltaic panel that the inspection and cleaning robot has driven on;

[0007] The operation and maintenance central control module is used to analyze the status of the photovoltaic panel according to the first image information and generate first status information;

[0008] It is also used to determine whether the photovoltaic panel needs to be cleaned according to the first status information, and if so, generate a photovoltaic panel cleaning instruction and send it to the inspection and cleaning robot;

[0009] The inspection and cleaning robot is also used to clean the photovoltaic panels according to the photovoltaic panel cleaning instructions;

[0010] The operation and maintenance central control module is further used to analyze the state of the photovoltaic panel according to the second image information and generate second state information;

[0011] It is also used to analyze whether the photovoltaic panel is damaged based on the first state information and the second state information of the same photovoltaic panel, and if so, generate damage information;

[0012] The operation and maintenance central control module is also used to generate operation and maintenance plans based on damage information.

[0013] Beneficial effects of the basic solution: In this solution, a patrol cleaning robot is used to move on the photovoltaic panels to perform daily cleaning of the photovoltaic panels and collect image information of the photovoltaic panels;

[0014] The collected image information includes the image of the next photovoltaic panel that the patrol cleaning robot will drive on, i.e., the first image information, and the image of the previous photovoltaic panel that the patrol cleaning robot has driven on, i.e., the second image information; the two image information can be used as the image information of a photovoltaic panel before cleaning and the image information of a photovoltaic panel after cleaning, respectively; therefore, the operation and maintenance central control module will first analyze the state of the photovoltaic panel according to the first image information and generate the first state information; and determine whether the photovoltaic panel needs to be cleaned according to the first state information, and if so, generate a photovoltaic panel cleaning instruction and send it to the patrol cleaning robot, and the patrol cleaning robot cleans the photovoltaic panel according to the photovoltaic panel cleaning instruction, thereby completing the daily cleaning of the photovoltaic panel;

[0015] The inspection cleaning robot is a mobile cleaning device. After cleaning one photovoltaic panel, it will move to the next photovoltaic panel. At this time, it can collect image information of the photovoltaic panel after cleaning, that is, the second image information;

[0016] The operation and maintenance central control module will analyze the status of the photovoltaic panel based on the second image information and generate second status information; and analyze whether the photovoltaic panel is damaged based on the first status information and the second status information of the same photovoltaic panel. If so, damage information is generated, and an operation and maintenance plan is generated accordingly for operation and maintenance, so as to find out whether the photovoltaic panel is damaged during the cleaning process, and comprehensively analyze the image information before and after cleaning, which can improve the accuracy of judging the damage of the photovoltaic panel, eliminate the need for additional inspections of the photovoltaic panel, save time and cost, and improve inspection efficiency.

[0017] Further, the inspection cleaning robot includes: a housing and a controller;

[0018] A universal roller is arranged at the bottom of the housing, and a motion motor is arranged inside. The output shaft of the motion motor is connected to the universal roller to drive the universal roller to rotate;

[0019] A cleaning motor is also provided at the bottom of the shell, and a cleaning brush is connected to the output shaft of the cleaning motor. A first camera device is provided on the front side of the shell, and a second camera device is provided on the rear side.

[0020] A first camera device is used to collect first image information of the photovoltaic panel at the front end of the housing and send it to the operation and maintenance central control module;

[0021] The second camera device is used to collect second image information of the photovoltaic panel at the rear end of the shell and send it to the operation and maintenance central control module.

[0022] The camera devices at the front and rear of the shell respectively collect the first image information and the second image information, which is convenient for collecting the images before and after cleaning.

[0023] Further, analyzing the state of the photovoltaic panel according to the first image information to generate the first state information includes:

[0024] Preprocessing the first image information;

[0025] The preprocessed first image information is input into the constructed neural network model, feature extraction is performed on the preprocessed first image information to obtain first image features, and the state of the photovoltaic panel is identified based on the extracted first image features to output first state information.

[0026] Further, the first status information includes: photovoltaic panel dirtiness status information and photovoltaic panel damage status information;

[0027] The constructed neural network model adopts a CNN neural network model; the preprocessed first image information is input into the CNN neural network model, the CNN neural network model extracts the first image features, and evaluates the degree of contamination of the photovoltaic panel based on the extracted first image features, generates a photovoltaic panel contamination evaluation value as the photovoltaic panel contamination status information, and at the same time analyzes whether the photovoltaic panel is damaged based on the extracted first image features, and if so, generates damage type information and damage area information as the photovoltaic panel damage status information.

[0028] The CNN neural network model can effectively extract image features, perform corresponding analysis, and obtain status information.

[0029] Further, generating a photovoltaic panel cleaning instruction according to the first state information and sending it to a controller of the patrol cleaning robot includes:

[0030] The preset dirtiness assessment range described in the photovoltaic panel dirtiness assessment value in the first state information is analyzed, and a photovoltaic panel cleaning instruction of a corresponding level is generated according to the preset range; and the higher the photovoltaic panel cleaning instruction level, the faster the cleaning motor speed.

[0031] Different levels of photovoltaic panel cleaning instructions enable the patrol cleaning robot to perform cleaning with different intensities to adapt to different cleaning needs.

[0032] Further, the range of the photovoltaic panel dirt evaluation value is [0, f], and the preset dirt evaluation ranges are: [0, a), [a, b), [b, c), [c, d), [d, f], and a < b < c < d < f, all of which are positive numbers. Among them, [a, b), [b, c), [c, d), [d, f] correspond to the photovoltaic panel cleaning instructions, including: the first-level photovoltaic panel cleaning instruction, the second-level photovoltaic panel cleaning instruction, the third-level photovoltaic panel cleaning instruction, and the fourth-level photovoltaic panel cleaning instruction;

[0033] After the photovoltaic panel dirt evaluation value is greater than or equal to a, the corresponding photovoltaic panel cleaning instruction is generated according to the obtained range of the photovoltaic panel dirt evaluation value and sent to the inspection and cleaning robot.

[0034] Further, analyzing the state of the photovoltaic panel according to the second image information to generate the second state information includes:

[0035] Preprocessing the second image information;

[0036] Input the preprocessed second image information into the constructed neural network model to extract features of the preprocessed second image information, obtain the second image features, and identify the state of the photovoltaic panel according to the extracted second image features, and output the second state information; where the second state information includes: the photovoltaic panel dirt state information and the photovoltaic panel damage state information;

[0037] Among them, the constructed neural network model uses a CNN neural network model; input the preprocessed second image information into the CNN neural network model, the CNN neural network model extracts the second image features, and evaluates the dirt degree of the photovoltaic panel according to the extracted second image features to generate the photovoltaic panel dirt evaluation value as the photovoltaic panel dirt state information. At the same time, according to the extracted second image features, analyze whether there is damage to the photovoltaic panel. If so, generate the damage type information and the damage area information as the photovoltaic panel damage state information.

[0038] Further, analyzing whether there is damage to the photovoltaic panel according to the first state information and the second state information of the same photovoltaic panel. If so, generating the damage information includes:

[0039] S101. Judge whether the photovoltaic panel dirt evaluation value in the second state information of the photovoltaic panel is less than the preset dirt evaluation threshold a. If so, execute S102; if not, execute S105;

[0040] S102. Judge whether there is photovoltaic panel damage state information in the second state information. If so, execute S103; if not, determine that the fault cause of the photovoltaic panel is dirt and execute S1011;

[0041] S103, determining whether the photovoltaic panel damage state information in the second state information of the photovoltaic panel is the same as the photovoltaic panel damage state information in the first state information, if so, executing S104; if not, executing S1010;

[0042] S104, determining that the photovoltaic panel is damaged, combining the photovoltaic panel damage status information in the first status information or the photovoltaic panel damage status information in the second status information and the location information of the photovoltaic panel as damage information, and executing S1011;

[0043] S105, determining whether the photovoltaic panel contamination assessment value in the second state information of the photovoltaic panel is less than the photovoltaic panel contamination assessment value in the first state information, if so, executing S106; if not, executing S109;

[0044] S106, judging whether the photovoltaic panel is cleaned, and the corresponding photovoltaic panel cleaning instruction level, if the photovoltaic panel is cleaned, and the photovoltaic panel cleaning instruction level is less than level four, then generating a level four photovoltaic panel cleaning instruction for cleaning the photovoltaic panel, and sending it to the controller, the controller cleans the photovoltaic panel again; if the photovoltaic panel is cleaned, and the photovoltaic panel cleaning instruction level is level four, executing S107; if the photovoltaic panel is not cleaned, executing S108;

[0045] S107, determining whether there is photovoltaic panel damage status information in the second status information, if yes, executing S103, if no, generating cleaning function failure information of the patrol cleaning robot, and providing a cleaning function failure prompt of the patrol cleaning robot, and executing S10;

[0046] S108, determining whether the photovoltaic panel dirtiness assessment value in the first state information is greater than a preset dirtiness assessment threshold a, and if so, generating cleaning function fault information of the patrol cleaning robot, and providing a cleaning function fault prompt for the patrol cleaning robot; if not, generating image acquisition function fault information of the patrol cleaning robot, and providing an image acquisition function fault prompt for the patrol cleaning robot;

[0047] S109, generating cleaning function failure information of the patrol cleaning robot, and providing a cleaning function failure prompt of the patrol cleaning robot;

[0048] S1010, determining that the photovoltaic panel is damaged, taking the photovoltaic panel damage status information in the first status information and the photovoltaic panel damage status information in the second status information, and the location information of the photovoltaic panel as damage information, and executing S1011;

[0049] S1011, end.

[0050] Based on the first status information and the second status information of the same photovoltaic panel, the operation and maintenance central control module not only analyzes whether the photovoltaic panel is damaged, but also analyzes whether the inspection and cleaning robot is faulty. It can determine whether the inspection and cleaning robot is faulty and determine the function of the fault, and make a comprehensive and all-round judgment on the damage of the photovoltaic panel, effectively improving the accuracy of the judgment on the damage of the photovoltaic panel.

[0051] Furthermore, generating an operation and maintenance plan based on the damage information includes: determining location information of the damaged photovoltaic panel based on the damage information, and generating an optimal operation and maintenance plan using an optimization algorithm based on the location information and location importance.

[0052] The second purpose of the present invention is to provide an intelligent fault operation and maintenance method for a photovoltaic new energy power generation system, which can clean the photovoltaic panels and detect whether the photovoltaic panels are damaged at the same time, thereby improving the accuracy of judging whether the photovoltaic panels are damaged. There is no need for additional inspections of the photovoltaic panels, saving time and cost, and improving inspection efficiency.

[0053] The present invention provides basic solution 2: an intelligent fault operation and maintenance method for a photovoltaic new energy power generation system, which adopts the above-mentioned intelligent fault operation and maintenance system for a photovoltaic new energy power generation system. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 The present invention is a logic block diagram of an embodiment of an intelligent fault operation and maintenance system for a photovoltaic new energy power generation system. DETAILED DESCRIPTION

[0055] The following is further described in detail through specific implementation methods:

[0056] The embodiment is basically as shown in the attached Figure 1 As shown: Intelligent fault operation and maintenance system for photovoltaic new energy power generation system, including: inspection and cleaning robot and operation and maintenance central control module;

[0057] The inspection and cleaning robot is used to move on the photovoltaic panel and collect image information of the photovoltaic panel; wherein the image information includes first image information and second image information, the first image information includes: the image of the next photovoltaic panel that the inspection and cleaning robot will drive on; the second image information includes: the image of the previous photovoltaic panel that the inspection and cleaning robot has driven on;

[0058] Specifically, the inspection cleaning robot includes: a housing and a controller;

[0059] A universal roller is arranged at the bottom of the shell, and a motion motor is arranged inside. The output shaft of the motion motor is connected to the universal roller, driving the universal roller to rotate, so that the inspection and cleaning robot can move on the photovoltaic panel;

[0060] A cleaning motor is also provided at the bottom of the shell, and a cleaning brush is connected to the output shaft of the cleaning motor. A first camera device is provided on the front side of the shell, and a second camera device is provided on the rear side.

[0061] A first camera device is used to collect first image information of the photovoltaic panel at the front end of the housing and send it to the operation and maintenance central control module;

[0062] A second camera device is used to collect second image information of the photovoltaic panel at the rear end of the housing and send it to the operation and maintenance central control module;

[0063] The operation and maintenance central control module is used to analyze the status of the photovoltaic panel according to the first image information and generate first status information;

[0064] Specifically, the first image information is preprocessed; the preprocessing includes, but is not limited to: image correction, image enhancement and denoising; image correction removes distortion in the image to make the image more realistic; image enhancement improves the contrast of the image to make the edge more obvious; noise removal uses filtering technology (such as median filtering, Gaussian filtering) to remove noise in the image.

[0065] Inputting the preprocessed first image information into the constructed neural network model, performing feature extraction on the preprocessed first image information to obtain first image features, and identifying the state of the photovoltaic panel according to the extracted first image features, and outputting first state information;

[0066] The first status information includes: photovoltaic panel dirtiness status information and photovoltaic panel damage status information;

[0067] The neural network model constructed therein adopts a CNN neural network model; the preprocessed first image information is input into the CNN neural network model, the CNN neural network model extracts the first image features, and evaluates the degree of contamination of the photovoltaic panel based on the extracted first image features, generates a photovoltaic panel contamination evaluation value as the photovoltaic panel contamination status information, and analyzes whether the photovoltaic panel is damaged based on the extracted first image features, and if so, generates damage type information and damage area information as the photovoltaic panel damage status information;

[0068] The operation and maintenance central control module is further used to generate a photovoltaic panel cleaning instruction according to the first status information and send it to the controller of the inspection and cleaning robot;

[0069] Specifically, analyze the preset soiling evaluation range of the PV panel soiling evaluation value in the first status information, and generate a PV panel cleaning instruction corresponding to the level according to the preset range; in this implementation, the range of the PV panel soiling evaluation value is [0, f], and the preset soiling evaluation ranges are: [0, a), [a, b), [b, c), [c, d), [d, f], and a < b < c < d < f, all of which are positive numbers. Among them, [a, b), [b, c), [c, d), [d, f] correspond to the PV panel cleaning instructions, including: the first-level PV panel cleaning instruction, the second-level PV panel cleaning instruction, the third-level PV panel cleaning instruction, and the fourth-level PV panel cleaning instruction; the higher the level of the PV panel cleaning instruction, the faster the cleaning motor speed and the stronger the cleaning force, so as to perform precise energy-saving cleaning; that is, when the PV panel soiling evaluation value is greater than or equal to a, it is determined that the PV panel is soiled and needs to be cleaned, and the corresponding PV panel cleaning instruction is generated according to the obtained range of the PV panel soiling evaluation value and sent to the controller.

[0070] The controller is used to control the opening and closing of the movement motor and the cleaning motor according to the PV panel cleaning instruction, as well as the speed of the cleaning motor, drive the cleaning brush to rotate, and clean the surface of the PV panel.

[0071] The operation and maintenance central control module is also used to analyze the status of the PV panel according to the second image information and generate the second status information.

[0072] Specifically, preprocess the second image information; input the preprocessed second image information into the constructed neural network model, extract the features of the preprocessed second image information, obtain the second image features, and identify the status of the PV panel according to the extracted second image features, and output the second status information; among them, the second status information includes: PV panel soiling status information and PV panel damage status information.

[0073] Among them, the constructed neural network model uses a CNN neural network model; input the preprocessed second image information into the CNN neural network model, the CNN neural network model extracts the second image features, and evaluates the soiling degree of the PV panel according to the extracted second image features, generates a PV panel soiling evaluation value as the PV panel soiling status information, and at the same time analyzes whether the PV panel is damaged according to the extracted second image features. If so, generate the damage type information and damage area information as the PV panel damage status information.

[0074] The operation and maintenance central control module is also used to analyze whether the PV panel is damaged according to the first status information and the second status information of the same PV panel. If so, generate damage information.

[0075] Specifically,

[0076] S101, determining whether the photovoltaic panel contamination assessment value in the second state information of the photovoltaic panel is less than a preset contamination assessment threshold a, if so, executing S102; if not, executing S105;

[0077] S102, determining whether there is photovoltaic panel damage status information in the second status information, if yes, executing S103; if no, determining that the photovoltaic panel failure cause is dirtiness, and executing S1011;

[0078] S103, judging whether the damage state information of the photovoltaic panel in the second state information of the photovoltaic panel is the same as the damage state information of the photovoltaic panel in the first state information, if so, executing S104; if not, executing S1010; wherein whether the damage state information of the photovoltaic panel in the second state information of the photovoltaic panel is the same as the damage state information of the photovoltaic panel in the first state information is judged according to whether the damage type information and the damage area information in the damage state information of the photovoltaic panel are the same, wherein if the type information is the same and the overlap degree of the damage area information meets the preset range, then they are determined to be the same, otherwise, they are determined to be different;

[0079] S104, determining that the photovoltaic panel is damaged, combining the photovoltaic panel damage status information in the first status information or the photovoltaic panel damage status information in the second status information, and the location information of the photovoltaic panel as damage information, and executing S1011; if the two information are the same, either one of the information can be used as the damage information, thereby reducing the amount of data transmission and saving system resources;

[0080] S105, determining whether the photovoltaic panel contamination assessment value in the second state information of the photovoltaic panel is less than the photovoltaic panel contamination assessment value in the first state information, if so, executing S106; if not, executing S109;

[0081] S106, judging whether the photovoltaic panel is cleaned, and the corresponding photovoltaic panel cleaning instruction level, if the photovoltaic panel is cleaned, and the photovoltaic panel cleaning instruction level is less than level four, then generating a level four photovoltaic panel cleaning instruction for cleaning the photovoltaic panel, and sending it to the controller, the controller cleans the photovoltaic panel again; if the photovoltaic panel is cleaned, and the photovoltaic panel cleaning instruction level is level four, executing S107; if the photovoltaic panel is not cleaned, executing S108;

[0082] S107, determining whether there is photovoltaic panel damage status information in the second status information, if yes, executing S103, if no, generating cleaning function failure information of the patrol cleaning robot, and providing a cleaning function failure prompt of the patrol cleaning robot, and executing S10;

[0083] S108, determining whether the photovoltaic panel dirtiness assessment value in the first state information is greater than a preset dirtiness assessment threshold a, and if so, generating cleaning function fault information of the patrol cleaning robot, and providing a cleaning function fault prompt for the patrol cleaning robot; if not, generating image acquisition function fault information of the patrol cleaning robot, and providing an image acquisition function fault prompt for the patrol cleaning robot;

[0084] S109, generating cleaning function failure information of the patrol cleaning robot, and providing a cleaning function failure prompt of the patrol cleaning robot;

[0085] S1010, determine that the photovoltaic panel is damaged, use the photovoltaic panel damage status information in the first status information and the photovoltaic panel damage status information in the second status information, as well as the location information of the photovoltaic panel, as damage information, and execute S1011; use both pieces of information as damage information, and the operation and maintenance personnel can make further detailed judgments.

[0086] S1011, end. The location information of the photovoltaic panel can be determined according to the location of the inspection and cleaning robot, and when collecting the first image information and the second image information, the location information of the photovoltaic panel in the first image information and the second image information is synchronously obtained;

[0087] The operation and maintenance central control module is also used to generate operation and maintenance plans based on damage information;

[0088] Specifically, based on the damage information, the location information of the damaged photovoltaic panel is determined, and based on the location information and location importance, an optimization algorithm is used to generate an optimal operation and maintenance plan.

[0089] This embodiment also provides an intelligent fault operation and maintenance method for a photovoltaic new energy power generation system, which adopts the above-mentioned intelligent fault operation and maintenance system for a photovoltaic new energy power generation system.

[0090] Embodiment 2

[0091] This embodiment is basically the same as the above embodiment, except that: a model storage module;

[0092] The model storage module is used to store damage analysis modules of several photovoltaic panels; the damage analysis model only analyzes one type of damage, for example, damage analysis model A only analyzes whether the surface of the photovoltaic panel is damaged, and damage analysis model B only analyzes whether the surface of the photovoltaic panel is corroded, etc. All damage analysis models use image analysis models, such as CNN, RCNN, GAN, etc., which are trained in advance with historical data to meet the requirements and then stored in the model storage module;

[0093] The operation and maintenance central control module is further used for detecting the first state information of the photovoltaic panel, and the first state information contains damage state information of the photovoltaic panel. When the state of the photovoltaic panel is analyzed according to the second image information and the second state information is generated, the corresponding damage analysis model is obtained in the model storage module according to the damage state information of the photovoltaic panel in the first state information, and a secondary analysis is performed on the damaged area of ​​the photovoltaic panel to improve the accuracy of the damage analysis;

[0094] Specifically, when the first state information of the photovoltaic panel is detected and the first state information contains the damaged state information of the photovoltaic panel, when the state of the photovoltaic panel is analyzed according to the second image information to generate the second state information:

[0095] According to the type information of the photovoltaic panel damage in the photovoltaic panel damage state information in the first state information, a damage analysis model of the corresponding type information is obtained in the model storage module, and according to the damage area information, an image of the damage area in the second image information is extracted as an input of the damage analysis model, and the photovoltaic panel damage state information output by the damage analysis model is obtained as a secondary analysis result;

[0096] If the secondary analysis result is the same as the damage status information of the photovoltaic panel in the first status information, the secondary analysis result and the location information of the photovoltaic panel will be used as damage information; wherein it is judged that the damage type information is the same, that is, the secondary analysis result also analyzes the same result as the damage type information of the damage status information of the photovoltaic panel in the first status information.

[0097] The damage analysis module is a more targeted analysis model. By performing a secondary analysis on the cleaned photovoltaic panels, it can more accurately determine whether there is real damage. During the secondary analysis, only images of the damaged area are extracted for analysis, which improves accuracy and reduces the amount of calculation.

[0098] Embodiment 3

[0099] This embodiment is basically the same as the above embodiment, except that:

[0100] The operation and maintenance central control module is further used to match the photovoltaic panel damage status information if both the first status information and the second status information of the same photovoltaic panel contain photovoltaic panel damage status information, and obtain the matching degree of the photovoltaic panel damage status information in the first status information and the second status information of the same photovoltaic panel;

[0101] The photovoltaic panel damage status information matching includes:

[0102] Determine whether the damage type information in the damage state information of the photovoltaic panel of the first state information and the damage type information in the damage state information of the photovoltaic panel of the second state information have the same damage type information. If so, calculate the proportion of the same damage type information in the total damage type information as the type matching degree; if not, directly determine that the type matching degree is zero;

[0103] Analyze the overlap of the damaged areas according to the damaged area information in the damaged state information of the photovoltaic panel in the first state information and the damaged area information in the damaged state information of the photovoltaic panel in the second state information, and calculate the proportion of the overlap area in the total area as the area matching degree;

[0104] The matching degree is calculated based on the type matching degree and the area matching degree and the comprehensive weight. Before the calculation, the type matching degree and the area matching degree can be normalized to facilitate data calculation.

[0105] If the matching degree is greater than or equal to the preset matching degree, the damage is determined, and the damage state information of the photovoltaic panel in the first state information and / or the damage state information of the photovoltaic panel in the second state information, as well as the position information of the photovoltaic panel, are combined as damage information; if the matching degree is greater than the preset matching degree, it means that the two damage analyses are similar, and the accuracy of the damage analysis can be confirmed;

[0106] If the matching degree is less than the preset matching degree, and the type matching degree is not zero, the images of the damaged areas in the first image information and the second image information are spliced ​​and fused to form a new image, and the fused image recognition model is used to recognize the new image to obtain new damage status information of the photovoltaic panel, which is combined with the position information of the photovoltaic panel and merged as damage information; wherein the splicing and fusion adopts a convolutional neural network, a multi-scale transformation or an image fusion model based on sparse representation; the fused image recognition model adopts an image analysis model, such as CNN, RCNN, GAN, etc., which is trained with historical data in advance before use; if the matching degree is less than the preset matching degree, it means that there must be a problem in the two damage analyses, resulting in inconsistency in the damage analysis, but considering that the first image information may have problems in the damage analysis due to uncleanness, it is also determined whether the type matching degree is zero. If it is not 0, it means that there are similar partial results in the two damage judgments, which are not necessarily caused by uncleanness. Therefore, the images are fused, and damage analysis is performed on the fused image, and the damage situation is analyzed more accurately based on the situation of the two images;

[0107] In other embodiments, it can also be set as follows: if the matching degree is less than a preset matching degree, and the confidence degree of one of the damages is greater than a preset confidence degree, the images of the damaged areas in the first image information and the second image information are spliced ​​and fused to form a new image, and the fused image recognition model is used to recognize the new image to obtain new photovoltaic panel damage status information, comprehensively analyze the three photovoltaic panel damage status information, calculate the matching degree of the photovoltaic panel damage status information in the first state information and the new photovoltaic panel damage status information, and the matching degree of the photovoltaic panel damage status information in the second state information and the new photovoltaic panel damage status information, obtain the photovoltaic panel damage status information with a large matching degree, and combine it with the position information of the photovoltaic panel as the damage information; the confidence degree of one of the damages is greater than the preset confidence, specifically: calculate the confidence degree of the photovoltaic panel damage status information in the first state information and the photovoltaic panel damage status information in the second state information, if there is a photovoltaic panel damage status information whose confidence degree is greater than the preset confidence, then perform subsequent fusion, when the confidence degree is greater than the preset confidence, it can be determined that one of the results is a relatively accurate result, but the matching degree of the two results is too low, so it can be fused and then analyzed and judged to improve the accuracy of the damage analysis. In other embodiments, it may also be determined whether only the confidence level of the photovoltaic panel damage status information in the first status information is greater than a preset confidence level.

[0108] The above is only an embodiment of the present invention. The common sense such as the known specific structure and characteristics in the scheme is not described in detail here. The ordinary technicians in the relevant field know all the common technical knowledge in the technical field of the invention before the application date or priority date, can obtain all the existing technologies in the field, and have the ability to apply the conventional experimental means before that date. The ordinary technicians in the relevant field can improve and implement this scheme in combination with their own abilities under the enlightenment given by this application. Some typical known structures or known methods should not become obstacles for ordinary technicians in the relevant field to implement this application. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several deformations and improvements can be made, which should also be regarded as the protection scope of the present invention, which will not affect the effect of the implementation of the present invention and the practicality of the patent. The protection scope required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. An intelligent fault operation and maintenance system for photovoltaic new energy power generation system, characterized in that: include: Inspection and cleaning robots and operation and maintenance central control modules; Inspection and cleaning robots are used to move on photovoltaic panels and collect image information of photovoltaic panels; The image information includes first image information and second image information, the first image information includes: an image of the next photovoltaic panel that the patrol cleaning robot will drive; the second image information includes: an image of the previous photovoltaic panel that the patrol cleaning robot has driven; The operation and maintenance central control module is used to analyze the status of the photovoltaic panel according to the first image information and generate first status information; It is also used to determine whether the photovoltaic panel needs to be cleaned according to the first status information, and if so, generate a photovoltaic panel cleaning instruction and send it to the inspection and cleaning robot; The inspection and cleaning robot is also used to clean the photovoltaic panels according to the photovoltaic panel cleaning instructions; The operation and maintenance central control module is further used to analyze the state of the photovoltaic panel according to the second image information and generate second state information; It is also used to analyze whether the photovoltaic panel is damaged based on the first state information and the second state information of the same photovoltaic panel, and if so, generate damage information; The operation and maintenance central control module is also used to generate operation and maintenance plans based on damage information.

2. The intelligent fault operation and maintenance system for photovoltaic new energy power generation system according to claim 1 is characterized in that: The inspection and cleaning robot comprises: a shell and a controller; A universal roller is arranged at the bottom of the housing, and a motion motor is arranged inside. The output shaft of the motion motor is connected to the universal roller to drive the universal roller to rotate; A cleaning motor is also provided at the bottom of the shell, and a cleaning brush is connected to the output shaft of the cleaning motor. A first camera device is provided on the front side of the shell, and a second camera device is provided on the rear side. A first camera device is used to collect first image information of the photovoltaic panel at the front end of the housing and send it to the operation and maintenance central control module; The second camera device is used to collect second image information of the photovoltaic panel at the rear end of the shell and send it to the operation and maintenance central control module. The camera devices at the front and rear of the shell respectively collect the first image information and the second image information, which is convenient for collecting the images before and after cleaning.

3. The intelligent fault operation and maintenance system for photovoltaic new energy power generation system according to claim 1 is characterized in that: The step of analyzing the state of the photovoltaic panel according to the first image information and generating the first state information comprises: Preprocessing the first image information; The preprocessed first image information is input into the constructed neural network model, feature extraction is performed on the preprocessed first image information to obtain first image features, and the state of the photovoltaic panel is identified based on the extracted first image features to output first state information.

4. The intelligent fault operation and maintenance system for photovoltaic new energy power generation system according to claim 3 is characterized in that: The first status information includes: photovoltaic panel dirtiness status information and photovoltaic panel damage status information; The constructed neural network model adopts a CNN neural network model; the preprocessed first image information is input into the CNN neural network model, the CNN neural network model extracts the first image features, and evaluates the degree of contamination of the photovoltaic panel based on the extracted first image features, generates a photovoltaic panel contamination evaluation value as the photovoltaic panel contamination status information, and at the same time analyzes whether the photovoltaic panel is damaged based on the extracted first image features, and if so, generates damage type information and damage area information as the photovoltaic panel damage status information.

5. The intelligent fault operation and maintenance system for photovoltaic new energy power generation system according to claim 4 is characterized in that: The method of generating a photovoltaic panel cleaning instruction according to the first state information and sending the instruction to a controller of the inspection cleaning robot includes: Analyze the preset dirt evaluation range of the photovoltaic panel dirt evaluation value in the first state information, and generate a photovoltaic panel cleaning instruction of the corresponding level according to the preset range; and the higher the level of the photovoltaic panel cleaning instruction, the faster the cleaning motor speed. Photovoltaic panel cleaning instructions of different levels enable the inspection and cleaning robot to perform cleaning with different intensities to meet different cleaning requirements.

6. The intelligent fault operation and maintenance system for photovoltaic new energy power generation system according to claim 5 is characterized in that: The range of the photovoltaic panel dirt evaluation value is [0, f], and the preset dirt evaluation ranges are: [0, a), [a, b), [b, c), [c, d), [d, f], and a < b < c < d < f, all of which are positive numbers. Among them, [a, b), [b, c), [c, d), [d, f] correspond to the photovoltaic panel cleaning instructions, including: the first-level photovoltaic panel cleaning instruction, the second-level photovoltaic panel cleaning instruction, the third-level photovoltaic panel cleaning instruction, and the fourth-level photovoltaic panel cleaning instruction; After the photovoltaic panel dirt evaluation value is greater than or equal to a, generate the corresponding photovoltaic panel cleaning instruction according to the obtained range of the photovoltaic panel dirt evaluation value and send it to the inspection and cleaning robot.

7. The intelligent fault operation and maintenance system for photovoltaic new energy power generation system according to claim 5 is characterized in that: Analyze the state of the photovoltaic panel according to the second image information to generate the second state information, including: Preprocess the second image information; Input the preprocessed second image information into the constructed neural network model, extract features from the preprocessed second image information to obtain the second image features, and identify the state of the photovoltaic panel according to the extracted second image features, and output the second state information; where the second state information includes: photovoltaic panel dirt state information and photovoltaic panel damage state information; Among them, the constructed neural network model uses a CNN neural network model; input the preprocessed second image information into the CNN neural network model, the CNN neural network model extracts the second image features, and evaluates the dirt degree of the photovoltaic panel according to the extracted second image features to generate a photovoltaic panel dirt evaluation value as the photovoltaic panel dirt state information. At the same time, according to the extracted second image features, analyze whether the photovoltaic panel is damaged. If so, generate the damage type information and damage area information as the photovoltaic panel damage state information.

8. The intelligent fault operation and maintenance system for photovoltaic new energy power generation system according to claim 7 is characterized in that: Analyze whether the photovoltaic panel is damaged according to the first state information and the second state information of the same photovoltaic panel. If so, generate the damage information, including: S101. Judge whether the photovoltaic panel dirt evaluation value in the second state information of the photovoltaic panel is less than the preset dirt evaluation threshold a. If so, execute S102; if not, execute S105; S102. Judge whether there is photovoltaic panel damage state information in the second state information. If so, execute S103; if not, determine that the fault cause of the photovoltaic panel is dirt and execute S1011; S103. Judge whether the photovoltaic panel damage state information in the second state information of the photovoltaic panel is the same as the photovoltaic panel damage state information in the first state information. If so, execute S104; if not, execute S1010; S104. Determine that the photovoltaic panel is damaged, and combine the photovoltaic panel damage state information in the first state information or the photovoltaic panel damage state information in the second state information, and the position information of the photovoltaic panel as the damage information, and execute S1011; S105, determining whether the photovoltaic panel contamination assessment value in the second state information of the photovoltaic panel is less than the photovoltaic panel contamination assessment value in the first state information, if so, executing S106; if not, executing S109; S106, judging whether the photovoltaic panel is cleaned, and the corresponding photovoltaic panel cleaning instruction level, if the photovoltaic panel is cleaned, and the photovoltaic panel cleaning instruction level is less than level four, then generating a level four photovoltaic panel cleaning instruction for cleaning the photovoltaic panel, and sending it to the controller, the controller cleans the photovoltaic panel again; if the photovoltaic panel is cleaned, and the photovoltaic panel cleaning instruction level is level four, executing S107; if the photovoltaic panel is not cleaned, executing S108; S107, determining whether there is photovoltaic panel damage status information in the second status information, if yes, executing S103, if no, generating cleaning function failure information of the patrol cleaning robot, and providing a cleaning function failure prompt of the patrol cleaning robot, and executing S10; S108, determining whether the photovoltaic panel dirtiness assessment value in the first state information is greater than a preset dirtiness assessment threshold a, and if so, generating cleaning function fault information of the patrol cleaning robot, and providing a cleaning function fault prompt for the patrol cleaning robot; if not, generating image acquisition function fault information of the patrol cleaning robot, and providing an image acquisition function fault prompt for the patrol cleaning robot; S109, generating cleaning function failure information of the patrol cleaning robot, and providing a cleaning function failure prompt of the patrol cleaning robot; S1010, determining that the photovoltaic panel is damaged, taking the photovoltaic panel damage status information in the first status information and the photovoltaic panel damage status information in the second status information, and the location information of the photovoltaic panel as damage information, and executing S1011; S1011, end.

9. The intelligent fault operation and maintenance system for photovoltaic new energy power generation system according to claim 1, characterized in that: The generating of the operation and maintenance plan according to the damage information includes: determining the location information of the damaged photovoltaic panel according to the damage information, and generating the optimal operation and maintenance plan by using an optimization algorithm according to the location information and the location importance.

10. An intelligent fault operation and maintenance method for a photovoltaic new energy power generation system, characterized in that: An intelligent fault operation and maintenance system for a photovoltaic new energy power generation system as described in any one of claims 1 to 9 is adopted.

Images