Monitoring alarm method and system of photovoltaic system

Abstract

The invention relates to the field of photovoltaic systems, in particular to a monitoring alarm method and system for a photovoltaic system. The method comprises the following steps: performing all-directional thermal imaging scanning on a photovoltaic module array to generate a photovoltaic module multi-directional thermal imaging view; performing temperature distribution analysis on the multi-azimuth thermal imaging view of the photovoltaic module to obtain temperature distribution data of the photovoltaic module; obtaining space structure data of the photovoltaic module based on the multi-azimuth thermal imaging view of the photovoltaic module; performing three-dimensional point cloud reconstruction on the space structure data of the photovoltaic module, and constructing a three-dimensional point cloud structure model of the photovoltaic module; performing abnormal temperature analysis on the three-dimensional temperature distribution field model on the basis of a preset temperature threshold value on the basis of the temperature distribution data of the photovoltaic module so as to generate a local abnormal hot spot region; and performing hot spot region accurate positioning on the three-dimensional point cloud structure model of the photovoltaic module based on the local abnormal hot spot region so as to generate abnormal hot spot module position data. According to the invention, efficient and accurate monitoring alarm of the photovoltaic system is realized.

Description

Technical Field

[0001] The present invention relates to the field of photovoltaic systems, and in particular to a monitoring and alarming method and system for photovoltaic systems. Background Art

[0002] As an important renewable energy power generation system, photovoltaic system has a wide range of application prospects. However, during the long-term operation of photovoltaic system, various abnormal conditions may occur, such as component failure, shadow coverage, pollution, etc. These abnormal conditions will lead to system performance degradation or even failure. However, traditional photovoltaic system monitoring methods often rely on manual observation and manual testing, which have the problems of inaccurate abnormal monitoring and low monitoring efficiency. Therefore, in order to meet the needs of modern photovoltaic system monitoring, an intelligent photovoltaic system monitoring and alarm method is needed. Summary of the invention

[0003] In order to solve the above technical problems, the present invention proposes a monitoring and alarm method and system for a photovoltaic system to solve at least one of the above technical problems.

[0004] To achieve the above object, the present invention provides a monitoring and alarm method for a photovoltaic system, comprising the following steps: Step S1: performing a full-range thermal imaging scan on the photovoltaic module array to generate a multi-directional thermal imaging view of the photovoltaic module; performing a temperature distribution analysis on the multi-directional thermal imaging view of the photovoltaic module to obtain temperature distribution data of the photovoltaic module; Step S2: obtaining the spatial structure data of the photovoltaic module based on the multi-directional thermal imaging view of the photovoltaic module; reconstructing the spatial structure data of the photovoltaic module by three-dimensional point cloud to construct a three-dimensional point cloud structure model of the photovoltaic module; and constructing a three-dimensional temperature distribution field model based on the temperature distribution data of the photovoltaic module; Step S3: performing abnormal temperature analysis on the three-dimensional temperature distribution field model based on a preset temperature threshold, thereby generating a local abnormal hot spot area; performing accurate positioning of the hot spot area on the three-dimensional point cloud structure model of the photovoltaic module based on the local abnormal hot spot area, thereby generating abnormal hot spot module position data; generating an abnormal temperature alarm signal based on the abnormal hot spot module position data; Step S4: acquiring electrical characteristic parameters of the photovoltaic module based on the sensor; performing multi-scale wavelet decomposition on the electrical characteristic parameters of the photovoltaic module to generate multiple time-frequency bands; performing frequency component distribution analysis on the multiple time-frequency bands to obtain frequency band component distribution data; Step S5: generating frequency band transient time-frequency feature data based on multiple time-frequency frequency bands; performing abnormal time-frequency change analysis on frequency band component distribution data based on the frequency band transient time-frequency feature data to obtain frequency band abnormal transient feature data; generating an abnormal operation state alarm signal based on the frequency band abnormal transient feature data; Step S6: Obtain a regional heat dynamic adjustment strategy based on the abnormal temperature alarm signal; generate an abnormal time-frequency fault diagnosis strategy based on the abnormal operating state alarm signal; construct a neural network model for the abnormal time-frequency fault diagnosis strategy and the regional heat dynamic adjustment strategy, and construct a photovoltaic system intelligent monitoring and diagnosis model to perform monitoring and alarm operations.

[0005] The present invention obtains thermal imaging views of photovoltaic components at different angles through omni-directional thermal imaging scanning, providing more comprehensive temperature information. The temperature distribution analysis performs quantitative analysis on the thermal conditions of photovoltaic components to obtain temperature distribution data, providing a basis for subsequent abnormality detection and alarm. The acquisition of spatial structure data of photovoltaic components and reconstruction of three-dimensional point clouds establish a real photovoltaic component model, providing accurate spatial information. The construction of a three-dimensional temperature distribution field model combines the geometric structure and temperature distribution of photovoltaic components, providing a basis for subsequent abnormality detection and positioning. Abnormal temperature analysis detects temperature anomalies in photovoltaic components, improving the accuracy of fault detection. By accurately locating the abnormal hot spot area and determining the specific location of the abnormal hot spot component, guidance is provided for subsequent maintenance and troubleshooting. An abnormal temperature alarm signal is generated to promptly notify relevant personnel so that corresponding maintenance measures can be taken to reduce the risk of failure. The electrical characteristic parameters of photovoltaic components are obtained to understand the power output of the components, providing basic data for subsequent abnormality detection. Multi-scale wavelet decomposition extracts the photovoltaic components. The different frequency components of the signal are conducive to analyzing abnormal conditions within different frequency ranges. The frequency component distribution analysis conducts quantitative analysis on the component distribution of different frequency bands, providing a basis for abnormal detection and alarm. The transient time-frequency characteristic data of the frequency band reflects the changes of the photovoltaic module signal in the time domain and frequency domain, which helps to capture short-term abnormal signals. The abnormal time-frequency change analysis detects the abnormal time-frequency change pattern in the frequency band component distribution data, improves the fault detection capability, generates abnormal operation status alarm signals to promptly notify relevant personnel so that they can take corresponding maintenance measures to reduce the risk of failure. The regional heat dynamic adjustment strategy adjusts the heat of the photovoltaic module according to the abnormal temperature alarm signal to improve the stability and efficiency of the system. The abnormal time-frequency fault diagnosis strategy diagnoses the fault of the photovoltaic module according to the abnormal operation status alarm signal, improves the accuracy and efficiency of fault detection, and constructs an intelligent monitoring and diagnosis model for the photovoltaic system to comprehensively consider the abnormal temperature and operation status information, realize intelligent monitoring and alarm of the photovoltaic system, and improve the reliability and maintainability of the system.

[0006] Preferably, step S1 comprises the following steps: Step S11: performing a full-range thermal imaging scan on the photovoltaic module array to generate a multi-directional thermal imaging view of the photovoltaic module; Step S12: performing component position identification on the multi-directional thermal imaging view of the photovoltaic component to obtain the photovoltaic component node; Step S13: Calculating the temperature value of the component node to obtain the temperature value of the photovoltaic component node; Step S14: performing temperature distribution analysis on the multi-directional thermal imaging views of the photovoltaic components based on the temperature values ​​of the photovoltaic component nodes to obtain temperature distribution data of the photovoltaic components. The present invention obtains thermal imaging views of photovoltaic components at different angles through omni-directional thermal imaging scanning, thereby providing more comprehensive temperature information. Through the thermal imaging view, the thermal distribution of the photovoltaic components can be understood in real time, and abnormal temperature areas can be found. The component position identification determines the specific position of the photovoltaic components in the photovoltaic array, and provides accurate node information for subsequent steps. Through the node information, the photovoltaic components are individually monitored and analyzed to achieve more refined monitoring and alarms. The temperature value calculation obtains the specific temperature value of each component node based on the thermal imaging view and node information of the photovoltaic components. Through the node temperature value, the temperature change of the photovoltaic components is monitored in real time to determine whether there is an abnormal hot spot or overheating. The temperature distribution analysis performs quantitative analysis on the thermal condition of the photovoltaic components to obtain temperature distribution data. The temperature distribution data reflects the thermal state of the photovoltaic components, including hot spot distribution, temperature gradient and other information, providing a basis for subsequent abnormal detection and alarms.

[0007] Preferably, the specific steps of step S2 are: Step S21: Performing a spatial structure analysis on the photovoltaic module multi-directional thermal imaging view to obtain the spatial structure data of the photovoltaic module; Step S22: performing three-dimensional feature point sampling processing on the multi-directional thermal imaging view of the photovoltaic module to obtain multi-directional feature points; Step S23: performing stereo matching on the multi-directional feature points to obtain spatially corresponding feature points; Step S24: reconstructing the three-dimensional point cloud of the photovoltaic module spatial structure data based on the spatial corresponding feature points to construct a three-dimensional point cloud structure model of the photovoltaic module; Step S25: Performing dynamic temperature interpolation processing on the three-dimensional point cloud structure model of the photovoltaic module based on the temperature distribution data of the photovoltaic module to construct a three-dimensional temperature distribution field model.

[0008] The present invention determines the relative positions and spatial layouts of photovoltaic components through multi-directional thermal imaging views through component spatial structure analysis. Through the spatial structure data of photovoltaic components, the overall structure of the photovoltaic system is understood, including information such as the distance and arrangement between components, providing a basis for subsequent steps. The three-dimensional feature point sampling process extracts representative feature points from the multi-directional thermal imaging views. Through the multi-directional feature points, the spatial features and shape features of the photovoltaic components are more accurately described, providing data support for subsequent steps. Stereo matching determines the corresponding relationship between the multi-directional feature points in space through the relationship between them. Through the obtained spatial corresponding feature points, different perspectives are established. The correspondence between the feature points under the 3D point cloud provides a data basis for subsequent steps. The 3D point cloud reconstruction restores the 3D shape and structure of the PV module through the spatial corresponding feature points. The 3D point cloud structure model of the PV module can provide more detailed component space information, including shape, size and relative position, etc., providing an accurate data basis for subsequent steps. The dynamic temperature interpolation processing associates the temperature distribution data of the PV module with the 3D point cloud structure model to realize the interpolation of temperature in 3D space. By constructing a 3D temperature distribution field model, the temperature distribution of the PV module can be intuitively understood, including hot spots, thermal gradients and other information, providing more comprehensive data support for the monitoring and alarm of the PV system.

[0009] Preferably, the specific steps of step S25 are: Calculate pixel coordinates of multi-directional thermal imaging views of photovoltaic modules based on photovoltaic module nodes to obtain two-dimensional pixel coordinates; Calculate the three-dimensional projection coordinates of the three-dimensional point cloud structure model of the photovoltaic module based on the two-dimensional pixel point coordinates to obtain the three-dimensional projection point coordinates; Based on the three-dimensional projection point coordinates, the temperature distribution data of the photovoltaic module is dynamically interpolated to the three-dimensional point cloud structure model of the photovoltaic module to obtain an initial three-dimensional temperature field model; Analyze the temperature difference between the structures of the photovoltaic module temperature distribution data to obtain the temperature difference change data of the module structure; According to the temperature difference change data of the component structure, the initial three-dimensional temperature field model is completed with optical flow details to construct a three-dimensional temperature distribution field model.

[0010] The present invention maps the hot spot position in the multi-directional thermal imaging view of the photovoltaic component to the two-dimensional pixel coordinates through pixel point coordinate calculation. The obtained two-dimensional pixel point coordinates provide the specific position information of the hot spot in the view, providing a data basis for subsequent steps. The three-dimensional projection coordinate calculation maps the two-dimensional pixel point coordinates to the three-dimensional point cloud structure model of the photovoltaic component. The three-dimensional projection point coordinates obtained by calculation map the position of the hot spot from the image space to the actual three-dimensional space, providing data support for subsequent steps. The dynamic temperature interpolation processing associates the temperature distribution data of the photovoltaic component with the three-dimensional point cloud structure model to realize the interpolation of temperature in the three-dimensional space. The initial three-dimensional temperature field obtained by the interpolation processing The model provides the temperature distribution of photovoltaic components in space and provides a data basis for the monitoring and alarm of the photovoltaic system. The temperature difference change analysis between structures compares the temperature differences between different components through the temperature distribution data of photovoltaic components. The obtained component structure temperature difference change data provides information on the temperature difference between different components, which is used to evaluate the thermal coupling between components. The optical flow detail completion processing uses the component structure temperature difference change data to correct and adjust the initial three-dimensional temperature field model to make it more accurate and complete. The constructed three-dimensional temperature distribution field model can provide detailed component temperature distribution information, including hot spots, thermal gradients, etc., providing more comprehensive data support for the monitoring and alarm of the photovoltaic system.

[0011] Preferably, the specific steps of step S3 are: Step S31: performing abnormal temperature analysis on the three-dimensional temperature distribution field model based on a preset temperature threshold to obtain an abnormal temperature area; Step S32: performing temperature gradient diffusion processing on the temperature abnormal area, thereby generating a local abnormal hot spot area; Step S33: performing regional morphological analysis on the local abnormal hotspot area to obtain morphological characteristics of the abnormal hotspot area; Step S34: performing interference edge segmentation on the local abnormal hot spot area based on the morphological characteristics of the abnormal hot spot area to obtain a component hot spot area; Step S35: accurately positioning the hot spot area of ​​the photovoltaic component three-dimensional point cloud structure model based on the component hot spot area, thereby generating abnormal hot spot component position data; Step S36: Perform abnormal temperature alarm processing based on the abnormal hot spot component position data to generate an abnormal temperature alarm signal.

[0012] The present invention analyzes the three-dimensional temperature distribution field model through a preset temperature threshold to determine whether there is an abnormal temperature area. The obtained abnormal temperature area indicates the area in the photovoltaic module where the temperature exceeds or is lower than the preset threshold, which is used for further abnormal detection and alarm processing. The temperature gradient diffusion processing expands the influence range of the temperature abnormal area to form a larger local abnormal hot spot area. The generated local abnormal hot spot area helps to more accurately locate and identify temperature abnormality problems and improve the sensitivity of abnormal detection. The regional morphological analysis analyzes and measures the shape, size, connectivity, etc. of the local abnormal hot spot area. The obtained abnormal hot spot area morphological characteristics provide the geometric shape and structural information of the abnormal hot spot area, which provides data basis for subsequent processing steps and interference The disturbance edge segmentation effectively separates the abnormal hot spot area from the surrounding environment to obtain a more accurate component hot spot area. The generated component hot spot area provides the specific contour and boundary information of the abnormal hot spot for further positioning and identification. Through the precise positioning of the hot spot area, the abnormal hot spot is associated with the specific position of the photovoltaic component. The generated abnormal hot spot component position data provides the accurate position information of the photovoltaic component where the abnormal hot spot is located, which is used for subsequent abnormal temperature alarm processing. Abnormal temperature alarm processing is performed based on the abnormal hot spot component position data. The corresponding alarm signal is generated according to the position and temperature information of the abnormal hot spot. The generated abnormal temperature alarm signal is used to promptly notify the operator or automation system to take corresponding measures for abnormal processing and troubleshooting.

[0013] Preferably, the specific steps of step S34 are: Identify edge feature points in local abnormal hotspot areas to extract edge feature points; Based on the morphological characteristics of the abnormal hot spot area, outlier edge point analysis is performed on the edge feature points to mark the outlier edge feature points; Based on the outlier edge feature points, the interference edge segmentation of the local abnormal hot spot area is performed to obtain a smooth hot spot contour line; The hot spot area of ​​the component is obtained by reconstructing the boundary of the local abnormal hot spot area through smoothing the hot spot contour.

[0014] The present invention extracts edge feature points with significant temperature changes from local abnormal hot spot areas through edge feature point recognition. The extracted edge feature points are helpful to further analyze the edge shape and boundary information of the abnormal hot spot area. Outlier edge point analysis identifies outlier edge feature points that are significantly different from surrounding edge points. Marking of outlier edge feature points is helpful to further screen and segment edge information of the abnormal hot spot area and improve the positioning accuracy of abnormal hot spots. Interference edge segmentation uses outlier edge feature points to effectively separate the abnormal hot spot area from the surrounding environment. The obtained smooth hot spot contour line more accurately describes the boundary shape of the abnormal hot spot area and improves the visualization effect of the abnormal hot spot. Boundary fracture reconstruction fills the boundary fracture of the abnormal hot spot area to obtain a more complete and continuous component hot spot area. The obtained component hot spot area provides accurate contour information of the abnormal hot spot for further abnormal temperature alarm processing and positioning.

[0015] Preferably, the specific steps of step S4 are: Step S41: obtaining electrical characteristic parameters of photovoltaic modules based on sensors; Step S42: performing smoothing and filtering processing on the electrical characteristic parameters of the photovoltaic module to obtain filtered electrical characteristic parameters; Step S43: performing multi-scale wavelet decomposition on the filter electrical characteristic parameters to generate multiple time-frequency bands; Step S44: calculating the time-frequency components of each frequency subband for each of the multiple time-frequency frequency bands to obtain the energy stability of each frequency subband; Step S45: performing frequency component distribution analysis on multiple time-frequency frequency bands based on the energy stationarity of each decomposition unit, thereby obtaining frequency band component distribution data.

[0016] The present invention obtains electrical characteristic parameters of a photovoltaic module through a sensor, including information such as voltage, current, and power. The electrical characteristic parameters reflect the real-time working state of the photovoltaic module and are an important basis for monitoring and analyzing the performance of a photovoltaic system. Smoothing and filtering remove noise and mutations in the electrical characteristic parameters to make them more stable and reliable. Obtaining the filtered electrical characteristic parameters is helpful to reduce false alarms and misjudgments, and improve the accuracy and reliability of photovoltaic system monitoring. Multi-scale wavelet decomposition decomposes the filtered electrical characteristic parameters into time-frequency bands of different scales and frequencies. The time-frequency bands represent the changes of the electrical characteristic parameters in different time scales and frequency ranges, which are conducive to revealing the dynamic characteristics and fault information of the photovoltaic module. The time-frequency component calculation is performed for each frequency sub-band to obtain the energy distribution of the frequency sub-band. The energy stability reflects the energy stability of the photovoltaic module in different frequency sub-bands, which is used to evaluate the working state and abnormal conditions of the photovoltaic module. The frequency component distribution data reflects the energy distribution of the photovoltaic module in different frequency bands, which is used to analyze the frequency characteristics and fault conditions of the photovoltaic module. By analyzing the frequency band component distribution data, the abnormal energy distribution in the frequency band is found, thereby judging whether the photovoltaic module has a fault or abnormal condition.

[0017] Preferably, the specific steps of step S5 are: Step S51: extracting transient time-frequency features from multiple time-frequency bands to generate frequency band transient time-frequency feature data; Step S52: performing frequency transient fluctuation analysis on the frequency band component distribution data based on the frequency band transient time-frequency characteristic data to obtain frequency band transient change data; Step S53: performing abnormal time-frequency change analysis on the frequency band transient change data to obtain abnormal transient characteristic data of the frequency band; Step S54: Perform component abnormal operation state alarm processing based on the frequency band abnormal transient characteristic data to generate an abnormal operation state alarm signal.

[0018] The present invention captures the instantaneous changes of the frequency band in time and frequency by extracting the transient time-frequency features, obtains the instantaneous energy, instantaneous frequency and other information of the frequency band by extracting the transient time-frequency features, further describes the dynamic characteristics of the photovoltaic module, analyzes the instantaneous changes of the frequency band in different time periods by analyzing the transient change data of the frequency band, detects the instantaneous change trend, amplitude fluctuation and other information of the frequency band, and provides a more detailed description of the dynamic characteristics of the frequency band, and detects the abnormal conditions in the transient change data of the frequency band by analyzing the abnormal time-frequency changes of the frequency band, such as mutations, abnormal fluctuations, etc., to find out the faults or abnormal operating states of the photovoltaic module, performs abnormal operating state alarm processing according to the abnormal transient characteristic data of the frequency band, quickly identifies and reports the abnormal conditions of the photovoltaic module, and generates abnormal operating state alarm signals for the monitoring and alarm systems of the photovoltaic system, so as to promptly notify the operation and maintenance personnel to troubleshoot and repair the faults, thereby improving the reliability and safety of the photovoltaic system.

[0019] Preferably, the specific steps of step S6 are: Step S61: dynamically adjusting the power supply of the hot spot component based on the abnormal temperature alarm signal to obtain a regional heat dynamic adjustment strategy; Step S62: identifying potential fault nodes based on the abnormal operation state alarm signal to obtain potential fault nodes of the photovoltaic components; Step S63: making a fault diagnosis decision for a potential fault node of a photovoltaic module to generate an abnormal time-frequency fault diagnosis strategy; Step S64: construct a neural network model for the abnormal time-frequency fault diagnosis strategy and the regional heat dynamic adjustment strategy, and build a photovoltaic system intelligent monitoring and diagnosis model to perform monitoring and alarm operations. The present invention adjusts the power supply of the photovoltaic module by dynamically adjusting the power supply of the hot spot module based on the abnormal temperature alarm signal, and effectively controls the temperature of the photovoltaic module by dynamically adjusting the power supply power, reduces the occurrence and development of the hot spot phenomenon, and improves the operating efficiency and life of the photovoltaic module. The generated regional heat dynamic adjustment strategy is optimized according to the heat distribution of different regions or modules, and further improves the overall performance of the photovoltaic system. By identifying potential fault nodes based on abnormal operating status alarm signals, the faulty components or nodes in the photovoltaic system are quickly located. The identification of potential fault nodes helps operation and maintenance personnel to quickly troubleshoot and locate faults, reduce fault handling time and maintenance costs, and The fault diagnosis decision is made at the fault node to determine the type, cause and severity of the fault. The abnormal time-frequency fault diagnosis strategy takes into account the abnormal characteristics in frequency and time to provide more accurate fault diagnosis results. The intelligent monitoring and diagnosis model of the photovoltaic system is constructed. The abnormal time-frequency fault diagnosis strategy and the regional heat dynamic adjustment strategy are comprehensively considered through the neural network model to achieve comprehensive monitoring and alarm functions. The intelligent monitoring and diagnosis model conducts comprehensive analysis and decision-making on multiple parameters and characteristics to improve the accuracy and efficiency of fault diagnosis. By executing monitoring and alarm operations, the intelligent monitoring and diagnosis model monitors the status of the photovoltaic system in real time and sends out alarm signals in time to help operation and maintenance personnel quickly respond to and handle abnormal situations, and ensure the safe and stable operation of the photovoltaic system.

[0020] In this specification, a monitoring and alarm system of a photovoltaic system is provided, which is used to execute the monitoring and alarm method of the photovoltaic system as described above, including: The temperature distribution module performs a full range of thermal imaging scans on the photovoltaic module array to generate a multi-directional thermal imaging view of the photovoltaic module; performs temperature distribution analysis on the multi-directional thermal imaging view of the photovoltaic module to obtain the temperature distribution data of the photovoltaic module; The three-dimensional structure module obtains the spatial structure data of the photovoltaic module based on the multi-directional thermal imaging view of the photovoltaic module; reconstructs the three-dimensional point cloud of the photovoltaic module spatial structure data to build a three-dimensional point cloud structure model of the photovoltaic module; and builds a three-dimensional temperature distribution field model based on the temperature distribution data of the photovoltaic module; The hot spot area module performs abnormal temperature analysis on the three-dimensional temperature distribution field model based on the preset temperature threshold, thereby generating local abnormal hot spot areas; based on the local abnormal hot spot areas, the three-dimensional point cloud structure model of the photovoltaic module is accurately positioned in the hot spot area, thereby generating abnormal hot spot module location data; based on the abnormal hot spot module location data, an abnormal temperature alarm signal is generated; The time-frequency decomposition module obtains the electrical characteristic parameters of the photovoltaic module based on the sensor; performs multi-scale wavelet decomposition on the electrical characteristic parameters of the photovoltaic module to generate multiple time-frequency bands; performs frequency component distribution analysis on multiple time-frequency bands to obtain frequency band component distribution data; The abnormal time-frequency module generates frequency band transient time-frequency feature data based on multiple time-frequency bands; performs abnormal time-frequency change analysis on the frequency band component distribution data based on the frequency band transient time-frequency feature data to obtain the frequency band abnormal transient feature data; generates an abnormal operation status alarm signal based on the frequency band abnormal transient feature data; The monitoring and diagnosis model module obtains the regional heat dynamic adjustment strategy based on the abnormal temperature alarm signal; generates the abnormal time-frequency fault diagnosis strategy based on the abnormal operation status alarm signal; constructs a neural network model for the abnormal time-frequency fault diagnosis strategy and the regional heat dynamic adjustment strategy, and constructs an intelligent monitoring and diagnosis model for the photovoltaic system to perform monitoring and alarm operations.

[0021] The present invention obtains thermal imaging views of multiple angles of the photovoltaic module array through omni-directional thermal imaging scanning, providing more comprehensive temperature information. The temperature distribution analysis performs detailed analysis and evaluation on the temperature conditions of the photovoltaic modules. The generated temperature distribution data of the photovoltaic modules is used as input for subsequent steps for further analysis and decision-making. The three-dimensional point cloud reconstruction obtains the spatial structure information of the photovoltaic modules according to the multi-directional thermal imaging views of the photovoltaic modules, realizes the restoration of the shape of the photovoltaic modules, and the constructed three-dimensional point cloud structure model of the photovoltaic modules provides accurate spatial information for subsequent analysis and positioning. The three-dimensional temperature distribution field model combines the temperature distribution data of the photovoltaic modules and comprehensively considers the spatial structure of the photovoltaic modules. The structure and temperature distribution of the photovoltaic module provide a basis for subsequent abnormal analysis. The abnormal temperature analysis detects the local abnormal hot spot area in the temperature distribution field model, that is, the area with potential faults or abnormalities. The hot spot area is accurately positioned by matching the abnormal hot spot area with the three-dimensional point cloud structure model of the photovoltaic module to determine the specific location of the abnormal hot spot component. The generated abnormal hot spot component location data is used for subsequent fault diagnosis and alarm processing, providing important information about the abnormal hot spot component. The time-frequency decomposition converts the electrical characteristic parameters of the photovoltaic module into time-frequency feature data in different frequency bands, extracts the frequency information of the electrical signal, and the multi-scale wavelet decomposition decomposes the electrical characteristic parameters at different scales. The frequency components in each time-frequency band are analyzed to obtain the time-frequency characteristic data in different frequency bands, providing more detailed frequency component information. The frequency component distribution analysis analyzes and counts the frequency components in each time-frequency band to obtain the frequency band component distribution data, providing a basis for subsequent abnormal time-frequency analysis. The frequency band transient time-frequency characteristic data captures the transient changes of the electrical characteristic parameters of the photovoltaic module in different frequency bands, providing richer time-frequency characteristic information. The abnormal time-frequency change analysis performs abnormal detection on the frequency band component distribution data, identifies the frequency band with abnormal time-frequency changes, further analyzes the cause and degree of the abnormality, and generates abnormal operation status alarm signals to remind operators in time. The personnel or system managers shall take corresponding maintenance or adjustment measures if there are abnormal photovoltaic modules. The regional heat dynamic adjustment strategy shall adjust the heat distribution of photovoltaic modules according to the abnormal temperature alarm signal, optimize the heat distribution and utilization efficiency of the photovoltaic system, and the abnormal time-frequency fault diagnosis strategy shall diagnose the electrical characteristics of photovoltaic modules according to the abnormal operation status alarm signal, and determine the cause and type of the fault or abnormality. The constructed photovoltaic system intelligent monitoring and diagnosis model comprehensively considers the abnormal temperature alarm and abnormal time-frequency fault diagnosis strategy, uses neural network and other technologies to build and analyze the model, realizes automated monitoring and alarm functions, and improves the stability and efficiency of the photovoltaic system. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 A schematic diagram of the steps of a photovoltaic system monitoring and alarm method according to the present invention; Figure 2Detailed implementation flow chart of step S1; Figure 3 Detailed implementation flow chart of step S2; Figure 4 Detailed implementation flow chart of step S3. DETAILED DESCRIPTION

[0023] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0024] The present application example provides a monitoring and alarming method and system for a photovoltaic system. The execution subjects of the monitoring and alarming method and system for the photovoltaic system include but are not limited to: mechanical equipment, data processing platform, cloud server node, network upload equipment, etc. equipped with the system can be regarded as general computing nodes of the present application, and the data processing platform includes but is not limited to: at least one of an audio and image management system, an information management system, and a cloud data management system.

[0025] See also Figures 1 to 4 The present invention provides a monitoring and alarm method for a photovoltaic system, and the monitoring and alarm method for a photovoltaic system comprises the following steps: Step S1: performing a full-range thermal imaging scan on the photovoltaic module array to generate a multi-directional thermal imaging view of the photovoltaic module; performing a temperature distribution analysis on the multi-directional thermal imaging view of the photovoltaic module to obtain temperature distribution data of the photovoltaic module; Step S2: obtaining the spatial structure data of the photovoltaic module based on the multi-directional thermal imaging view of the photovoltaic module; reconstructing the spatial structure data of the photovoltaic module by three-dimensional point cloud to construct a three-dimensional point cloud structure model of the photovoltaic module; and constructing a three-dimensional temperature distribution field model based on the temperature distribution data of the photovoltaic module; Step S3: performing abnormal temperature analysis on the three-dimensional temperature distribution field model based on a preset temperature threshold, thereby generating a local abnormal hot spot area; performing accurate positioning of the hot spot area on the three-dimensional point cloud structure model of the photovoltaic module based on the local abnormal hot spot area, thereby generating abnormal hot spot module position data; generating an abnormal temperature alarm signal based on the abnormal hot spot module position data; Step S4: acquiring electrical characteristic parameters of the photovoltaic module based on the sensor; performing multi-scale wavelet decomposition on the electrical characteristic parameters of the photovoltaic module to generate multiple time-frequency bands; performing frequency component distribution analysis on the multiple time-frequency bands to obtain frequency band component distribution data; Step S5: generating frequency band transient time-frequency feature data based on multiple time-frequency frequency bands; performing abnormal time-frequency change analysis on frequency band component distribution data based on the frequency band transient time-frequency feature data to obtain frequency band abnormal transient feature data; generating an abnormal operation state alarm signal based on the frequency band abnormal transient feature data; Step S6: Obtain a regional heat dynamic adjustment strategy based on the abnormal temperature alarm signal; generate an abnormal time-frequency fault diagnosis strategy based on the abnormal operating state alarm signal; construct a neural network model for the abnormal time-frequency fault diagnosis strategy and the regional heat dynamic adjustment strategy, and construct a photovoltaic system intelligent monitoring and diagnosis model to perform monitoring and alarm operations.

[0026] The present invention obtains thermal imaging views of photovoltaic components at different angles through omni-directional thermal imaging scanning, providing more comprehensive temperature information. The temperature distribution analysis performs quantitative analysis on the thermal conditions of photovoltaic components to obtain temperature distribution data, providing a basis for subsequent abnormality detection and alarm. The acquisition of spatial structure data of photovoltaic components and reconstruction of three-dimensional point clouds establish a real photovoltaic component model, providing accurate spatial information. The construction of a three-dimensional temperature distribution field model combines the geometric structure and temperature distribution of photovoltaic components, providing a basis for subsequent abnormality detection and positioning. Abnormal temperature analysis detects temperature anomalies in photovoltaic components, improving the accuracy of fault detection. By accurately locating the abnormal hot spot area and determining the specific location of the abnormal hot spot component, guidance is provided for subsequent maintenance and troubleshooting. An abnormal temperature alarm signal is generated to promptly notify relevant personnel so that corresponding maintenance measures can be taken to reduce the risk of failure. The electrical characteristic parameters of photovoltaic components are obtained to understand the power output of the components, providing basic data for subsequent abnormality detection. Multi-scale wavelet decomposition extracts the photovoltaic components. The different frequency components of the signal are conducive to analyzing abnormal conditions within different frequency ranges. The frequency component distribution analysis conducts quantitative analysis on the component distribution of different frequency bands, providing a basis for abnormal detection and alarm. The transient time-frequency characteristic data of the frequency band reflects the changes of the photovoltaic module signal in the time domain and frequency domain, which helps to capture short-term abnormal signals. The abnormal time-frequency change analysis detects the abnormal time-frequency change pattern in the frequency band component distribution data, improves the fault detection capability, generates abnormal operation status alarm signals to promptly notify relevant personnel so that they can take corresponding maintenance measures to reduce the risk of failure. The regional heat dynamic adjustment strategy adjusts the heat of the photovoltaic module according to the abnormal temperature alarm signal to improve the stability and efficiency of the system. The abnormal time-frequency fault diagnosis strategy diagnoses the fault of the photovoltaic module according to the abnormal operation status alarm signal, improves the accuracy and efficiency of fault detection, and constructs an intelligent monitoring and diagnosis model for the photovoltaic system to comprehensively consider the abnormal temperature and operation status information, realize intelligent monitoring and alarm of the photovoltaic system, and improve the reliability and maintainability of the system.

[0027] In the embodiment of the present invention, refer to Figure 1 , is a schematic flow chart of the steps of a monitoring and alarming method for a photovoltaic system of the present invention. In this example, the steps of the monitoring and alarming method for a photovoltaic system include: Step S1: performing a full-range thermal imaging scan on the photovoltaic module array to generate a multi-directional thermal imaging view of the photovoltaic module; performing a temperature distribution analysis on the multi-directional thermal imaging view of the photovoltaic module to obtain temperature distribution data of the photovoltaic module; In this embodiment, multiple high-resolution thermal imaging cameras are deployed around the photovoltaic module array, and the complete thermal imaging data of the module surface is obtained by shooting from multiple perspectives. The collected thermal imaging data is pre-processed by correction and registration to remove interference factors such as noise and shadows in the thermal imaging. The pre-processed thermal imaging data is analyzed using image processing technology to extract the temperature distribution characteristics of each photovoltaic module surface, including local hot spot areas, temperature gradients, abnormal temperature differences and other indicators. Based on the results of temperature analysis, the temperature distribution data of the photovoltaic module is generated. The temperature information is displayed in the form of images or data tables to ensure that the temperature distribution data is visualized and easy to understand.

[0028] Step S2: obtaining the spatial structure data of the photovoltaic module based on the multi-directional thermal imaging view of the photovoltaic module; reconstructing the spatial structure data of the photovoltaic module by three-dimensional point cloud to construct a three-dimensional point cloud structure model of the photovoltaic module; and constructing a three-dimensional temperature distribution field model based on the temperature distribution data of the photovoltaic module; In this embodiment, computer vision technology and image processing algorithms are used to extract spatial structure data of photovoltaic modules, which include information such as the position, size, and shape of the photovoltaic modules. According to the specific selection of visual technology, feature point extraction, contour detection, edge detection, and other methods are used to obtain the spatial structure data of the photovoltaic modules, and the spatial structure data of the photovoltaic modules is converted into three-dimensional point cloud data. This is achieved by using computer vision and three-dimensional reconstruction technology. According to the spatial structure data of the photovoltaic modules, the three-dimensional coordinate points of the photovoltaic modules are calculated, and a corresponding point cloud data set is generated. Common three-dimensional reconstruction technologies include stereo matching, structured light scanning, laser scanning, and other methods. Based on the obtained three-dimensional point cloud data, point cloud processing and model construction are performed to generate a three-dimensional point cloud structure model of the photovoltaic module. Computer-aided design software or three-dimensional modeling software is used to process the point cloud data and construct the model. By connecting the points in the point cloud, a three-dimensional surface model of the photovoltaic module is obtained, including information such as the shape and geometric features of the photovoltaic module. The temperature distribution information is associated with the three-dimensional point cloud structure model of the photovoltaic module to construct a three-dimensional temperature distribution field model of the photovoltaic module. The temperature information is mapped to the corresponding point cloud coordinates to form the temperature attribute of each point in the point cloud model.

[0029] Step S3: performing abnormal temperature analysis on the three-dimensional temperature distribution field model based on a preset temperature threshold, thereby generating a local abnormal hot spot area; performing accurate positioning of the hot spot area on the three-dimensional point cloud structure model of the photovoltaic module based on the local abnormal hot spot area, thereby generating abnormal hot spot module position data; generating an abnormal temperature alarm signal based on the abnormal hot spot module position data; In this embodiment, the temperature in the model is analyzed, and the temperature value of each point is compared with the preset temperature threshold to determine whether there is abnormal temperature. When the temperature of a point exceeds the threshold, it is marked as an abnormal hot spot. The points marked as abnormal hot spots are clustered to form local abnormal hot spot areas. The clustering algorithm uses common spatial clustering algorithms, such as K-means, DBSCAN, etc. These clustering algorithms cluster according to the spatial position and temperature value of the points, and classify adjacent abnormal hot spots into the same hot spot area. Based on the local abnormal hot spot area, the three-dimensional point cloud structure model of the photovoltaic module is accurately located in the hot spot area. By matching the features of the abnormal hot spot area with the features in the point cloud model, the specific position of the abnormal hot spot in the photovoltaic module is determined, and this is achieved using point cloud processing and registration algorithms, such as ICP (Iterative Closest The hot spot location data is generated based on the location results of the hot spot area. These data include the number, coordinates, size of the hot spot area and other information of the PV module. The location information of the abnormal hot spot module is associated with other systems for further analysis and processing. Based on the abnormal hot spot module location data, an abnormal temperature alarm signal is generated. According to system requirements, the alarm signal is displayed in a visual form and sent to the monitoring system or the corresponding alarm device is triggered, so that timely measures can be taken to deal with the abnormal situation.

[0030] Step S4: acquiring electrical characteristic parameters of the photovoltaic module based on the sensor; performing multi-scale wavelet decomposition on the electrical characteristic parameters of the photovoltaic module to generate multiple time-frequency bands; performing frequency component distribution analysis on the multiple time-frequency bands to obtain frequency band component distribution data; In this embodiment, appropriate sensors or measuring devices are used to obtain electrical characteristic parameters of the photovoltaic module, including voltage, current, power, etc. of the photovoltaic module. The sensor is directly connected to the circuit of the photovoltaic module, or the electrical characteristic parameters are obtained through wireless communication and other means. Multi-scale wavelet decomposition is performed on the obtained electrical characteristic parameters of the photovoltaic module. Multi-scale wavelet decomposition is a signal processing technology used to decompose a signal into components of different scales and frequencies. Wavelet transform technology, such as discrete wavelet transform (DWT) or continuous wavelet transform (CWT), is used to decompose the electrical characteristic parameters, which will generate a series of time-frequency frequencies. Bands, each frequency band corresponds to a different scale and frequency component, and frequency component distribution analysis is performed on multiple time-frequency bands. By performing statistics and analysis on each time-frequency band, the distribution of frequency components is obtained, including the amplitude, phase, energy and other information of the frequency components. Various signal processing and spectrum analysis methods, such as Fourier analysis and power spectral density estimation, are used to obtain the distribution data of frequency components. The frequency component distribution data are integrated to obtain frequency band component distribution data. The frequency component data in each frequency band are merged and sorted to form frequency band component distribution data, which are used for subsequent analysis, modeling and judgment.

[0031] Step S5: generating frequency band transient time-frequency feature data based on multiple time-frequency frequency bands; performing abnormal time-frequency change analysis on frequency band component distribution data based on the frequency band transient time-frequency feature data to obtain frequency band abnormal transient feature data; generating an abnormal operation state alarm signal based on the frequency band abnormal transient feature data; In this embodiment, based on multiple time-frequency bands, frequency band transient time-frequency feature data is generated, which is completed by applying a suitable time-frequency analysis method to each frequency band, such as short-time Fourier transform (STFT), wavelet packet transform (WPT), etc. The time-frequency feature data includes a spectrum diagram, a time-frequency diagram, transient feature parameters, etc. Based on the frequency band transient time-frequency feature data, the frequency band component distribution data is analyzed for abnormal time-frequency changes. By comparing the time-frequency feature data at the current moment with the reference data under normal conditions, various abnormality detection and time-frequency analysis methods are used, such as difference analysis, statistical analysis, machine learning algorithms, etc. To identify abnormal time-frequency changes in the frequency band, according to the results of the analysis of abnormal time-frequency changes in the frequency band, the abnormal transient characteristic data of the frequency band is obtained, including abnormal spectrum characteristics, abnormal time-frequency distribution, abnormal transient characteristic parameters, etc. By extracting and sorting the abnormal transient characteristic data, the abnormal situation of the frequency band is accurately described. Based on the abnormal transient characteristic data of the frequency band, an abnormal operating status alarm signal is generated. According to system requirements, the alarm signal is displayed in a visual form and sent to the monitoring system or the corresponding alarm device is triggered. The alarm signal helps operation and maintenance personnel to promptly discover and handle abnormal operating conditions to ensure the safety and performance of the photovoltaic system.

[0032] Step S6: Obtain a regional heat dynamic adjustment strategy based on the abnormal temperature alarm signal; generate an abnormal time-frequency fault diagnosis strategy based on the abnormal operating state alarm signal; construct a neural network model for the abnormal time-frequency fault diagnosis strategy and the regional heat dynamic adjustment strategy, and construct a photovoltaic system intelligent monitoring and diagnosis model to perform monitoring and alarm operations.

[0033] In this embodiment, according to the abnormal temperature alarm signal, the thermal anomaly problem existing in the photovoltaic component or photovoltaic system is analyzed, which involves real-time monitoring and analysis of temperature sensor data. According to the location and distribution of the abnormal temperature, a regional thermal dynamic adjustment strategy is formulated, such as adjusting the spacing between components, optimizing the heat dissipation system, etc., to reduce heat accumulation and improve system efficiency. According to the abnormal operation state alarm signal, the fault or abnormal situation existing in the photovoltaic system is analyzed, which involves comprehensive analysis of electrical characteristic parameters, frequency band component distribution data and frequency band abnormal transient characteristic data. According to the type and mode of the abnormal alarm, an abnormal time-frequency fault diagnosis strategy is formulated, such as identifying specific fault modes, inferring fault causes, etc., to quickly and accurately diagnose and locate the fault. Based on the abnormal time-frequency fault diagnosis strategy and regional thermal The intelligent monitoring and diagnosis model of the photovoltaic system is constructed by adopting machine learning and deep learning technologies, such as neural networks, to train the model to automatically learn and identify abnormal modes and fault modes. By using abnormal time-frequency fault diagnosis strategies and regional heat dynamic adjustment strategies as the input and reference of the model, an end-to-end intelligent monitoring and diagnosis model is established. The constructed intelligent monitoring and diagnosis model of the photovoltaic system is deployed to the actual monitoring system, and the monitoring alarm operation is performed. The model receives real-time data streams, including electrical characteristic parameters of photovoltaic modules, temperature data, frequency band component distribution data, etc. Through the reasoning and judgment of the model, the system status is monitored in real time, faults are diagnosed, and corresponding alarm signals are generated. These alarm signals are used to alert operation and maintenance personnel, trigger maintenance operations or notify relevant parties.

[0034] In this embodiment, refer to Figure 2 , is a flowchart of detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include: Step S11: performing a full-range thermal imaging scan on the photovoltaic module array to generate a multi-directional thermal imaging view of the photovoltaic module; Step S12: performing component position identification on the multi-directional thermal imaging view of the photovoltaic component to obtain the photovoltaic component node; Step S13: Calculating the temperature value of the component node to obtain the temperature value of the photovoltaic component node; Step S14: performing temperature distribution analysis on the multi-directional thermal imaging views of the photovoltaic components based on the temperature values ​​of the photovoltaic component nodes to obtain temperature distribution data of the photovoltaic components. In this embodiment, according to actual conditions, scanning parameters of the thermal imaging device are set, such as scanning speed, scanning mode and scanning range. The scanning range should cover the entire photovoltaic module array to obtain all-round thermal imaging data. The photovoltaic module array is scanned in all directions using the thermal imaging device. Each photovoltaic module is scanned at an appropriate distance and angle to obtain its thermal imaging image. The stability of the device is ensured during the scanning process to obtain a clear thermal imaging image. The thermal imaging image of each photovoltaic module is recorded and sorted. Each thermal imaging image should contain the thermal distribution information of the photovoltaic module and its related position information. Computer vision technology, such as image segmentation and target detection algorithms, is used to analyze the pre-processed thermal imaging view to identify the position of the photovoltaic module. The goal is to accurately locate and mark each photovoltaic module in the image. Photovoltaic module nodes are generated based on the identified photovoltaic module positions. Each photovoltaic module node represents a photovoltaic module, including its position information and other related attributes. The thermal imaging is performed according to the characteristics of the thermal imaging device and environmental conditions. The image is temperature calibrated, and the actual temperature value corresponding to each pixel value is determined by referring to a standard temperature source or an ambient temperature sensor. The temperature of the thermal imaging image area corresponding to each photovoltaic module node is calculated, and the temperature value of the photovoltaic module node is obtained by extracting the pixel value in the area and converting it into an actual temperature value. The calculated photovoltaic module node temperature value is recorded and sorted. Each photovoltaic module node should have a corresponding temperature value. Based on the photovoltaic module node temperature value, it is visualized on a multi-directional thermal imaging view. The temperature values ​​of different temperature ranges are mapped to different colors using color mapping and other methods to display the temperature distribution of the photovoltaic module. The visualized multi-directional thermal imaging view of the photovoltaic module is analyzed to obtain the temperature distribution data of the photovoltaic module. By observing the shape, area and change trend of the temperature distribution, the abnormal or uneven temperature distribution is identified. The analyzed photovoltaic module temperature distribution data is recorded and sorted. These data include the temperature range, average temperature, maximum temperature, temperature gradient and other information of each photovoltaic module.

[0035] In this embodiment, refer to Figure 3 , is a flowchart of detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include: Step S21: Performing a spatial structure analysis on the photovoltaic module multi-directional thermal imaging view to obtain spatial structure data of the photovoltaic module; Step S22: performing three-dimensional feature point sampling processing on the multi-directional thermal imaging view of the photovoltaic module to obtain multi-directional feature points; Step S23: performing stereo matching on the multi-directional feature points to obtain spatially corresponding feature points; Step S24: reconstructing the three-dimensional point cloud of the photovoltaic module spatial structure data based on the spatial corresponding feature points to construct a three-dimensional point cloud structure model of the photovoltaic module; Step S25: Performing dynamic temperature interpolation processing on the three-dimensional point cloud structure model of the photovoltaic module based on the temperature distribution data of the photovoltaic module to construct a three-dimensional temperature distribution field model.

[0036] In this embodiment, the multi-directional thermal imaging views of the photovoltaic components are aligned to ensure that images from different perspectives can correspond to the same position in space. This is achieved by feature point matching, image registration, or camera posture estimation. The features of the photovoltaic components are extracted from the aligned multi-directional thermal imaging views. The features include edges, corners, textures, and other features that are stable at different perspectives. The extracted features are used to perform spatial structure analysis to determine the spatial position, relative posture, and size of the photovoltaic components. The three-dimensional structure of the photovoltaic components is inferred by triangulation, stereo geometry relationship calculation, and other methods. The analyzed spatial structure data of the photovoltaic components are recorded and organized. These data include position coordinates, posture parameters, size information, and the like of the photovoltaic components. Three-dimensional feature points are extracted from the multi-directional thermal imaging views of the photovoltaic components. Feature extraction algorithms, such as SIFT (Scale Invariant Feature Transform) or SURF (Speeded Up Robust Features), are used to detect and extract feature points with good stability and distinctiveness. The feature points extracted at different perspectives are matched to establish a corresponding relationship between them. A feature descriptor matching algorithm, such as a matching method based on a feature descriptor, is used. Methods (such as nearest neighbor matching based on distance measurement) are used to achieve this, and the three-dimensional coordinates of the matched feature points are sampled. Through triangulation or stereo geometry calculation, the corresponding relationship between feature points under multiple perspectives is used to estimate the three-dimensional coordinates of the feature points. A suitable stereo matching algorithm is selected, such as disparity calculation, stereo matching algorithm, etc. Commonly used algorithms include region-based stereo matching algorithm, feature-based stereo matching algorithm, etc. The selected stereo matching algorithm is used to stereo match the multi-directional feature points to determine their corresponding relationship in three-dimensional space. The goal of stereo matching is to find the corresponding points of each feature point under the left and right perspectives, and the matched spatial corresponding feature points are recorded and sorted. Each feature point should contain its two-dimensional pixel coordinates under the left and right perspectives and the corresponding three-dimensional spatial coordinates. According to the three-dimensional coordinate information of the spatial corresponding feature points, it is reconstructed into a three-dimensional point cloud structure model of the photovoltaic module. The point cloud structure model is a model composed of a large number of discrete three-dimensional points that represents the surface shape of an object. The reconstructed three-dimensional point cloud structure model of the photovoltaic module is recorded and sorted, and the point cloud structure model is saved in a common point cloud data format, such as PLY (Polygon File Format) or OBJ (Wavefront OBJ).

[0037] In this embodiment, the specific steps of step S25 are: Calculate pixel coordinates of multi-directional thermal imaging views of photovoltaic modules based on photovoltaic module nodes to obtain two-dimensional pixel coordinates; Calculate the three-dimensional projection coordinates of the three-dimensional point cloud structure model of the photovoltaic module based on the two-dimensional pixel point coordinates to obtain the three-dimensional projection point coordinates; Based on the three-dimensional projection point coordinates, the temperature distribution data of the photovoltaic module is dynamically interpolated to the three-dimensional point cloud structure model of the photovoltaic module to obtain an initial three-dimensional temperature field model; Analyze the temperature difference between the structures of the photovoltaic module temperature distribution data to obtain the temperature difference change data of the module structure; According to the temperature difference change data of the component structure, the initial three-dimensional temperature field model is completed with optical flow details to construct a three-dimensional temperature distribution field model.

[0038] In this embodiment, pixel coordinates are calculated for each node in the multi-directional thermal imaging view of the photovoltaic component, and the pixel position of the node on the image is calculated by the pixel coordinate system and the image resolution. The calculated two-dimensional pixel coordinates are recorded and sorted. Each node should contain its pixel coordinates on the image. The two-dimensional pixel coordinates are converted into three-dimensional projection coordinates by using camera parameters and the image coordinate system. This is achieved by methods such as camera calibration parameters (such as camera intrinsic parameters, camera extrinsic parameters) and inverse perspective projection. A suitable dynamic temperature interpolation algorithm is selected, such as the Kriging interpolation algorithm, the radial basis function interpolation algorithm, etc. The selected dynamic temperature interpolation algorithm is used to perform temperature interpolation processing according to the projection point coordinates and the corresponding temperature data on the three-dimensional point cloud structure model of the photovoltaic component to obtain an initial three-dimensional temperature field model. The initial three-dimensional interpolated temperature field model is converted into the three-dimensional projection coordinates. The temperature field model is recorded and organized, and the initial three-dimensional temperature field model is saved using common three-dimensional data representation formats, such as grid data formats (such as VTK, OBJ, etc.) or voxel data formats (such as Voxelize, HDF5, etc.). According to the structural design and temperature distribution data of the photovoltaic module, the temperature difference between different structural parts is calculated, and this is achieved by comparing the temperature values ​​of different nodes or regions. The temperature difference changes between different nodes or regions are analyzed, including the size, distribution and change trend of the temperature difference, etc., and a suitable optical flow detail completion algorithm is selected, such as an interpolation algorithm based on the optical flow method, an interpolation algorithm based on the gradient, etc. The selected optical flow detail completion algorithm is used to complete the details of the initial three-dimensional temperature field model according to the temperature difference change data of the module structure, and this is achieved by adjusting the temperature value or gradient in the initial temperature field model according to the temperature difference change data.

[0039] In this embodiment, refer to Figure 4 , is a flowchart of detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include: Step S31: performing abnormal temperature analysis on the three-dimensional temperature distribution field model based on a preset temperature threshold to obtain an abnormal temperature area; Step S32: performing temperature gradient diffusion processing on the temperature abnormal area, thereby generating a local abnormal hot spot area; Step S33: performing regional morphological analysis on the local abnormal hotspot area to obtain morphological characteristics of the abnormal hotspot area; Step S34: performing interference edge segmentation on the local abnormal hot spot area based on the morphological characteristics of the abnormal hot spot area to obtain a component hot spot area; Step S35: accurately positioning the hot spot area of ​​the photovoltaic component three-dimensional point cloud structure model based on the component hot spot area, thereby generating abnormal hot spot component position data; Step S36: Perform abnormal temperature alarm processing based on the abnormal hot spot component position data to generate an abnormal temperature alarm signal.

[0040] In this embodiment, according to the needs and application scenarios, a temperature threshold is preset to determine whether there is an abnormal temperature area. The threshold is determined according to the design specifications, working conditions and safety standards of the photovoltaic module. In the three-dimensional temperature distribution field model, the temperature value of each node or area is compared to determine whether it exceeds the preset temperature threshold. If the temperature value exceeds the threshold, the node or area is marked as an abnormal temperature area. In the abnormal temperature area, the temperature gradient of each node or area is calculated. The temperature gradient is obtained by calculating the derivative or difference of the temperature field. According to the size and direction of the temperature gradient, the abnormal temperature area is subjected to temperature gradient diffusion processing. The diffusion is achieved by propagating the heat of the high temperature area to the surrounding low temperature area. Regional morphological methods, such as corrosion, expansion, opening operation, closing operation, etc., are applied to process the local abnormal hot spot area. These methods change the shape, size and connectivity of the area. According to the processed local abnormal hot spot area, its morphological features are extracted, such as area, perimeter, aspect ratio, etc. These features are used to describe the shape and size of the abnormal hot spot area. According to the morphological features of the abnormal hot spot area, an appropriate image processing algorithm or segmentation method is applied to perform interference edge segmentation on the local abnormal hot spot area, which is achieved by threshold segmentation, edge detection or regional growth, etc., and regional positioning is performed in the three-dimensional point cloud structure model of the photovoltaic component. Point cloud processing algorithms, such as nearest neighbor search, point cloud registration or projection, are used to correspond the component hot spot area to the point cloud model. According to safety standards and application requirements, the conditions for triggering abnormal temperature alarms are set, for example, the temperature is set to exceed a certain threshold or the number of abnormal hot spots exceeds a certain limit, etc. According to the abnormal hot spot component position data, it is judged whether the conditions for abnormal temperature alarms are met. If the conditions are met, an abnormal temperature alarm signal is generated.

[0041] In this embodiment, the specific steps of step S34 are: Identify edge feature points in local abnormal hotspot areas to extract edge feature points; Based on the morphological characteristics of the abnormal hot spot area, outlier edge point analysis is performed on the edge feature points to mark the outlier edge feature points; Based on the outlier edge feature points, the interference edge segmentation of the local abnormal hot spot area is performed to obtain a smooth hot spot contour line; The hot spot area of ​​the component is obtained by reconstructing the boundary of the local abnormal hot spot area through smoothing the hot spot contour.

[0042] In this embodiment, an appropriate edge detection algorithm, such as the Canny edge detection algorithm, is used to perform edge detection on the local abnormal hot spot area. The algorithm can identify edge information in the image. Edge feature points are extracted from the edge image. A corner detection algorithm (such as Harris corner detection) or other feature point extraction methods are used to identify significant edge feature points. According to the morphological features of the abnormal hot spot area, such as area, perimeter, aspect ratio, etc., morphological features are extracted from the edge feature points. The morphological features are analyzed and edge feature points with outlier features are marked. Outlier edge feature points usually represent abnormal edges in abnormal or abnormal hot spot areas. The local abnormal hot spot area is subjected to interference edge segmentation using the position information of the outlier edge feature points and the edge segmentation algorithm. This is achieved by analyzing the spatial relationship of the edge feature points, using a regional growing algorithm or a gradient-based segmentation method. Based on the edge information obtained by the interference edge segmentation, a smooth hot spot contour line is generated. Using techniques such as curve fitting or smoothing filtering, the edges are optimized and smoothed to obtain a more accurate hot spot contour line. The smooth hot spot contour line is subjected to boundary break detection to find out the existing boundary break points or line segments. These breaks are caused by edge detection or interference edge segmentation. Based on the detected boundary break points or line segments, boundary break reconstruction is performed through interpolation, curve fitting or boundary repair algorithms. In this way, the discontinuity in the hot spot contour is repaired to obtain a complete component hot spot area.

[0043] In this embodiment, step S4 includes the following steps: Step S41: obtaining electrical characteristic parameters of photovoltaic modules based on sensors; Step S42: performing smoothing and filtering processing on the electrical characteristic parameters of the photovoltaic module to obtain filtered electrical characteristic parameters; Step S43: performing multi-scale wavelet decomposition on the filter electrical characteristic parameters to generate multiple time-frequency bands; Step S44: calculating the time-frequency components of each frequency subband for each of the multiple time-frequency frequency bands to obtain the energy stability of each frequency subband; Step S45: performing frequency component distribution analysis on multiple time-frequency frequency bands based on the energy stationarity of each decomposition unit, thereby obtaining frequency band component distribution data.

[0044] In this embodiment, the photovoltaic module is monitored to obtain the current, voltage or other relevant electrical characteristic parameters of the photovoltaic module. These sensors should be able to accurately measure the electrical output of the photovoltaic module. An appropriate filtering algorithm is selected, such as moving average filtering, median filtering or low-pass filtering, to filter the electrical characteristic parameters of the photovoltaic module. The purpose of filtering is to remove noise or instability and obtain smooth electrical characteristic parameters. Wavelet analysis methods, such as discrete wavelet transform (DWT) or continuous wavelet transform (CWT), are used to perform multi-scale decomposition on the filtered electrical characteristic parameters, which will generate multiple time-frequency bands, each frequency band representing signal components of different scales and frequencies, and the frequency sub-bands within each time-frequency band are decomposed. Time-frequency analysis, such as using short-time Fourier transform (STFT) or wavelet packet analysis, calculates the energy of each frequency sub-band. The energy represents the strength or amplitude of the signal in the sub-band. Perform statistical analysis on the energy of each frequency sub-band, compare the energy changes in different time periods, and evaluate the energy stability of each frequency sub-band by calculating the mean, variance or other statistical indicators of the energy. According to the energy stability of each frequency sub-band, multiple time-frequency frequency bands are analyzed for frequency component distribution. This is achieved by clustering, segmenting or thresholding the energy stability, dividing the frequency bands into different components or regions. Each component or region represents the characteristics of a frequency component, and the components are grouped according to factors such as the energy stability and the frequency range.

[0045] In this embodiment, the specific steps of step S5 are: Step S51: extracting transient time-frequency features from multiple time-frequency bands to generate frequency band transient time-frequency feature data; Step S52: performing frequency transient fluctuation analysis on the frequency band component distribution data based on the frequency band transient time-frequency characteristic data to obtain frequency band transient change data; Step S53: performing abnormal time-frequency change analysis on the frequency band transient change data to obtain abnormal transient characteristic data of the frequency band; Step S54: Perform component abnormal operation state alarm processing based on the frequency band abnormal transient characteristic data to generate an abnormal operation state alarm signal.

[0046] In this embodiment, for each time-frequency frequency band, an appropriate time-frequency analysis method (such as short-time Fourier transform) is used to extract transient time-frequency features, which include transient frequency, transient amplitude, transient phase, etc. In each time-frequency frequency band, the time-frequency transient features of the signal are calculated and organized into frequency band transient time-frequency feature data, which will be used for subsequent frequency band transient change analysis. The frequency band transient time-frequency feature data is used to perform frequency transient fluctuation analysis on the frequency band component distribution data, including calculating the transient frequency change rate, the transient amplitude change rate, etc. By comparing the frequency band transient time-frequency feature data in different time periods, the frequency band transient time-frequency feature data is evaluated. The transient changes of the frequency band will provide a basis for the subsequent abnormal time-frequency change analysis. The abnormal time-frequency change analysis is performed on the frequency band transient change data. Statistical methods, pattern recognition methods or machine learning methods are used to detect and identify abnormal time-frequency change patterns. According to the results of the abnormal time-frequency change analysis, the abnormal transient feature data of the frequency band are extracted. These feature data reflect the abnormal situation of the frequency band in terms of transient changes. Based on the abnormal transient feature data of the frequency band, appropriate anomaly detection algorithms or rules are designed to judge the abnormal operating status of the component. When the anomaly detection algorithm or rule is triggered, an abnormal operating status alarm signal is generated.

[0047] In this embodiment, the specific steps of step S6 are: Step S61: dynamically adjusting the power supply of the hot spot component based on the abnormal temperature alarm signal to obtain a regional heat dynamic adjustment strategy; Step S62: identifying potential fault nodes based on the abnormal operation state alarm signal to obtain potential fault nodes of the photovoltaic components; Step S63: making a fault diagnosis decision for a potential fault node of a photovoltaic module to generate an abnormal time-frequency fault diagnosis strategy; Step S64: construct a neural network model for the abnormal time-frequency fault diagnosis strategy and the regional heat dynamic adjustment strategy, and build a photovoltaic system intelligent monitoring and diagnosis model to perform monitoring and alarm operations. In this embodiment, the temperature of the photovoltaic module is monitored and a threshold value of abnormal temperature is set. When the temperature of the photovoltaic module exceeds the set threshold value, an abnormal temperature alarm signal is triggered. Based on the abnormal temperature alarm signal, a hot spot module power supply dynamic adjustment strategy is adopted. The temperature of the hot spot area is adjusted by controlling the power supply of the photovoltaic module to reduce the impact of the hot spot on the performance of the photovoltaic module. Different adjustment strategies are adopted according to specific circumstances, such as reducing the power supply in the hot spot area, adjusting the layout or spacing between photovoltaic modules, etc. This will help achieve regional dynamic heat adjustment to improve the efficiency and life of the photovoltaic module. The operating status of the photovoltaic module is monitored, and a rule or algorithm for the abnormal operating status is set. When the operating status of the photovoltaic module is abnormal, an abnormal operating status alarm signal is triggered. Based on the abnormal operating status alarm signal, potential fault nodes are identified. The faulty photovoltaic module is determined by analyzing the characteristics, correlation and timing information in the alarm signal. According to the identification result of the potential fault node, the potential fault node of the photovoltaic module, that is, the location of the faulty module, is obtained. Or numbering, based on the location of the potential fault node of the PV module and other related information, make fault diagnosis decisions, including using rules, models or algorithms to determine the specific fault type of the potential fault node, and generate abnormal time-frequency fault diagnosis strategies based on the results of the fault diagnosis decisions, which will provide fault diagnosis suggestions or guidance for the potential fault nodes, helping operators or maintenance personnel to take corresponding repair measures, and based on the abnormal time-frequency fault diagnosis strategies and regional heat dynamic adjustment strategies, collect relevant data and features, including abnormal alarm data, temperature data, power data, etc., use neural network model building technology to input data and features into the model, and perform model training and optimization, which will help build an intelligent monitoring and diagnosis model for the PV system to realize automated monitoring and alarm operations. The built intelligent monitoring and diagnosis model analyzes and predicts real-time data, identifies abnormal situations, and generates corresponding monitoring alarms. The model makes predictions and decisions based on the input abnormal time-frequency fault diagnosis strategies and regional heat dynamic adjustment strategies to perform effective monitoring and alarm operations.

[0048] In this embodiment, a monitoring and alarming system for a photovoltaic system is provided, which is used to execute the monitoring and alarming method for a photovoltaic system as described above, including: The temperature distribution module performs a full range of thermal imaging scans on the photovoltaic module array to generate a multi-directional thermal imaging view of the photovoltaic module; performs temperature distribution analysis on the multi-directional thermal imaging view of the photovoltaic module to obtain the temperature distribution data of the photovoltaic module; The three-dimensional structure module obtains the spatial structure data of the photovoltaic module based on the multi-directional thermal imaging view of the photovoltaic module; reconstructs the three-dimensional point cloud of the photovoltaic module spatial structure data to build a three-dimensional point cloud structure model of the photovoltaic module; and builds a three-dimensional temperature distribution field model based on the temperature distribution data of the photovoltaic module; The hot spot area module performs abnormal temperature analysis on the three-dimensional temperature distribution field model based on the preset temperature threshold, thereby generating local abnormal hot spot areas; based on the local abnormal hot spot areas, the three-dimensional point cloud structure model of the photovoltaic module is accurately positioned in the hot spot area, thereby generating abnormal hot spot module location data; based on the abnormal hot spot module location data, an abnormal temperature alarm signal is generated; The time-frequency decomposition module obtains the electrical characteristic parameters of the photovoltaic module based on the sensor; performs multi-scale wavelet decomposition on the electrical characteristic parameters of the photovoltaic module to generate multiple time-frequency bands; performs frequency component distribution analysis on multiple time-frequency bands to obtain frequency band component distribution data; The abnormal time-frequency module generates frequency band transient time-frequency feature data based on multiple time-frequency bands; performs abnormal time-frequency change analysis on the frequency band component distribution data based on the frequency band transient time-frequency feature data to obtain the frequency band abnormal transient feature data; generates an abnormal operation status alarm signal based on the frequency band abnormal transient feature data; The monitoring and diagnosis model module obtains the regional heat dynamic adjustment strategy based on the abnormal temperature alarm signal; generates the abnormal time-frequency fault diagnosis strategy based on the abnormal operation status alarm signal; constructs a neural network model for the abnormal time-frequency fault diagnosis strategy and the regional heat dynamic adjustment strategy, and constructs an intelligent monitoring and diagnosis model for the photovoltaic system to perform monitoring and alarm operations.

[0049] The present invention obtains thermal imaging views of multiple angles of the photovoltaic module array through omni-directional thermal imaging scanning, providing more comprehensive temperature information. The temperature distribution analysis performs detailed analysis and evaluation on the temperature conditions of the photovoltaic modules. The generated temperature distribution data of the photovoltaic modules is used as input for subsequent steps for further analysis and decision-making. The three-dimensional point cloud reconstruction obtains the spatial structure information of the photovoltaic modules according to the multi-directional thermal imaging views of the photovoltaic modules, realizes the restoration of the shape of the photovoltaic modules, and the constructed three-dimensional point cloud structure model of the photovoltaic modules provides accurate spatial information for subsequent analysis and positioning. The three-dimensional temperature distribution field model combines the temperature distribution data of the photovoltaic modules and comprehensively considers the spatial structure of the photovoltaic modules. The structure and temperature distribution of the photovoltaic module provide a basis for subsequent abnormal analysis. The abnormal temperature analysis detects the local abnormal hot spot area in the temperature distribution field model, that is, the area with potential faults or abnormalities. The hot spot area is accurately positioned by matching the abnormal hot spot area with the three-dimensional point cloud structure model of the photovoltaic module to determine the specific location of the abnormal hot spot component. The generated abnormal hot spot component location data is used for subsequent fault diagnosis and alarm processing, providing important information about the abnormal hot spot component. The time-frequency decomposition converts the electrical characteristic parameters of the photovoltaic module into time-frequency feature data in different frequency bands, extracts the frequency information of the electrical signal, and the multi-scale wavelet decomposition decomposes the electrical characteristic parameters at different scales. The frequency components in each time-frequency band are analyzed to obtain the time-frequency characteristic data in different frequency bands, providing more detailed frequency component information. The frequency component distribution analysis analyzes and counts the frequency components in each time-frequency band to obtain the frequency band component distribution data, providing a basis for subsequent abnormal time-frequency analysis. The frequency band transient time-frequency characteristic data captures the transient changes of the electrical characteristic parameters of the photovoltaic module in different frequency bands, providing richer time-frequency characteristic information. The abnormal time-frequency change analysis performs abnormal detection on the frequency band component distribution data, identifies the frequency band with abnormal time-frequency changes, further analyzes the cause and degree of the abnormality, and generates abnormal operation status alarm signals to remind operators in time. The personnel or system managers shall take corresponding maintenance or adjustment measures if there are abnormal photovoltaic modules. The regional heat dynamic adjustment strategy shall adjust the heat distribution of photovoltaic modules according to the abnormal temperature alarm signal, optimize the heat distribution and utilization efficiency of the photovoltaic system, and the abnormal time-frequency fault diagnosis strategy shall diagnose the electrical characteristics of photovoltaic modules according to the abnormal operation status alarm signal, and determine the cause and type of the fault or abnormality. The constructed photovoltaic system intelligent monitoring and diagnosis model comprehensively considers the abnormal temperature alarm and abnormal time-frequency fault diagnosis strategy, uses neural network and other technologies to build and analyze the model, realizes automated monitoring and alarm functions, and improves the stability and efficiency of the photovoltaic system.

[0050] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0051] The above is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A photovoltaic system monitoring and alarm method, characterized in that: The following steps are involved: Step S1: performing a full-range thermal imaging scan on the photovoltaic module array to generate a multi-directional thermal imaging view of the photovoltaic module; performing a temperature distribution analysis on the multi-directional thermal imaging view of the photovoltaic module to obtain temperature distribution data of the photovoltaic module; Step S2: obtaining the spatial structure data of the photovoltaic module based on the multi-directional thermal imaging view of the photovoltaic module; reconstructing the spatial structure data of the photovoltaic module by three-dimensional point cloud to construct a three-dimensional point cloud structure model of the photovoltaic module; constructing a three-dimensional temperature distribution field model based on the temperature distribution data of the photovoltaic module; and constructing a three-dimensional point cloud structure model of the module; Step S3: performing abnormal temperature analysis on the three-dimensional temperature distribution field model based on a preset temperature threshold, thereby generating a local abnormal hot spot area; Based on the local abnormal hot spot area, the hot spot area of ​​the three-dimensional point cloud structure model of the photovoltaic module is accurately located, thereby generating abnormal hot spot component location data; based on the abnormal hot spot component location data, an abnormal temperature alarm signal is generated; Step S4: acquiring electrical characteristic parameters of the photovoltaic module based on the sensor; performing multi-scale wavelet decomposition on the electrical characteristic parameters of the photovoltaic module to generate multiple time-frequency bands; performing frequency component distribution analysis on the multiple time-frequency bands to obtain frequency band component distribution data; Step S5: generating frequency band transient time-frequency feature data based on multiple time-frequency frequency bands; performing abnormal time-frequency change analysis on frequency band component distribution data based on the frequency band transient time-frequency feature data to obtain frequency band abnormal transient feature data; generating an abnormal operation state alarm signal based on the frequency band abnormal transient feature data; Step S6: Obtain a regional heat dynamic adjustment strategy based on the abnormal temperature alarm signal; generate an abnormal time-frequency fault diagnosis strategy based on the abnormal operating state alarm signal; construct a neural network model for the abnormal time-frequency fault diagnosis strategy and the regional heat dynamic adjustment strategy, and construct a photovoltaic system intelligent monitoring and diagnosis model to perform monitoring and alarm operations.

2. The photovoltaic system monitoring and alarm method according to claim 1, characterized in that: The specific steps of step S1 are: Step S11: performing a full-range thermal imaging scan on the photovoltaic module array to generate a multi-directional thermal imaging view of the photovoltaic module; Step S12: performing component position identification on the multi-directional thermal imaging view of the photovoltaic component to obtain the photovoltaic component node; Step S13: Calculating the temperature value of the component node to obtain the temperature value of the photovoltaic component node; Step S14: performing temperature distribution analysis on the multi-directional thermal imaging views of the photovoltaic components based on the temperature values ​​of the photovoltaic component nodes to obtain temperature distribution data of the photovoltaic components.

3. The photovoltaic system monitoring and alarm method according to claim 1, characterized in that: The specific steps of step S2 are: Step S21: Performing a spatial structure analysis on the photovoltaic module multi-directional thermal imaging view to obtain the spatial structure data of the photovoltaic module; Step S22: performing three-dimensional feature point sampling processing on the multi-directional thermal imaging view of the photovoltaic module to obtain multi-directional feature points; Step S23: performing stereo matching on the multi-directional feature points to obtain spatially corresponding feature points; Step S24: reconstructing the three-dimensional point cloud of the photovoltaic module spatial structure data based on the spatial corresponding feature points to construct a three-dimensional point cloud structure model of the photovoltaic module; Step S25: Performing dynamic temperature interpolation processing on the three-dimensional point cloud structure model of the photovoltaic module based on the temperature distribution data of the photovoltaic module to construct a three-dimensional temperature distribution field model.

4. The photovoltaic system monitoring and alarm method according to claim 3, characterized in that: The specific steps of step S25 are: Calculate pixel coordinates of multi-directional thermal imaging views of photovoltaic modules based on photovoltaic module nodes to obtain two-dimensional pixel coordinates; Calculate the three-dimensional projection coordinates of the three-dimensional point cloud structure model of the photovoltaic module based on the two-dimensional pixel point coordinates to obtain the three-dimensional projection point coordinates; Based on the three-dimensional projection point coordinates, the temperature distribution data of the photovoltaic module is dynamically interpolated to the three-dimensional point cloud structure model of the photovoltaic module to obtain an initial three-dimensional temperature field model; Analyze the temperature difference between the structures of the photovoltaic module temperature distribution data to obtain the temperature difference change data of the module structure; According to the temperature difference change data of the component structure, the initial three-dimensional temperature field model is completed with optical flow details to construct a three-dimensional temperature distribution field model.

5. The photovoltaic system monitoring and alarm method according to claim 1, characterized in that: The specific steps of step S3 are: Step S31: performing abnormal temperature analysis on the three-dimensional temperature distribution field model based on a preset temperature threshold to obtain an abnormal temperature area; Step S32: performing temperature gradient diffusion processing on the temperature abnormal area, thereby generating a local abnormal hot spot area; Step S33: performing regional morphological analysis on the local abnormal hotspot area to obtain morphological characteristics of the abnormal hotspot area; Step S34: performing interference edge segmentation on the local abnormal hot spot area based on the morphological characteristics of the abnormal hot spot area to obtain a component hot spot area; Step S35: accurately positioning the hot spot area of ​​the photovoltaic component three-dimensional point cloud structure model based on the component hot spot area, thereby generating abnormal hot spot component position data; Step S36: Perform abnormal temperature alarm processing based on the abnormal hot spot component position data to generate an abnormal temperature alarm signal.

6. The photovoltaic system monitoring and alarm method according to claim 1, characterized in that: The specific steps of step S34 are: Identify edge feature points in local abnormal hotspot areas to extract edge feature points; Based on the morphological characteristics of the abnormal hot spot area, outlier edge point analysis is performed on the edge feature points to mark the outlier edge feature points; Based on the outlier edge feature points, the interference edge segmentation of the local abnormal hot spot area is performed to obtain a smooth hot spot contour line; The hot spot area of ​​the component is obtained by reconstructing the boundary of the local abnormal hot spot area through smoothing the hot spot contour.

7. The photovoltaic system monitoring and alarm method according to claim 1, characterized in that: The specific steps of step S4 are: Step S41: obtaining electrical characteristic parameters of photovoltaic modules based on sensors; Step S42: performing smoothing and filtering processing on the electrical characteristic parameters of the photovoltaic module to obtain filtered electrical characteristic parameters; Step S43: performing multi-scale wavelet decomposition on the filter electrical characteristic parameters to generate multiple time-frequency bands; Step S44: calculating the time-frequency components of each frequency subband for each of the multiple time-frequency frequency bands to obtain the energy stability of each frequency subband; Step S45: performing frequency component distribution analysis on multiple time-frequency frequency bands based on the energy stationarity of each decomposition unit, thereby obtaining frequency band component distribution data.

8. The photovoltaic system monitoring and alarm method according to claim 1, characterized in that: The specific steps of step S5 are: Step S51: extracting transient time-frequency features from multiple time-frequency bands to generate frequency band transient time-frequency feature data; Step S52: performing frequency transient fluctuation analysis on the frequency band component distribution data based on the frequency band transient time-frequency characteristic data to obtain frequency band transient change data; Step S53: performing abnormal time-frequency change analysis on the frequency band transient change data to obtain abnormal transient characteristic data of the frequency band; Step S54: Perform component abnormal operation state alarm processing based on the frequency band abnormal transient characteristic data to generate an abnormal operation state alarm signal.

9. The photovoltaic system monitoring and alarm method according to claim 1, characterized in that: The specific steps of step S6 are: Step S61: dynamically adjusting the power supply of the hot spot component based on the abnormal temperature alarm signal to obtain a regional heat dynamic adjustment strategy; Step S62: identifying potential fault nodes based on the abnormal operation state alarm signal to obtain potential fault nodes of the photovoltaic components; Step S63: making a fault diagnosis decision for a potential fault node of a photovoltaic module to generate an abnormal time-frequency fault diagnosis strategy; Step S64: construct a neural network model for the abnormal time-frequency fault diagnosis strategy and the regional heat dynamic adjustment strategy, and build a photovoltaic system intelligent monitoring and diagnosis model to perform monitoring and alarm operations.

10. A photovoltaic system monitoring and alarm system, characterized in that: The method for performing the monitoring and alarming method of the photovoltaic system according to claim 1 comprises: The temperature distribution module performs a full range of thermal imaging scans on the photovoltaic module array to generate a multi-directional thermal imaging view of the photovoltaic module; performs temperature distribution analysis on the multi-directional thermal imaging view of the photovoltaic module to obtain the temperature distribution data of the photovoltaic module; The three-dimensional structure module obtains the spatial structure data of the photovoltaic module based on the multi-directional thermal imaging view of the photovoltaic module; reconstructs the three-dimensional point cloud of the photovoltaic module spatial structure data to build a three-dimensional point cloud structure model of the photovoltaic module; and builds a three-dimensional temperature distribution field model based on the temperature distribution data of the photovoltaic module; The hot spot area module performs abnormal temperature analysis on the three-dimensional temperature distribution field model based on the preset temperature threshold, thereby generating local abnormal hot spot areas; based on the local abnormal hot spot areas, the three-dimensional point cloud structure model of the photovoltaic module is accurately positioned in the hot spot area, thereby generating abnormal hot spot module location data; based on the abnormal hot spot module location data, an abnormal temperature alarm signal is generated; The time-frequency decomposition module obtains the electrical characteristic parameters of the photovoltaic module based on the sensor; performs multi-scale wavelet decomposition on the electrical characteristic parameters of the photovoltaic module to generate multiple time-frequency bands; performs frequency component distribution analysis on multiple time-frequency bands to obtain frequency band component distribution data; The abnormal time-frequency module generates frequency band transient time-frequency feature data based on multiple time-frequency bands; performs abnormal time-frequency change analysis on the frequency band component distribution data based on the frequency band transient time-frequency feature data to obtain the frequency band abnormal transient feature data; generates an abnormal operation status alarm signal based on the frequency band abnormal transient feature data; The monitoring and diagnosis model module obtains the regional heat dynamic adjustment strategy based on the abnormal temperature alarm signal; generates the abnormal time-frequency fault diagnosis strategy based on the abnormal operation status alarm signal; constructs a neural network model for the abnormal time-frequency fault diagnosis strategy and the regional heat dynamic adjustment strategy, and constructs an intelligent monitoring and diagnosis model for the photovoltaic system to perform monitoring and alarm operations.

Images