Method and system for predicting deformation of photovoltaic panel

Abstract

The invention provides a photovoltaic panel deformation prediction method and system. Stress distribution data in the photovoltaic panel are collected in real time through the first sensor module, and data on the surface of the photovoltaic panel are collected in real time through the second sensor module. Establishing a deformation monitoring reference surface of the photovoltaic panel through a level gauge module, outputting inclination angle data in real time, aligning the stress distribution data and the inclination angle data through a timestamp synchronization algorithm, and eliminating time sequence deviation through a dynamic time warping algorithm; and constructing a deformation prediction model for the alignment stress distribution data and the alignment inclination angle data, and triggering multi-stage early warning according to the relationship between the deformation quantity of the photovoltaic panel and a preset threshold. Multidirectional comprehensive and accurate monitoring is realized by collecting the internal stress data, the surface data and the inclination angle data of the photovoltaic panel; and corresponding measures can be taken in time according to a prediction result, and comprehensive and accurate deformation prediction of the photovoltaic panel in a complex environment is realized to reduce loss.

Description

Technical Field

[0001] The present invention relates to the field of photovoltaic technology, and particularly to a method and system for predicting the deformation of a photovoltaic panel. Background Art

[0002] With the continuous growth of the global demand for clean energy, photovoltaic power generation, as a sustainable energy solution, has been widely applied. The photovoltaic panel power generation method has the characteristics of no exhaustion risk, cleanliness, safety, and no noise. Among them, photovoltaic panels have a wide range of applications, are not restricted by the geographical distribution of resources, and also have the advantages of easy installation, short construction period, and reliable power supply system during use. The performance and stability of photovoltaic panels will directly affect the power generation efficiency and the reliability of the system. During actual operation, photovoltaic panels will be affected by various factors, such as temperature changes, uneven illumination, wind force, snow accumulation, etc. These factors will cause the deformation of photovoltaic panels.

[0003] Traditional prediction of photovoltaic panel deformation mostly relies on a single sensor. For example, strain gauges are used to monitor the deformation of photovoltaic panels in real time. However, strain gauges can only measure the local strain at the pasted position. For the overall deformation of a large area of photovoltaic panels, it is difficult to comprehensively and accurately describe the deformation state of the entire photovoltaic panel only by a small number of strain gauges. In addition, due to the influence of complex factors such as temperature, illumination, and wind force on photovoltaic panels, their deformation characteristics are extremely complex. Traditional prediction of photovoltaic panel deformation through a single sensor can only obtain local or single-dimensional information and cannot achieve comprehensive and accurate deformation prediction in a complex environment. Summary of the Invention

[0004] The present invention provides a method and system for predicting the deformation of a photovoltaic panel to solve the problem in the prior art that the prediction of photovoltaic panel deformation through a single sensor can only obtain local or single-dimensional information and cannot achieve comprehensive and accurate deformation prediction in a complex environment.

[0005] The present invention provides a method for predicting the deformation of a photovoltaic panel, including:

[0006] Collecting stress distribution data inside the photovoltaic panel in real time through a first sensor module;

[0007] Collecting data on the surface of the photovoltaic panel in real time through a second sensor module;

[0008] Establishing a deformation monitoring reference plane for the photovoltaic panel through a level module and outputting tilt angle data in real time, where the tilt angle data is the tilt angle value between the surface of the photovoltaic panel and the deformation monitoring reference plane;

[0009] Align the stress distribution data and the tilt angle data through a timestamp synchronization algorithm and eliminate the timing deviation through a dynamic time warping algorithm, wherein the stress distribution data correspondingly generates aligned stress distribution data, and the tilt angle data correspondingly generates aligned tilt angle data;

[0010] Construct a deformation prediction model for the aligned stress distribution data and the aligned tilt angle data, and trigger multi-level warnings according to the relationship between the deformation amount of the photovoltaic panel in the deformation prediction model and a preset threshold.

[0011] According to the photovoltaic panel deformation prediction method provided by the present invention, the real-time acquisition of the stress distribution data inside the photovoltaic panel by the first sensor module specifically includes: embedding a flexible fiber optic grating array into the internal encapsulation layer of the photovoltaic panel in a serpentine path, and collecting the stress distribution data inside the photovoltaic panel through the flexible fiber optic grating array.

[0012] According to the photovoltaic panel deformation prediction method provided by the present invention, after collecting the stress distribution data inside the photovoltaic panel, it includes: performing mean filtering on the stress distribution data.

[0013] According to the photovoltaic panel deformation prediction method provided by the present invention, the real-time acquisition of the surface data of the photovoltaic panel by the second sensor module specifically includes:

[0014] Obtain the surface data of the photovoltaic panel through the laser range finder in the second sensor module, and perform smoothing processing on the surface data of the photovoltaic panel through a filtering algorithm.

[0015] According to the photovoltaic panel deformation prediction method provided by the present invention, the establishment of the deformation monitoring reference plane of the photovoltaic panel by the level module and the real-time output of the tilt angle data specifically includes: the level module collects the tilt angle data of the photovoltaic panel before use at a set time interval, and calculates the average value of the tilt angle data, and the plane corresponding to the average value is the deformation monitoring reference plane.

[0016] According to the photovoltaic panel deformation prediction method provided by the present invention, the alignment of the stress distribution data and the tilt angle data through the timestamp synchronization algorithm includes:

[0017] S40: The first sensor module, the second sensor module, and the level module select a unified time reference source through the Network Time Protocol;

[0018] S41: The first sensor module, the second sensor module, and the level module record the timestamps of data collection when collecting data, and verify the timestamps corresponding to the first sensor module, the second sensor module, and the level module respectively;

[0019] S42: Determine whether the time errors among the first sensor module, the second sensor module, and the level module are within a set threshold range. If the time error exceeds the set threshold, return to S40.

[0020] According to the photovoltaic panel deformation prediction method provided by the present invention, constructing the deformation prediction model for the alignment stress distribution data and the alignment tilt angle data specifically includes: integrating the alignment stress distribution data and the alignment tilt angle data through a vector splicing algorithm to form a unified data set. The integration method of the vector splicing algorithm is: where X is the vector of the unified data set after integration, x1 is the vector composed of the alignment stress distribution data, and x2 is the vector composed of the alignment tilt angle data. Among them, the alignment stress distribution data has n1 dimensions, which is expressed as The alignment tilt angle data has n2 dimensions, which is expressed as

[0021] Train and evaluate the deformation prediction model through a long short-term memory network.

[0022] According to the photovoltaic panel deformation prediction method provided by the present invention, the multi-level early warning is a three-level early warning mechanism. When the deformation amount of the photovoltaic panel exceeds the first preset threshold, trigger a first-level early warning to send out a status abnormal signal; when the deformation amount of the photovoltaic panel exceeds the second preset threshold, trigger a second-level early warning for sound and light alarm; when the deformation amount of the photovoltaic panel exceeds the third preset threshold, trigger a third-level early warning to deactivate the photovoltaic panel.

[0023] According to the photovoltaic panel deformation prediction method provided by the present invention, the tilt angle value calculation formula is: Δθ = |θ measured - θ base |, where Δθ is the tilt angle value, θ measured is the actual tilt angle of the photovoltaic panel measured by the level module, and θ base is the tilt angle of the deformation monitoring reference plane of the level module.

[0024] The present invention also provides a photovoltaic panel deformation prediction system, including:

[0025] A first sensor module, which is used to collect the stress distribution data inside the photovoltaic panel in real time, and the first sensor module is embedded in the internal encapsulation layer of the photovoltaic panel;

[0026] A second sensor module, which is used to collect the data on the surface of the photovoltaic panel in real time, and the second sensor module is arranged on the surface of the photovoltaic panel;

[0027] A level module, which is used to establish a deformation monitoring reference plane of the photovoltaic panel and output tilt angle data in real time. Among them, the level module is communicatively connected to the second sensor module and obtains data on the surface of the photovoltaic panel. The tilt angle data is the tilt angle value between the surface of the photovoltaic panel and the deformation monitoring reference plane. The level module is arranged on the side of the photovoltaic panel;

[0028] A timestamp synchronization algorithm module. The first sensor module, the second sensor module and the level module are respectively connected to the timestamp synchronization algorithm module. Through the timestamp synchronization algorithm module, the stress distribution data and the tilt angle data are aligned and the timing deviation is eliminated through the dynamic time warping algorithm. Among them, the stress distribution data correspondingly generates aligned stress distribution data, and the tilt angle data correspondingly generates aligned tilt angle data;

[0029] A construction module, which is connected to the timestamp synchronization algorithm module. The construction module constructs a deformation prediction model for the aligned stress distribution data and the aligned tilt angle data, and triggers multi-level early warnings according to the relationship between the deformation amount of the photovoltaic panel and a preset threshold.

[0030] The present invention provides a method and system for predicting the deformation of a photovoltaic panel. By the first sensor module, the stress distribution data inside the photovoltaic panel is collected in real time. By the second sensor module, the data on the surface of the photovoltaic panel is collected in real time. By the level module, the deformation monitoring reference plane of the photovoltaic panel is established and the tilt angle data is output in real time. And by the timestamp synchronization algorithm, the stress distribution data and the tilt angle data are aligned and the timing deviation is eliminated through the dynamic time warping algorithm. Among them, the stress distribution data correspondingly generates aligned stress distribution data, and the tilt angle data correspondingly generates aligned tilt angle data. In addition, a deformation prediction model is constructed for the aligned stress distribution data and the aligned tilt angle data, and multi-level early warnings are triggered according to the relationship between the deformation amount of the photovoltaic panel in the deformation prediction model and a preset threshold. Among them, the tilt angle data is the tilt angle value between the surface of the photovoltaic panel and the deformation monitoring reference plane. The present invention realizes multi-faceted, comprehensive and accurate monitoring by collecting the internal stress data, surface data and tilt angle data of the photovoltaic panel, and can timely take countermeasures according to the prediction results, realizing comprehensive and accurate deformation prediction of the photovoltaic panel in a complex environment and reducing losses. Description of the Drawings

[0031] To more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0032] Figure 1 It is a flowchart of the method for predicting the deformation of a photovoltaic panel provided by an embodiment of the present invention;

[0033] Figure 2 It is a flowchart corresponding to S4 in the method for predicting the deformation of a photovoltaic panel provided by an embodiment of the present invention;

[0034] Figure 3 It is a schematic structural diagram of the system for predicting the deformation of a photovoltaic panel provided by an embodiment of the present invention. Detailed implementation manners

[0035] To make the objectives, technical solutions, and advantages of the present invention clearer, the following will clearly and completely describe the technical solutions in the present invention with reference to the drawings in the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0036] The present invention provides a method for predicting the deformation of a photovoltaic panel, including:

[0037] Real-time collecting stress distribution data inside the photovoltaic panel through a first sensor module;

[0038] Real-time collecting data on the surface of the photovoltaic panel through a second sensor module;

[0039] Establishing a deformation monitoring reference plane for the photovoltaic panel through a level module and real-time outputting tilt angle data, where the tilt angle data is the tilt angle value between the surface of the photovoltaic panel and the deformation monitoring reference plane;

[0040] Aligning the stress distribution data and the tilt angle data through a timestamp synchronization algorithm and eliminating the timing deviation through a dynamic time warping algorithm, where the stress distribution data correspondingly generates aligned stress distribution data, and the tilt angle data correspondingly generates aligned tilt angle data;

[0041] Constructing a deformation prediction model for the aligned stress distribution data and the aligned tilt angle data, and triggering multi-level early warnings according to the relationship between the deformation amount of the photovoltaic panel in the deformation prediction model and a preset threshold.

[0042] Among them, the real-time acquisition of the stress distribution data inside the photovoltaic panel by the first sensor module specifically includes: embedding a flexible fiber optic grating array into the internal encapsulation layer of the photovoltaic panel in a serpentine path, and acquiring the stress distribution data inside the photovoltaic panel through the flexible fiber optic grating array.

[0043] Among them, after acquiring the stress distribution data inside the photovoltaic panel, it includes: performing mean filtering on the stress distribution data.

[0044] Among them, the real-time acquisition of the surface data of the photovoltaic panel by the second sensor module specifically includes: obtaining the data on the surface of the photovoltaic panel through the laser range finder sensor in the second sensor module, and performing smoothing processing on the data on the surface of the photovoltaic panel through a filtering algorithm.

[0045] Among them, the establishment of the deformation monitoring reference plane of the photovoltaic panel by the level module and the real-time output of the tilt angle data specifically includes: the level module acquiring the tilt angle data of the photovoltaic panel before use according to a set time interval, and calculating the average value of the tilt angle data, and the plane corresponding to the average value is the deformation monitoring reference plane.

[0046] Among them, the alignment of the stress distribution data and the tilt angle data through the timestamp synchronization algorithm includes:

[0047] S40: The first sensor module, the second sensor module, and the level module select a unified time reference source through the Network Time Protocol;

[0048] S41: The first sensor module, the second sensor module, and the level module record the timestamps of data acquisition when acquiring data, and verify the timestamps corresponding to the first sensor module, the second sensor module, and the level module respectively;

[0049] S42: Determine whether the time errors between the first sensor module, the second sensor module, and the level module are within the set threshold range. If the time error exceeds the set threshold, return to S40.

[0050] Among them, the construction of the deformation prediction model for the aligned stress distribution data and the aligned tilt angle data specifically includes: integrating the aligned stress distribution data and the aligned tilt angle data through a vector splicing algorithm to form a unified data set, and the integration method of the vector splicing algorithm is: Among them, X is the vector of the unified data set after integration, x1 is the vector composed of the aligned stress distribution data, x2 is the vector composed of the aligned tilt angle data. Among them, if the aligned stress distribution data has n1 dimensions, it is expressed as If the alignment tilt angle data has n2 dimensions, it is expressed as

[0051] The deformation prediction model is trained and evaluated through a long short-term memory network.

[0052] Among them, the multi-level early warning is a three-level early warning mechanism. When the deformation amount of the photovoltaic panel exceeds the first preset threshold, a first-level early warning is triggered to send out a status anomaly signal; when the deformation amount of the photovoltaic panel exceeds the second preset threshold, a second-level early warning is triggered for audible and visual alarms; when the deformation amount of the photovoltaic panel exceeds the third preset threshold, a third-level early warning is triggered to deactivate the photovoltaic panel.

[0053] Among them, the calculation formula for the tilt angle value is: Δθ = |θ measured - θ base |, where Δθ is the tilt angle value, θ measured is the actual tilt angle of the photovoltaic panel measured by the level gauge module, and θ base is the tilt angle of the deformation monitoring reference plane of the level gauge module.

[0054] The present invention also provides a prediction system for the deformation of a photovoltaic panel, including:

[0055] A first sensor module, which is used to collect the stress distribution data inside the photovoltaic panel in real time, and the first sensor module is embedded in the internal encapsulation layer of the photovoltaic panel;

[0056] A second sensor module, which is used to collect the data on the surface of the photovoltaic panel in real time, and the second sensor module is arranged on the surface of the photovoltaic panel;

[0057] A level gauge module, which is used to establish a deformation monitoring reference plane of the photovoltaic panel and output the tilt angle data in real time. Among them, the level gauge module is communicatively connected to the second sensor module and obtains the data on the surface of the photovoltaic panel. The tilt angle data is the tilt angle value between the surface of the photovoltaic panel and the deformation monitoring reference plane, and the level gauge module is arranged on the side of the photovoltaic panel;

[0058] A timestamp synchronization algorithm module, the first sensor module, the second sensor module and the level gauge module are respectively connected to the timestamp synchronization algorithm module. The stress distribution data and the tilt angle data are aligned through the timestamp synchronization algorithm module, and the timing deviation is eliminated through the dynamic time warping algorithm. Among them, the aligned stress distribution data is correspondingly generated from the stress distribution data, and the aligned tilt angle data is correspondingly generated from the tilt angle data;

[0059] A construction module is connected to the timestamp synchronization algorithm module. The construction module constructs a deformation prediction model for the aligned stress distribution data and the aligned tilt angle data, and triggers multi-level warnings according to the relationship between the deformation amount of the photovoltaic panel and a preset threshold.

[0060] The present invention provides a method and system for predicting the deformation of a photovoltaic panel. The internal stress distribution data of the photovoltaic panel is collected in real time by a first sensor module, the surface data of the photovoltaic panel is collected in real time by a second sensor module, a deformation monitoring reference plane of the photovoltaic panel is established by a level module and the tilt angle data is output in real time, and the data of the first sensor module, the second sensor module, and the level module are aligned by a timestamp synchronization algorithm, and the timing deviation is eliminated by a dynamic time warping algorithm to construct a deformation prediction model. Multi-level warnings are triggered according to the relationship between the deformation amount of the photovoltaic panel and a preset threshold. Among them, the tilt angle data is the tilt angle value between the surface of the photovoltaic panel and the deformation monitoring reference plane. In addition, the first sensor module is embedded in the internal encapsulation layer of the photovoltaic panel, the second sensor module is arranged on the surface of the photovoltaic panel, and the level module is arranged on the side of the photovoltaic panel. The present invention realizes multi-faceted, comprehensive and accurate monitoring by collecting internal stress, surface data and tilt angle data, and can take countermeasures in time according to the prediction results, realizing comprehensive and accurate deformation prediction in a complex environment and reducing losses.

[0061] In this embodiment, please refer to Figure 1 , Figure 2 and Figure 3 , which are respectively the flowchart of the method for predicting the deformation of a photovoltaic panel provided by an embodiment of the present invention, the flowchart corresponding to S4 in the method for predicting the deformation of a photovoltaic panel provided by an embodiment of the present invention, and the structural schematic diagram of the prediction system 1 of the deformation of a photovoltaic panel provided by an embodiment of the present invention.

[0062] Specifically, in this embodiment, as Figure 1 , Figure 2 and Figure 3 shown, an embodiment of the present invention provides a method for predicting the deformation of a photovoltaic panel, including the following steps:

[0063] S1: The internal stress distribution data of the photovoltaic panel is collected in real time by the first sensor module 10, where the first sensor module 10 is embedded in the internal encapsulation layer of the photovoltaic panel;

[0064] S2: The surface data of the photovoltaic panel is collected in real time by the second sensor module 11, where the second sensor module 11 is arranged on the surface of the photovoltaic panel;

[0065] S3: Establish a deformation monitoring reference plane for the photovoltaic panel through the level meter module 12 and output tilt angle data in real time. Among them, the level meter module 12 is communicatively connected to the second sensor module 11 and obtains data on the surface of the photovoltaic panel. The tilt angle data is the tilt angle value between the surface of the photovoltaic panel and the deformation monitoring reference plane. The level meter module 12 is arranged on the side of the photovoltaic panel;

[0066] S4: Align the stress distribution data and the tilt angle data through a timestamp synchronization algorithm and eliminate the timing deviation through a dynamic time warping algorithm. Among them, the aligned stress distribution data is correspondingly generated from the stress distribution data, and the aligned tilt angle data is correspondingly generated from the tilt angle data;

[0067] S5: Construct a deformation prediction model for the aligned stress distribution data and the aligned tilt angle data, and trigger multi-level warnings according to the relationship between the deformation amount of the photovoltaic panel in the deformation prediction model and a preset threshold.

[0068] In this embodiment, S1 specifically includes: real-time collecting the stress distribution data inside the photovoltaic panel through a flexible fiber Bragg grating array. The first sensor module 10 includes a plurality of flexible fiber Bragg grating sensors, and the plurality of flexible fiber Bragg grating sensors constitute the flexible fiber Bragg grating array. The flexible fiber Bragg grating array is embedded in the internal encapsulation layer of the photovoltaic panel in a serpentine path. In this embodiment, the flexible fiber Bragg grating array arranged in a serpentine path can cover a larger range inside the photovoltaic panel, can comprehensively capture stress changes at various positions, accurately obtain the internal stress distribution data, and timely discover potential stress concentration points. In addition, the flexible fiber Bragg grating sensor has a fast response characteristic and can collect stress data in real time, providing timely information for predicting deformation. The flexible fiber Bragg grating array can also adapt to the shape and structure of the photovoltaic panel, and the serpentine embedding method further enhances the adhesion to the internal encapsulation layer of the photovoltaic panel, reduces the impact on the original structure of the photovoltaic panel, can effectively resist external environmental interference, and ensure the stability and reliability of stress data collection. In addition, in this embodiment, after collecting the stress distribution data inside the photovoltaic panel, it further includes: performing mean filtering on the stress distribution data to clean the data.

[0069] In this embodiment, the stress distribution data reflects the stress condition inside the photovoltaic panel. It consists of information in multiple dimensions, such as stress values at different positions and in different directions. Assuming that the stress distribution data has n1 dimensions, it can be expressed as These data are collected in real time through the flexible fiber Bragg grating array of the first sensor module 10 embedded in the internal encapsulation layer of the photovoltaic panel.

[0070] In this embodiment, S2 specifically includes: obtaining data on the surface of the photovoltaic panel through a number of laser rangefinder sensors, and smoothing the data on the surface of the photovoltaic panel through a filtering algorithm. Among them, a number of the laser rangefinder sensors are arranged in the second sensor module 11. In this embodiment, the laser rangefinder sensor emits a laser beam to the surface of the photovoltaic panel. After the laser beam hits the surface of the photovoltaic panel, it reflects back. The laser rangefinder sensor can accurately calculate the distance between the laser rangefinder sensor and the corresponding point on the surface of the photovoltaic panel by measuring the time difference between laser emission and reception and combining the propagation speed of the laser in the air. A number of the laser rangefinder sensors can measure the surface of the photovoltaic panel from different angles and positions, so as to obtain the distance data of different points on the surface of the photovoltaic panel. For example, measurements can be made at regular time intervals, and the surface data at different times can be recorded to form a time series of surface data for analyzing the change trend of the surface of the photovoltaic panel over time.

[0071] In this embodiment, S3 includes: real-time monitoring and outputting the tilt angle value of the photovoltaic panel through a triangular reference network. Among them, the level module 12 includes at least three electronic levels, and the three electronic levels form the triangular reference network. The triangular reference network covers multiple positions of the photovoltaic panel. This comprehensive monitoring information helps to accurately obtain the tilt angle value data.

[0072] In this embodiment, preferably, establishing the deformation monitoring reference plane of the photovoltaic panel specifically includes: in the state before the photovoltaic panel is installed and put into use, starting the level module 12 to collect data. The level module 12 is located on the side of the photovoltaic panel, and the tilt angle data is recorded at regular time intervals (such as every 5 minutes). The collection duration is set to 24 - 48 hours to cover the minute changes that may be caused by different time periods (such as daytime lighting, nighttime cooling, etc.); then, statistical analysis is performed on the large amount of collected tilt angle data, and the average value of these data is calculated. The plane corresponding to this average value is the deformation monitoring reference plane.

[0073] In this embodiment, the tilt angle value calculation formula is:

[0074] Δθ = |θ measured -θ base |, where Δθ is the tilt angle value, θ measured is the actual tilt angle of the photovoltaic panel obtained by the level module 12 according to the data on the surface of the photovoltaic panel, and θ base is the tilt angle of the deformation monitoring reference plane obtained by the level module 12. Through the tilt angle value, the deformation degree of the photovoltaic panel relative to the normal state can be accurately described, providing a key basis for accurately judging whether the photovoltaic panel is deformed and the severity of the deformation.

[0075] In this embodiment, S4 specifically includes the following steps:

[0076] S40: The first sensor module 10, the second sensor module 11, and the level module 12 select a unified time reference source through the Network Time Protocol (NTP). In this embodiment, specifically, NTP obtains accurate time from an atomic clock time source and distributes the time information to the first sensor module 10, the second sensor module 11, and the level module 12 for time calibration, thereby selecting a unified time reference source to ensure that their respective times are consistent with the standard time;

[0077] S41: The first sensor module 10, the second sensor module 11, and the level module 12 record the timestamps of data collection simultaneously when collecting data, and verify the timestamps corresponding to the first sensor module 10, the second sensor module 11, and the level module 12 respectively. In this embodiment, specifically, during the data collection stage, whenever the first sensor module 10 collects the internal stress distribution data of the photovoltaic panel, the second sensor module 11 collects the surface data of the photovoltaic panel, and the level module 12 collects the tilt angle data, the timestamps of data collection will be recorded simultaneously. The timestamps are accurate to the millisecond or even microsecond level and are used to mark the accurate time when the data is generated. After the data collection is completed, the timestamps corresponding to the first sensor module 10, the second sensor module 11, and the level module 12 are verified. The verification process includes checking whether the format of the timestamp is correct, whether the timestamp is within a reasonable time range, and whether the deviation between the timestamp and the current system time is within the allowable range, etc.; for example: set an allowable error range for the timestamp. If the timestamp of the first sensor module 10, the second sensor module 11, or the level module 12 exceeds this error range, it is determined that the timestamp is abnormal and the cause needs to be further investigated.

[0078] S42: Determine whether the time errors between the first sensor module 10, the second sensor module 11, and the level module 12 are within the set threshold range. If the time error exceeds the set threshold, return to S40. Among them, if the time error exceeds the set threshold, it means that there is a problem with the time synchronization between the first sensor module 10, the second sensor module 11, and the level module 12, which may affect subsequent data fusion and analysis. At this time, it will return to S40, and the first sensor module 10, the second sensor module 11, and the level module 12 will perform time calibration again through the Network Time Protocol to re-select a unified time reference source to ensure the accuracy of time synchronization.

[0079] In this embodiment, constructing a deformation prediction model for the aligned stress distribution data and the aligned tilt angle data specifically includes: integrating the time-aligned data of the aligned stress distribution data and the aligned tilt angle data through a vector splicing algorithm to form a unified data set, and the integration method of the vector splicing algorithm is:

[0080]

[0081] where X is the vector of the unified data set after integration, x1 is the vector composed of the aligned stress distribution data, x2 is the vector composed of the aligned tilt angle data, and if the aligned stress distribution data has n1 dimensions, it is expressed as If the aligned tilt angle data has n2 dimensions, it is expressed as The vector splicing algorithm fuses the originally scattered data reflecting the state of the photovoltaic panel from different angles into a unified data set, providing more comprehensive data support for subsequent model training.

[0082] The deformation prediction model is trained and evaluated through a long short-term memory network. Accurate deformation prediction results can provide strong support for the operation and maintenance management of a photovoltaic power station. Operation and maintenance personnel can arrange maintenance plans in advance according to the prediction results, take timely measures to repair potential deformation problems, avoid the decline in power generation efficiency and equipment damage caused by severe deformation of photovoltaic panels, reduce operation and maintenance costs, and improve the economic benefits and reliability of the photovoltaic power station. Among them, the long short-term memory network (LSTM) is a type of recurrent neural network that can effectively process time series data and solve the problem of long-term dependence. Training and evaluating the deformation prediction model through the long short-term memory network includes building an LSTM model. In this embodiment, specifically, the number of nodes in the input layer is determined according to the integrated data dimension n. Several hidden layers are set in the middle. Neurons in the hidden layer process and transmit information through special gating mechanisms (forget gate, input gate, and output gate). Finally, the number of nodes in the output layer corresponds to the dimension related to the deformation quantity to be predicted. The integrated unified data set is divided into a training set, a validation set, and a test set according to a certain ratio. The training set is used to train the LSTM model. During the training process, the LSTM model continuously adjusts its internal parameters (such as weights and biases) to minimize the error between the predicted value and the actual value. For example, the mean square error (MSE) is used as the loss function, and the LSTM model parameters are updated through the backpropagation algorithm. The training process will perform multiple rounds of iteration. In each round of iteration, the LSTM model learns from the training set data and continuously optimizes the prediction ability of the model. The validation set is used to evaluate the LSTM model during the training process, observing the prediction performance of the LSTM model on the validation set, such as indicators like accuracy and mean square error. The hyperparameters of the LSTM model (such as the number of neurons in the hidden layer, learning rate, etc.) are adjusted according to the evaluation results to prevent the LSTM model from overfitting or underfitting. When the LSTM model performs stably on the validation set, the test set is used to test the final model to obtain the prediction performance of the LSTM model on unknown data and evaluate the generalization ability of the LSTM model.

[0083] In this embodiment, the multi-level early warning is a three-level early warning mechanism. When the deformation amount of the photovoltaic panel exceeds the first preset threshold, a first-level early warning is triggered to send out a status abnormal signal; when the deformation amount of the photovoltaic panel exceeds the second preset threshold, a second-level early warning is triggered for audible and visual alarms; when the deformation amount of the photovoltaic panel exceeds the third preset threshold, a third-level early warning is triggered to deactivate the photovoltaic panel. In this embodiment, preferably, the first preset threshold is set to 5 mm. When the deformation amount of the photovoltaic panel calculated by the deformation prediction model exceeds 5 mm, the first-level early warning is triggered. The first-level early warning mainly reminds the operation and maintenance personnel by sending out a status abnormal signal. This signal can be a prominent prompt message displayed on the monitoring system interface, such as a red flashing warning icon, accompanied by a text description "The status of the photovoltaic panel is abnormal, please pay attention", or it can send notifications to relevant operation and maintenance personnel via text messages, emails, etc., informing them which specific photovoltaic panel or group of photovoltaic panels has an abnormal status, and the specific form is not limited; the second preset threshold is set to 10 mm. When the deformation amount of the photovoltaic panel exceeds the second preset threshold of 10 mm, the second-level early warning is activated. In addition to continuously displaying the abnormal information, the second-level early warning will also activate the audible and visual alarm device. In the control room of the photovoltaic power station, a loud alarm sound will sound, accompanied by strong flashing lights, to attract the high attention of the operation and maintenance personnel. This audible and visual combined alarm method can quickly attract the attention of the operation and maintenance personnel in a noisy working environment, ensuring that they can take measures in time; the third preset threshold is set to 15 mm. When the deformation amount of the photovoltaic panel exceeds the third preset threshold of 15 mm, the third-level early warning is triggered. The third-level early warning will not only continuously issue audible and visual alarms, but also automatically trigger the control system to immediately deactivate the problematic photovoltaic panel. By cutting off the circuit connection of the photovoltaic panel, its continuous operation is prevented, avoiding serious accidents such as the rupture and fire of the photovoltaic panel caused by excessive deformation. At the same time, detailed fault information is recorded and a professional maintenance team is notified for emergency handling.

[0084] The present invention also provides a prediction system 1 for the deformation of a photovoltaic panel, including:

[0085] A first sensor module 10, which is used to collect the stress distribution data inside the photovoltaic panel in real time, and the first sensor module 10 is embedded in the internal encapsulation layer of the photovoltaic panel;

[0086] A second sensor module 11, which is used to collect the data on the surface of the photovoltaic panel in real time, and the second sensor module 11 is arranged on the surface of the photovoltaic panel;

[0087] A level module 12 is used to establish a deformation monitoring reference plane for the photovoltaic panel and output tilt angle data in real time. Among them, the level module 12 is communicatively connected to the second sensor module 11 and obtains data on the surface of the photovoltaic panel. The tilt angle data is the tilt angle value between the surface of the photovoltaic panel and the deformation monitoring reference plane. The level module 12 is provided on the side of the photovoltaic panel;

[0088] A timestamp synchronization algorithm module 13. The first sensor module 10, the second sensor module 11, and the level module 12 are respectively connected to the timestamp synchronization algorithm module 13. Through the timestamp synchronization algorithm module 13, the stress distribution data and the tilt angle data are aligned, and the timing deviation is eliminated through the dynamic time warping algorithm. Among them, the aligned stress distribution data is correspondingly generated from the stress distribution data, and the aligned tilt angle data is correspondingly generated from the tilt angle data;

[0089] A construction module 14. The construction module 14 is connected to the timestamp synchronization algorithm module 13. The construction module 14 constructs a deformation prediction model for the aligned stress distribution data and the aligned tilt angle data, and triggers multi-level warnings according to the relationship between the deformation amount of the photovoltaic panel and a preset threshold. The prediction system 1 for the deformation of the photovoltaic panel can be combined with the prediction method for the photovoltaic panel. For specific details, please refer to the above content and will not be elaborated one by one.

[0090] The present invention provides a method and system for predicting the deformation of a photovoltaic panel. By using the first sensor module 10 to collect the internal stress distribution data of the photovoltaic panel in real time, the second sensor module 11 to collect the surface data of the photovoltaic panel in real time, the level module 12 to establish a deformation monitoring reference plane for the photovoltaic panel and output tilt angle data in real time, and by using the timestamp synchronization algorithm to align the data of the first sensor module 10, the second sensor module 11, and the level module 12, and eliminating the timing deviation through the dynamic time warping algorithm, a deformation prediction model is constructed, and multi-level warnings are triggered according to the relationship between the deformation amount of the photovoltaic panel and a preset threshold. Among them, the tilt angle data is the tilt angle value between the surface of the photovoltaic panel and the deformation monitoring reference plane. In addition, the first sensor module 10 is embedded in the internal encapsulation layer of the photovoltaic panel, the second sensor module 11 is provided on the surface of the photovoltaic panel, and the level module 12 is provided on the side of the photovoltaic panel. The present invention realizes multi-faceted, comprehensive and accurate monitoring by collecting internal stress, surface data and tilt angle data, and can timely take countermeasures according to the prediction results, realizing comprehensive and accurate prediction of the deformation of the photovoltaic panel in a complex environment and reducing losses.

[0091] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0092] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0093] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features. And these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting photovoltaic panel deformation, characterized in that: include: The stress distribution data inside the photovoltaic panel is collected in real time by the first sensor module; Collecting data on the surface of the photovoltaic panel in real time through a second sensor module; Establishing a deformation monitoring reference plane of the photovoltaic panel through a leveling instrument module and outputting tilt angle data in real time, wherein the tilt angle data is a tilt angle value between the surface of the photovoltaic panel and the deformation monitoring reference plane; Align the stress distribution data and the tilt angle data by a timestamp synchronization algorithm and eliminate timing deviation by a dynamic time warping algorithm, wherein the stress distribution data generates aligned stress distribution data, and the tilt angle data generates aligned tilt angle data; A deformation prediction model is constructed for the alignment stress distribution data and the alignment tilt angle data, and a multi-level warning is triggered according to the relationship between the deformation amount of the photovoltaic panel in the deformation prediction model and a preset threshold.

2. The photovoltaic panel deformation prediction method according to claim 1, characterized in that: The real-time collection of stress distribution data inside the photovoltaic panel by the first sensor module specifically includes: embedding a flexible fiber grating array into the internal packaging layer of the photovoltaic panel in a serpentine path, and collecting the stress distribution data inside the photovoltaic panel by the flexible fiber grating array.

3. The photovoltaic panel deformation prediction method according to claim 2, characterized in that: The collecting of the internal stress distribution data of the photovoltaic panel includes: performing mean filtering on the stress distribution data.

4. The photovoltaic panel deformation prediction method according to claim 1, characterized in that: The real-time collection of the surface data of the photovoltaic panel by the second sensor module specifically includes: The data on the surface of the photovoltaic panel is acquired by the laser ranging sensor in the second sensor module, and the data on the surface of the photovoltaic panel is smoothed by a filtering algorithm.

5. The photovoltaic panel deformation prediction method according to claim 4, characterized in that: The method of establishing the deformation monitoring reference plane of the photovoltaic panel and outputting the tilt angle data in real time by means of the level meter module specifically includes: the level meter module collects the tilt angle data of the photovoltaic panel before use according to a set time interval, and calculates the average value of the tilt angle data, and the plane corresponding to the average value is the deformation monitoring reference plane.

6. The photovoltaic panel deformation prediction method according to claim 1, characterized in that: The aligning the stress distribution data and the tilt angle data by using a timestamp synchronization algorithm comprises: S40: The first sensor module, the second sensor module and the leveling instrument module select a unified time reference source through a network time protocol; S41: the first sensor module, the second sensor module and the leveling module record the timestamp of data collection when collecting data, and verify the timestamps corresponding to the first sensor module, the second sensor module and the leveling module respectively; S42: Determine whether the time error between the first sensor module, the second sensor module and the level module is within a set threshold range; if the time error exceeds the set threshold, return to S40.

7. The photovoltaic panel deformation prediction method according to claim 1, characterized in that: The step of constructing a deformation prediction model for the alignment stress distribution data and the alignment tilt angle data specifically includes: integrating the alignment stress distribution data and the alignment tilt angle data through a vector stitching algorithm to form a unified data set, wherein the vector stitching algorithm is integrated in the following manner: Wherein, X is the vector of the unified data set after integration, x1 is the vector formed by the alignment stress distribution data, and x2 is the vector formed by the alignment tilt angle data. The alignment stress distribution data has n1 dimensions, which is expressed as The alignment tilt angle data has n2 dimensions, which can be expressed as The deformation prediction model is trained and evaluated through a long short-term memory network.

8. The photovoltaic panel deformation prediction method according to claim 7, characterized in that: The multi-level warning is a three-level warning mechanism. When the deformation of the photovoltaic panel exceeds the first preset threshold, the first-level warning is triggered to send out a state abnormality signal; when the deformation of the photovoltaic panel exceeds the second preset threshold, the second-level warning sound and light alarm is triggered; When the deformation amount of the photovoltaic panel exceeds a third preset threshold, a third-level warning is triggered to disable the photovoltaic panel.

9. The photovoltaic panel deformation prediction method according to claim 1, characterized in that: The calculation formula of the tilt angle value is: Δθ=|θ measured -θ base |, where Δθ is the tilt angle value, θ measured is the actual tilt angle of the photovoltaic panel measured by the level module, θ base The tilt angle of the deformation monitoring reference surface of the level module.

10. A photovoltaic panel deformation prediction system, characterized in that: include: A first sensor module, the first sensor module is used to collect stress distribution data inside the photovoltaic panel in real time, and the first sensor module is embedded in the internal packaging layer of the photovoltaic panel; A second sensor module, the second sensor module is used to collect data on the surface of the photovoltaic panel in real time, and the second sensor module is arranged on the surface of the photovoltaic panel; A leveling module, the leveling module is used to establish a deformation monitoring reference plane of the photovoltaic panel and output tilt angle data in real time, wherein the leveling module is connected to the second sensor module in communication and obtains data on the surface of the photovoltaic panel, the tilt angle data is the tilt angle value between the surface of the photovoltaic panel and the deformation monitoring reference plane, and the leveling module is arranged on the side of the photovoltaic panel; A timestamp synchronization algorithm module, wherein the first sensor module, the second sensor module and the level module are respectively connected to the timestamp synchronization algorithm module, and the stress distribution data and the tilt angle data are aligned by the timestamp synchronization algorithm module, and the timing deviation is eliminated by a dynamic time warping algorithm, wherein the stress distribution data generates aligned stress distribution data, and the tilt angle data generates aligned tilt angle data; A construction module is connected to the timestamp synchronization algorithm module, and the construction module constructs a deformation prediction model for the alignment stress distribution data and the alignment tilt angle data, and triggers a multi-level warning according to the relationship between the photovoltaic panel deformation amount and a preset threshold.

Images