A flexible foldable solar panel breakage sensing system and method

Abstract

The application discloses a flexible foldable solar cell panel damage sensing system and method, and relates to the technical field of flexible photoelectric device state monitoring. In the system, a flexible electronic sensing layer is arranged between a photovoltaic unit and a flexible substrate; the flexible electronic sensing layer acquires original signals of a plurality of monitoring points on the flexible foldable solar cell panel; the plurality of monitoring points cover key stress areas of the flexible foldable solar cell panel; a signal acquisition and preprocessing module acquires the plurality of original signals, and pre-processes the plurality of original signals to obtain a plurality of pre-processed original signals; a feature recognition and damage determination algorithm module extracts a feature index group of each pre-processed original signal, and determines a detection result by using a weighted center positioning algorithm combined with a convolutional neural network classifier based on the plurality of feature index groups. The flexible electronic sensing layer is integrated in the foldable flexible solar cell panel to realize structural health monitoring and self-diagnosis of the flexible foldable solar cell panel.

Description

Technical Field

[0001] This application relates to the field of condition monitoring technology for flexible optoelectronic devices, and in particular to a damage sensing system and method for flexible folding solar panels. Background Technology

[0002] Flexible solar panels, with their lightweight and foldable characteristics, are widely used in spacecraft deployable wings, drone power systems, and portable photovoltaic devices. Currently, most flexible solar panels on the market use polyimide, PET, and other materials as substrates. These substrates possess a certain degree of plasticity, but after repeated folding and unfolding operations, or when subjected to external forces such as wind loads and vibrations, flexible solar panels are prone to problems such as microcracks, delamination, or breakage of conductive paths. Traditional electrical methods mainly include I-V curve scanning and EL testing, which require external equipment for detection and cannot achieve real-time monitoring. Optical / infrared imaging methods can detect obvious cracks, but are not sensitive enough to micro-damage and struggle to detect early minor breaks. Acoustic and vibration methods are easily affected by structural noise and environmental factors, making them unsuitable for detecting flexible structures. Existing detection solutions are insufficient to meet the detection requirements of flexible, foldable solar panels, specifically in three aspects. First, there is the real-time requirement; existing solutions struggle to effectively monitor the dynamic process of panel folding and unfolding. Secondly, there are compatibility requirements; the detection element needs to conformally fit with the flexible battery, a condition that existing elements struggle to meet. Thirdly, there are low power consumption and lightweight requirements; the embedded detection system should not impose an additional burden on the battery panel, and existing systems are significantly inadequate in this regard. Therefore, there is an urgent need for a damage detection method that can be embedded in flexible structures to achieve online sensing and intelligent judgment. Summary of the Invention

[0003] The purpose of this application is to provide a damage sensing system and method for flexible foldable solar panels. By integrating a flexible strain / resistance sensor array into the foldable flexible solar panel, the system can monitor, accurately locate, and identify the damage type of the flexible foldable solar panel in real time, thereby realizing structural health monitoring and self-diagnosis of the flexible foldable solar panel and improving the reliability and service life of the flexible foldable solar panel system.

[0004] To achieve the above objectives, this application provides the following solution:

[0005] In a first aspect, this application provides a damage detection system for a flexible foldable solar panel, which is applied to a flexible foldable solar panel, the flexible foldable solar panel comprising: a photovoltaic unit and a flexible substrate;

[0006] The flexible, unfoldable solar panel damage detection system includes:

[0007] The flexible electronic sensing layer, the signal acquisition and preprocessing module, and the feature recognition and damage assessment algorithm module are connected in sequence.

[0008] The flexible electronic sensing layer is disposed between the photovoltaic unit and the flexible substrate; the flexible electronic sensing layer is used to acquire the original signals of multiple monitoring points on the flexible folding solar panel; the multiple monitoring points cover the key stress areas of the flexible folding solar panel;

[0009] The signal acquisition and preprocessing module is used to acquire multiple raw signals and preprocess the multiple raw signals to obtain multiple preprocessed raw signals;

[0010] The feature recognition and damage determination algorithm module is used to extract feature index groups for each preprocessed original signal. Based on multiple feature index groups, a weighted center localization algorithm combined with a convolutional neural network classifier is used to determine the detection result. The feature index group is determined by performing wavelet packet decomposition on the original signal, calculating the energy of multiple frequency bands after decomposition, selecting the top preset number of frequency bands with the highest energy contribution rate for signal reconstruction, and performing principal component analysis on the reconstructed signal. The detection result includes whether damage has occurred, and when damage occurs, the location, type, and severity level of the damage.

[0011] Optionally, the flexible electronic sensing layer specifically includes:

[0012] Flexible strain / resistance sensing array and conductive interconnect network;

[0013] The flexible strain / resistance sensing array is used to acquire the raw signals from multiple monitoring points on the flexible folded solar panel; the raw signals are strain / resistance signals.

[0014] The conductive interconnect network is used to connect the flexible strain / resistance sensing array and the signal acquisition and preprocessing module.

[0015] Optionally, the conductive interconnect network is fabricated using silver paste printing.

[0016] Optionally, the flexible folding solar panel damage detection system further includes:

[0017] Temperature sensor;

[0018] The temperature sensor is placed in the environment where the flexible folding solar panel is located, and the temperature sensor is connected to the signal acquisition and preprocessing module; the temperature sensor is used to acquire the ambient temperature.

[0019] Optionally, the signal acquisition and preprocessing module specifically includes:

[0020] Signal acquisition unit and data preprocessing unit;

[0021] The signal acquisition unit employs a multi-channel synchronous sampling circuit; the multi-channel synchronous sampling circuit is used to control the synchronous acquisition of raw signals from multiple monitoring points on the flexible folding solar panel;

[0022] The data preprocessing unit specifically includes:

[0023] Filtering subunit, data normalization subunit, and temperature compensation subunit;

[0024] The filtering subunit employs a 5th-order Butterworth low-pass filter; the filtering subunit is used to denoise the original signal to obtain the denoised original signal.

[0025] The data normalization subunit is used to normalize the original signal after denoising based on the standard signal to obtain the normalized original signal.

[0026] The temperature compensation subunit is used to perform temperature compensation on the normalized original signal based on the ambient temperature using a temperature compensation formula, so as to obtain the preprocessed original signal.

[0027] Optionally, the temperature compensation formula is:

[0028] ;

[0029] in, This is the temperature-compensated signal; The signal before temperature compensation; Ambient temperature; The material temperature coefficient; This is a reference temperature.

[0030] Optionally, the feature recognition and damage determination algorithm module specifically includes:

[0031] Feature extraction unit and damage determination unit;

[0032] The feature extraction unit is used to extract a set of feature indicators for each preprocessed original signal; the set of feature indicators includes the resistance change rate, strain gradient, and energy spectrum; wherein the energy spectrum is obtained by Fourier transform;

[0033] The damage determination unit is used to determine the detection result based on multiple feature index groups, using a weighted center localization algorithm combined with a convolutional neural network classifier; the damage location is determined by the weighted center localization algorithm; and the damage type is determined by the trained convolutional neural network classifier.

[0034] The damage determination unit is also used to perform drift compensation of the signal by using adaptive Kalman filtering when running the weighted center localization algorithm and the convolutional neural network classifier.

[0035] Optionally, the convolutional neural network classifier specifically includes:

[0036] The first convolutional layer, the first ReLU activation-max pooling module, the second convolutional layer, the second ReLU activation-max pooling module, the first fully connected layer, the second fully connected layer, and the output layer are connected in sequence.

[0037] The first convolutional layer uses eight 3×1 convolutional kernels;

[0038] The second convolutional layer uses 16 3×1 convolutional kernels;

[0039] The number of neurons in the first fully connected layer is 32;

[0040] The number of neurons in the second fully connected layer is 3;

[0041] The output layer uses the Softmax function.

[0042] Optionally, the flexible folding solar panel damage detection system further includes:

[0043] Output and communication interface module;

[0044] The output and communication interface module is connected to the external terminal and the feature recognition and damage determination algorithm module, respectively.

[0045] The output and communication interface module is used to transmit the detection results to an external terminal via Bluetooth.

[0046] Secondly, this application provides a method for detecting damage to a flexible folding solar panel, the method being applied to the aforementioned flexible folding solar panel damage detection system, the method comprising:

[0047] Acquire raw signals from multiple monitoring points on a flexible, foldable solar panel; the multiple monitoring points cover the key stress areas of the flexible, foldable solar panel.

[0048] The original signals are preprocessed to obtain multiple preprocessed original signals;

[0049] Extract feature index groups from each preprocessed original signal; the feature index groups are determined by performing wavelet packet decomposition on the original signal, calculating the energy of multiple frequency bands after decomposition, selecting the first preset number of frequency bands with the highest energy contribution rate for signal reconstruction, and performing principal component analysis on the reconstructed signal.

[0050] Based on multiple feature index groups, a weighted center localization algorithm combined with a convolutional neural network classifier is used to determine the detection results; the detection results include whether damage has occurred, and when damage occurs, the location of the damage, the type of damage, and the severity level of the damage.

[0051] According to the specific embodiments provided in this application, the following technical effects are disclosed:

[0052] This application provides a damage detection system and method for flexible folding solar panels. Addressing the issues of microcracks, delamination, and fatigue damage that easily occur in flexible folding solar panels under repeated folding and harsh operating conditions, and the low sensitivity and difficulty in real-time location of existing detection methods, this application adopts a structural integrated design. A flexible electronic sensing layer is set in the structural layer between the flexible substrate and the photovoltaic unit of the solar panel. This layer works in conjunction with the solar panel's operating system to collect mechanical stress signals and electrical parameter changes at various monitoring points of the solar panel in real time under folding motion and load. The collected raw signals are then transmitted to a data processing module. First, the raw signals undergo noise reduction and filtering preprocessing. Then, a feature extraction algorithm is used to extract stress distribution features, electrical signal abrupt change features, and signal trend features. Subsequently, a multi-point data fusion algorithm is used to cross-validate and correlate the feature data from multiple sensing nodes, achieving automatic identification, severity determination, and precise location of damage defects such as microcrack initiation, delamination, and fatigue damage propagation in the solar panel.

[0053] When setting up a flexible electronic sensing layer, a flexible strain sensing array or flexible resistance sensing array adapted to the folding direction and stress concentration areas of the solar panel is embedded at a preset array density. The sensing array has good conformality with the solar panel structure, a detection sensitivity of up to 0.01% strain, and a response time of less than 10ms. It features high sensitivity, strong real-time performance, high structural integration, and strong anti-interference capabilities, effectively avoiding energy output interruptions caused by potential damage. It is suitable for online self-diagnostic applications in scenarios such as aerospace deployable wings, UAV photovoltaic systems, and wearable energy devices, providing a guarantee for the reliable operation of flexible energy devices. Attached Figure Description

[0054] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0055] Figure 1 This is a schematic diagram of the flexible folding solar panel structure layer in the embodiment;

[0056] Figure 2This is a schematic diagram showing the placement and interconnection of the flexible electronic sensors in the embodiment;

[0057] Figure 3 This is a block diagram of the signal acquisition and feature extraction module in the embodiment;

[0058] Figure 4 This is a flowchart of the damage detection and classification algorithm in the embodiment;

[0059] Figure 5 This is the result of sensing the damage type of the flexible folding solar panel in the example;

[0060] Figure 6 This is the result of sensing the location of damage to the flexible folding solar panel in the example. Detailed Implementation

[0061] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0062] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0063] In one exemplary embodiment, a damage detection system for a flexible foldable solar panel is provided. The flexible foldable solar panel includes a photovoltaic unit and a flexible substrate. The photovoltaic unit is used to realize photoelectric conversion, and the flexible substrate is made of polyimide material to provide support for the photovoltaic unit.

[0064] like Figure 1As shown, the flexible foldable solar panel damage detection system includes: a temperature sensor, and a flexible electronic sensing layer, a signal acquisition and preprocessing module, a feature recognition and damage assessment algorithm module, and an output and communication interface module connected in sequence; the flexible electronic sensing layer is disposed between the photovoltaic unit and the flexible substrate; the flexible electronic sensing layer is used to acquire the raw signals from multiple monitoring points on the flexible foldable solar panel; the multiple monitoring points cover the key stress areas of the flexible foldable solar panel; the temperature sensor is disposed in the environment where the flexible foldable solar panel is located, and the temperature sensor is connected to the signal acquisition and preprocessing module; the temperature sensor is used to acquire the ambient temperature. The signal acquisition and preprocessing module acquires multiple raw signals and preprocesses them to obtain multiple preprocessed raw signals. The feature recognition and damage determination algorithm module extracts feature index groups from each preprocessed raw signal. Based on multiple feature index groups, a weighted center localization algorithm combined with a convolutional neural network classifier is used to determine the detection result. The feature index groups are determined by wavelet packet decomposition of the raw signal, calculating the energy of multiple frequency bands after decomposition, selecting the top preset number of frequency bands with the highest energy contribution rate for signal reconstruction, and performing principal component analysis on the reconstructed signal. The detection result includes whether damage has occurred, and if damage has occurred, the location, type, and severity level of the damage. The output and communication interface module connects to both the external terminal and the feature recognition and damage determination algorithm module. The output and communication interface module transmits the detection result to the external terminal via Bluetooth. The damage severity level is determined based on the output of the convolutional neural network classifier to determine whether damage has occurred, and when damage is determined to have occurred, the severity level classification of the damage is output simultaneously. The damage location is determined using a weighted center localization algorithm based on the spatial distribution of the abnormal feature index groups.

[0065] The flexible electronic sensing layer specifically includes a flexible strain / resistance sensing array and a conductive interconnect network. The flexible strain / resistance sensing array acquires raw signals from multiple monitoring points on the flexible, unfolded solar panel; the raw signals are strain / resistance signals. The conductive interconnect network connects the flexible strain / resistance sensing array and the signal acquisition and preprocessing module. The conductive interconnect network is fabricated using silver paste printing. A schematic diagram of the flexible electronic sensor placement and interconnection is shown below. Figure 2 As shown.

[0066] The signal acquisition and preprocessing module specifically includes: a signal acquisition unit and a data preprocessing unit; the signal acquisition unit employs a multi-channel synchronous sampling circuit; the multi-channel synchronous sampling circuit is used to control the synchronous acquisition of raw signals from multiple monitoring points on the flexible folding solar panel. The signal acquisition and preprocessing module is as follows... Figure 3As shown, the data preprocessing unit specifically includes: a filtering subunit, a data normalization subunit, and a temperature compensation subunit. The filtering subunit uses a 5th-order Butterworth low-pass filter. The filtering subunit is used to denoise the original signal, obtaining the denoised original signal. The data normalization subunit is used to normalize the denoised original signal based on a standard signal, obtaining the normalized original signal. The temperature compensation subunit is used to perform temperature compensation on the normalized original signal based on the ambient temperature using a temperature compensation formula, obtaining the preprocessed original signal. The temperature compensation formula is:

[0067] .

[0068] in, This is the temperature-compensated signal; The signal before temperature compensation; Ambient temperature; The material temperature coefficient; This is a reference temperature.

[0069] The feature recognition and damage determination algorithm module specifically includes a feature extraction unit and a damage determination unit. The feature extraction unit is used to extract feature index groups for each preprocessed original signal. The feature index groups include resistance change rate, strain gradient, and energy spectrum. The energy spectrum is obtained through Fourier transform. The damage determination unit is used to determine the detection result based on multiple feature index groups, using a weighted center localization algorithm combined with a convolutional neural network classifier. The damage location is determined by the weighted center localization algorithm. The damage type is determined by the trained convolutional neural network classifier. The damage determination unit is also used to perform drift compensation of the signal by using adaptive Kalman filtering when running the weighted center localization algorithm and the convolutional neural network classifier.

[0070] The convolutional neural network classifier specifically includes: a first convolutional layer, a first ReLU activation-max pooling module, a second convolutional layer, a second ReLU activation-max pooling module, a first fully connected layer, a second fully connected layer, and an output layer connected in sequence; the first convolutional layer uses 8 3×1 convolutional kernels; the second convolutional layer uses 16 3×1 convolutional kernels; the first fully connected layer has 32 neurons; the second fully connected layer has 3 neurons; and the output layer uses the Softmax function.

[0071] The flexible foldable solar panel damage detection system provided in this embodiment includes a flexible electronic sensing layer, a signal acquisition and preprocessing module, a feature recognition and damage assessment algorithm module, and an output and communication interface module. The flexible solar panel module comprises photovoltaic units and a flexible substrate. The flexible electronic sensing layer consists of a sensing array and a conductive interconnect network. The sensing array is used to acquire strain or resistance signals from the solar panel, and the conductive interconnect network is printed with silver paste to connect the sensing array to the signal acquisition module. The signal acquisition and preprocessing module uses a low-power microcontroller as its core, along with a signal conditioning circuit, to complete signal acquisition and preliminary processing. The feature recognition and damage assessment algorithm module is integrated into the microcontroller and analyzes the processed signals using a preset algorithm. The output and communication interface module uses a Bluetooth module to transmit the damage detection results to an external terminal. The resistance of each sensing unit in the flexible electronic sensing layer changes with the strain of the flexible foldable solar panel; this change is directly related to the structural state of the solar panel. When a local area of ​​a solar panel develops cracks or fatigue damage, the corresponding sensing unit will exhibit characteristic anomalies in resistance or voltage due to structural deformation. By collecting the signal differences and time series changes from multiple sensing units, accurate detection of solar panel damage can be achieved.

[0072] The specific detection steps of the flexible folding solar panel damage detection system provided in this embodiment are as follows:

[0073] Sensor array layout: A two-dimensional array of sensor units is laid out along the folding hinge and stress concentration area of ​​the flexible folding solar panel. The spacing between adjacent sensor units is set to 5mm to ensure coverage of the critical stress areas of the solar panel.

[0074] Signal acquisition: The resistance, voltage or strain data of each sensing unit are acquired through a multi-channel synchronous sampling circuit. The sampling frequency is set to 100Hz and the sampling accuracy is controlled within 0.1% to ensure the real-time performance and accuracy of signal acquisition.

[0075] Data preprocessing: The acquired raw signals are sequentially filtered, normalized, and temperature compensated. A 5th-order Butterworth low-pass filter is used to remove high-frequency noise. Data normalization is performed by comparing with a standard signal. A ambient temperature is collected using a temperature sensor, and the influence of temperature on the sensor signal is eliminated by combining it with a preset temperature compensation formula.

[0076] Feature extraction: The resistance change rate ΔR / R0, strain gradient, energy spectrum and other feature indicators are extracted from the preprocessed signal. ΔR / R0 reflects the relative change in resistance of the sensing unit, the strain gradient reflects the strain difference in different regions, and the energy spectrum is obtained by Fourier transform to reflect the frequency characteristics of the signal.

[0077] Damage assessment: A weighted center localization algorithm combined with a convolutional neural network classifier is used for damage assessment. The weighted center localization algorithm calculates the damage location based on the degree of signal anomaly of each sensing unit, while the convolutional neural network classifier learns from the extracted feature indicators to identify the damage type. The damage detection and classification algorithm flow is as follows: Figure 4 .

[0078] Status output: Through the output and communication interface module, information such as the location, type, and severity level of the injury is transmitted to the external terminal. The severity level is divided into three levels: mild, moderate, and severe, based on the degree of abnormality of the characteristic indicators.

[0079] Feature denoising is achieved using wavelet packet decomposition combined with principal component analysis (PCA). Wavelet packet decomposition decomposes the signal into different frequency bands, and the bands containing the main features are selected for reconstruction. PCA reduces the dimensionality of the reconstructed signal features, retaining principal components with a contribution rate exceeding 95% to reduce noise interference. A convolutional neural network (CNN) is used for pattern recognition. The network input is the extracted feature index matrix, which is processed through three convolutional layers and two fully connected layers to output the damage type identification result. Adaptive Kalman filtering is employed for signal drift compensation. By adjusting the filter gain in real time, drift phenomena during long-term sensor signal acquisition are eliminated, ensuring signal stability.

[0080] The following solutions can be considered when implementing the flexible folding solar panel damage detection system provided in this application:

[0081] Option 1: Strain-type sensor network solution.

[0082] Metal foil strain gauges are embedded in the folding hinges and stress concentration areas of the flexible battery substrate. The strain gauges are made of copper-nickel alloy with a thickness controlled within 5μm. They are arranged in a two-dimensional array with a 5mm spacing. The strain gauge leads are connected to the signal acquisition module via a conductive interconnect network printed with silver paste. The signal acquisition module uses a Wheatstone bridge to convert the resistance change of the strain gauges into a voltage signal. The Wheatstone bridge's supply voltage is set to 3.3V. The output signal is amplified by an instrumentation amplifier and then acquired by the microcontroller's ADC interface at a sampling frequency of 100Hz. The acquired voltage signal is preprocessed. First, a 5th-order Butterworth low-pass filter is used to filter out high-frequency noise above 10Hz. Then, temperature compensation is performed based on the ambient temperature collected by a temperature sensor. The strain gradient characteristics of the signal are extracted. When the flexible solar panel is folded more than 500 times, the frequency of the strain signal will drift by more than 0.5Hz, and the strain gradient will produce a sudden change of more than 5%. Based on these two characteristics, the system determines that the solar panel has fatigue damage and transmits the damage information to an external terminal via a Bluetooth module.

[0083] Option 2: Resistive conductive composite film solution.

[0084] A conductive composite film was prepared using silver nanowires / PDMS composite material. The silver nanowires had a diameter of 50 nm and a length of 10 μm. The silver nanowires were dispersed in a PDMS matrix at a mass fraction of 5%. The composite material was then printed onto the encapsulation layer surface of a flexible solar panel using a blade coating method, with the coating thickness controlled at 10 μm, forming a continuous resistive network. The resistive network was divided into sensing units with a 5 mm × 5 mm grid, and the initial resistance of each sensing unit was controlled at approximately 1 kΩ. A conventional multi-channel resistance acquisition chip was used to acquire the resistance values ​​of each sensing unit, with an acquisition accuracy of 0.01 Ω. The acquired resistance data was transmitted to a microcontroller via an I2C interface. When a crack appeared in the solar panel, the conductive composite film in the corresponding area would break, causing a sharp increase in the resistance value of the sensing unit in that area, reducing the connectivity of the resistive network, and the rate of resistance change was directly proportional to the length of the crack. The microcontroller calculates the damage location based on the resistance change rate of each sensing unit and a weighted center positioning algorithm. It then maps the resistance change rate data into grayscale values ​​to generate a damage heatmap, which is displayed in real time through the output interface for precise location of the peeling area of ​​the solar panel.

[0085] Option 3: Signal analysis algorithm implementation scheme.

[0086] The signal analysis algorithm is implemented on a microcontroller-based embedded system. First, the acquired raw signal undergoes wavelet packet decomposition, using the db4 wavelet as the wavelet basis. The decomposition layer is set to three layers. Energy calculation is performed on the eight frequency bands after decomposition, and the three frequency bands with the highest energy contribution rates are selected for signal reconstruction. Principal component analysis is then performed on the reconstructed signal to calculate the covariance matrix of each feature. Principal components with eigenvalues ​​greater than 1 are selected to form a dimensionality-reduced feature vector. This feature vector is then input into a convolutional neural network (CNN) for damage classification. The input layer of the CNN is a 1×16 feature vector. The first convolutional layer uses eight 3×1 convolutional kernels with ReLU activation. After max pooling, the vector is input into the second convolutional layer, which uses sixteen 3×1 convolutional kernels, also undergoing ReLU activation and max pooling. The processed features are then input into two fully connected layers with 32 and 3 neurons respectively. Finally, the output layer uses the Softmax function to output the probabilities of three typical damage types: crack, delamination, and local stress concentration. The type with the highest probability is selected as the damage identification result. During the algorithm's operation, an adaptive Kalman filter is used to compensate for signal drift. The noise covariance Q during the filtering process is set to 0.01, and the observation noise covariance R is dynamically adjusted according to the fluctuation of the real-time signal to ensure signal stability and improve the accuracy of damage classification.

[0087] Experimental results and model validation curves are as follows Figure 5 As shown, the sensing results of the damaged location of the flexible folding solar panel are as follows: Figure 6 As shown, this application can identify microcrack-level (<100μm) damage with high sensitivity detection capability; it can operate continuously during the unfolding process and in the working state, meeting the requirements of real-time online monitoring; the flexible electronic sensing layer is conformally bonded to the flexible solar panel, without affecting the folding performance and light absorption performance of the solar panel, and has good flexibility compatibility; it adopts an integrated electronic ink circuit with power consumption of less than 10mW, meeting the requirements of low power consumption and lightweight; based on the data-driven model, it can realize automatic damage classification and lifetime prediction, achieving intelligent diagnostic effect.

[0088] In another embodiment, a method for detecting damage to a flexible, foldable solar panel applied to the aforementioned flexible, foldable solar panel damage detection system is provided, the method comprising:

[0089] Step 1: Acquire raw signals from multiple monitoring points on the flexible folding solar panel; multiple monitoring points cover the key stress areas of the flexible folding solar panel;

[0090] Step 2: Preprocess multiple raw signals to obtain multiple preprocessed raw signals;

[0091] Step 3: Extract the feature index group of each preprocessed original signal; the feature index group is determined by performing wavelet packet decomposition on the original signal, calculating the energy of multiple frequency bands after decomposition, selecting the first preset number of frequency bands with the highest energy contribution rate for signal reconstruction, and performing principal component analysis on the reconstructed signal.

[0092] Step 4: Based on multiple feature index groups, the weighted center localization algorithm combined with a convolutional neural network classifier is used to determine the detection results; the detection results include whether damage has occurred, and when damage occurs, the location of the damage, the type of damage, and the severity level of the damage.

[0093] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0094] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A flexible, foldable solar panel damage detection system, characterized in that, The flexible foldable solar panel damage detection system is applied to a flexible foldable solar panel, which includes: photovoltaic units and a flexible substrate; The flexible, unfoldable solar panel damage detection system includes: The flexible electronic sensing layer, the signal acquisition and preprocessing module, and the feature recognition and damage assessment algorithm module are connected in sequence. The flexible electronic sensing layer is disposed between the photovoltaic unit and the flexible substrate; the flexible electronic sensing layer is used to acquire the original signals of multiple monitoring points on the flexible folding solar panel; the multiple monitoring points cover the key stress areas of the flexible folding solar panel; The signal acquisition and preprocessing module is used to acquire multiple raw signals and preprocess the multiple raw signals to obtain multiple preprocessed raw signals; The feature recognition and damage determination algorithm module is used to extract feature index groups for each preprocessed original signal. Based on multiple feature index groups, a weighted center localization algorithm combined with a convolutional neural network classifier is used to determine the detection result. The feature index group is determined by performing wavelet packet decomposition on the original signal, calculating the energy of multiple frequency bands after decomposition, selecting the top preset number of frequency bands with the highest energy contribution rate for signal reconstruction, and performing principal component analysis on the reconstructed signal. The detection result includes whether damage has occurred, and when damage occurs, the location, type, and severity level of the damage.

2. The flexible folding solar panel damage detection system according to claim 1, characterized in that, The flexible electronic sensing layer specifically includes: Flexible strain / resistance sensing array and conductive interconnect network; The flexible strain / resistance sensing array is used to acquire the raw signals from multiple monitoring points on the flexible folded solar panel; the raw signals are strain / resistance signals. The conductive interconnect network is used to connect the flexible strain / resistance sensing array and the signal acquisition and preprocessing module.

3. The flexible folding solar panel damage detection system according to claim 2, characterized in that, The conductive interconnect network is fabricated using silver paste printing.

4. The flexible folding solar panel damage detection system according to claim 1, characterized in that, The flexible, unfoldable solar panel damage detection system also includes: Temperature sensor; The temperature sensor is placed in the environment where the flexible folding solar panel is located, and the temperature sensor is connected to the signal acquisition and preprocessing module; the temperature sensor is used to acquire the ambient temperature.

5. The flexible folding solar panel damage detection system according to claim 4, characterized in that, The signal acquisition and preprocessing module specifically includes: Signal acquisition unit and data preprocessing unit; The signal acquisition unit employs a multi-channel synchronous sampling circuit; the multi-channel synchronous sampling circuit is used to control the synchronous acquisition of raw signals from multiple monitoring points on the flexible folding solar panel; The data preprocessing unit specifically includes: Filtering subunit, data normalization subunit, and temperature compensation subunit; The filtering subunit employs a 5th-order Butterworth low-pass filter; the filtering subunit is used to denoise the original signal to obtain the denoised original signal. The data normalization subunit is used to normalize the original signal after denoising based on the standard signal to obtain the normalized original signal. The temperature compensation subunit is used to perform temperature compensation on the normalized original signal based on the ambient temperature using a temperature compensation formula, so as to obtain the preprocessed original signal.

6. The flexible folding solar panel damage detection system according to claim 5, characterized in that, The temperature compensation formula is as follows: ； in, This is the temperature-compensated signal; The signal before temperature compensation; Ambient temperature; The material temperature coefficient; This is a reference temperature.

7. The flexible folding solar panel damage detection system according to claim 1, characterized in that, The feature recognition and damage determination algorithm module specifically includes: Feature extraction unit and damage determination unit; The feature extraction unit is used to extract a set of feature indicators for each preprocessed original signal; the set of feature indicators includes the resistance change rate, strain gradient, and energy spectrum; wherein the energy spectrum is obtained by Fourier transform; The damage determination unit is used to determine the detection result based on multiple feature index groups, using a weighted center localization algorithm combined with a convolutional neural network classifier; the damage location is determined by the weighted center localization algorithm; and the damage type is determined by the trained convolutional neural network classifier. The damage determination unit is also used to perform drift compensation of the signal by using adaptive Kalman filtering when running the weighted center localization algorithm and the convolutional neural network classifier.

8. The flexible folding solar panel damage detection system according to claim 7, characterized in that, The convolutional neural network classifier specifically includes: The first convolutional layer, the first ReLU activation-max pooling module, the second convolutional layer, the second ReLU activation-max pooling module, the first fully connected layer, the second fully connected layer, and the output layer are connected in sequence. The first convolutional layer uses eight 3×1 convolutional kernels; The second convolutional layer uses 16 3×1 convolutional kernels; The number of neurons in the first fully connected layer is 32; The number of neurons in the second fully connected layer is 3; The output layer uses the Softmax function.

9. The flexible folding solar panel damage detection system according to claim 1, characterized in that, The flexible, unfoldable solar panel damage detection system also includes: Output and communication interface module; The output and communication interface module is connected to the external terminal and the feature recognition and damage determination algorithm module, respectively. The output and communication interface module is used to transmit the detection results to an external terminal via Bluetooth.

10. A method for detecting damage to a flexible, foldable solar panel, characterized in that, The method is applied to the flexible folding solar panel damage detection system as described in any one of claims 1-9, and the method includes: Acquire raw signals from multiple monitoring points on a flexible, foldable solar panel; the multiple monitoring points cover the key stress areas of the flexible, foldable solar panel. The original signals are preprocessed to obtain multiple preprocessed original signals; Extract feature index groups from each preprocessed original signal; the feature index groups are determined by performing wavelet packet decomposition on the original signal, calculating the energy of multiple frequency bands after decomposition, selecting the first preset number of frequency bands with the highest energy contribution rate for signal reconstruction, and performing principal component analysis on the reconstructed signal. Based on multiple feature index groups, a weighted center localization algorithm combined with a convolutional neural network classifier is used to determine the detection results; the detection results include whether damage has occurred, and when damage occurs, the location of the damage, the type of damage, and the severity level of the damage.

Images