Photovoltaic module hot spot defect automatic identification method based on thermal infrared imaging

Abstract

The application discloses a photovoltaic module hot spot defect automatic identification method based on thermal infrared imaging, relates to the technical field of defect identification, and improves the environmental adaptation precision and multi-frequency thermal wave penetration coverage dimension in the photovoltaic module hot spot defect detection process, realizes efficient and accurate multi-frequency excitation parameter configuration and environmental thermal noise suppression, significantly improves the defect identification accuracy and fault diagnosis reliability level in the photovoltaic module operation process, effectively prevents the insufficient response of fixed excitation parameter setting to actual thermal physical changes, avoids the depth deviation and defect misjudgment caused by experience judgment, effectively improves the pertinence of the final heat source distribution field reconstruction and the defect identification effect, determines the defect belonging to the packaging layer which has an influence on the performance of the photovoltaic module, and ensures that the defect layered identification measure is highly consistent with the actual heat source depth distribution risk.

Description

Technical Field

[0001] This invention relates to the field of defect identification technology, and more specifically to an automatic identification method for hot spot defects in photovoltaic modules based on thermal infrared imaging. Background Technology

[0002] As a core technology in the operation and maintenance management of photovoltaic power plants, hot spot defect detection of photovoltaic modules requires accurate location and identification of defects at different depths and differentiated diagnostic analysis at different encapsulation levels under multi-layer encapsulation structures. Therefore, it is essential to build an intelligent hot spot defect depth inversion system based on environmental adaptive excitation parameter optimization.

[0003] In existing technologies, hot spot detection of photovoltaic modules relies on fixed excitation parameter configuration and a single-frequency thermal imaging mode. This cannot cover the dynamic changes in material thermal properties and the adjustment law of heat wave conduction depth under different ambient temperatures and irradiance conditions. Although existing technologies have introduced basic infrared temperature monitoring mechanisms and hot spot area marking functions, due to the single excitation parameter setting and high dependence on empirical threshold configuration, it is difficult to achieve accurate judgment of the target module's real environmental response state, personalized thermal diffusivity matching, and hidden deep defects. When faced with complex multi-layer encapsulation structures or uneven thermal property distribution caused by environmental parameter fluctuations, problems such as insufficient excitation energy, missed detection of deep defects, and confusion between surface and deep defects are prone to occur.

[0004] Furthermore, existing technologies lack a differentiated depth inversion prediction mechanism for the differences in thermal properties and thermal wave frequency domain response characteristics of different encapsulation layer materials during the detection process. This makes it impossible to effectively optimize heat source location and identify personalized defect boundaries that vary depending on the encapsulation layer. In addition, there are shortcomings such as weak temperature-dependent thermal property correction capability, insufficient calculation of multi-frequency thermal wave penetration depth, lack of amplitude attenuation and phase delay coupling analysis, improper handling of ill-conditioned nature of heat source depth inversion equation, and low accuracy of defect layer identification. As a result, high-precision heat source depth values ​​cannot be effectively inverted and quality guaranteed. At the same time, there is a lack of adaptive boundary discrimination strategies for different encapsulation layer depth ranges, and it is impossible to adaptively adjust the defect attribution to the encapsulation layer based on the dynamic changes in the real-time main peak position parameters of the heat source depth distribution histogram and the dispersion of heat source intensity. Summary of the Invention

[0005] The purpose of this invention is to provide an automatic identification method for hot spot defects in photovoltaic modules based on thermal infrared imaging, which solves the problems existing in the background technology.

[0006] To solve the above technical problems, the present invention adopts the following technical solution: The present invention provides an automatic identification method for hot spot defects in photovoltaic modules based on thermal infrared imaging, including: Step 1: Obtain the encapsulation material type of the target photovoltaic module to be detected, collect the current ambient temperature and ambient irradiance, and construct an environmental adaptive excitation parameter model based on the encapsulation material type of the target photovoltaic module, the current ambient temperature and ambient irradiance, thereby determining the modulation frequency and modulation excitation power value of the target photovoltaic module.

[0007] Step 2: Based on the modulation frequencies and modulation excitation power values ​​of the target photovoltaic module, perform multi-frequency phase-locked thermal imaging tests on the target photovoltaic module to obtain temperature amplitude images and temperature phase images of the target photovoltaic module at each modulation frequency.

[0008] Step 3: Based on the current ambient temperature of the target photovoltaic module, construct the heat wave conduction transfer function of each encapsulation layer of the target photovoltaic module, and combine the temperature amplitude image and temperature phase image of the target photovoltaic module at each modulation frequency to perform heat source depth inversion solution, thereby reconstructing the heat source distribution field of the target photovoltaic module.

[0009] Step 4: Based on the heat source distribution field of the target photovoltaic module, identify the abnormal heat source regions of the target photovoltaic module and their depth and intensity distribution characteristics, and distinguish the defect regions of each encapsulation layer of the target photovoltaic module.

[0010] Step 5: Send the defective areas of each encapsulation layer of the target photovoltaic module to the operation and maintenance management personnel.

[0011] The beneficial effects of this invention are as follows: This invention improves the environmental adaptability accuracy and multi-frequency thermal wave penetration coverage dimension in the detection of hot spot defects in photovoltaic modules. By constructing the temperature dependence relationship of the thermal diffusivity coefficient between the target photovoltaic module encapsulation material type and the current ambient temperature, and by accurately calculating the critical modulation frequency and modulation excitation power value of each encapsulation layer interface, it can adapt to the differentiated thermal property response requirements and thermal wave penetration depth adjustment characteristics under different environmental conditions. This achieves efficient and accurate multi-frequency excitation parameter configuration and environmental thermal noise suppression. By improving the analysis accuracy of the heat source depth inversion model and the speed of encapsulation layer defect location, it significantly improves the defect identification accuracy and fault diagnosis reliability level in the operation and maintenance process of photovoltaic modules. By constructing a multi-modulation frequency temperature response prediction mechanism, it can predict the thermal conductivity differences of different encapsulation layers. The calculation strategy of dynamically adjusting the heat wave conduction transfer function and regularized constraint conditions based on the value and thickness difference effectively prevents insufficient response of fixed excitation parameter settings to actual thermal property changes, avoids depth deviation and defect misjudgment caused by experience judgment, and can adjust the heat source depth inversion iteration process in real time according to the amplitude and phase coupling information of the surface temperature feature vector and constraint conditions. This enables intelligent allocation and differentiated investment of computing resources, effectively improving the pertinence of the final heat source distribution field reconstruction and the defect identification effect. Through accurate identification and analysis of the main peak position and standard deviation of heat source intensity in the histogram of heat source depth distribution, combined with the determination of the depth range of each encapsulation layer and the peak sharpness threshold discrimination mechanism, the encapsulation layer to which the defect affecting the performance of photovoltaic modules belongs is determined, ensuring that the defect layer identification measures are highly consistent with the actual heat source depth distribution risk. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] Reference Figure 1As shown, the present invention provides an automatic identification method for hot spot defects in photovoltaic modules based on thermal infrared imaging, including: Step 1: Obtain the encapsulation material type of the target photovoltaic module to be detected, collect the current ambient temperature and ambient irradiance, and construct an environmental adaptive excitation parameter model based on the encapsulation material type of the target photovoltaic module, the current ambient temperature and ambient irradiance, thereby determining the modulation frequency and modulation excitation power value of the target photovoltaic module.

[0016] In one specific embodiment, the encapsulation material type of the target photovoltaic module to be tested is obtained, and the current ambient temperature and ambient irradiance are collected. The specific method is as follows: the encapsulation material type of the target photovoltaic module to be tested is obtained from the local database, and the current ambient temperature and ambient irradiance are obtained through sensors.

[0017] It should be noted that the local database is used to store the encapsulation material type of the target photovoltaic module to be tested, the temperature dependence curve of the thermal diffusivity of the encapsulation material type and the thickness sequence of each encapsulation layer, the thermal capacity parameter corresponding to the encapsulation material type, the surface absorptivity corresponding to the encapsulation material type, the preset signal-to-noise ratio threshold and the heat source excitation efficiency coefficient, the theoretical formula for the thermal wave penetration depth of periodic thermal excitation in an infinite half-space medium, the standard thermal properties parameters of each encapsulation layer of the target photovoltaic module in each temperature range, the thermal contact coefficient of each encapsulation layer interface of the target photovoltaic module, the physical constraint range of heat source depth and the non-negative constraint condition of heat source intensity, the regularization coefficient and the iterative convergence threshold, the heat source intensity deviation coefficient threshold, the ideal heat source intensity value of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency under the current ambient temperature, the number of depth segments, the depth range of each encapsulation layer of the target photovoltaic module, and the peak sharpness discrimination threshold.

[0018] In a specific embodiment of the present invention, an environmental adaptive excitation parameter model is constructed to determine the modulation frequency and modulation excitation power value of the target photovoltaic module. The specific method is as follows: the thermal diffusivity temperature dependence curve and the thickness sequence of each encapsulation layer corresponding to the encapsulation material type are obtained from the local database, and the current ambient temperature is mapped on the thermal diffusivity temperature dependence curve to obtain the current effective thermal diffusivity of the encapsulation material type of the target photovoltaic module.

[0019] It should be noted that the thermal diffusivity temperature dependence curve refers to the relationship between the thermal diffusivity of the encapsulation material and temperature. The horizontal axis represents the temperature value, and the vertical axis represents the thermal diffusivity value at the corresponding temperature. Different encapsulation material types correspond to different thermal diffusivity temperature dependence curves. The thickness sequence of each encapsulation layer refers to the numerical sequence formed by arranging the thickness values ​​of each encapsulation layer of the target photovoltaic module from the surface to the bottom in sequence. For example, the encapsulation layers of a photovoltaic module from the outside to the inside include the front glass layer, the front encapsulant layer, the cell layer, the rear encapsulant layer, and the backsheet layer.

[0020] Based on the current effective thermal diffusivity of the encapsulation material type of the target photovoltaic module and the thickness sequence of each encapsulation layer, a physical correlation formula between the thermal wave penetration depth and the modulation frequency of the target photovoltaic module is established. Based on this formula, the critical modulation frequency at which the thermal wave penetrates to the interface of each encapsulation layer of the target photovoltaic module is calculated and used as the modulation frequency of the target photovoltaic module.

[0021] The thermal capacity parameters corresponding to the type of encapsulation material are obtained from the local database. Based on the current ambient irradiance, the environmental thermal noise level of the target photovoltaic module is calculated, and the modulation excitation power value required for the target photovoltaic module to overcome environmental thermal noise interference is determined accordingly.

[0022] In one specific embodiment, the environmental thermal noise level of the target photovoltaic module is calculated, and the modulation excitation power value required for the target photovoltaic module to overcome environmental thermal noise interference is determined accordingly. The specific method is as follows: the surface absorptivity corresponding to the encapsulation material type is obtained from the local database; the current environmental irradiance is multiplied by the surface absorptivity to obtain the radiative heat flux density absorbed by the module surface; the radiative heat flux density is divided by the heat capacity parameter corresponding to the encapsulation material type to obtain the temperature change rate per unit time, and this is used as the environmental thermal noise level; a preset signal-to-noise ratio threshold and heat source excitation efficiency coefficient are obtained from the local database; the environmental thermal noise level is multiplied by the signal-to-noise ratio threshold to obtain the minimum temperature response amplitude; the minimum temperature response amplitude is multiplied by the heat capacity parameter and then divided by the heat source excitation efficiency coefficient to obtain the modulation excitation power value required for the target photovoltaic module to overcome environmental thermal noise interference.

[0023] In a specific embodiment of the present invention, a physical correlation formula between the thermal wave penetration depth and the modulation frequency of the target photovoltaic module is established, and the critical modulation frequency at which the thermal wave penetrates to each encapsulation layer interface of the target photovoltaic module is calculated accordingly. The specific method is as follows: the theoretical formula for the thermal wave penetration depth of periodic thermal excitation in an infinite half-space medium is obtained from a local database, and the thermal diffusivity parameter in the theoretical formula is assigned a value according to the current effective thermal diffusivity of the encapsulation material type of the target photovoltaic module, thereby obtaining the physical correlation formula between the thermal wave penetration depth and the modulation frequency of the target photovoltaic module.

[0024] It should be noted that the theoretical formula for the penetration depth of heat waves in an infinite half-space medium under periodic thermal excitation is a classic formula in heat conduction theory. It is expressed as the penetration depth of heat waves equals the square root of the product of the thermal diffusivity and the modulation frequency. This formula describes the physical relationship between the depth value corresponding to the decay of the heat wave amplitude to 1 / e of the surface amplitude under periodic thermal excitation, the modulation frequency, and the thermal diffusivity of the material. The formula shows that the lower the modulation frequency, the greater the penetration depth of the heat waves, and the higher the thermal diffusivity, the greater the penetration depth of the heat waves.

[0025] Based on the thickness sequence of each encapsulation layer of the target photovoltaic module, and combined with the physical correlation formula between the thermal wave penetration depth and the modulation frequency of the target photovoltaic module, the modulation frequency value that makes the thermal wave penetration depth equal to the cumulative thickness of each encapsulation layer is calculated, thereby obtaining the critical modulation frequency at which the thermal wave penetrates to the interface of each encapsulation layer of the target photovoltaic module.

[0026] In one specific embodiment, the modulation frequency value that makes the heat wave penetration depth equal to the cumulative thickness of each encapsulation layer is calculated. The specific method is as follows: Based on the thickness sequence of each encapsulation layer of the target photovoltaic module, the cumulative thickness of each encapsulation layer is calculated. The cumulative thickness of the first encapsulation layer is equal to the thickness of the first encapsulation layer, the cumulative thickness of the second encapsulation layer is equal to the sum of the thicknesses of the first and second encapsulation layers, and so on until the cumulative thickness of all encapsulation layers is calculated. Based on the physical correlation formula between heat wave penetration depth and modulation frequency, the heat wave penetration depth is set to a certain cumulative thickness value of an encapsulation layer. The formula is transformed to express the modulation frequency as the product of the thermal diffusivity coefficient divided by pi and the square of the cumulative thickness of the encapsulation layer. The current effective thermal diffusivity coefficient and the cumulative thickness of the encapsulation layer are substituted into the formula to calculate the corresponding critical modulation frequency value. The above calculation process is repeated for each cumulative thickness of the encapsulation layer to obtain the critical modulation frequency of each encapsulation layer interface of the target photovoltaic module.

[0027] Step 2: Based on the modulation frequencies and modulation excitation power values ​​of the target photovoltaic module, perform multi-frequency phase-locked thermal imaging tests on the target photovoltaic module to obtain temperature amplitude images and temperature phase images of the target photovoltaic module at each modulation frequency.

[0028] In a specific embodiment of the present invention, multi-frequency phase-locked thermal imaging tests are performed to obtain temperature amplitude images and temperature phase images of the target photovoltaic module at each modulation frequency. The specific method is as follows: based on each modulation frequency and modulation excitation power value of the target photovoltaic module, an external modulation heat source is controlled to apply periodic sinusoidal modulation heating to the target photovoltaic module, and high frame rate continuous thermal infrared image acquisition is performed simultaneously to obtain the temperature time sequence signal of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency.

[0029] In one specific embodiment, periodic sinusoidal modulation heating is applied. The specific method is as follows: based on the modulation frequency and modulation excitation power value of the target photovoltaic module, a corresponding sinusoidal power modulation signal is generated for each modulation frequency, so that the power signal changes periodically between zero and the modulation excitation power value.

[0030] Phase-locked correlation analysis is performed on the temperature time-series signals of each pixel position in the thermal infrared images of the target photovoltaic module at various modulation frequencies to extract the temperature amplitude and temperature phase values ​​of each pixel position in the thermal infrared images of the target photovoltaic module at various modulation frequencies. These values ​​are then combined through spatial pixel reconstruction to generate temperature amplitude images and temperature phase images of the target photovoltaic module at various modulation frequencies.

[0031] In a specific embodiment of the present invention, phase-locked correlation analysis is performed to extract the temperature amplitude and temperature phase values ​​of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency. The specific method is as follows: based on each modulation frequency of the target photovoltaic module, in-phase reference signals and quadrature reference signals corresponding to each modulation frequency of the target photovoltaic module are constructed respectively, wherein the in-phase reference signal is a standard sine wave signal and the quadrature reference signal is a standard cosine wave signal with a phase lead of 90 degrees.

[0032] In one specific embodiment, the in-phase reference signal and quadrature reference signal corresponding to each modulation frequency of the target photovoltaic module are constructed as follows: The in-phase reference signal is constructed by multiplying twice pi by the modulation frequency and the sampling time to obtain the phase angle, taking the sine function value of the phase angle as the in-phase reference signal value at the sampling time, and traversing all sampling times to generate a complete in-phase reference signal sequence. The quadrature reference signal is constructed by multiplying twice pi by the modulation frequency and the sampling time to obtain the phase angle, taking the cosine function value of the phase angle as the quadrature reference signal value at the sampling time, and traversing all sampling times to generate a complete quadrature reference signal sequence.

[0033] The temperature time-series signal of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency is subjected to discrete Fourier correlation operation with the in-phase reference signal and the quadrature reference signal of the corresponding modulation frequency, so as to obtain the in-phase component and the quadrature component of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency.

[0034] In one specific embodiment, discrete Fourier correlation is performed to obtain the in-phase and quadrature components of each pixel position in the thermal infrared image of the target photovoltaic module at various modulation frequencies. Specifically, for a pixel position in the thermal infrared image of the target photovoltaic module at a certain modulation frequency, the temperature time-series signal sequence of that pixel position, as well as the in-phase reference signal sequence and the quadrature reference signal sequence corresponding to that modulation frequency, are obtained. The in-phase component is calculated by comparing the temperature value of each sampling point in the temperature time-series signal sequence with the corresponding sampling value in the in-phase reference signal sequence. The signal values ​​of each sampling point are multiplied point by point, and the product results of all sampling points are summed. The summation result is then divided by the total number of sampling points and multiplied by 2 to obtain the in-phase component. The quadrature component is calculated as follows: the temperature value of each sampling point in the temperature time series signal sequence is multiplied point by point with the signal value of the corresponding sampling point in the quadrature reference signal sequence, and the product results of all sampling points are summed. The summation result is then divided by the total number of sampling points and multiplied by 2 to obtain the quadrature component. Thus, the in-phase component and quadrature component of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency are obtained.

[0035] Complex vectors are constructed based on the in-phase and quadrature components of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency. The magnitude of the complex vector is used as the temperature amplitude value of the corresponding pixel position, and the argument of the complex vector is used as the temperature phase value of the corresponding pixel position. Thus, the temperature amplitude and temperature phase values ​​of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency are obtained.

[0036] In one specific embodiment, a complex vector is constructed by: based on the in-phase component and quadrature component of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency, the in-phase component is taken as the real part of the complex number, the quadrature component is taken as the imaginary part of the complex number, and the real part and the imaginary part are combined to construct a complex vector.

[0037] It should be noted that phase-locked thermal imaging technology is an existing technology. Its principle is to perform correlation operations between the temperature time-series signal and a reference signal of known frequency, extract the signal component synchronized with the excitation frequency, and filter out asynchronous noise. When performing phase-locked correlation analysis, the sampling frame rate needs to be much higher than the modulation frequency, and the sampling duration needs to cover multiple complete excitation cycles.

[0038] Step 3: Based on the current ambient temperature of the target photovoltaic module, construct the heat wave conduction transfer function of each encapsulation layer of the target photovoltaic module, and combine the temperature amplitude image and temperature phase image of the target photovoltaic module at each modulation frequency to perform heat source depth inversion solution, thereby reconstructing the heat source distribution field of the target photovoltaic module.

[0039] In a specific embodiment of the present invention, the thermal wave conduction transfer function of each encapsulation layer of the target photovoltaic module is constructed by: obtaining the standard thermal property parameters of each encapsulation layer of the target photovoltaic module in each temperature range from the local database, and mapping the current thermal property parameters of each encapsulation layer of the target photovoltaic module according to the current ambient temperature.

[0040] It should be noted that standard thermophysical parameters include standard thermal conductivity, standard specific heat capacity, and standard density.

[0041] The thermal contact coefficients of each encapsulation layer interface of the target photovoltaic module are obtained from the local database. Based on the current thermal properties of each encapsulation layer of the target photovoltaic module, the temperature jump and heat transfer at the adjacent interfaces of each encapsulation layer of the target photovoltaic module are calculated.

[0042] In one specific embodiment, the temperature jump variable and heat transfer amount at the adjacent interfaces of each encapsulation layer of the target photovoltaic module are calculated as follows: based on the current thermal properties of each encapsulation layer of the target photovoltaic module, for a certain adjacent encapsulation layer interface, the heat flux density through the interface is calculated. The heat flux density is equal to the current thermal conductivity of the encapsulation layer on the upper side of the interface multiplied by the temperature gradient inside the encapsulation layer. The temperature gradient is the temperature difference between two adjacent points inside the encapsulation layer divided by the distance between the two points. The temperature jump variable at the interface is obtained by dividing the heat flux density through the interface by the thermal contact coefficient of the interface. The temperature jump variable represents the temperature discontinuity on both sides of the interface. The heat transfer amount is equal to the heat flux density value through the interface.

[0043] Based on the temperature jump and heat transfer at the adjacent interfaces of each encapsulation layer of the target photovoltaic module, a physical model of the medium heat transfer of each encapsulation layer of the target photovoltaic module is constructed.

[0044] In one specific embodiment, a physical model of the medium heat transfer of each encapsulation layer of the target photovoltaic module is constructed. The specific method is as follows: based on the current thermal properties of each encapsulation layer of the target photovoltaic module, a one-dimensional Fourier heat conduction equation is established for each encapsulation layer as the temperature field control relationship inside the encapsulation layer. The equation is expressed as the rate of change of temperature with respect to time is equal to the thermal diffusivity multiplied by the second-order spatial rate of change of temperature with respect to depth. The temperature jump variable at the adjacent interface of each encapsulation layer of the target photovoltaic module is used as the temperature discontinuity boundary condition at the junction of adjacent encapsulation layers. The heat flow transfer at the adjacent interface of each encapsulation layer of the target photovoltaic module is used as the heat flow continuity boundary condition at the junction of adjacent encapsulation layers. The heat conduction equation, temperature discontinuity boundary condition and heat flow continuity boundary condition of each encapsulation layer are combined to construct the physical model of the medium heat transfer of each encapsulation layer of the target photovoltaic module.

[0045] The thermal wave conduction transfer function of each encapsulation layer of the target photovoltaic module is obtained by performing a frequency domain Laplace transform on the physical model of the dielectric heat transfer of each encapsulation layer.

[0046] In one specific embodiment, a frequency-domain Laplace transform is performed to obtain the heat wave conduction transfer function of each encapsulation layer of the target photovoltaic module. The specific method is as follows: based on the physical model of the heat transfer of each encapsulation layer of the target photovoltaic module, a Laplace transform is performed on the time-domain temperature variable of the heat conduction equation of each encapsulation layer in the model. The transformation method is to replace the time-domain representation of temperature with the complex frequency-domain representation of temperature, and replace the first derivative of temperature with time with the product of the complex frequency variable and the complex frequency-domain representation of temperature. The complex frequency variable is taken as the product of twice pi and the modulation frequency and then multiplied by the imaginary unit. After the transformation, the heat conduction equation of each encapsulation layer is transformed from a time-domain partial differential equation into a complex frequency-domain ordinary differential equation. The temperature discontinuity boundary condition and the heat flow continuity boundary condition are also processed by Laplace transform. The transformed complex frequency-domain equation system is solved to obtain the complex representation of the surface temperature response generated by a unit intensity heat source at each depth position inside the module at each modulation frequency. This complex representation is used as the heat wave conduction transfer function of each encapsulation layer of the target photovoltaic module. The modulus of the transfer function represents the proportional relationship between the surface temperature amplitude and the heat source intensity.

[0047] In a specific embodiment of the present invention, a heat source depth inversion solution is performed to reconstruct the heat source distribution field of the target photovoltaic module. The specific method is as follows: based on the temperature amplitude value and temperature phase value of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency, a vector combination is performed to construct the surface temperature feature vector of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency.

[0048] In one specific embodiment, a surface temperature feature vector is constructed for each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency. The specific method is as follows: for a certain pixel position in the thermal infrared image of the target photovoltaic module, the temperature amplitude value and temperature phase value of the pixel position at each modulation frequency are obtained, and the modulation frequencies are arranged and numbered in order from low to high. Let there be a total of N modulation frequencies. The method for constructing the surface temperature feature vector is as follows: Create an empty vector with dimension 2N. Fill the first position of the vector with the temperature amplitude value at the first modulation frequency, fill the second position with the temperature phase value at the first modulation frequency, fill the third position with the temperature amplitude value at the second modulation frequency, fill the fourth position with the temperature phase value at the second modulation frequency, and so on until the temperature amplitude value at the Nth modulation frequency is filled into the (2N-1)th position and the temperature phase value is filled into the 2Nth position. This completes the construction of the surface temperature feature vector for that pixel position. By repeating this process, the surface temperature feature vectors of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency can be obtained.

[0049] The physical constraint range of heat source depth and the non-negative constraint condition of heat source intensity are obtained from the local database. Based on the heat wave conduction transfer function of each encapsulation layer of the target photovoltaic module, and combined with the surface temperature feature vector of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency, a regularized least squares iterative solution with constraints is performed to obtain the heat source depth value and heat source intensity value of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency, and these are used as the heat source distribution field of the target photovoltaic module.

[0050] It should be noted that the physical constraint range of heat source depth refers to the reasonable range of values ​​for heat source depth. The lower limit of the heat source depth value is zero, which is the position on the surface of the component. The upper limit of the heat source depth value is obtained by summing the thickness values ​​of all encapsulation layers in the thickness sequence of each encapsulation layer, which is the total thickness of the component. The physical constraint range of heat source depth is represented as a closed interval from zero to the total thickness of the component. This constraint ensures that the heat source depth value obtained by inversion is located inside the component rather than in an invalid region outside the component. The non-negative constraint condition of heat source intensity refers to the mathematical constraint that the heat source intensity value must be greater than or equal to zero. Because hot spot defects manifest as local abnormal heating phenomena, the defect area generates additional heat relative to the normal area, and its heat source intensity value cannot be negative physically.

[0051] In one specific embodiment, a constrained regularized least squares iterative solution is performed to obtain the heat source depth and heat source intensity values ​​of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency. The specific method is as follows: Based on the heat wave conduction transfer function of each encapsulation layer of the target photovoltaic module, the heat source depth value to be determined is set to 'd' and the heat source intensity value to be determined is set to 'Q' for a certain pixel position. For the measured values ​​of temperature amplitude and temperature phase corresponding to the i-th modulation frequency in the surface temperature feature vector, the heat source depth value 'd' and the i-th modulation frequency are substituted into the heat wave conduction transfer function to obtain the transfer function value. The magnitude of the transfer function value is multiplied by the heat source intensity value 'Q' to obtain the theoretical value of temperature amplitude at that modulation frequency. The argument of the transfer function value is used as the theoretical value of temperature phase at that modulation frequency.

[0052] The sum of the squared differences between the theoretical and measured temperature amplitude values ​​at each modulation frequency is added to the sum of the squared differences between the theoretical and measured temperature phase values ​​at each modulation frequency. Finally, the product of the regularization coefficient and the sum of the squared values ​​of the heat source depth and the heat source intensity is added as the regularization penalty term.

[0053] The regularization coefficient and iteration convergence threshold are obtained from the local database. Initial guesses for the heat source depth and heat source intensity are set. The partial derivatives of the objective function with respect to the heat source depth and heat source intensity are calculated as gradient components. The current heat source depth and heat source intensity are updated along their respective negative gradient directions. If the updated heat source depth is less than zero, it is set to zero; if it is greater than the total component thickness, it is set to the total component thickness. If the updated heat source intensity is less than zero, it is set to zero. The absolute value of the difference between the objective function value of the current iteration and the previous iteration is calculated. If the absolute value is less than the convergence threshold, the iteration is stopped and the current heat source depth and heat source intensity are output. The above solution process is repeated for each pixel position to obtain the heat source distribution field.

[0054] Step 4: Based on the heat source distribution field of the target photovoltaic module, identify the abnormal heat source regions of the target photovoltaic module and their depth and intensity distribution characteristics, and distinguish the defect regions of each encapsulation layer of the target photovoltaic module.

[0055] In a specific embodiment of the present invention, the method for identifying abnormal heat source regions of a target photovoltaic module and their depth and intensity distribution characteristics is as follows: based on the heat source intensity value of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency, the heat source intensity deviation coefficient of each pixel position in the thermal infrared image of the target photovoltaic module is calculated, the heat source intensity deviation coefficient threshold is obtained from the local database, and if the heat source intensity deviation coefficient of a certain pixel position in the thermal infrared image of the target photovoltaic module is greater than the heat source intensity deviation coefficient threshold, the pixel position is marked as the target pixel position, thereby filtering each target pixel position in the thermal infrared image of the target photovoltaic module, and performing connected component spatial clustering analysis on adjacent target pixel positions to identify each abnormal heat source region of the target photovoltaic module.

[0056] In one specific embodiment, the method for calculating the heat source intensity deviation coefficient at each pixel location in the thermal infrared image of the target photovoltaic module is as follows: Obtain the ideal heat source intensity value at each pixel location in the thermal infrared image of the target photovoltaic module at each modulation frequency under the current ambient temperature from a local database. Where N represents the number of each modulation frequency, M is a positive integer greater than 2, and P represents the number of each pixel. Q is a positive integer greater than 2, and is based on the heat source intensity value of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency. Calculate the heat source intensity deviation coefficient at each pixel location in the thermal infrared image of the target photovoltaic module. .

[0057] In one specific embodiment, connected domain spatial clustering analysis is performed to identify the abnormal heat source regions of the target photovoltaic module. The specific method is as follows: the abnormal heat source regions of the target photovoltaic module can be identified by using existing connected domain spatial clustering analysis technology.

[0058] Based on the heat source distribution field of the target photovoltaic module, the heat source depth value and heat source intensity value of each pixel position in each abnormal heat source region of the target photovoltaic module are extracted. Based on this, the heat source depth distribution histogram and heat source intensity spatial distribution map of each abnormal heat source region of the target photovoltaic module are constructed respectively, and these are used as the depth distribution feature and intensity distribution feature of each abnormal heat source region of the target photovoltaic module.

[0059] In one specific embodiment, a histogram of heat source depth distribution and a spatial distribution map of heat source intensity are constructed for each abnormal heat source region of the target photovoltaic module. The specific method is as follows: For a certain abnormal heat source region, the heat source depth value and heat source intensity value of each pixel position in the region are extracted from the heat source distribution field. The number of depth segments is obtained from the local database. According to the thickness sequence of each encapsulation layer, all encapsulation layer thickness values ​​are summed sequentially to obtain the total thickness of the module. The total thickness of the module is divided by the number of depth segments to obtain the segment width value of a single depth. A counting array with the same length as the number of depth segments is created and all elements are initialized to zero. The heat source depth value of each pixel position in the region is traversed. The depth value is divided by the bin width and rounded down to obtain the segment index number to which the depth value belongs. The count value at the index number position in the counting array is incremented by 1. After the statistics are completed, a heat source depth distribution histogram is drawn with the center depth value of each segment as the x-axis and the corresponding count value as the y-axis.

[0060] The method for constructing the spatial distribution map of heat source intensity is as follows: determine the minimum row coordinate, maximum row coordinate, minimum column coordinate, and maximum column coordinate of all pixel positions in the abnormal heat source region to determine the minimum bounding rectangle of the region, create an image matrix with the same row and column dimensions as the bounding rectangle and initialize all elements to zero, traverse each pixel position in the region, and fill the corresponding row and column coordinate positions of the pixel position into the image matrix.

[0061] In a specific embodiment of the present invention, the defective regions of each encapsulation layer of the target photovoltaic module are distinguished by the following method: based on the histogram of the heat source depth distribution and the spatial distribution map of the heat source intensity of each abnormal heat source region of the target photovoltaic module, the main peak depth position and peak sharpness of each abnormal heat source region of the target photovoltaic module are analyzed.

[0062] The depth range and peak sharpness discrimination threshold of each encapsulation layer of the target photovoltaic module are obtained from the local database. Based on the main peak depth position and peak sharpness of each abnormal heat source region of the target photovoltaic module, if the main peak depth position of an abnormal heat source region of the target photovoltaic module is within the depth range of a certain encapsulation layer and the peak sharpness exceeds the peak sharpness discrimination threshold, then the abnormal heat source region is marked as a defect region of the encapsulation layer, thereby filtering each defect region of each encapsulation layer of the target photovoltaic module.

[0063] In a specific embodiment of the present invention, the main peak depth position and peak sharpness of each abnormal heat source region of the target photovoltaic module are extracted. The specific method is as follows: peak detection is performed on the heat source depth distribution histogram of each abnormal heat source region of the target photovoltaic module, the depth value corresponding to the maximum peak value in the heat source depth distribution histogram of each abnormal heat source region is identified, and it is used as the main peak depth position of each abnormal heat source region of the target photovoltaic module.

[0064] Based on the spatial distribution map of heat source intensity in each abnormal heat source region of the target photovoltaic module, the standard deviation of heat source intensity in each abnormal heat source region of the target photovoltaic module is calculated, and it is used as the peak sharpness of each abnormal heat source region of the target photovoltaic module.

[0065] In one specific embodiment, the standard deviation of heat source intensity of each abnormal heat source region of the target photovoltaic module is calculated. The specific method is as follows: based on the spatial distribution map of heat source intensity of each abnormal heat source region of the target photovoltaic module, for a certain abnormal heat source region, the heat source intensity value of each pixel position in the abnormal heat source region is extracted, and the standard deviation of heat source intensity of each abnormal heat source region of the target photovoltaic module is obtained according to the existing standard deviation calculation method.

[0066] Step 5: Send the defective areas of each encapsulation layer of the target photovoltaic module to the operation and maintenance management personnel.

[0067] All formulas in this manual are dimensionless and calculated numerically. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0068] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. An automatic identification method for hot spot defects in photovoltaic modules based on thermal infrared imaging, characterized in that, include: Step 1: Obtain the encapsulation material type of the target photovoltaic module to be tested, collect the current ambient temperature and ambient irradiance, and construct an environmental adaptive excitation parameter model based on the encapsulation material type of the target photovoltaic module, the current ambient temperature and ambient irradiance, so as to determine the modulation frequency and modulation excitation power value of the target photovoltaic module; The method for constructing an environmental adaptive excitation parameter model to determine the modulation frequencies and modulation excitation power values ​​of the target photovoltaic module is as follows: The thermal diffusivity temperature dependence curve and the thickness sequence of each encapsulation layer corresponding to the encapsulation material type are obtained from the local database. The current ambient temperature is then mapped onto the thermal diffusivity temperature dependence curve to obtain the current effective thermal diffusivity of the encapsulation material type of the target photovoltaic module. Based on the current effective thermal diffusivity of the encapsulation material type of the target photovoltaic module and the thickness sequence of each encapsulation layer, a physical correlation formula between the thermal wave penetration depth and the modulation frequency of the target photovoltaic module is established, and the critical modulation frequency at which the thermal wave penetrates to the interface of each encapsulation layer of the target photovoltaic module is calculated accordingly, thereby generating each modulation frequency of the target photovoltaic module. The thermal capacity parameters corresponding to the type of encapsulation material are obtained from the local database. Based on the current ambient irradiance, the environmental thermal noise level of the target photovoltaic module is calculated, and the modulation excitation power value required for the target photovoltaic module to overcome environmental thermal noise interference is determined accordingly. Step 2: Based on the modulation frequencies and modulation excitation power values ​​of the target photovoltaic module, perform multi-frequency phase-locked thermal imaging tests on the target photovoltaic module to obtain temperature amplitude images and temperature phase images of the target photovoltaic module at each modulation frequency; Step 3: Based on the current ambient temperature of the target photovoltaic module, construct the heat wave conduction transfer function of each encapsulation layer of the target photovoltaic module, and combine the temperature amplitude image and temperature phase image of the target photovoltaic module at each modulation frequency to perform heat source depth inversion solution, thereby reconstructing the heat source distribution field of the target photovoltaic module; The specific method for constructing the heat wave conduction transfer function of each encapsulation layer of the target photovoltaic module is as follows: The standard thermal properties of each encapsulation layer of the target photovoltaic module in each temperature range are obtained from the local database, and the current thermal properties of each encapsulation layer of the target photovoltaic module are mapped according to the current ambient temperature. The thermal contact coefficients of each encapsulation layer interface of the target photovoltaic module are obtained from the local database. Based on the current thermal properties of each encapsulation layer of the target photovoltaic module, the temperature jump and heat transfer at the adjacent interfaces of each encapsulation layer of the target photovoltaic module are calculated. Based on the temperature jump and heat transfer at the adjacent interfaces of each encapsulation layer of the target photovoltaic module, a physical model of the medium heat transfer of each encapsulation layer of the target photovoltaic module is constructed. The thermal transfer physical model of each encapsulation layer of the target photovoltaic module is subjected to frequency domain Laplace transform to obtain the thermal wave conduction transfer function of each encapsulation layer of the target photovoltaic module. Step 4: Based on the heat source distribution field of the target photovoltaic module, identify the abnormal heat source regions of the target photovoltaic module and their depth and intensity distribution characteristics, and distinguish the defect regions of each encapsulation layer of the target photovoltaic module. Step 5: Send the defective areas of each encapsulation layer of the target photovoltaic module to the operation and maintenance management personnel.

2. The method for automatic identification of hot spot defects in photovoltaic modules based on thermal infrared imaging according to claim 1, characterized in that, The specific method for establishing the physical relationship between the heat wave penetration depth and modulation frequency of the target photovoltaic module, and calculating the critical modulation frequency at which heat waves penetrate to each encapsulation layer interface of the target photovoltaic module, is as follows: The theoretical formula for the thermal wave penetration depth of periodic thermal excitation in an infinite half-space medium is obtained from the local database. The thermal diffusivity parameter in the theoretical formula is assigned a value according to the current effective thermal diffusivity of the encapsulation material type of the target photovoltaic module, thereby obtaining the physical relationship formula between the thermal wave penetration depth of the target photovoltaic module and the modulation frequency. Based on the thickness sequence of each encapsulation layer of the target photovoltaic module, and combined with the physical correlation formula between the thermal wave penetration depth and the modulation frequency of the target photovoltaic module, the modulation frequency value that makes the thermal wave penetration depth equal to the cumulative thickness of each encapsulation layer is calculated, thereby obtaining the critical modulation frequency at which the thermal wave penetrates to the interface of each encapsulation layer of the target photovoltaic module.

3. The method for automatic identification of hot spot defects in photovoltaic modules based on thermal infrared imaging according to claim 1, characterized in that, The specific method for performing multi-frequency phase-locked thermal imaging tests to obtain temperature amplitude and temperature phase images of the target photovoltaic module at various modulation frequencies is as follows: Based on the modulation frequency and modulation excitation power value of the target photovoltaic module, the external modulation heat source is controlled to apply periodic sinusoidal modulation heating to the target photovoltaic module, and high frame rate continuous thermal infrared image acquisition is performed simultaneously to obtain the temperature time sequence signal of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency. Phase-locked correlation analysis is performed on the temperature time-series signals of each pixel position in the thermal infrared images of the target photovoltaic module at various modulation frequencies to extract the temperature amplitude and temperature phase values ​​of each pixel position in the thermal infrared images of the target photovoltaic module at various modulation frequencies. These values ​​are then combined through spatial pixel reconstruction to generate temperature amplitude images and temperature phase images of the target photovoltaic module at various modulation frequencies.

4. The method for automatic identification of hot spot defects in photovoltaic modules based on thermal infrared imaging according to claim 3, characterized in that, The specific method for performing phase-locked correlation analysis to extract the temperature amplitude and temperature phase values ​​of each pixel in the thermal infrared image of the target photovoltaic module at various modulation frequencies is as follows: Based on the modulation frequencies of the target photovoltaic module, in-phase reference signals and quadrature reference signals corresponding to each modulation frequency of the target photovoltaic module are constructed respectively. The in-phase reference signal is a standard sine wave signal, and the quadrature reference signal is a standard cosine wave signal with a phase lead of 90 degrees. The temperature time-series signal of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency is subjected to discrete Fourier correlation operation with the in-phase reference signal and the quadrature reference signal of the corresponding modulation frequency, so as to obtain the in-phase component and the quadrature component of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency. Complex vectors are constructed based on the in-phase and quadrature components of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency. The magnitude of the complex vector is used as the temperature amplitude value of the corresponding pixel position, and the argument of the complex vector is used as the temperature phase value of the corresponding pixel position. Thus, the temperature amplitude and temperature phase values ​​of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency are obtained.

5. The method for automatic identification of hot spot defects in photovoltaic modules based on thermal infrared imaging according to claim 3, characterized in that, The specific method for performing deep inversion of heat sources to reconstruct the heat source distribution field of the target photovoltaic module is as follows: The surface temperature feature vector of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency is constructed by combining the temperature amplitude and temperature phase values ​​of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency. The physical constraint range of heat source depth and the non-negative constraint condition of heat source intensity are obtained from the local database. Based on the heat wave conduction transfer function of each encapsulation layer of the target photovoltaic module, and combined with the surface temperature feature vector of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency, a regularized least squares iterative solution with constraints is performed to obtain the heat source depth value and heat source intensity value of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency, and these are used as the heat source distribution field of the target photovoltaic module.

6. The method for automatic identification of hot spot defects in photovoltaic modules based on thermal infrared imaging according to claim 5, characterized in that, The specific method for identifying the abnormal heat source regions of the target photovoltaic module and their depth and intensity distribution characteristics is as follows: Based on the heat source intensity values ​​of each pixel position in the thermal infrared image of the target photovoltaic module at each modulation frequency, the heat source intensity deviation coefficient of each pixel position in the thermal infrared image of the target photovoltaic module is calculated. The heat source intensity deviation coefficient threshold is obtained from the local database. If the heat source intensity deviation coefficient of a certain pixel position in the thermal infrared image of the target photovoltaic module is greater than the heat source intensity deviation coefficient threshold, the pixel position is marked as the target pixel position. This process is used to filter the target pixel positions in the thermal infrared image of the target photovoltaic module. Connected domain spatial clustering analysis is performed on adjacent target pixel positions to identify the abnormal heat source regions of the target photovoltaic module. Based on the heat source distribution field of the target photovoltaic module, the heat source depth value and heat source intensity value of each pixel position in each abnormal heat source region of the target photovoltaic module are extracted. Based on this, the heat source depth distribution histogram and heat source intensity spatial distribution map of each abnormal heat source region of the target photovoltaic module are constructed respectively, and these are used as the depth distribution feature and intensity distribution feature of each abnormal heat source region of the target photovoltaic module.

7. The method for automatic identification of hot spot defects in photovoltaic modules based on thermal infrared imaging according to claim 6, characterized in that, The specific method for distinguishing the defect regions of each encapsulation layer of the target photovoltaic module is as follows: Based on the histogram of heat source depth distribution and the spatial distribution map of heat source intensity of each abnormal heat source region of the target photovoltaic module, the main peak depth position and peak sharpness of each abnormal heat source region of the target photovoltaic module are analyzed. The depth range and peak sharpness discrimination threshold of each encapsulation layer of the target photovoltaic module are obtained from the local database. Based on the main peak depth position and peak sharpness of each abnormal heat source region of the target photovoltaic module, if the main peak depth position of an abnormal heat source region of the target photovoltaic module is within the depth range of a certain encapsulation layer and the peak sharpness exceeds the peak sharpness discrimination threshold, then the abnormal heat source region is marked as a defect region of the encapsulation layer, thereby filtering each defect region of each encapsulation layer of the target photovoltaic module.

8. The method for automatic identification of hot spot defects in photovoltaic modules based on thermal infrared imaging according to claim 7, characterized in that, The specific method for extracting the main peak depth position and peak sharpness of each abnormal heat source region of the target photovoltaic module is as follows: Peak detection is performed on the histogram of heat source depth distribution in each abnormal heat source region of the target photovoltaic module. The depth value corresponding to the maximum peak value in the histogram of heat source depth distribution in each abnormal heat source region is identified and used as the main peak depth position of each abnormal heat source region of the target photovoltaic module. Based on the spatial distribution map of heat source intensity in each abnormal heat source region of the target photovoltaic module, the standard deviation of heat source intensity in each abnormal heat source region of the target photovoltaic module is calculated, and it is used as the peak sharpness of each abnormal heat source region of the target photovoltaic module.

Images