A method for detecting photoluminescence defects of photovoltaic modules in bright environment

Abstract

The present invention discloses a method for detecting photoluminescent defects in photovoltaic modules under bright conditions, comprising: connecting a programmable power supply to an industrial computer, setting a modulation frequency and modulation current on the host computer of the industrial computer to generate a square wave modulated electrical signal; connecting the current output of the programmable power supply to a light source module, and driving the light source to generate a square wave modulated optical signal; aiming the light source at the photovoltaic module to be tested, optically modulating the photovoltaic module; aiming an optical imaging module at the photovoltaic module to be tested, and imaging the composite radiation light on the surface of the photovoltaic module; caching the composite radiation light image on the surface of the photovoltaic module as an image sequence, applying a defect saliency enhancement algorithm to the image sequence to filter out the DC component in the ambient light, extracting defect information that is submerged by the ambient light, and obtaining a photovoltaic module defect image; and displaying the photovoltaic module defect image on the host computer. The present invention has the advantages of strong resistance to ambient light interference, fast defect display rate, convenient operation, and high reliability.

Description

Technical Field

[0001] The present invention belongs to the field of automatic optical detection, and in particular relates to a method for detecting photoluminescence defects of photovoltaic modules in a bright environment. Background Art

[0002] In recent years, photovoltaic power generation has gained widespread recognition as a clean, environmentally friendly, and renewable energy technology. With the development of related industries and the advancement of industrialization, defect detection technology for photovoltaic modules has become increasingly mature and efficient. Photovoltaic modules, also known as solar panels, are mostly made of silicon semiconductor materials. When stimulated by external light, they gain energy, and the non-equilibrium minority carriers within the semiconductor continuously recombine with majority carriers to produce light. Based on this characteristic, photoluminescence (PL) technology has emerged as a promising technology for photovoltaic module defect detection.

[0003] PL technology is a type of automated optical inspection (AOI) technology. It uses a dedicated digital camera to convert PV module radiation into digital images, with differences in the image's grayscale values ​​reflecting the module's defect level. The PL phenomenon in PV modules is relatively weak, but during the day, ambient light with the same wavelength as the PV module's composite radiation exists. As the intensity of ambient light increases, it easily overwhelms the PL phenomenon in PV modules. Therefore, PL inspections performed in bright daylight conditions cannot effectively capture defect information in PV modules. Consequently, current PL technology is typically applied in low-light environments. Consequently, most photoluminescence equipment, both domestically and internationally, is only used to inspect process wafers in PV module production, such as silicon ingots, silicon wafers, or cells before they are assembled into PV modules. These objects are typically inspected in sealed darkrooms. Consequently, PL-related technical solutions, both domestically and internationally, cannot detect PV module defects in bright daylight conditions, and are limited to low-light environments. After photovoltaic modules are mass-produced and put into outdoor use, regular on-site inspection and maintenance of distributed photovoltaic modules has become a market demand. Currently, contact electroluminescence (EL) technology is usually used for outdoor inspection. Relatively speaking, EL technology has low detection efficiency and does not have the fast and non-contact technical advantages of PL.

[0004] Therefore, improving the adaptability of PL technology to ambient light and realizing PL detection of photovoltaic modules in bright environments will be a research focus in this field in the future. Summary of the Invention

[0005] The purpose of the present invention is to provide a method for detecting photoluminescence defects of photovoltaic modules in a bright environment, which can carry out photoluminescence detection of photovoltaic modules under daytime ambient light, improve the adaptability of photoluminescence detection of photovoltaic modules to ambient light, and increase the significance of defects of photovoltaic modules in a bright environment.

[0006] The technical solution for achieving the purpose of the present invention is: a method for detecting photoluminescence defects of photovoltaic modules in a bright environment, comprising the following steps:

[0007] Step 1: Connect the programmable power supply to the industrial computer, set the modulation frequency and modulation current on the host computer of the industrial computer, and the programmable power supply generates a square wave modulation electrical signal;

[0008] Step 2: Connect the current output of the programmable power supply to the light source module, and drive the light source to generate a square wave modulated light signal;

[0009] Step 3: Aim the light source at the photovoltaic module to be tested, and light modulate the photovoltaic module;

[0010] Step 4: Align the optical imaging module with the photovoltaic module to be tested, and image the composite radiation light on the surface of the photovoltaic module;

[0011] Step 5: Cache the composite radiation light image of the photovoltaic module surface that meets the modulation characteristics in the form of an image sequence. Apply a defect saliency enhancement algorithm to the image sequence to filter out the DC component in the ambient light, extract the defect information submerged by the ambient light, and obtain the photovoltaic module defect image.

[0012] Step 6: Connect the output end of the optical imaging module to the input end of the industrial computer, and display the defect image of the photovoltaic module through the host computer.

[0013] Furthermore, the program-controlled power supply is a programmable power supply, the modulation frequency is higher than 1 Hz, and the square wave duty cycle is 50%.

[0014] Furthermore, the step 1 specifically includes:

[0015] Step 1-1, connect the industrial computer and the programmable power supply via the serial communication line;

[0016] Step 1-2: Configure the communication port, slave address, modulation high / low current and modulation frequency on the host computer of the industrial computer;

[0017] Steps 1-3: Generate a square wave modulated electrical signal that satisfies the following formula:

[0018]

[0019] Where I low It is a low current modulated output by the programmable power supply, I high is the high current modulated output of the programmable power supply, T0 is the period of the modulated output of the programmable power supply, and N is a set natural number.

[0020] Furthermore, the light source module adopts an 850nm LED array light source.

[0021] Furthermore, the square wave modulated optical signal satisfies the following formula:

[0022]

[0023] Where α corresponds to I low The electro-optical conversion coefficient of the light source module, β corresponds to I high The electro-optical conversion coefficient of the light source module.

[0024] Furthermore, the step 3 specifically includes:

[0025] Step 3-1: Evenly illuminate the entire surface of the photovoltaic module with the light source;

[0026] Step 3-2: Photoluminescence effect occurs on the surface of the photovoltaic module, and near-infrared light with a central wavelength of 1150nm is radiated and meets the modulation characteristics, with a modulation frequency of 5Hz.

[0027] Furthermore, the optical imaging module uses an InGaAs camera, which is placed in front of the center point of the photovoltaic module, and the acquisition frequency is set to 80fps.

[0028] Furthermore, the step 5 specifically includes the steps of:

[0029] Step 5-1: Cache an n-frame image sequence, where n is at least the number of image frames captured by the camera within one light modulation cycle, satisfying the following formula:

[0030]

[0031] Where F camera is the camera frame rate, M power It is the modulation frequency of the programmable power supply;

[0032] Step 5-2: Calculate the grayscale mean G of the corresponding pixels of the n-frame image x :

[0033]

[0034] Where X(n) is the grayscale value of a pixel in the X-th frame image;

[0035] Step 5-3: Obtain the grayscale mean G of the corresponding pixel points of the n-frame image x Deviation value G s (n):

[0036] G s (n) = X(n) - G x

[0037] Step 5-4: Deviation value G s (n) Perform mathematical operations to determine the mean absolute error, mean square error, or root mean square error; the mean absolute error is:

[0038]

[0039] The mean square error is:

[0040]

[0041] The root mean square error is:

[0042]

[0043] Step 5-5: Perform defect saliency enhancement algorithm processing based on mean absolute error, mean square error or root mean square error, and perform adaptive piecewise linear stretching on the image after the defect saliency enhancement algorithm.

[0044] Furthermore, the n is ≥16.

[0045] Furthermore, the optical imaging module is connected to the InGaAs and the industrial computer via a USB3.0 data cable, and data interface communication is performed based on the Superspeed USB3.0 protocol.

[0046] Compared with the prior art, the present invention has the following significant advantages: (1) It is the first to propose a method of realizing photoluminescence defect detection by modulating light to excite photovoltaic modules. Different from the traditional DC light excitation, this method performs a defect saliency enhancement algorithm on the acquired image sequence with modulation characteristics. The algorithm involves multi-frame processing of the image, which can effectively resist the interference of ambient light on photoluminescence imaging and realize photoluminescence defect detection of photovoltaic modules in bright daylight environment; (2) It is the first to propose a defect saliency enhancement algorithm based on mean absolute error / mean square error / root mean square error for the image sequence. The defect degree of a certain part of the photovoltaic module is characterized by the deviation relative to the mean. After the defect saliency enhancement based on mean absolute error / mean square error / root mean square error is obtained, the defect saliency enhancement algorithm is realized. The defect information presented is stable, and the algorithm is easy to describe in hardware, and is suitable for integration into hardware to realize related technical research such as real-time enhancement of photovoltaic module defect significance; (3) It is proposed for the first time to stimulate photovoltaic modules with non-visible light adjacent to 380nm or 760nm for human eyes. The human eye has a low visual rate for light in this band, which can avoid irritating the human eye after the light is reflected by the glass surface of the photovoltaic module; (4) Compared with the existing technology, the present invention can improve the adaptability of photovoltaic module photoluminescence detection to ambient light and improve the defect significance of photovoltaic modules in bright environments. Under 1210lux, the existing conventional methods can no longer present the defect information of photovoltaic modules. Relatively speaking, the present invention is not limited to 1210lux and can effectively present the defect information of photovoltaic modules as the ambient illumination increases. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1This is a flow chart of the method for detecting photoluminescence defects of photovoltaic modules in a bright environment according to the present invention.

[0048] Figure 2 This is a flow chart of the defect saliency enhancement algorithm described in the present invention.

[0049] Figure 3 This is an experimental effect diagram of the photovoltaic module photoluminescence (PL) defect detection method in a bright environment described in the present invention. DETAILED DESCRIPTION

[0050] This invention provides a method for detecting photoluminescence (PL) defects in photovoltaic modules under bright conditions. This method primarily addresses the problem in the field of automated optical inspection, and more specifically, in the field of optical detection of photovoltaic module defects. Domestic and international photoluminescence-related technical solutions are unable to detect photovoltaic module defects under bright daylight conditions, suffering from the technical limitation of only being able to operate in low-light environments. The method primarily involves acquiring a sequence of PL images of photovoltaic modules under bright conditions and designing a defect saliency enhancement algorithm for these image sequences with modulation characteristics to extract information about photovoltaic module defects under bright conditions. The method proposes that the spectrum of the excitation light source for visible light-excited photovoltaic modules should be less than 380 or greater than 760, but close to these two indicators.

[0051] The present invention is further described in detail below with reference to the accompanying drawings. It should be noted that this example is based on the present technical solution and concretizes the relevant modules in the invention content, including: 1) the programmable power supply module is concretized as a programmable power supply; 2) the light source module is concretized as an 850nm LED array light source; 3) the optical imaging module is concretized as an InGaAs short-wave infrared camera; 4) the defect significance enhancement algorithm is concretized as the mean absolute error; and 5) the high-speed data line is concretized as a USB3.0 ultra-high-speed data line.

[0052] Combine Figure 1 The method for detecting photoluminescence (PL) defects of photovoltaic modules under bright environment specifically includes the following steps:

[0053] Step 1: Connect the communication input of the programmable power supply to the industrial computer, set the modulation frequency and modulation high / low current on the host computer of the industrial computer, and then the programmable power supply generates a square wave modulated electrical signal;

[0054] Step 2: Connect the current output of the programmable power supply to an 850nm LED array light source, and then drive the light source to generate a square wave modulated light signal;

[0055] Step 3: Aim the light source at the photovoltaic module to be tested, and then light modulate the photovoltaic module;

[0056] Step 4: Aim the InGaAs short-wave infrared camera at the photovoltaic module to be tested, and then image the composite radiation light on the surface of the photovoltaic module;

[0057] Step 5: Cache the composite radiation light image of the photovoltaic module that meets the modulation characteristics in the form of an image sequence. Then, use the defect saliency enhancement algorithm to filter out the DC component in the ambient light and extract the defect information that is submerged by the ambient light.

[0058] Step 6: The output end of the InGaAs short-wave infrared camera and the input end of the industrial computer are connected via a USB3.0 ultra-high-speed data cable, and then the host computer displays the defect image of the photovoltaic module.

[0059] Furthermore, provide a reasonable explanation for step 1 and briefly describe the specific implementation steps. A programmable power supply interacts with a computer via a data communication serial port. The computer typically configures the power supply's port number, slave address, maximum output current / voltage, minimum output current / voltage, modulation frequency, and other parameters. It is typically used to drive high-power devices or equipment. Due to its ease of control and adjustable power, it can drive LED arrays of varying illumination areas to stimulate photovoltaic panels of varying sizes.

[0060] Step 1-1, connect the computer and the programmable power supply via a serial communication line;

[0061] Step 1-2: Configure the communication port, slave address, modulation high / low current, and modulation frequency to 5Hz on the host computer.

[0062] Step 1-3: The output is turned on and modulation begins. The modulated electrical signal satisfies the following formula:

[0063]

[0064] Where I low The low current output of the power supply is modulated, I high is the high current modulated output of the power supply, T0 is the period of the modulated output of the power supply, T0 = 200ms, and N is a set natural number.

[0065] Furthermore, a reasonable explanation for step 2 is provided, along with a brief description of the specific implementation steps. PL defect detection technology requires that the photovoltaic module be stimulated by an external light source. In addition to its energy-saving and high-brightness characteristics, LEDs also offer high-speed switching, emitting square-wave light signals in rapid response to the periodic level changes of the modulated power supply. This makes them an ideal light source for the system. To prevent visible light from irritating the human eye after reflection from the photovoltaic module's glass surface, the system uses a near-infrared LED array with a low human visibility and a central wavelength of 850nm as the excitation light source.

[0066] Step 2-1: Connect the 850nm LED array light source with the programmable power supply turned off;

[0067] Step 2-2: Turn on the programmable power supply, and the modulated optical signal approximately satisfies the following formula:

[0068]

[0069] Where α corresponds to I low The electro-optical conversion coefficient of the LED, β corresponds to I high The electro-optical conversion coefficient of the LED.

[0070] Furthermore, a reasonable explanation is given for step 3, and the specific implementation steps are briefly described. The technical method of the light modulation photovoltaic module is a switching light excitation method for the photovoltaic module. The photoluminescence phenomenon of the photovoltaic module meets the switching modulation characteristics. The photoluminescence of the photovoltaic module is a composite radiation luminescence phenomenon. Since there is ambient light with the same composite radiation band as the photovoltaic module during the day, in a short period of time, the daytime ambient light can be regarded as a DC component. Compared with the photoluminescence phenomenon in a dark environment, the DC component directly causes the overall grayscale value of the image to increase, resulting in a decrease in contrast. When the ambient light illumination increases to a certain level, the ambient light submerges the photoluminescence phenomenon, and the image will have a contrast value that is too small to present defect information. Taking light modulation and demodulation for the photovoltaic module can effectively resist the interference of daytime ambient light on the photoluminescence detection of the photovoltaic module.

[0071] Step 3-1: Evenly illuminate the entire surface of the photovoltaic module with an 850nm LED light source;

[0072] Step 3-2: Photoluminescence effect occurs on the surface of the photovoltaic module, and composite radiation emits near-infrared light with a central wavelength of 1150nm and meets the modulation characteristics, with a modulation frequency of 5Hz;

[0073] Furthermore, provide a reasonable explanation for step 4 and briefly describe the specific implementation steps. As the core of the imaging system, the detector directly determines the system's operating environment and detection targets. This system uses an InGaAs short-wave infrared focal plane array camera with a spectral response range of 0.9-1.7 μm. Compared to traditional silicon- or germanium-based CCDs, InGaAs detectors have a higher responsivity in the wavelength range around 1150 nm, making them more easily able to capture photovoltaic module defect information.

[0074] Step 4-1: Place the InGaAs camera directly in front of the center of the photovoltaic module;

[0075] Step 4-2: Turn on the InGaAs camera, set the acquisition frequency to 80fps, adjust the focal length to image the entire photovoltaic module, and adjust the lens aperture to adapt to the ambient light to prevent overexposure;

[0076] Step 4-3: Start imaging and collecting the ambient light and the composite radiation light of the photovoltaic module.

[0077] Furthermore, a reasonable explanation is given for step 5, and the specific implementation steps are briefly described. Figure 2 The defect saliency enhancement algorithm processes multiple frames of a continuously acquired modulated image sequence. Because ambient light causes the mean value of corresponding pixels in the image sequence to increase, the grayscale values ​​of different defects have different deviations from their mean. Therefore, this method characterizes the degree of defectivity at a specific location in a photovoltaic module by using deviation from the mean. Specifically, varying degrees of defects in a photovoltaic module appear as differences in grayscale values ​​in the digital image. In the modulated image sequence, the grayscale values ​​of defect-free locations have larger deviations from their mean. Depending on the degree of defect, the grayscale values ​​of defective locations have smaller or even no deviation from their mean. This method uses one of the following three methods: mean absolute error, mean square error, or root mean square error. After the defect saliency enhancement algorithm is applied, only the amplitude characteristics of defects of varying degrees are retained in the image. The image grayscale is now concentrated in darker areas. An adaptive piecewise linear stretching algorithm is used to map the overall image grayscale to a larger dynamic range, making the image grayscale distribution uniform and improving image quality.

[0078] Step 5-1: Cache n frames of image sequence. For a camera with a frame rate of 80 fps and a modulation frequency of 5 Hz, n is ≥ 16, that is, the number of cached image sequences is at least the number of image frames captured by the camera within one light modulation cycle.

[0079] Step 5-2: Calculate the grayscale mean G of the corresponding pixels of the n-frame image x , as shown below:

[0080]

[0081] Where X(n) is the grayscale value of pixel X in the n-th frame image, and n is the number of image sequences.

[0082] Step 5-3: Obtain the grayscale mean G of the corresponding pixel points of the n-frame image x Deviation value G s (n), as shown in the following formula:

[0083] G s (n) = X(n) - G x

[0084] Step 5-4: For the deviation value G s (n) Perform corresponding mathematical operations to achieve mean absolute error, mean square error, or root mean square error. Specifically, the mean absolute error of a multi-frame image sequence is as follows:

[0085]

[0086] The mean square error of a multi-frame image sequence is shown as follows:

[0087]

[0088] The root mean square error of a multi-frame image sequence is shown as follows:

[0089]

[0090] Step 5-5: Perform adaptive piecewise linear stretching on the image after the defect saliency enhancement algorithm.

[0091] Furthermore, provide a reasonable explanation for step 6 and briefly describe the specific implementation steps. USB 3.0 is a full-duplex interface, which allows data to be exchanged bidirectionally between the master and slave devices at the same time. It is often used for plug-and-play audio transmission.

[0092] Step 6-1. Connect the InGaAs camera to the industrial computer via a USB 3.0 ultra-high-speed data cable.

[0093] Step 6-2: Display PV module defect information through the host computer interface.

[0094] Figure 3 The effect diagram of the conventional detection method and the detection effect diagram of the method of the present invention are combined Figure 3 The results of conventional detection methods and the detection results of the present invention show that the conventional method is unable to reveal photovoltaic module defect information under 1210 lux. In contrast, the present invention can effectively reveal photovoltaic module defect information regardless of the ambient illumination level, regardless of 1210 lux, and as the ambient illumination increases. The present invention has the advantages of strong resistance to interference from ambient light, fast defect detection rate, convenient operation, and high reliability.

Claims

1. A method for detecting photoluminescence defects of photovoltaic modules in a bright environment, characterized in that: The following steps are involved: Step 1: Connect the programmable power supply to the industrial computer, set the modulation frequency and modulation current on the host computer of the industrial computer, and the programmable power supply generates a square wave modulation electrical signal; Step 2: Connect the current output of the programmable power supply to the light source module, and drive the light source to generate a square wave modulated light signal; Step 3: Aim the light source at the photovoltaic module to be tested, and light modulate the photovoltaic module; Step 4: Align the optical imaging module with the photovoltaic module to be tested, and image the composite radiation light on the surface of the photovoltaic module; Step 5: Cache the composite radiation light image of the photovoltaic module surface that meets the modulation characteristics in the form of an image sequence. Apply a defect saliency enhancement algorithm to the image sequence to filter out the DC component in the ambient light, extract the defect information submerged by the ambient light, and obtain the photovoltaic module defect image. Step 6: Connect the output terminal of the optical imaging module to the input terminal of the industrial computer, and display the defect image of the photovoltaic module through the host computer; The optical imaging module uses an InGaAs camera, which is placed in front of the center of the photovoltaic module and the acquisition frequency is set to 80fps; The step 5 specifically includes the following steps: Step 5-1: Cache an n-frame image sequence, where n is at least the number of image frames captured by the camera within one light modulation cycle, satisfying the following formula: Where F camera is the camera frame rate, M power It is the modulation frequency of the programmable power supply; Step 5-2: Calculate the grayscale mean G of the corresponding pixels of the n-frame image x : Where X(n) is the grayscale value of a pixel in the n-th frame image; Step 5-3: Obtain the grayscale mean G of the corresponding pixel points of the n-frame image x Deviation value G s (n): G s (n)=X(n)-G x Step 5-4: For the deviation value G s (n) Perform mathematical operations to determine the mean absolute error, mean square error, or root mean square error; the mean absolute error is: The mean square error is: The root mean square error is: Step 5-5: Perform defect saliency enhancement algorithm processing based on mean absolute error, mean square error or root mean square error, and perform adaptive piecewise linear stretching on the image after the defect saliency enhancement algorithm.

2. The method for detecting photoluminescence defects of photovoltaic modules in a bright environment according to claim 1, characterized in that: The program-controlled power supply is a programmable power supply, the modulation frequency is higher than 1 Hz, and the square wave duty cycle is 50%.

3. The method for detecting photoluminescence defects of photovoltaic modules in a bright environment according to claim 1, characterized in that: The step 1 specifically includes: Step 1-1, connect the industrial computer and the programmable power supply via the serial communication line; Step 1-2: Configure the communication port, slave address, modulation high / low current and modulation frequency on the host computer of the industrial computer; Steps 1-3: Generate a square wave modulated electrical signal that satisfies the following formula: Where I low It is a low current modulated output by the programmable power supply, I high is the high current modulated output of the programmable power supply, T0 is the period of the modulated output of the programmable power supply, and N is a set natural number.

4. The method for detecting photoluminescence defects of photovoltaic modules in a bright environment according to claim 1, wherein: The light source module in step 2 adopts an 850nm LED array light source.

5. The method for detecting photoluminescence defects of photovoltaic modules in a bright environment according to claim 3, characterized in that: The square wave modulated optical signal in step 2 satisfies the following formula: Where α corresponds to I low The electro-optical conversion coefficient of the light source module, β corresponds to I high The electro-optical conversion coefficient of the light source module.

6. The method for detecting photoluminescence defects of photovoltaic modules in a bright environment according to claim 1, characterized in that: The step 3 specifically includes: Step 3-1: Evenly illuminate the entire surface of the photovoltaic module with the light source; Step 3-2: Photoluminescence effect occurs on the surface of the photovoltaic module, and near-infrared light with a central wavelength of 1150nm is radiated and meets the modulation characteristics, with a modulation frequency of 5Hz.

7. The method for detecting photoluminescence defects of photovoltaic modules in a bright environment according to claim 1, characterized in that: The n is ≥16.

8. The method for detecting photoluminescence defects of photovoltaic modules in a bright environment according to claim 1, wherein: The optical imaging module is connected to the InGaAs and industrial computer via a USB3.0 data cable, and data interface communication is performed based on the Superspeed USB3.0 protocol.

Images