Physical separation control system for silicon wafer and glass of retired photovoltaic module

Abstract

The invention discloses a physical separation control system for silicon wafers and glass of an out-of-service photovoltaic module. The physical separation control system comprises a feeding assembly, a sorting assembly, a sensing assembly and a control assembly. The sorting assembly and the sensing assembly communicate with the control assembly. The feeding assembly is used for conveying the mixed materials to the sorting assembly for sorting; the sensing assembly is used for collecting a real-time image when the mixed materials on the sorting assembly are sorted, and judging whether silicon wafers and glass in the mixed materials are successfully sorted or not according to the real-time image; and the control assembly is used for controlling the sorting assembly to adjust sorting parameters when the sensing assembly judges that the silicon wafers and the glass in the mixed materials fail to be sorted, and controlling the sorted silicon wafers and the glass materials to be quickly cut out and discharged when the sensing assembly judges that the silicon wafers and the glass materials in the mixed materials are successfully sorted. According to the invention, a machine vision technology is combined with an automatic control system, so that rapid physical separation and treatment of the silicon wafer and the glass of the retired photovoltaic module can be realized, and automation of the production process is further realized.

Description

Technical Field

[0001] The present application relates to the field of recycling of retired photovoltaic modules, and in particular to a control system for physical separation of silicon wafer glass from retired photovoltaic modules. Background Art

[0002] With the rapid development of solar photovoltaic power generation, a large number of retired photovoltaic modules will need to be recycled. In the process of recycling retired photovoltaic modules, efficient separation and recovery of silicon wafers and glass materials is an important step in realizing the recycling and reuse of photovoltaic modules. The traditional retired photovoltaic module silicon wafer glass physical sorting equipment has a low degree of automation in the parameter operation process, often relying on manual operation, with low separation efficiency and unable to guarantee the consistency of the quality of the separated products. Therefore, it is necessary to develop an intelligent control system for retired photovoltaic module silicon wafer glass physical separation equipment to achieve efficient separation of silicon wafers and glass materials in retired photovoltaic modules. Summary of the invention

[0003] The present application provides a physical separation control system for silicon wafer glass of retired photovoltaic modules, which is used to solve the technical problem that the physical sorting equipment for silicon wafer glass of retired photovoltaic modules relies on manual operation and has low separation efficiency.

[0004] In order to solve the above technical problems, the technical solution proposed in this application is: A physical separation control system for silicon wafer glass of retired photovoltaic modules, comprising: A feeding component, a sorting component, a sensing component and a control component; the sorting component and the sensing component both establish communication with the control component; The feeding component is used to convey the mixed material to the sorting component for sorting; The sensing component is used to collect real-time images of the mixed material being sorted on the sorting component, and determine whether the silicon wafers and glass in the mixed material are successfully sorted according to the real-time images; The control component is used to control the sorting component to adjust the sorting parameters when the sensing component determines that the silicon wafers and glass in the mixed material have failed to be sorted.

[0005] Preferably, the perception component includes an image acquisition component and an image processing module, wherein the image acquisition component is used to acquire the real-time image and send the real-time image to the image processing module; The image processing module is used to input the real-time image into a pre-trained sorting and recognition model to obtain the position information of the segmentation boundary of the silicon wafer area at the material conveying end of the sorting platform and the segmentation boundary of the glass area.

[0006] Preferably, the sorting and recognition model includes an encoder, a decoder and a boundary extraction module; The encoder includes a residual backbone network, a dilated convolutional network, a spatial pyramid pooling module and an attention mechanism; the dilated convolutional network is used to extract multiple global feature maps of different scales on the input image in parallel through the residual backbone network; the spatial pyramid pooling module is used to enhance the local semantic information corresponding to the multiple global feature maps of different scales; the attention mechanism is used to calculate the correlation between multiple global feature maps and generate an attention weight map; The decoder includes a feature fusion module, an upsampling module and a prediction module, and a batch regularization module; the feature fusion module is used to fuse the multiple global feature maps of different scales according to the attention weight map to obtain a fused feature map; the upsampling module and the prediction module are used to perform linear interpolation upsampling on the fused feature map, and perform channel adjustment to obtain segmentation images and segmentation masks of silicon wafer materials and glass materials; the batch regularization module is used to perform batch regularization operations on the convolutional layers in the decoder; The boundary extraction module is used to extract boundaries according to the segmentation images and segmentation masks of the silicon wafer material and the glass material, and calculate the coordinates of the separation boundary based on the extracted boundaries.

[0007] Preferably, the sorting component includes a material platform, a vibration voltage frequency adjustment component, a height adjustment component, a first sorting plate and a second sorting plate; the vibration voltage frequency adjustment component and the height adjustment component both establish communication with the control component; The material platform is used to receive and sort the mixed materials transported by the feeding assembly; The vibration voltage frequency adjustment component is used to control the vibration of the material platform; The height adjustment component is used to adjust the posture of the material platform; The first sorting plate is used to intercept the silicon wafer material sorted by the material platform; The second sorting plate is used to intercept the glass material sorted by the material platform; The control component is used to control the vibration voltage and frequency adjustment component and / or the height adjustment component to adjust the vibration voltage, vibration frequency and / or posture of the material platform to successfully sort the mixed material when the sensing component determines that the silicon wafers and glass in the mixed material have failed to be sorted, and when the mixed material is successfully sorted, control the first sorting plate and the second sorting plate to quickly intercept the material.

[0008] Preferably, the vibration voltage frequency adjustment component includes a plurality of sorting vibrators arranged at the bottom of the material platform, and the plurality of sorting vibrators are spaced apart along the transport direction of the material platform; the plurality of sorting vibrators are all in communication with the control component; and / or The height adjustment device comprises a first three-phase motor and a second three-phase motor; The first three-phase motor is used to adjust the front-to-rear longitudinal height of the material platform in a first direction; the first direction is the material transportation direction of the material platform; The second three-phase motor is used to adjust the left and right longitudinal height of the material platform in a second direction, and the second direction is a direction that is in the same horizontal plane as the first direction and is perpendicular to each other.

[0009] Preferably, the perception component includes a lighting module and an image display module; the image display module is connected to the image processing module, the image acquisition component includes a plurality of industrial cameras for collecting real-time photos of the end of material transportation, and the lighting module is used to provide fill light for the industrial camera.

[0010] Preferably, the sorting component further includes a first position adjustment component and a second position adjustment component; the first position adjustment component and the second position adjustment component are both connected to the control component; The first position adjustment component is used to adjust the position of the first sorting plate; The second position adjustment component is used to adjust the position of the second sorting plate.

[0011] Preferably, the sorting recognition model is also used to determine the segmentation position of the silicon wafer and the glass on the real-time image, and the predicted falling position of the silicon wafer and the glass at the end of the material platform; The control component is also used to compare the first falling position of the silicon wafer with the first real-time position of the first sorting plate, and when the deviation between the first falling position and the first real-time position is greater than a preset threshold, control the first position adjustment component to adjust the position of the first sorting plate so that the deviation between the first falling position and the first real-time position is within the preset threshold; and / or The control component is also used to compare the second falling position of the glass with the second real-time position of the second sorting plate. When the deviation between the second falling position and the second real-time position is greater than a preset threshold, the second position adjustment component is controlled to adjust the position of the second sorting plate so that the deviation between the second falling position and the second real-time position is within the preset threshold.

[0012] Preferably, the feeding assembly comprises: a feeding hopper and a feeding vibrator; The feeding hopper is used to transfer the input materials to the material platform; The feed vibrator is used to vibrate the feed hopper.

[0013] Preferably, the first position adjustment component and the second position adjustment component are both stepping motors; The control component includes a stepper motor driver, a stepper motor controller, a feeding vibration controller, a sorting vibration controller, a three-phase motor controller, and a PLC controller; the PLC controller establishes communication with the stepper motor controller, the feeding vibration controller, the sorting vibration controller, and the three-phase motor controller; the stepper motor controller is connected to the stepper motor through the stepper motor driver; the feeding vibration controller is connected to the feeding vibrator; the sorting vibration controller is connected to the sorting vibrator; the three-phase motor controller is connected to the first three-phase motor and the second three-phase motor.

[0014] This application has the following beneficial effects: This application combines machine vision technology with an automated control system to achieve rapid physical separation and processing of retired photovoltaic module silicon wafers and glass, thereby automating the production process. This not only improves production efficiency, but also helps improve product quality and reduce production cycles.

[0015] In addition to the above-described purposes, features and advantages, the present application also has other purposes, features and advantages. The present application will be further described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings: Figure 1 This is a workflow diagram of a physical separation system for silicon wafer glass of retired photovoltaic modules based on automatic control and machine vision in a preferred embodiment of the present application; Figure 2 This is an image processing flow chart of a physical separation system of silicon wafer glass of retired photovoltaic modules based on automatic control and machine vision in a preferred embodiment of the present application; Figure 3 This is an IO allocation diagram of a physical separation system for silicon wafer glass of retired photovoltaic modules based on automatic control and machine vision in a preferred embodiment of the present application; Figure 4 This is a schematic diagram of the spacing distribution of the vibrators of the sorting platform along the material transportation direction in the preferred embodiment of the present application; Figure 5 This is a schematic diagram of the height difference between the front and rear of the sorting platform adjusted by the first three-phase motor in the preferred embodiment of the present application; Figure 6 This is a schematic diagram of the left and right height difference of the sorting platform adjusted by the second three-phase motor in the preferred embodiment of the present application; Figure 7 This is a simplified structural diagram of the sorting and identification model in the preferred embodiment of the present application. DETAILED DESCRIPTION

[0017] The embodiments of the present application are described in detail below with reference to the accompanying drawings; however, the present application can be implemented in many different ways as defined and covered by the claims.

[0018] Example: The present application aims to provide a control system and method for the physical separation of silicon wafers and glass from retired photovoltaic modules, so as to solve the problems in the prior art that the entire separation process cannot be monitored in real time, operational errors that occur during the separation process cannot be identified, and the technical problems of separating silicon wafer particles and glass particles cannot be effectively solved.

[0019] To achieve the above-mentioned purpose, the present application provides a physical separation control system for silicon wafer glass of retired photovoltaic modules, including five parts: feeding control, sorting control, visual collection control, material separation control, and control center, which sequentially realize the functions of automatic vibration feeding, silicon wafer glass oscillation separation, product visual image feature extraction, automatic interception and collection of silicon wafer glass products, centralized control and management, etc.; The feeding control part includes a feeding hopper, a feeding vibrator and a feeding vibration controller. The feeding vibrator is arranged above the rear of the device load-bearing frame and connected to the feeding vibration controller inside the control cabinet. The main function is to deliver materials.

[0020] like Figure 4-6 As shown, the sorting control part includes three identical sorting vibrators connected in parallel to the same sorting vibration controller, the sorting vibrators are arranged below the material platform, and the sorting vibration controller is arranged inside the control cabinet; the sorting part also includes two three-phase motors, namely the first three-phase motor and the second three-phase motor; the ends of the two three-phase motors are respectively connected to a reducer; the two three-phase motors with reducers at the ends respectively adjust the front and rear longitudinal height and the left and right longitudinal height of the material platform; the material platform is surrounded by metal bending and forms a passage for material separation; the main function is to convey the sorted materials; wherein, the front and rear longitudinal height of the material platform is adjusted mainly to adjust the front and rear height difference of the material platform and then adjust the longitudinal inclination angle of the material platform; the left and right longitudinal height of the material platform is adjusted mainly to adjust the left and right height difference of the material platform and then adjust the horizontal inclination angle of the material platform.

[0021] The visual acquisition control part is mainly composed of four parts: lighting module, image acquisition component, image display module and image processing module. The lighting module contains an industrial camera light source; the image acquisition component contains two industrial cameras, which are set directly above the entire material sorting platform; the image processing module uses a single-chip microcomputer as the control core, and is connected to the controller for data interaction to form a feedback path.

[0022] The material sorting control part includes two sorting plates, namely the first sorting plate and the second sorting plate, two stepper motors and a stepper motor driver; the sorting plate is located above the material passage, welded with the stepper motor shaft column and rotates coaxially; the stepper motor is connected to the stepper motor driver; the stepper motor driver is respectively connected to the PLC controller in the control cabinet; the outlets under the sorting plates are respectively placed with material recovery containers; their function is to hold the recycled materials after sorting.

[0023] The control center is a control cabinet. The cabinet contains PLC controller, HMI, feeding vibration controller, sorting vibration controller, stepper motor driver, stepper motor controller, AC contactor, thermal relay, intermediate relay, air switch, socket, switch power supply, button, knob, etc.; it is responsible for data collection and processing, intelligent decision-making and control, automated control process, system monitoring and management, security, network communication, optimization and upgrade, etc., to ensure the efficient, stable and safe operation of the whole system.

[0024] In a preferred embodiment, the control cabinet is installed above the load-bearing frame of the device; the electric control element is arranged inside the control cabinet; the communication connection of the display device is arranged on the surface of the control cabinet; It is further preferred that the vibrator of the feeding part and the conveying part is controlled by controlling the power switch of the vibrator through the digital output signal of the PLC controller to start and stop the vibrator; and the voltage signal and frequency signal of the vibrator are controlled by the analog output signal of the controller to control the output voltage and vibration frequency of the vibrator.

[0025] In the preferred embodiment, the three-phase motor control of the material platform of the conveying part uses the controller output signal to control the normally open contacts of the AC contactor, and controls the on and off of the three-phase power supply to control the operation and stop of the motor by designing a self-locking interlocking circuit, thereby realizing the start and stop of the operation of the material platform; the controller output signal is used to turn on and off the forward and reverse interlocking contactor to control the forward and reverse rotation of the three-phase motor, thereby realizing the lifting and lowering of the sorting platform and adjusting the horizontal inclination angle of the material platform; a knob conversion switch is connected to the three-phase motor control circuit, and the two working modes of local control and remote control can be switched at any time.

[0026] In the preferred embodiment, the stepper motor control of the sorting plate part uses the controller IO interface to control the pulse signal of the stepper motor driver, and the speed and number of movement steps of the stepper motor are controlled by the frequency and number of pulses of the pulse signal. The controller can set parameters such as movement direction, acceleration and deceleration, starting position and target position as needed, and realize precise positioning and motion control of the stepper motor through logic control, thereby realizing precise adjustment of the angle of the sorting plate.

[0027] In a preferred embodiment, Figure 3As shown, the C2 port of the feeding vibration controller sdvc31 of the vibrating feeding part is connected to the DQa.0 of the controller to realize the start and stop control of the controller; the D1 and D2 ports of the feeding vibration controller sdvc31 of the vibrating feeding part are connected in series to the intermediate relay, and the normally open contact of the intermediate relay is connected in series to the common end and DIa.0 of the controller to realize the real-time monitoring of the operating parameters such as the voltage and frequency of the vibrator of the feeding part by the controller; the C2 port of the sorting vibration controller sdvc34-ur of the vibrating transmission part is connected to the DQa.1 of the controller to realize the start and stop control of the controller; the D1 and D2 ports of the sorting vibration controller sdvc31 of the vibrating transmission part are connected in series to the intermediate relay, and the normally open contact of the intermediate relay is connected in series to the common end and DIa.1 of the controller to realize the control The device monitors the operating parameters of the vibrator in the transmission part, such as voltage and frequency, in real time; the lifting motor 1 (i.e. the first three-phase motor) of the material platform part occupies DIa.6, DIa.7, DIb.0 and DIb.1 of the controller respectively, realizing remote switching, forward start and stop, reverse start and stop, and fault alarm; the lifting motor 2 (i.e. the second three-phase motor) of the material platform part occupies DIb.2, DIb.3, DIb.4 and DIb.5 of the controller respectively, realizing remote switching, forward start and stop, reverse start and stop, and fault alarm; the Pend+ and ALM+ of the two stepper motor drivers in the sorting board part are connected to DIa.2, DIa.3, DIa.4 and DIa.5 of the controller respectively, realizing the fault alarm of the stepper motor and the prompt of the angle of the sorting board.

[0028] In the preferred solution, the image processing module of the visual part has two main functions: first, to identify the current separation effect of the product and predict the separation position of the silicon wafer and glass for adjusting the voltage and frequency values; second, to identify the actual falling position of the product at the end of the material platform for adjusting the position of the sorting plate.

[0029] In a preferred embodiment, Figure 2 , Figure 7 As shown in the figure, the image processing module is based on a pre-trained deep learning model, which improves the DeepLabV3+ algorithm based on the residual network (ResNet) and dilated convolutions; thus obtaining a fast, lightweight and highly accurate sorting and recognition algorithm. The improved DeepLabV3+ model not only utilizes the powerful feature extraction capability of ResNet, but also increases the receptive field through dilated convolutions and maintains the resolution of the feature map; Among them, the improved DeepLabV3+ algorithm framework still uses the encoder-decoder structure. The encoder part is responsible for extracting the high-level semantic information of the image, while the decoder part is responsible for restoring the spatial information to obtain more accurate segmentation results.

[0030] The encoder includes five parts: residual backbone network, dilated convolution network, ASPP module (spatial pyramid pooling module) and attention mechanism. The backbone network adopts the improved ResNet residual network to extract the basic features of the image. By introducing the residual block, the gradient vanishing problem in the deep neural network is effectively alleviated, and the training efficiency and performance of the network are improved. On the basis of the backbone network, dilated convolution is introduced to expand the receptive field while maintaining the resolution of the feature map. Dilated convolution inserts holes (i.e. skips some pixels) between the convolution kernels, so that the convolution kernel can capture a wider range of contextual information without increasing the amount of calculation. The ASPP module contains multiple parallel dilated convolutions, each with a different dilation rate. These dilated convolutions act on the input feature map in parallel and can capture features at different scales. The ASPP module also contains a global average pooling operation and a standard 1x1 convolution to capture the most basic local information. Through the ASPP module, the encoder can generate a feature map with rich multi-scale contextual information. The attention mechanism is introduced to enhance the network's ability to focus on important features. The attention mechanism can generate an attention weight map by calculating the correlation between feature maps, thereby highlighting important features and suppressing irrelevant features, thereby improving the segmentation accuracy of silicon-glass oscillation separation visual images.

[0031] The decoder mainly includes three parts: feature fusion module, upsampling module and prediction module, and batch normalization module. The decoder receives high-level semantic feature maps from the encoder and low-level feature maps in the backbone network. First, the low-level feature map is subjected to channel dimension reduction using 1x1 convolution, and then the high-level semantic feature map from ASPP is interpolated and upsampled to make it the same size as the low-level feature map. Then, the two are concatenated and sent to a set of 3x3 convolution blocks for processing to fuse feature information of different scales. After feature fusion, the decoder performs linear interpolation upsampling again to restore the resolution of the feature map to the same as the original image. Finally, a 1x1 convolution is used to adjust the channel to obtain the final segmentation result. In addition, after each convolution layer of the network, a batch normalization operation is introduced to accelerate model convergence and prevent gradient disappearance. By normalizing each small batch of data, the output of the network is more stable and the generalization ability of the network is improved.

[0032] In the preferred solution, the model also incorporates the attention mechanism to further enhance the ability to capture key information. In order to accelerate model training and enhance the generalization ability of the model, the batch normalization method is used, and the network structure is optimized to improve the convergence speed and segmentation accuracy of the model, thereby generating accurate segmentation lines; According to the image segmentation results, the subsequent separation positions of the silicon wafer and the glass are predicted, and the voltage and frequency values ​​are adaptively adjusted according to the predicted separation positions and the current position of the sorting plate.

[0033] In the preferred solution, the image processing module uses the improved DeepLabV3+ algorithm for image segmentation. By accurately segmenting the image at the end of the material platform, we can accurately identify the actual falling position of the product.

[0034] In the preferred solution, the controller uses Siemens 1200 series, which is used to collect equipment working indicators and send command signals. The DI interface of the controller is connected in series with the normally open contact of the thermal relay to ensure the safe and stable operation of the motor and fault alarm. The display device uses Siemens series as a human-computer interaction interface to view the working status of the equipment in real time and control the start and stop of the equipment. The core of the working index lies in the voltage and frequency setting of the sorting vibrator, among which there is a known and clear correlation between the specific values ​​of voltage and frequency and the movement state of silicon wafers and glass during the separation process.

[0035] In addition, to achieve the above purpose, Figure 1 As shown, the workflow of the physical separation system of silicon wafer glass of retired photovoltaic modules in this application is as follows: Step 1: The system starts and the material reaches the entrance of the vibrating feeder; Step 2: The vibrating feeder is started according to the initial default value, and the control center obtains and displays the current frequency and voltage of the feeding vibrator in real time; Step 3: The material reaches the material platform, the sorting vibrator operates according to the default initial value, and the material platform is at the default initial tilt angle; Step 4: The material is transported to the visual collection area to achieve the initial separation of silicon wafers and glass; Step 5: The image acquisition component captures the image of the material platform in real time through the industrial camera, and sends it to the image processing module for processing and analysis to obtain the predicted separation position coordinate value; the controller reads the predicted coordinate and compares it with the current sorting plate coordinate to check whether the position deviation is within the allowable range. If so, execute step 7, otherwise, execute step 6; Step 6: When the system detects that the position deviation of the silicon wafer or glass on the material platform exceeds the preset allowable range, the model predictive control (MPC) algorithm will be triggered immediately. The core of this MPC algorithm is to use its built-in system dynamic model, which deeply simulates the physical behavior characteristics of the sorting vibrator, material platform and silicon wafer or glass separation to ensure the high accuracy of the model. During the operation of the algorithm, the current system state will be fully considered, including the actual position of the silicon wafer or glass on the material platform, as well as the key control variables - voltage and frequency. At the same time, the MPC algorithm also clearly defines the prediction time domain and control time domain: the prediction time domain is used to define the time span of the algorithm to predict the future state of the system, while the control time domain limits the time range considered by the algorithm when optimizing the control input. Based on the above information, the MPC algorithm carefully constructs an optimization problem to find a series of optimal control input sequences to minimize future position deviations or meet other preset cost function requirements. Solved by advanced optimization algorithms, MPC can output a series of optimal control input sequences that can guide the system to operate in the optimal state in a specific time period in the future. Subsequently, the system intelligently selects the control input required at the current moment from the optimal control input sequence, that is, the precise voltage and frequency values, and converts them into adjustment instructions. These instructions are then sent to the sorting vibrator controller, which immediately performs the corresponding adjustment actions after receiving the instructions. To ensure the accuracy and effectiveness of the adjustment, the system will continue to monitor the feedback of the controller until it is confirmed that the voltage and frequency values ​​have been set correctly. Once the adjustment is completed and verified, the system will automatically return to step 5 and restart the position prediction and deviation check process to ensure that the precise separation of silicon wafers or glass can be continuously achieved; Step 7: The image processing module identifies the falling position of the product; Step 8: Convert the actual falling position of the product into a machine coordinate system that can be understood by the PLC to obtain the falling coordinates of the product; Step 9: The controller converts the product falling coordinates into control signals and transmits them to the stepper motor driver through the I / O module; Step 10: The stepper motor driver adjusts the pulse frequency and pulse number of the stepper motor according to the signal, thereby accurately controlling the placement angle of the sorting plate; Step 11: Use a position sensor to detect the current position of the sorting plate; Step 12: The PLC receives the data from the sensor and processes the data according to a predetermined algorithm or logic to obtain the coordinates of the scoring board; Step 13: The controller compares the current sorting plate coordinates and the product drop coordinates to form a closed-loop control system to ensure positioning accuracy. If they are the same, execute step 14; Step 14: The separated silicon wafer glass materials fall into the material recovery container below due to automatic interception by the sorting plate, and the collection is completed.

[0036] Further preferred is: the control method for physical separation of silicon wafer glass of retired photovoltaic modules, wherein the control center monitors the operating status of the entire system in real time, including the vibration voltage, frequency, inclination angle of the sorting plate, horizontal inclination angle and longitudinal inclination angle of the sorting platform, etc. of the vibrator, and displays it in real time in the HMI.

[0037] It is further preferred that: the initial position coordinates of the sorting plate are the origin of the coordinate system of the controller; Further preferred is: in a control method for physical separation of silicon wafer glass from retired photovoltaic modules, initial values ​​of the vibrator's vibration voltage, frequency, platform horizontal inclination angle, etc. are all optimal values ​​summarized through experiments, which can achieve better separation effects.

[0038] In summary, the present application aims to provide a control system and method for physical separation of silicon wafers and glass from retired photovoltaic modules. By providing a control system and method for physical separation of silicon wafers and glass from retired photovoltaic modules, efficient separation of silicon wafers and glass during oscillation separation can be achieved, the dividing boundary between silicon wafers and glass products can be accurately identified, and the oscillation separation equipment can be controlled to perform stable automatic interception and separation of silicon wafers and glass particles, thus realizing automation and intelligent control of the entire physical separation process of silicon wafers and glass.

[0039] This application has the following advantages: (1) High separation efficiency: Traditional separation of silicon wafers from retired photovoltaic modules usually requires manual operation, which is slow and prone to errors. Machine vision technology can achieve fast and accurate image recognition and analysis, thereby improving separation efficiency.

[0040] (2) Reduce labor costs: The use of machine vision technology can reduce the need for manual labor, thereby reducing labor costs. Since the machine vision system can automatically perform the separation process without direct human involvement, it can reduce related training and management costs.

[0041] (3) Improved safety: In the traditional silicon wafer glass separation process, operators may face certain safety risks, such as being scratched by glass fragments. The physical separation control system and method based on machine vision can avoid these potential dangers and protect the safety of operators.

[0042] (4) Automated production: Machine vision technology combined with automated control systems can achieve rapid physical separation and processing of retired photovoltaic module silicon wafers and glass, thereby automating the production process. This not only improves production efficiency, but also helps to improve product quality and reduce production cycles.

[0043] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A physical separation control system for silicon wafer glass of retired photovoltaic modules, characterized in that: include: A feeding component, a sorting component, a sensing component and a control component; the sorting component and the sensing component both establish communication with the control component; The feeding component is used to convey the mixed material to the sorting component for sorting; The sensing component is used to collect real-time images of the mixed material being sorted on the sorting component, and determine whether the silicon wafers and glass in the mixed material are successfully sorted according to the real-time images; The control component is used to control the sorting component to adjust the sorting parameters when the sensing component determines that the silicon wafers and glass in the mixed material have failed to be sorted.

2. The retired photovoltaic module silicon wafer glass physical separation control system according to claim 1, characterized in that: The perception component includes an image acquisition component and an image processing module, wherein the image acquisition component is used to acquire the real-time image and send the real-time image to the image processing module; The image processing module is used to input the real-time image into a pre-trained sorting and recognition model to obtain the position information of the segmentation boundary of the silicon wafer area at the material conveying end of the sorting platform and the segmentation boundary of the glass area.

3. The retired photovoltaic module silicon wafer glass physical separation control system according to claim 2, characterized in that: The sorting and recognition model includes an encoder, a decoder and a boundary extraction module; The encoder includes a residual backbone network, a dilated convolutional network, a spatial pyramid pooling module, and an attention mechanism; the dilated convolutional network is used to extract multiple global feature maps of different scales on the input image in parallel through the residual backbone network; The spatial pyramid pooling module is used to enhance the local semantic information corresponding to the multiple global feature maps of different scales; the attention mechanism is used to calculate the correlation between the multiple global feature maps and generate an attention weight map; The decoder includes a feature fusion module, an upsampling module and a prediction module, and a batch regularization module; the feature fusion module is used to fuse the multiple global feature maps of different scales according to the attention weight map to obtain a fused feature map; the upsampling module and the prediction module are used to perform linear interpolation upsampling on the fused feature map, and perform channel adjustment to obtain segmentation images and segmentation masks of silicon wafer materials and glass materials; the batch regularization module is used to perform batch regularization operations on the convolutional layers in the decoder; The boundary extraction module is used to extract boundaries according to the segmentation images and segmentation masks of the silicon wafer material and the glass material, and calculate the coordinates of the separation boundary based on the extracted boundaries.

4. The retired photovoltaic module silicon wafer glass physical separation control system according to claim 3, characterized in that: The sorting component includes a material platform, a vibration voltage frequency adjustment component, a height adjustment component, a first sorting plate and a second sorting plate; the vibration voltage frequency adjustment component and the height adjustment component both establish communication with the control component; The material platform is used to receive and sort the mixed materials transported by the feeding assembly; The vibration voltage frequency adjustment component is used to control the vibration of the material platform; The height adjustment component is used to adjust the posture of the material platform; The first sorting plate is used to intercept the silicon wafer material sorted by the material platform; The second sorting plate is used to intercept the glass material sorted by the material platform; The control component is used to control the vibration voltage and frequency adjustment component and / or the height adjustment component to adjust the vibration voltage, vibration frequency and / or posture of the material platform to successfully sort the mixed material when the sensing component determines that the silicon wafers and glass in the mixed material have failed to be sorted, and when the mixed material is successfully sorted, control the first sorting plate and the second sorting plate to quickly intercept the material.

5. The retired photovoltaic module silicon wafer glass physical separation control system according to claim 4, characterized in that: The vibration voltage frequency adjustment component includes a plurality of sorting vibrators arranged at the bottom of the material platform, and the plurality of sorting vibrators are spaced apart along the transport direction of the material platform; the plurality of sorting vibrators are all in communication with the control component; and / or The height adjustment device comprises a first three-phase motor and a second three-phase motor; The first three-phase motor is used to adjust the front-to-rear longitudinal height of the material platform in a first direction; the first direction is the material transportation direction of the material platform; The second three-phase motor is used to adjust the left and right longitudinal height of the material platform in a second direction, and the second direction is a direction that is in the same horizontal plane as the first direction and is perpendicular to each other.

6. The retired photovoltaic module silicon wafer glass physical separation control system according to claim 4, characterized in that: The perception component includes a lighting module and an image display module; the image display module is connected to the image processing module, the image acquisition component includes multiple industrial cameras for collecting real-time photos of the end of material transportation, and the lighting module is used to provide fill light for the industrial camera.

7. The retired photovoltaic module silicon wafer glass physical separation control system according to claim 5, characterized in that: The sorting component further includes a first position adjustment component and a second position adjustment component; the first position adjustment component and the second position adjustment component are both connected to the control component; The first position adjustment component is used to adjust the position of the first sorting plate; The second position adjustment component is used to adjust the position of the second sorting plate.

8. The retired photovoltaic module silicon wafer glass physical separation control system according to claim 7, characterized in that: The sorting recognition model is also used to determine the segmentation position of the silicon wafer and the glass on the real-time image, and the predicted falling position of the silicon wafer and the glass at the end of the material platform; The control component is also used to compare the first falling position of the silicon wafer with the first real-time position of the first sorting plate, and when the deviation between the first falling position and the first real-time position is greater than a preset threshold, control the first position adjustment component to adjust the position of the first sorting plate so that the deviation between the first falling position and the first real-time position is within the preset threshold; and / or The control component is also used to compare the second falling position of the glass with the second real-time position of the second sorting plate. When the deviation between the second falling position and the second real-time position is greater than a preset threshold, the second position adjustment component is controlled to adjust the position of the second sorting plate so that the deviation between the second falling position and the second real-time position is within the preset threshold.

9. The retired photovoltaic module silicon wafer glass physical separation control system according to claim 8, characterized in that: The feeding assembly comprises: a feeding hopper and a feeding vibrator; The feeding hopper is used to transfer the input materials to the material platform; The feed vibrator is used to vibrate the feed hopper.

10. The retired photovoltaic module silicon wafer glass physical separation control system according to claim 9, characterized in that: The first position adjustment component and the second position adjustment component are both stepper motors; The control component includes a stepper motor driver, a stepper motor controller, a feeding vibration controller, a sorting vibration controller, a three-phase motor controller, and a PLC controller; the PLC controller establishes communication with the stepper motor controller, the feeding vibration controller, the sorting vibration controller, and the three-phase motor controller; the stepper motor controller is connected to the stepper motor through the stepper motor driver; the feeding vibration controller is connected to the feeding vibrator; the sorting vibration controller is connected to the sorting vibrator; the three-phase motor controller is connected to the first three-phase motor and the second three-phase motor.

Images