Cleaning system and method
Through deep learning image processing model, the stains on the inner wall of the polysilicon reduction furnace bell cover are analyzed, and the movement of the cleaning parts is automatically adjusted, which solves the problem of incomplete cleaning of the inner wall of the bell cover, improves the cleaning efficiency and finished product quality, and reduces energy consumption.
Patent Information
- Application Number
- CN202411367522.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-29
- Publication Date
- 2024-11-22
AI Technical Summary
Incomplete cleaning of the inner wall of the polysilicon reduction furnace bell cover leads to frequent operation, waste of water resources and low production efficiency, affecting the quality of the polysilicon finished product.
Using a cleaning system that includes controlling the cleaning table, cleaning components, visual feedback device and control system, the screen information is obtained through the visual feedback device, and the deep learning image processing model is used to analyze the location, shape and severity of the stain, generate the cleaning motion speed and direction, and automatically Adjust the movement of the cleaning parts for a thorough cleaning.
The cleaning automation level has been improved, the frequency of operators has been reduced, manpower has been saved, and energy consumption growth and finished product quality problems caused by multiple cleanings have been avoided.
Smart Images

Figure CN118988839A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of polysilicon reduction furnaces, and in particular to a cleaning system and method. Background Art
[0002] In the process of polysilicon production by polysilicon production enterprises, the reduction furnace is the place where the hydrogen reduction reaction occurs. It is the core equipment for directly producing polysilicon and is also the key link in determining the system's production capacity and energy consumption.
[0003] In the polysilicon production process, in order to make hydrogen and trichlorosilane undergo hydrogen reduction reaction, a high temperature of about 1000℃ is required inside the bell jar. As the production progresses, after each batch of reaction is completed and the furnace is opened, it is found that the silver layer on the inner wall of the bell jar of the reduction furnace has a reduced finish, obvious discoloration (blackening, yellowing), and there are reaction deposits on the inner wall to varying degrees that contaminate the inner wall. As the production continues, this phenomenon will become more serious. Over time, the reduction of the finish of the inner wall of the bell jar of the reduction furnace and the pollution of the inner wall will cause an increase in energy and material consumption in the production process, and the quality of the polysilicon products will be reduced or abnormal.
[0004] At present, there are the following problems in cleaning the inner wall of the bell jar of a polysilicon reduction furnace: after cleaning, the bell jar needs to be lifted to check the cleaning condition of the inner wall. If the cleaning is not thorough, it needs to be cleaned multiple times. This process leads to frequent operations, waste of manpower, waste of water resources for multiple cleanings, and affects production efficiency. Summary of the invention
[0005] The object of the present invention is to provide a cleaning system and method to solve the above technical problems.
[0006] To achieve the above object, the present invention provides a cleaning system, comprising:
[0007] Control the cleaning table to receive the waste liquid after cleaning the products to be cleaned and to support the products to be cleaned;
[0008] A cleaning component, used for cleaning the product to be cleaned;
[0009] A visual feedback device for obtaining image information of the product to be cleaned;
[0010] A control system, used for analyzing the information fed back by the visual feedback device and sending a control signal to the adjustment device according to the analysis result;
[0011] The adjusting device is used to drive the cleaning component to move according to the control signal.
[0012] The present invention also provides a cleaning method, comprising:
[0013] Determine whether the product to be cleaned needs to be cleaned according to the acquired image information of the product to be cleaned;
[0014] If the product to be cleaned needs to be cleaned, the movement speed and direction of the cleaning stains are generated according to the image information;
[0015] The product to be cleaned is cleaned according to the moving speed and moving direction of the stains to be cleaned.
[0016] Technical effects and advantages of the present invention:
[0017] 1. By improving the automation level of the device, the frequency of on-site operations by operators is reduced, saving manpower.
[0018] 2. Solve the problem that the inner wall of the bell jar is not thoroughly cleaned in the original technology, which leads to increased power consumption of the reduction furnace in the later production process and pollutes the quality of polysilicon products.
[0019] Other features and advantages of the present invention will be described in the following description, and partly become obvious from the description, or be understood by implementing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures pointed out in the description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] Figure 1 Schematic diagram of the cleaning system. DETAILED DESCRIPTION
[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0022] It should be noted that the structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification for people familiar with this technology to understand and read, and are not used to limit the limiting conditions for the implementation of the present invention, so they have no substantial technical significance. Any modification of the structure, change of the proportion relationship or adjustment of the size should still fall within the scope of the technical content disclosed by the present invention without affecting the effects and purposes that can be achieved by the present invention. At the same time, the terms such as "upper", "lower", "left", "right", "middle" and "one" quoted in this specification are only for the convenience of description, and are not used to limit the scope of the implementation of the present invention. The change or adjustment of their relative relationship should also be regarded as the scope of the implementation of the present invention without substantially changing the technical content.
[0023] The present invention discloses a cleaning system, such as Figure 1The system includes: a control cleaning station 101, cleaning components (cleaning self-rotating nozzle 105 and cleaning liquid delivery pipe), an adjustment device (rotating motor 102, lead screw 103), a visual feedback device 104 and a control system 106.
[0024] Among them, the cleaning system can be applied to multiple cleaning scenarios, and the present application preferably cleans the bell jar of a polysilicon reduction furnace; J4 is a high-pressure cleaning liquid, and J6 is a cleaning waste liquid.
[0025] Among them, the control system 106 is preferably a PLC control system, and the visual feedback device 104 is preferably a visual camera.
[0026] Specifically, the control system 106 includes an image processing model unit based on deep learning, and the operation steps are as follows:
[0027] S1, preprocessing.
[0028] The live image is scaled to the same size to match the model input requirements. At the same time, the image brightness, contrast and color are adjusted to improve image quality. Filters are used to remove image noise and retain important features.
[0029] S2, feature extraction and cleaning requirement judgment.
[0030] Convolutional neural networks are used to extract features from images, including the location, shape, color, and texture of stains, and classify stains based on the features and grade the severity of the stains, including: mild, moderate, and severe. The convolutional neural networks used are models pre-trained on large image datasets, such as ResNet and EfficientNet, and then fine-tuned to adapt to the cleaning environment of the silver bell jar of the polysilicon reduction furnace.
[0031] In a specific embodiment, an image I is input to a pre-trained deep learning model M (the model may be any in the prior art). The model outputs stain features, including location (x, y), type t, and severity s.
[0032] The formula means:
[0033] Where N represents the number of stains; x and y represent the location information of the stains; t represents the type of stains; and s represents the severity of the stains.
[0034] Next, the type and severity of the stains determine whether to clean the surface. That is, the extracted features are converted into data information (including the weight of each stain and the overall cleaning need score) for decision-making and control. Specifically, it includes:
[0035] S21, calculate the weight ω of each stain i, weights can be determined based on the type and severity of the stain.
[0036] S22, using the formula: ω i =f(t i ,s i ).
[0037] S23, calculate the overall cleaning needs score:
[0038] S24, determining whether the cleaning demand score is greater than a preset value.
[0039] S25, if the cleaning demand score is greater than or equal to the preset value, enter S3; if the cleaning demand score is less than the preset value, end cleaning.
[0040] S3, determining the movement path, movement speed and movement direction of the cleaning device according to the data information.
[0041] S31, distributing the stain weights on the grid of the entire cleaning area to form a weight matrix W.
[0042] S32, determining the starting point and end point of the cleaning path, and the order of cleaning.
[0043] The stain weights are distributed over the grid of the entire clean area to form a weight matrix W.
[0044] Let W be the weight matrix, where W ij Represents the weight of the grid (i, j).
[0045] The calculation of direction D and speed V is based on the weight center and weight distribution.
[0046] Direction D calculation:
[0047] Speed V calculation: Where k is the speed proportional constant, which should be adjusted according to specific equipment and requirements.
[0048] S4, sends a control signal to the rotating motor 102 to control the cleaning component to move according to D and V.
[0049] In addition, the operator can view the internal condition of the cleaned bell jar in real time through the visual feedback device 104, so as to adjust the cleaning method in time.
[0050] In addition, the high-pressure cleaning self-rotating nozzle 105 has the function of maintaining the self-rotation of the cleaning process by providing high pressure when the cleaning liquid J4 is connected to the high-pressure cleaning self-rotating nozzle 105 through a cleaning liquid delivery pipe (preferably a high-pressure hose).
[0051] In addition, by the rise or fall and self-rotation of the lead screw 103, the cleaning of each part inside the bell jar can be ensured.
[0052] In order to better understand the cleaning system, the cleaning steps of the silver bell of the polysilicon reduction furnace are explained below: the visual feedback device 104 transmits the real-time image of the inside of the bell to be cleaned to the control system 106; the control system 106 analyzes the transmitted image to extract preset data information, and analyzes based on the data information, and then sends a control signal to the rotating motor 102 according to the analysis result. The upper part of the rotating motor 102 is connected to a lead screw 103, and the rotating motor 102 receives the control signal and drives the lead screw 103 to move in a certain direction and speed. The lead screw 103 drives the upper part of the cleaning platform 101 to control the rise or fall of the cleaning parts in the reduction furnace bell 90 until the control system 106 determines that the cleaning meets the preset standards.
[0053] The present invention also provides a cleaning method, comprising:
[0054] Determine whether the product to be cleaned needs to be cleaned according to the acquired image information of the product to be cleaned;
[0055] If the product to be cleaned needs to be cleaned, the movement speed and direction of the cleaning stains are generated according to the image information;
[0056] The product to be cleaned is cleaned according to the moving speed and moving direction of the stains to be cleaned.
[0057] Since the content protected by this method is similar to that protected by the above system, we will not introduce it in detail here. Please refer to the discussion section of the above method for details.
[0058] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A cleaning system, characterized in that: include: Control the cleaning table to receive the waste liquid after cleaning the products to be cleaned and to support the products to be cleaned; A cleaning component, used for cleaning the product to be cleaned; A visual feedback device for obtaining image information of the product to be cleaned; A control system, used for analyzing the information fed back by the visual feedback device and sending a control signal to the adjustment device according to the analysis result; The adjusting device is used to drive the cleaning component to move according to the control signal.
2. The system according to claim 1, characterized in that The visual feedback device is a visual camera.
3. The system according to claim 1, characterized in that The adjusting device comprises: a motor and a lead screw; The motor drives the lead screw to move up and down or rotate according to the control signal sent by the control system; the lead screw is used to drive the cleaning component to move.
4. The system according to claim 1, characterized in that The control system includes: an image processing model unit based on deep learning; Among them, the deep learning image processing model unit performs in-depth analysis on the image information sent by the visual feedback device to obtain the stain feature formula.
5. The system according to claim 4, characterized in that The stain characteristic formula is: Among them, N represents the number of stains; x and y represent the location information of the stains; t represents the type of stains; s represents the severity of the stains; and M represents the stain feature set.
6. The system according to claim 4, characterized in that The control system further includes: a judgment unit, which is used to calculate the information in the stain characteristic formula to obtain the weight of each stain and the overall cleaning demand score, and judge whether the product to be cleaned needs to be cleaned according to the overall cleaning demand score.
7. The system according to claim 6, characterized in that The weight formula of the stain is as follows: ω i =f(t i ,s i ) Among them, ω i represents the weight of each stain; t represents the type of stain; s represents the severity of the stain; and f represents the weight function.
8. The system according to claim 6, characterized in that The overall cleaning needs scoring formula is as follows: Where C represents the overall cleaning need score; N represents the number of stains; ω i Represents the weight of each stain.
9. The system according to claim 6, characterized in that The cleaning components include: a high-pressure cleaning self-rotating nozzle and a cleaning liquid delivery pipe; The cleaning liquid delivery pipe is used to deliver cleaning liquid to the high-pressure cleaning self-rotating nozzle, and the high-pressure cleaning self-rotating nozzle is used to spray cleaning liquid onto the product to be cleaned.
10. The system according to claim 9, characterized in that The control system further comprises: a determination unit, which is used to determine the movement direction and movement speed of the high-pressure cleaning self-rotating nozzle according to the information in the stain characteristic formula.
11. The system according to claim 10, characterized in that The formula for the direction of motion is as follows: Where D represents the movement angle; W represents the weight matrix formed by distributing the stain weights on the grid of the entire cleaning area; W ij represents the weight of the grid (i, j).
12. The system according to claim 11, characterized in that The formula for movement speed is as follows: Where V represents the speed; k is the speed proportional constant; W ij Represents the weight of the grid (i, j).
13. The system according to claim 1, characterized in that The cleaning system is used to clean the bell jar of a polysilicon reduction furnace.
14. A cleaning method, characterized in that: include: Determine whether the product to be cleaned needs to be cleaned according to the acquired image information of the product to be cleaned; If the product to be cleaned needs to be cleaned, the movement speed and direction of the cleaning stains are generated according to the image information; The product to be cleaned is cleaned according to the moving speed and moving direction of the stains to be cleaned.
Citation Information
Patent Citations
Bogie cleaning robot and system
CN110216686A
Visual cleaning method and device for crystal furnace
CN113399319A
Sand blasting cleaning system adopting visual servo control
CN114749451A
Distribution network overhead line insulation terminal dirt cleaning equipment
CN115365196A
Cleaning robot and cleaning method
CN116551702A
