Full melting control method and system based on CCD (Charge Coupled Device)

Abstract

The invention belongs to the technical field of CCD image processing, and particularly relates to a CCD-based full melting control method and system, and the method comprises the following steps: S101, obtaining a to-be-detected image of a melt in a single crystal furnace collected by a CCD; s102, detecting the image obtained in the step S101 based on an image segmentation model, and segmenting a target area; s103, calculating a solid-liquid ratio in the furnace based on the obtained pixel result of each silicon material type; s104, based on the calculated area of the blocky silicon material, estimating the volume of the blocky silicon material in the furnace in combination with power and temperature changes; s105, according to the obtained melting volume change rate and the obtained time change rate, the amplitude needing to be changed for adjusting the power of the heater is calculated; by collecting data in real time, process parameters are dynamically adjusted according to the actual solid-liquid percentage, and various variables are flexibly dealt with; by accurately regulating and controlling the heating power and the heating time, the energy consumption is greatly reduced, the service life of equipment is prolonged, and the maintenance cost is reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of CCD image processing, and specifically relates to a full melting control method and system based on CCD. Background Art

[0002] Full melting refers to using the crucible peripheral heater to melt the silicon material into uniform liquid silicon material. In the melting process of single crystal growth, the melting of silicon material is a slow and long process. For a long time after starting heating, due to the low temperature of the single crystal furnace chamber, the radiation of the silicon material and the crucible wall is not sufficient to emit light; as the silicon material gradually melts upward and the furnace chamber temperature rises, the light in the furnace becomes brighter, and then the silicon material melts; according to the situation of the unmolten silicon material on the liquid silicon material surface, it can be divided into four stages: a small amount of melting stage, a violent melting stage, a stage with suspended matter, and a complete melting stage.

[0003] In the prior art, when melting materials in a single crystal furnace, it mainly relies on the artificial observation window to understand the melting situation of the silicon material in the single crystal furnace, and relies on the experience of the molten state of the silicon material to control the time point of reducing the power of the bottom heater. Devices with an automatic melting function set a large enough heater power within a sufficient time to ensure that the silicon material in the furnace is fully melted.

[0004] Currently, in the prior art, manually adjusting the temperature in the furnace usually relies on the operator's observation and judgment of temperature changes, and cannot respond to small fluctuations in the furnace temperature in real time. Especially in the stage of rapid heating or cooling, it is difficult for manual adjustment to keep up with the changes in thermodynamics in a timely manner; and this method is limited by subjective judgment and the operator's experience, and cannot meet the high-precision and high-efficiency requirements of silicon raw material melting detection, resulting in problems during the melting process of silicon raw materials, such as incomplete melting or uneven melting, which seriously affects the quality and performance of single crystal materials.

[0005] Therefore, the present invention provides a full melting control method and system based on CCD. Summary of the Invention

[0006] In order to make up for the deficiencies of the prior art and solve at least one of the technical problems proposed in the background art.

[0007] The technical solution adopted by the present invention to solve its technical problems is as follows: A full melting control method based on CCD according to the present invention includes the following steps:

[0008] S101: Obtain an image of the molten material to be detected in the single crystal furnace collected by the CCD;

[0009] S102: Based on the image segmentation model, detect the image obtained in S101 and segment out the target area;

[0010] S103: Calculate the solid-liquid ratio in the furnace based on the pixel results of each type of silicon material obtained in S102;

[0011] S104: Estimate the volume of the bulk silicon material in the furnace by combining the area of the bulk silicon material calculated in S103 with the power and temperature changes;

[0012] S105: Calculate the amplitude of the change required for adjusting the heater power by obtaining the melting volume change rate and the time change rate.

[0013] Preferably, the CCD in S101 collects images of the melting process of the silicon material in the single crystal furnace for subsequent processing; the melting process is from the moment when the solid bulk silicon material is first put into the single crystal furnace until the end of the entire melting process.

[0014] Preferably, the method for obtaining the segmentation model in S102 is as follows:

[0015] S201: Mark the types of silicon materials by using software in the way of depicting contour points on the image;

[0016] The types of silicon materials include liquid silicon material, bulk silicon material, and crystalline silicon material;

[0017] S202: Generate a dedicated data set from the marked images in the form of digital coding;

[0018] S203: Provide the data set for training the segmentation model based on deep learning, and the segmentation model can be obtained after the training is completed.

[0019] Preferably, S301: Based on the area of the bulk silicon material in the furnace obtained in S103, establish a model for calculating the volume of the silicon material;

[0020] S302: Calculate the total heat in the furnace by the heater power and the heating time in the furnace;

[0021] Among them, the calculation expression for calculating the total heat input into the furnace is:

[0022] Q = P·t

[0023] In the above formula, Q is the total heat input into the furnace, P is the heating power, and t is the heating time;

[0024] S303: Deduce whether it melts according to the temperature change, and deduce the volume according to the area and height;

[0025] The calculation expression for deducing the temperature change through heat is:

[0026]

[0027] Among them, ΔT is the temperature change in the furnace, ρ is the density of the massive silicon material in the furnace, and c is the specific heat capacity of the massive silicon material in the furnace; the height h of the massive silicon material in the furnace can be deduced from the available variables; the volume of the massive silicon material in the furnace can be estimated as V = A·h;

[0028] S304: Describe the volume change of the silicon material through the density difference between the solid and liquid states of silicon;

[0029] Among them, during the melting process, the calculation expression for the volume change amount ΔV is:

[0030]

[0031] In the above formula, V 0 is the initial volume of the silicon material, ρ 液 and ρ 固 are the densities of the silicon material in the liquid state and the solid state respectively.

[0032] Preferably, the method of estimating the volume of the massive silicon material in the furnace by combining power and temperature change in S104 is calculated based on the thermal field model of silicon material melting, and the specific steps are as follows:

[0033] S401: It is necessary to establish a thermal field model for the furnace environment, including the positions and sizes of the heater, crucible, and silicon material;

[0034] S402: Use numerical simulation software to solve the thermal field model, obtain the distribution of the temperature field in the furnace through simulation calculation, and calculate the time and heat required for silicon material melting according to the change of the temperature field;

[0035] S403: Given the time and heat required for silicon material melting, the volume change during the silicon material melting process can be calculated in combination with the physical properties of the silicon material;

[0036] S404: Under the condition that the temperature field distribution in the furnace is uniform, estimate the volume of the massive silicon material through geometric calculation or proportional relationship.

[0037] Preferably, the expression for the amplitude change required for calculating and adjusting the heater power in S105 is:

[0038]

[0039] Among them, P is the heater power; ρ is the density of silicon; V is the current volume of the silicon material; c is the specific heat capacity of the silicon material; L f is the latent heat of fusion of the silicon material; is the temperature change rate of the silicon material; is the volume change rate of the silicon material.

[0040] Preferably, the heating power is equal to the working power of the heaters on both sides and at the bottom of the crucible.

[0041] Preferably, a full melting control system based on CCD, the execution module of the above control method includes: an image processing module, a volume solving module, and a predicted power module;

[0042] The image processing module is used to collect the image of the CCD during the melting process of the single crystal furnace, preprocess the image and then send it to the segmentation model, and the segmentation model will output the pixels corresponding to the massive silicon material, crystalline silicon material and liquid silicon material;

[0043] The volume solving module is used to obtain the number of pixels of the massive silicon material, crystalline silicon material and liquid silicon material, and calculate the actual area values of the massive silicon material and liquid silicon material in the single crystal furnace;

[0044] The power adjustment module is used to estimate the adjustment amplitude of the heater power according to the volume, volume change rate and time change rate of the massive silicon material.

[0045] The beneficial effects of the present invention are as follows:

[0046] 1. For a full melting control method and system based on CCD of the present invention, by collecting data in real time, dynamically adjusting process parameters according to the actual solid-liquid percentage, flexibly coping with various variables; by precisely controlling the heating power and heating time, providing sufficient energy only when needed, greatly reducing energy consumption, avoiding overheating, increasing the service life of the equipment, and reducing maintenance costs.

[0047] 2. For a full melting control method and system based on CCD of the present invention, the solid-liquid interface and solid-liquid percentage are monitored in real time. By adjusting the heating power and the position of the thermal shield, the shape of the solid-liquid interface is ensured to be stable, reducing the thermal stress and dislocation in the crystal, improving the overall quality and uniformity of the single crystal; by dynamically adjusting the heating time according to the solid-liquid percentage detected by the experiment, when the target ratio is reached, it can enter the next stage, avoiding unnecessary waiting time, improving production efficiency, and increasing the output capacity of the single crystal furnace; by detecting and adjusting the heater power in real time, it can ensure that the melting temperature is maintained within the range suitable for the stable growth of the crystal. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The present invention will be further described below with reference to the drawings.

[0049] Figure 1 is a schematic diagram of the flow of a full melting control method based on CCD and a schematic diagram of the execution module in the present invention;

[0050] Figure 2 is a schematic diagram of the flow of the acquisition method of the segmentation model in the present invention;

[0051] Figure 3It is a schematic flow chart of the method for calculating the solid-liquid ratio in the furnace in the present invention;

[0052] Figure 4 It is a schematic diagram of the calculation method of the thermal field model in the present invention; Specific embodiments

[0053] In order to make the technical means, creative features, achieved purposes and effects of the present invention easy to understand, the present invention will be further described below in conjunction with specific embodiments.

[0054] As Figure 1 shown, a full melting control method based on CCD described in an embodiment of the present invention includes the following steps:

[0055] S101: Obtain an image of the molten material to be detected in the single crystal furnace collected by the CCD;

[0056] S102: Based on the image segmentation model, detect the image obtained in S101 and segment the target area;

[0057] S103: Based on the pixel results of each silicon material type obtained in S102, calculate the solid-liquid ratio in the furnace;

[0058] S104: Based on the area of the bulk silicon material calculated in S103, estimate the volume of the bulk silicon material in the furnace in combination with the power and temperature changes;

[0059] S105: Calculate the amplitude of the change required for adjusting the heater power by obtaining the melting volume change rate and the time change rate.

[0060] The full melting control method based on CCD provided by the present invention needs to first obtain the images during the melting process of the single crystal furnace by the CCD camera, and use a segmentation model for detection based on the acquired images to segment the bulk silicon material region, the crystalline silicon material region, and the liquid silicon material region; according to the obtained region results of each silicon material type, calculate the solid-liquid ratio in the single crystal furnace; according to the calculated area of the bulk silicon material, estimate the volume of the bulk silicon material in the furnace in combination with the power and temperature changes; finally, through the obtained melting volume change rate and time change rate, calculate the amplitude of the change required for adjusting the heater power; through the full melting control method based on CCD, the solid-liquid interface and the solid-liquid percentage are monitored in real time, and by adjusting the heating power and the position of the thermal shield, the shape of the solid-liquid interface is ensured to be stable, reducing the thermal stress and dislocation in the crystal, improving the overall quality and uniformity of the single crystal; by precisely controlling the heating power and heating time, sufficient energy is provided when needed, significantly reducing energy consumption, avoiding overheating, increasing the service life of the equipment, and reducing maintenance costs; by collecting data in real time and dynamically adjusting process parameters according to the actual solid-liquid percentage, various variables can be flexibly responded to; by dynamically adjusting the heating time according to the solid-liquid percentage detected by experiments, when the target ratio is reached, the next stage can be entered, avoiding unnecessary waiting time, improving production efficiency, and increasing the output capacity of the single crystal furnace.

[0061] As Figure 1 shown, the image of the melting process in the single crystal furnace collected by the CCD in S101 is used for subsequent processing; the melting process is from the moment when the solid bulk silicon material is first put into the single crystal furnace until the end of the entire melting process.

[0062] The present invention collects the images of the melting process in the single crystal furnace for subsequent processing, where the melting process is from the moment when the solid bulk silicon material is first put into the single crystal furnace until the end of the entire melting process;

[0063] Among them, when the silicon material is first put into the single crystal furnace, the image of the furnace collected by the CCD does not emit light because the temperature of the silicon material is low, so the CCD image collected in the single crystal furnace is in a completely black state; as the heater power gradually increases, the temperature of the silicon material also gradually rises, and the brightness of the image of the single crystal furnace collected by the CCD begins to approach the normal situation after it starts to emit light.

[0064] As Figure 2 shown, the acquisition method of the segmentation model in S102 is as follows:

[0065] S201: Mark the silicon material types by using software in the way of depicting contour points on the image;

[0066] The silicon material types include liquid silicon material, bulk silicon material, and crystalline silicon material;

[0067] S202: Generate a dedicated dataset from the annotated images in the form of digital coding;

[0068] S203: Provide the dataset for the training of a deep learning-based segmentation model, and the segmentation model can be obtained after the training is completed.

[0069] The segmentation model provided by the present invention is based on the deep learning image processing method; through S101, a large number of images of the melting process in the single crystal furnace are collected by CCD, and each image in the furnace is annotated using a dedicated software annotation tool; the annotation is to mark which areas belong to liquid silicon material, which belong to massive silicon material, and which belong to crystalline silicon material by depicting contour points on the image. Finally, a dedicated dataset is generated from the annotated images in the form of digital coding and provided for the training of a deep learning-based segmentation model, and the segmentation model can be obtained after the training is completed.

[0070] After the original image passes through this model, in the output result image, it is marked which pixel points are liquid silicon material areas, which pixel points are massive silicon materials, and which pixels are crystalline silicon materials. By obtaining the pixel category images of each silicon material, the next calculation can be carried out.

[0071] As Figure 3 shown, the method for calculating the solid-liquid ratio in the furnace in S103 is as follows:

[0072] S301: Based on S103, obtain the area of the massive silicon material in the furnace and establish a model for calculating the volume of the silicon material;

[0073] S302: Calculate the total heat in the furnace through the heater power and heating time in the furnace;

[0074] Among them, the calculation expression for the total heat input into the furnace is:

[0075] Q = P·t

[0076] In the above formula, Q is the total heat input into the furnace, P is the heating power, and t is the heating time;

[0077] S303: Deduce whether it melts according to the temperature change, and deduce the volume according to the area and height;

[0078] The calculation expression for deducing the temperature change through the heat is:

[0079]

[0080] Among them, ΔT is the temperature change in the furnace, ρ is the density of the massive silicon material in the furnace, and c is the specific heat capacity of the massive silicon material in the furnace; the height h of the massive silicon material in the furnace can be deduced through the available variables; the volume of the massive silicon material in the furnace can be estimated as V = A·h;

[0081] S304: Describe the volume change of silicon materials through the density difference between the solid state and the liquid state of silicon.

[0082] Among them, during the melting process, the calculation expression of the volume change amount ΔV is:

[0083]

[0084] In the above formula, V 0 is the initial volume of the silicon material, ρ 液 and ρ 固 are the densities of the silicon material in the liquid state and the solid state respectively.

[0085] As Figure 4 shown, the method of estimating the volume of bulk silicon material in the furnace by combining power and temperature changes in S104 is calculated based on the thermal field model of silicon material melting. The specific steps are as follows:

[0086] S401: It is necessary to establish a thermal field model for the furnace environment, including the positions and sizes of the heater, crucible, and silicon material.

[0087] S402: Use numerical simulation software to solve the thermal field model, obtain the distribution of the temperature field in the furnace through simulation calculation, and calculate the time and heat required for silicon material melting according to the change of the temperature field.

[0088] S403: Given the time and heat required for silicon material melting, combined with the physical properties of the silicon material, the volume change during the melting process of the silicon material can be calculated.

[0089] S404: Under the condition that the temperature field distribution in the furnace is uniform, estimate the volume of bulk silicon material through geometric calculation or proportional relationship.

[0090] The present invention estimates the volume of bulk silicon material in the furnace through the thermal field model, calculates the area of the bulk material in the furnace through S103, but does not know how much bulk silicon material exists under the liquid in the furnace, and the heater power cannot be continuously output according to the area of the bulk material. In particular, at the moment when liquid silicon material, bulk silicon material, and crystalline silicon material coexist during the mid-stage of heating to melt the material, it is not known how much bulk silicon material exists under the liquid silicon material.

[0091] If only most of the bulk silicon material floats on the liquid silicon material and there is not a large amount of bulk silicon material actually under the liquid silicon material, and the heater power output is too large, it will cause waste of energy; if a small part of the bulk silicon material floats on the liquid silicon material and there is a large amount of bulk silicon material actually under the liquid silicon material, and the heater power output is too small, it will cause low efficiency of melting the material.

[0092] Therefore, only obtaining the area of the massive silicon material cannot accurately give the appropriate heater power. It is necessary to estimate how much massive silicon material still exists under the liquid silicon material. The method for specifically estimating the volume of the bulk material in the furnace is calculated based on the thermal field model of silicon material melting.

[0093] As Figure 1 shown, the expression for the amplitude change required to calculate and adjust the heater power in S105 is:

[0094]

[0095] where P is the heater power; ρ is the density of silicon; V is the current volume of the silicon material; c is the specific heat capacity of the silicon material; L f is the latent heat of fusion of the silicon material; is the temperature change rate of the silicon material; is the volume change rate of the silicon material.

[0096] As Figure 3 shown, the heating power is equal to the working power of the heaters on both sides and the bottom of the crucible.

[0097] The heating power provided by the present invention is the working power of the heaters on both sides and the bottom of the crucible.

[0098] As Figure 1 shown, a full melting control system based on CCD, the execution module of the above control method includes: an image processing module, a volume solving module, and a predicted power module;

[0099] The image processing module is used to collect the image of the CCD during the melting process of the single crystal furnace, preprocess the image and then send it to the segmentation model, and the segmentation model will output the pixels corresponding to the massive silicon material, crystalline silicon material, and liquid silicon material;

[0100] The volume solving module is used to obtain the number of pixels of the massive silicon material, crystalline silicon material, and liquid silicon material, and calculate the actual area values of the massive silicon material and liquid silicon material in the single crystal furnace;

[0101] The power adjustment module is used to estimate the adjustment amplitude of the heater power according to the volume, volume change rate, and time change rate of the massive silicon material.

[0102] Working principle: It is necessary to first obtain the images of the molten material in the single crystal furnace collected by the CCD camera. Among them, when the silicon material is initially put into the single crystal furnace, the images of the furnace collected by the CCD are completely black because the silicon material has a low temperature and does not emit light. As the power of the heater gradually increases, the temperature of the silicon material also gradually rises. After the silicon material starts to emit light, the brightness of the images of the single crystal furnace collected by the CCD begins to approach the normal situation. Based on the collected images, a segmentation model is used for detection to segment the blocky silicon material area, the crystalline silicon material area, and the liquid silicon material area. According to the obtained regional results of each type of silicon material, the solid-liquid ratio in the single crystal furnace is calculated. According to the calculated area of the blocky silicon material, combined with the power and temperature changes, the volume of the blocky silicon material in the furnace is estimated. Finally, through the obtained melting volume change rate and time change rate, the amplitude of the change required for adjusting the heater power is calculated. Through the full melting control method based on the CCD, the solid-liquid interface and the solid-liquid percentage are monitored in real time. By adjusting the heating power and the position of the thermal shield, the shape of the solid-liquid interface is ensured to be stable, reducing the thermal stress and dislocation in the crystal, and improving the overall quality and uniformity of the single crystal. By precisely controlling the heating power and heating time, sufficient energy is provided when needed, significantly reducing energy consumption, avoiding overheating, increasing the service life of the equipment, and reducing maintenance costs. By collecting data in real time and dynamically adjusting the process parameters according to the actual solid-liquid percentage, various variables can be flexibly responded to. By dynamically adjusting the heating time according to the solid-liquid percentage detected by experiments, when the target ratio is reached, it can be transferred to the next stage, avoiding unnecessary waiting time, improving production efficiency, and increasing the output capacity of the single crystal furnace.

[0103] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and the descriptions in the specification only illustrate the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A CCD-based full melt control method, characterized in that: The following steps are involved: S101: Acquire an image of the melt to be detected in the single crystal furnace collected by a CCD; S102: Detect the image acquired in S101 based on the image segmentation model and segment the target area; S103: Calculate the solid-liquid ratio in the furnace based on the pixel results of each type of silicon material obtained in S102; S104: based on the area of ​​the block silicon material calculated in S103, the volume of the block silicon material in the furnace is estimated in combination with the power and temperature changes; S105: Calculate the amplitude of the change required to adjust the heater power by using the acquired melting volume change rate and time change rate.

2. A CCD-based full melting control method according to claim 1, characterized in that: The CCD in S101 collects images of the melting process in the single crystal furnace for subsequent processing; the melting process is from the moment when the solid block silicon material is first put into the single crystal furnace to the moment when the entire melting process ends.

3. A CCD-based full melting control method according to claim 1, characterized in that: The method for obtaining the segmentation model in S102 is: S201: marking the type of silicon material by drawing contour points on the image using software; The types of silicon materials include liquid silicon materials, block silicon materials and crystalline silicon materials; S202: Generate a dedicated data set from the labeled images in the form of digital coding; S203: Provide the data set to a segmentation model training based on deep learning, and obtain the segmentation model after the training is completed.

4. A CCD-based full melting control method according to claim 1, characterized in that: The method for calculating the solid-liquid ratio in the furnace in S103 is: S301: Based on S103, the area of ​​the block silicon material in the furnace is obtained, and a model for calculating the volume of the silicon material is established; S302: Calculate the total heat in the furnace according to the heater power and heating time in the furnace; Among them, the calculation expression for the total heat input into the furnace is: Q=P·t In the above formula, Q is the total heat input into the furnace, P is the heating power, and t is the heating time; S303: Calculate whether it is melted according to the temperature change, and calculate the volume according to the area and height; The calculation expression for calculating temperature change by heat is: Among them, ΔT is the temperature change in the furnace, ρ is the density of the bulk silicon material in the furnace, and c is the specific heat capacity of the bulk silicon material in the furnace; the height h of the bulk silicon material in the furnace can be calculated through the available variables; the volume of the bulk silicon material in the furnace can be estimated as V = A·h; S304: describing the volume change of the silicon material by the density difference between the solid state and the liquid state of silicon; Among them, during the melting process, the calculation expression of the volume change ΔV is: In the above formula, V0 is the initial volume of silicon material, ρ 液 and ρ 固 They are the densities of silicon material in liquid and solid states respectively.

5. A CCD-based full melting control method according to claim 1, characterized in that: The method for estimating the volume of the block silicon material in the furnace by combining the power and temperature changes in S104 is calculated based on the thermal field model of the melting of the silicon material, and the specific steps are: S401: It is necessary to establish a thermal field model for the furnace environment, including the location and size of the heater, crucible, and silicon material; S402: using numerical simulation software to solve the thermal field model, obtain the distribution of the temperature field in the furnace through simulation calculation, and calculate the time and heat required for melting the silicon material according to the change of the temperature field; S403: When the time and heat required for melting the silicon material are known, the volume change during the melting process of the silicon material can be calculated in combination with the physical properties of the silicon material; S404: Under the condition that the temperature field in the furnace is uniformly distributed, the volume of the block silicon material is estimated through geometric calculation or proportional relationship.

6. A CCD-based full melting control method according to claim 1, characterized in that: The amplitude of the change required to adjust the heater power in S105 is expressed as: Where P is the heater power; ρ is the density of silicon; V is the current volume of silicon material; c is the specific heat capacity of silicon material; L f is the melting latent heat of silicon material; is the silicon material temperature change rate; is the volume change rate of silicon material.

7. A CCD-based full melting control method according to claim 4, characterized in that: The heating power is equal to the working power of the heaters on both sides and the bottom of the crucible.

8. A CCD-based full melt control system, applicable to a CCD-based full melt control method according to any one of claims 1 to 7, characterized in that: The execution modules of the above control method include: an image processing module, a volume solving module, and a power prediction module; The image processing module is used to collect CCD images during the single crystal furnace melting process, and pre-process the images before sending them to the segmentation model. The segmentation model will output pixels corresponding to bulk silicon material, crystalline silicon material and liquid silicon material; The volume solving module is used to obtain the number of pixels of the block silicon material, the crystalline silicon material and the liquid silicon material, and calculate the actual area value of the block silicon material and the liquid silicon material in the single crystal furnace; The power adjustment module is used to estimate the adjustment amplitude of the heater power according to the volume, volume change rate and time change rate of the block silicon material.

Images