Single Crystal Furnace Finishing Control Method and Device

Abstract

This invention proposes a method and apparatus for controlling the finishing process of a single crystal furnace, relating to the field of single crystal silicon production technology. The method includes: obtaining the historical crystal tail contour diameter at historical finishing times of the single crystal furnace, and characteristic process parameters related to the historical crystal tail contour diameter; constructing a crystal tail contour diameter prediction model based on the historical crystal tail contour diameter and characteristic process parameters; predicting the real-time crystal tail contour diameter corresponding to the real-time characteristic process parameters; determining the optimal process parameters for real-time finishing of the single crystal furnace based on the real-time crystal tail contour diameter and the real-time crystal length, and using the optimal process parameters as the target characteristic process parameters for the finishing process of the single crystal furnace at the next moment. Thus, based on the real-time crystal tail contour diameter predicted by the crystal tail contour diameter prediction model, the optimal process parameters for real-time finishing of the single crystal furnace are accurately determined, realizing the rationality and efficiency of the single crystal furnace finishing process and improving the finishing success rate.

Description

Technical Field

[0001] This invention relates to the field of monocrystalline silicon production technology, and in particular to a method and apparatus for controlling the finishing process of a monocrystalline furnace. Background Technology

[0002] The single-crystal furnace finishing process, as the final step in crystal growth, involves gradually reducing the crystal diameter to a critical point before it leaves the solution surface. This aims to prevent dislocation back extension caused by thermal stress from damaging the single-crystal structure and resulting in the loss of the constant-diameter portion. However, during finishing, as the diameter gradually shrinks, the aperture is obscured by the constant-diameter portion of the crystal, making it impossible to measure the tail diameter by capturing the aperture image. This makes it difficult to observe and control the tail shape in real time.

[0003] In related technologies, there are two main methods for controlling the finishing process of a single crystal furnace. One method is based on preset control parameters according to process experience. The disadvantage is that it cannot guarantee that the set finishing process is optimal, resulting in material waste. Furthermore, the control strategy is relatively rigid, making it difficult to correct tail shape deviations in real time, and its consistency and universality are poor. The other method is based on real-time control parameters to correct the tail shape according to crystal weight deviation. The disadvantage is that crystal weight is greatly affected by historical diameter, has strong lag, and is highly dependent on the accuracy of the weight sensor. Using its deviation as the control target can easily cause significant system overshoot, increasing the risk of tail breakage. Therefore, a more reliable single crystal furnace finishing control method is urgently needed. Summary of the Invention

[0004] A first aspect of this invention proposes a method for controlling the finishing process of a single crystal furnace. The method includes: acquiring a first dataset of historical finishing times of the single crystal furnace, wherein the first dataset includes historical crystal tail contour diameters and characteristic process parameters related to the historical crystal tail contour diameters; constructing a crystal tail contour diameter prediction model based on the historical crystal tail contour diameters and the characteristic process parameters; inputting real-time characteristic process parameters into the crystal tail contour diameter prediction model to predict the real-time crystal tail contour diameter corresponding to the real-time characteristic process parameters; and determining the optimal process parameters for real-time finishing of the single crystal furnace based on the real-time crystal tail contour diameter and the real-time crystal length, so as to use the optimal process parameters as the target characteristic process parameters for the finishing process of the single crystal furnace at the next moment.

[0005] In one embodiment of the present invention, the step of obtaining the first dataset at the historical end of a single crystal furnace, wherein the first dataset includes the historical crystal tail contour diameter and the characteristic process parameters related to the historical crystal tail contour diameter, includes: obtaining the production data at the historical end of the single crystal furnace within a preset period, and performing abnormal data screening on the production data to obtain the screened production backup data; and constructing the first dataset at the historical end of the single crystal furnace based on the mapping relationship between the historical crystal tail contour diameter and the characteristic process parameters in the production backup data.

[0006] In one embodiment of the present invention, the step of constructing a crystal tail profile diameter prediction model based on the historical crystal tail profile diameter and the characteristic process parameters includes: calculating the historical crystal tail profile derived diameter based on the characteristic process parameters; using the characteristic process parameters and the historical crystal tail profile derived diameter as input variables, and using the historical crystal tail profile diameter as an output variable, to construct the crystal tail profile diameter prediction model.

[0007] In one embodiment of the present invention, determining the optimal process parameters for real-time end-of-life processing of the single crystal furnace based on the real-time crystal tail profile diameter and the real-time crystal length, and using the optimal process parameters as the target characteristic process parameters for the next end-of-life processing of the single crystal furnace, includes: ending the single crystal furnace end-of-life processing when the real-time crystal length is greater than a preset minimum length, the real-time crystal tail profile diameter is less than a preset maximum end diameter, and the real-time crystal length is greater than the real-time crystal tail profile diameter; and constructing a cost function for constraints on the single crystal furnace heater power and crystal pulling speed when the real-time crystal length is less than or equal to the preset minimum length, or the real-time crystal tail profile diameter is greater than or equal to the preset maximum end diameter, or the real-time crystal length is less than or equal to the real-time crystal tail profile diameter, to solve for the optimal crystal pulling speed and the optimal heater power, and using the optimal crystal pulling speed and the optimal heater power as the target characteristic process parameters for the next end-of-life processing of the single crystal furnace.

[0008] In one embodiment of the present invention, the step of constructing a cost function for the single crystal furnace heater power and crystal pulling speed constraints when the real-time crystal length is less than or equal to a preset minimum length, or the real-time crystal tail profile diameter is greater than or equal to a preset maximum tail diameter, or the real-time crystal length is less than or equal to the real-time crystal tail profile diameter, to solve for the optimal crystal pulling speed and optimal heater power, and using the optimal crystal pulling speed and optimal heater power as the target characteristic process parameters at the next moment of the single crystal furnace's final stage, includes: when the real-time crystal length is less than or equal to a preset minimum length, or the real-time crystal tail profile diameter is greater than or equal to a preset maximum tail diameter, or the real-time crystal length is less than or equal to the real-time crystal tail profile diameter... In the case of a diameter, based on real-time characteristic process parameters, the search range for the optimal crystal pulling speed and optimal heater power at the next moment of the single crystal furnace's end-of-life is determined, as well as the characteristic candidate process parameters at the next moment of the single crystal furnace's end-of-life. The crystal tail profile diameter prediction model is called, and the characteristic candidate process parameters at the next moment are used as input to obtain the predicted candidate diameter corresponding to the characteristic candidate process parameters. Based on the predicted candidate diameter, search range, and real-time characteristic process parameters, a cost function constraining the single crystal furnace heater power and crystal pulling speed is constructed to solve for the optimal crystal pulling speed and optimal heater power, and the optimal crystal pulling speed and optimal heater power are used as the target characteristic process parameters at the next moment of the single crystal furnace's end-of-life.

[0009] In one embodiment of the present invention, a second dataset is obtained at the end of a single crystal furnace. The second dataset includes real-time characteristic process parameters for each end of multiple cycles within a preset time period, and the measured crystal tail contour diameter corresponding to each real-time characteristic process parameter. The measured crystal tail contour diameter is divided into different classes according to the diameter curve. The average end-of-cycle loss of each tail contour is calculated. The class with the smallest end-of-cycle loss is the optimal tail contour class, and the target diameter curve with the optimal tail contour class as the cluster center is the optimal diameter curve. Based on the single crystal furnace power curve and crystal pulling speed curve corresponding to the optimal tail contour class, the optimal single crystal furnace power curve and optimal crystal pulling speed curve corresponding to the optimal diameter curve are fitted. A crystal tail contour diameter prediction model is obtained when the difference between the optimal single crystal furnace power curve, optimal crystal pulling speed curve, and optimal diameter curve at multiple end-of-cycle cycles is less than a preset difference threshold. The crystal tail contour diameter prediction model is used as the target crystal tail contour diameter prediction model at the end of a single crystal furnace.

[0010] This invention proposes a single-crystal furnace tail-end control method. It obtains the historical crystal tail contour diameter at previous tail-end times and the related characteristic process parameters. Based on the historical crystal tail contour diameter and characteristic process parameters, a crystal tail contour diameter prediction model is constructed. The real-time crystal tail contour diameter corresponding to the real-time characteristic process parameters is predicted. Based on the real-time crystal tail contour diameter and real-time crystal length, the optimal process parameters for real-time tail-end times in the single-crystal furnace are determined. These optimal process parameters are then used as the target characteristic process parameters for the next tail-end times in the single-crystal furnace. Thus, based on the real-time crystal tail contour diameter predicted by the tail-end diameter prediction model, the optimal process parameters for real-time tail-end times in the single-crystal furnace are accurately determined, achieving rationality and efficiency in the single-crystal furnace tail-end process and improving the tail-end success rate.

[0011] A second aspect of the present invention provides a single crystal furnace finishing control device, the device comprising: an acquisition module for acquiring a first dataset of historical finishing times of the single crystal furnace, wherein the first dataset includes historical crystal tail contour diameters and characteristic process parameters related to the historical crystal tail contour diameters; a construction module for constructing a crystal tail contour diameter prediction model based on the historical crystal tail contour diameters and the characteristic process parameters; a prediction module for inputting real-time characteristic process parameters into the crystal tail contour diameter prediction model to predict the real-time crystal tail contour diameter corresponding to the real-time characteristic process parameters; and a determination module for determining the optimal process parameters for real-time finishing times of the single crystal furnace based on the real-time crystal tail contour diameter and the real-time crystal length, so as to use the optimal process parameters as the target characteristic process parameters for the finishing times of the single crystal furnace at the next moment.

[0012] In one embodiment of the present invention, the acquisition module is specifically used to: acquire production data of the single crystal furnace at the historical end of the production within a preset period, and perform abnormal data screening on the production data to obtain the screened production backup data; and construct a first dataset of the single crystal furnace at the historical end of the production based on the mapping relationship between the historical crystal tail contour diameter and the characteristic process parameters in the production backup data.

[0013] In one embodiment of the present invention, the construction module is specifically used to: calculate the derived diameter of the historical crystal tail profile based on the characteristic process parameters; and construct the crystal tail profile diameter prediction model by using the characteristic process parameters and the derived diameter of the historical crystal tail profile as input variables and the historical crystal tail profile diameter as output variables.

[0014] In one embodiment of the present invention, the determining module includes: a determining unit, configured to determine that the single crystal furnace finishing process ends when the real-time crystal length is greater than a preset minimum length, the real-time crystal tail profile diameter is less than a preset maximum tail diameter, and the real-time crystal length is greater than the real-time crystal tail profile diameter; and a solving unit, configured to construct a cost function for the single crystal furnace heater power and crystal pulling speed constraints when the real-time crystal length is less than or equal to the preset minimum length, or the real-time crystal tail profile diameter is greater than or equal to the preset maximum tail diameter, or the real-time crystal length is less than or equal to the real-time crystal tail profile diameter, in order to solve for the optimal crystal pulling speed and the optimal heater power, and to use the optimal crystal pulling speed and the optimal heater power as the target characteristic process parameters for the single crystal furnace finishing process at the next moment.

[0015] In one embodiment of the present invention, the solving unit is specifically configured to: determine the search range of the optimal crystal pulling speed and the optimal heater power at the next moment of the single crystal furnace's end-of-life, and the candidate process parameters for the next moment of the single crystal furnace's end-of-life, based on the real-time characteristic process parameters, when the real-time crystal length is less than or equal to a preset minimum length, or the real-time crystal tail profile diameter is greater than or equal to a preset maximum tail diameter, or the real-time crystal length is less than or equal to the real-time crystal tail profile diameter; call the crystal tail profile diameter prediction model, and take the candidate process parameters for the next moment of the single crystal furnace as input to obtain the predicted candidate diameter corresponding to the candidate process parameters; construct a cost function for the single crystal furnace heater power and crystal pulling speed constraints based on the predicted candidate diameter, the search range, and the real-time characteristic process parameters, so as to solve for the optimal crystal pulling speed and the optimal heater power, and use the optimal crystal pulling speed and the optimal heater power as the target characteristic process parameters for the next moment of the single crystal furnace's end-of-life.

[0016] In one embodiment of the present invention, the apparatus further includes: a partitioning module for acquiring a second dataset at the end of a single crystal furnace, wherein the second dataset includes real-time characteristic process parameters for each end of multiple end-of-life cycles within a preset time period, and the measured crystal tail contour diameter corresponding to each real-time characteristic process parameter, and partitioning the measured crystal tail contour diameter into different classes of tail contours according to the diameter curve; a calculation module for calculating the average end-of-life loss of each tail contour, wherein the class with the smallest end-of-life loss is the optimal tail contour class, and the target diameter curve with the optimal tail contour class as the cluster center is the optimal diameter curve; a fitting module for fitting the optimal single crystal furnace power curve and the optimal crystal pulling speed curve corresponding to the optimal diameter curve according to the single crystal furnace power curve and the crystal pulling speed curve corresponding to the optimal tail contour class; and a generation module for acquiring a crystal tail contour diameter prediction model corresponding to when the difference between the optimal single crystal furnace power curve, the optimal crystal pulling speed curve, and the optimal diameter curve at multiple end-of-life cycles is less than a preset difference threshold, and using the crystal tail contour diameter prediction model as the target crystal tail contour diameter prediction model at the end of a single crystal furnace.

[0017] This invention proposes a single-crystal furnace finishing control device. It acquires the historical crystal tail contour diameter at previous finishing times and the related characteristic process parameters. Based on the historical crystal tail contour diameter and characteristic process parameters, a crystal tail contour diameter prediction model is constructed. The real-time crystal tail contour diameter corresponding to the real-time characteristic process parameters is predicted. Based on the real-time crystal tail contour diameter and real-time crystal length, the optimal process parameters for real-time finishing of the single-crystal furnace are determined. These optimal process parameters are then used as the target characteristic process parameters for the next finishing time of the single-crystal furnace. Thus, based on the real-time crystal tail contour diameter predicted by the crystal tail contour diameter prediction model, the optimal process parameters for real-time finishing of the single-crystal furnace are accurately determined, achieving rationality and efficiency in the single-crystal furnace finishing process and improving the finishing success rate.

[0018] A third aspect of the present invention provides an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the method described in the first aspect.

[0019] A fourth aspect of the present invention provides a non-transitory computer-readable storage medium storing computer instructions for causing the computer to perform the method described in the first aspect.

[0020] A fifth aspect of the present invention provides a computer program product that, when executed by a processor, implements the method described in the first aspect.

[0021] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0022] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0023] Figure 1 This is a flowchart illustrating a single crystal furnace finishing control method provided in an embodiment of the present invention;

[0024] Figure 2 This is a schematic flowchart of another single crystal furnace finishing control method provided in an embodiment of the present invention;

[0025] Figure 3 This is a schematic diagram of the structure of a single crystal furnace finishing control device provided in an embodiment of the present invention. Detailed Implementation

[0026] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0027] It should be noted that the acquisition, storage, use, and processing of data in the technical solution of this invention all comply with the relevant provisions of national laws and regulations.

[0028] The single crystal furnace finishing control method and apparatus of the present invention are described below with reference to the accompanying drawings.

[0029] Figure 1 This is a flowchart illustrating a single crystal furnace finishing control method provided in an embodiment of the present invention.

[0030] like Figure 1 As shown, the method includes the following steps:

[0031] Step 101: Obtain the first dataset from the historical end-of-life of the single crystal furnace. The first dataset includes the historical crystal tail profile diameter and the characteristic process parameters related to the historical crystal tail profile diameter.

[0032] In some embodiments, obtaining a first dataset at the historical end of a single crystal furnace, wherein the first dataset includes the historical crystal tail contour diameter and characteristic process parameters related to the historical crystal tail contour diameter, can be implemented by obtaining production data at the historical end of a single crystal furnace within a preset period, and screening the production data for abnormal data to obtain the screened production backup data; based on the mapping relationship between the historical crystal tail contour diameter and characteristic process parameters in the production backup data, the first dataset at the historical end of a single crystal furnace is constructed, thereby ensuring the reliability of the first dataset.

[0033] Specifically, within a preset sampling period, production data from the historical final stages of single-crystal furnace production under the first finalization process can be acquired in batches. The finalized crystals are then fixed, and the tail contour of the crystal is sampled using methods including, but not limited to, large-scale profile rulers and profilers to obtain crystal length and measurement diameter sequences. Based on identifiers such as production time, furnace number, and single-crystal number, the corresponding production data for the measured crystal is determined, and the measurement diameter sequence is stitched into the production data using crystal length as a guide. The batch production data is then integrated, and preprocessing operations such as deleting duplicate records, removing abnormal parameter values, and linearly interpolating missing parameter values ​​with the most recent record are performed to improve data quality and form the first dataset.

[0034] Among them, the characteristic process parameters related to the diameter of the tail profile of the historical crystal include, but are not limited to, the power of the single crystal furnace heater, the liquid temperature position, the crucible pulling speed, the crystal length, the crystal weight, and the crystal pulling speed. This embodiment does not specifically limit these parameters.

[0035] Step 102: Based on the historical crystal tail profile diameter and characteristic process parameters, construct a crystal tail profile diameter prediction model.

[0036] In some embodiments, one way to construct a crystal tail profile diameter prediction model is to analyze the mapping relationship between feature process parameters and historical crystal tail profile diameters based on a Long Short-Term Memory (LSTM) network, that is, the long-term and short-term dependencies between the feature process parameter sequence and the historical crystal tail profile diameter sequence, and encapsulate them to obtain a crystal tail profile diameter prediction model.

[0037] In other embodiments, one implementation of constructing a crystal tail profile diameter prediction model based on historical crystal tail profile diameter and characteristic process parameters can calculate the derived diameter of the historical crystal tail profile based on the characteristic process parameters; the characteristic process parameters and the derived diameter of the historical crystal tail profile are used as input variables, and the historical crystal tail profile diameter is used as the output variable to construct the crystal tail profile diameter prediction model, ensuring the reliability of the crystal tail profile diameter prediction model.

[0038] Specifically, the characteristic process parameters can be processed using the frustum method to calculate the tail profile diameter of historical candidate crystals. For example, the calculation method for the tail profile diameter of historical candidate crystals using the frustum method can be as follows:

[0039]

[0040] in, The diameter of the candidate crystal tail at time t is calculated using the frustum weight formula; M(t) and L(t) are the tail crystal weight and crystal length at time t, respectively; τ is the differential time window, τ>0; initial value Take the diameter of equal diameter.

[0041] Step 103: Input the real-time feature process parameters into the crystal tail profile diameter prediction model to predict the real-time crystal tail profile diameter corresponding to the real-time feature process parameters.

[0042] In some embodiments, the real-time characteristic process parameters can be the real-time characteristic process parameters of any target single crystal furnace. By using the real-time characteristic process parameters of the target single crystal furnace, the real-time finishing process of the target single crystal furnace can be determined, thereby accurately predicting the real-time crystal tail profile diameter corresponding to the real-time characteristic process parameters. This eliminates reliance on experience and objectively and accurately quantifies parameter output, achieving precise control over the tail shape and minimizing finishing losses.

[0043] Step 104: Based on the real-time crystal tail profile diameter and real-time crystal length, determine the optimal process parameters for the real-time closing of the single crystal furnace, so as to use the optimal process parameters as the target characteristic process parameters for the closing of the single crystal furnace at the next moment.

[0044] In some embodiments, the optimal process parameters for real-time finishing of the single crystal furnace are determined based on the real-time crystal tail profile diameter and the real-time crystal length. One implementation method for using these optimal process parameters as the target characteristic process parameters for single crystal furnace finishing is as follows: When the real-time crystal length is greater than a preset minimum length, the real-time crystal tail profile diameter is less than a preset maximum end diameter, and the real-time crystal length is greater than the real-time crystal tail profile diameter, the single crystal furnace finishing process ends, and the real-time characteristic process parameter corresponding to the real-time crystal tail profile diameter is used as the target characteristic process parameter for single crystal furnace finishing. When the real-time crystal length is less than or equal to the preset minimum length, or the real-time crystal tail profile diameter is greater than or equal to the preset maximum end diameter, or the real-time crystal length is less than or equal to the real-time crystal tail profile diameter, a cost function constraining the single crystal furnace heater power and crystal pulling speed is constructed to solve for the optimal crystal pulling speed and optimal heater power. The optimal crystal pulling speed and optimal heater power are then used as the target characteristic process parameters for single crystal furnace finishing at the next moment. This accurately determines the optimal process parameters for real-time finishing of the single crystal furnace, ensuring the high efficiency of single crystal furnace finishing.

[0045] Specifically, when the real-time crystal length is less than or equal to the preset minimum length, or the real-time crystal tail profile diameter is greater than or equal to the preset maximum tail diameter, or the real-time crystal length is less than or equal to the real-time crystal tail profile diameter, the search range for the optimal crystal pulling speed and optimal heater power at the next moment of the single crystal furnace's end-of-life is determined based on the real-time characteristic process parameters, as well as the characteristic candidate process parameters at the next moment of the single crystal furnace's end-of-life. The crystal tail profile diameter prediction model is called, and the characteristic candidate process parameters at the next moment are used as input to obtain the predicted candidate diameter corresponding to the characteristic candidate process parameters. Based on the predicted candidate diameter, search range, and real-time characteristic process parameters, a cost function for the single crystal furnace heater power and crystal pulling speed constraints is constructed to solve for the optimal crystal pulling speed and optimal heater power, and the optimal crystal pulling speed and optimal heater power are used as the target characteristic process parameters at the next moment of the single crystal furnace's end-of-life.

[0046] Optionally, the minimum real-time crystal length at the preset termination point is L. min and the maximum end diameter D of the real-time crystal tail profile max In this case, if the length of the tail crystal is greater than the preset minimum length L, min And the predicted diameter is smaller than the preset maximum end diameter D max If the real-time crystal length is greater than the maximum tail diameter of the real-time crystal, then the crystallization process ends and a segment is taken for crystallization; if the real-time crystal length is less than or equal to the minimum length L... min Or, the real-time crystal tail profile diameter is greater than or equal to the preset maximum tail diameter D. max If the real-time crystal length is less than or equal to the real-time crystal tail profile diameter, the search range for the optimal heater power and crystal pulling speed at the next moment is determined based on the set real-time characteristic process parameters.

[0047] P best (t+1)∈[max(P set (t+1)-ΔP,P min ),min(P set (t+1)+ΔP,P max )]

[0048] V best (t+1)∈[max(V set (t+1)-ΔV,V min ),min(V set (t+1)+ΔV,V max )]

[0049] Among them, P best (t+1) and V best(t+1) represent the optimal output power and crystal pulling speed at time t+1, respectively; P set (t+1) and V set (t+1) represents the set heater power and crystal pulling speed at time t+1, respectively; ΔP and ΔV are the preset maximum variations in power and pulling speed, respectively. If the number of iterations for optimization in the current process is small, they should be larger; otherwise, they should be smaller. min and V min These are the preset minimum power and pulling speed at the end of the pull, respectively, P max and V max These are the preset maximum power and pulling speed for the final pull.

[0050] Then, the crystal tail profile diameter prediction model is invoked, and the feature candidate process parameters at the next time step are used as input feature variables to obtain the predicted candidate diameter D corresponding to the feature candidate process parameters. pred (P,V); and based on the predicted candidate diameter, search range, and real-time characteristic process parameters, a cost function for the single crystal furnace heater power and crystal pulling speed constraints is constructed. Simultaneously, the following optimization problem is established to determine the optimal crystal pulling speed V. best (t+1) and optimal heater power P best Solving for (t+1):

[0051]

[0052] Among them, D set (t+1) is the set crystal tail profile diameter at time t+1; λ1 and λ2 are the penalty term coefficients for the single crystal furnace heater power and crystal pulling speed, respectively;

[0053] Among them, the optimal crystal pulling speed V best (t+1) and optimal heater power P best The solution to (t+1) can be obtained by methods including but not limited to enumeration, simulated annealing, and genetic algorithms. This embodiment does not make any specific limitations on these methods.

[0054] The single-crystal furnace finishing control method of this invention obtains the historical crystal tail contour diameter at the historical finishing time of the single-crystal furnace, as well as the characteristic process parameters related to the historical crystal tail contour diameter; based on the historical crystal tail contour diameter and characteristic process parameters, a crystal tail contour diameter prediction model is constructed; the real-time crystal tail contour diameter corresponding to the real-time characteristic process parameters is predicted; based on the real-time crystal tail contour diameter and the real-time crystal length, the optimal process parameters for the real-time finishing of the single-crystal furnace are determined, and the optimal process parameters are used as the target characteristic process parameters for the finishing of the single-crystal furnace at the next moment. Thus, based on the real-time crystal tail contour diameter predicted by the crystal tail contour diameter prediction model, the optimal process parameters for the real-time finishing of the single-crystal furnace are accurately determined, realizing the rationality and efficiency of the single-crystal furnace finishing process and improving the finishing success rate.

[0055] To clearly illustrate the above embodiment, this embodiment also provides a method for controlling the finishing process of a single crystal furnace. Figure 2 This is a flowchart illustrating another single-crystal furnace finishing control method provided in an embodiment of the present invention.

[0056] like Figure 2 As shown, the method may include the following steps:

[0057] Step 201: Obtain the first dataset from the historical end-of-life of the single crystal furnace. The first dataset includes the historical crystal tail profile diameter and the characteristic process parameters related to the historical crystal tail profile diameter.

[0058] Step 202: Based on the mapping relationship between the historical crystal tail profile diameter and characteristic process parameters, construct a crystal tail profile diameter prediction model.

[0059] Step 203: Input the real-time feature process parameters into the crystal tail profile diameter prediction model to predict the real-time crystal tail profile diameter corresponding to the real-time feature process parameters.

[0060] Step 204: Based on the real-time crystal tail profile diameter and the preset profile diameter range threshold, determine the optimal process parameters for the real-time closing of the single crystal furnace, so as to use the optimal process parameters as the target characteristic process parameters for the closing of the single crystal furnace at the next moment.

[0061] It should be noted that the specific implementation methods of steps 201 to 204 can be found in the relevant descriptions in the above embodiments.

[0062] Step 205: Obtain the second dataset at the end of the single crystal furnace. The second dataset includes the real-time characteristic process parameters at each end within a preset time period, as well as the measured crystal tail contour diameter corresponding to each real-time characteristic process parameter. The measured crystal tail contour diameter is then divided into different types of tail contours according to the diameter curve.

[0063] In some embodiments, clustering algorithms can be used to define curve similarity with dynamic time warping, and tail contours of real-time crystals can be divided into different classes according to the diameter curve.

[0064] The preset time period can be determined based on the single crystal furnace finishing cycle, but is not limited to this.

[0065] Step 206: Calculate the average tail loss of each tail contour. The class with the smallest tail loss is the optimal tail contour class, and the target diameter curve with the optimal tail contour class as the cluster center is the optimal diameter curve.

[0066] In some embodiments, one method for calculating the average termination loss of each tail profile can be:

[0067] Loss = M mid +M tail +w*T tail

[0068] Among them, M mid The reverse shear weight represents the weight removed when a crack, fracture, or broken edge at the tail of the crystal causes dislocations to extend backward to a segment of equal diameter; M tail T represents the weight of the crystal tail. tail is the finishing time for crystal production; w is the average effective output of crystals per hour.

[0069] Step 207: Based on the single crystal furnace power curve and crystal pulling speed curve corresponding to the optimal tail profile type, fit the optimal single crystal furnace power curve and optimal crystal pulling speed curve corresponding to the optimal diameter curve.

[0070] In some embodiments, based on the single crystal furnace power curve and crystal pulling speed curve corresponding to the optimal tail profile class, the least squares method can be used to fit the optimal single crystal furnace power curve and optimal crystal pulling speed curve corresponding to the optimal diameter curve, but this is not limited to this.

[0071] Step 208: Obtain the crystal tail profile diameter prediction model corresponding to the optimal single crystal furnace power curve, optimal crystal pulling speed curve, and optimal diameter curve when the difference is less than a preset difference threshold during multiple finishing processes, and use the crystal tail profile diameter prediction model as the target crystal tail profile diameter prediction model during single crystal furnace finishing processes.

[0072] Optionally, after initially obtaining the optimal single crystal furnace power curve, optimal crystal pulling speed curve, and optimal diameter curve, these are used as the set process parameters for crystal preparation. The corresponding measured crystal tail profile diameter is collected. Based on the set process parameters and the corresponding measured crystal tail profile diameter, a training set, a test set, and a validation set are established. The crystal tail profile diameter prediction model is iteratively optimized until the difference between the optimal single crystal furnace power curve, optimal crystal pulling speed curve, and optimal diameter curve at the end approaches 0. The crystal tail profile diameter prediction model at this point is then used as the target crystal tail profile diameter prediction model.

[0073] In some embodiments, the optimal single crystal furnace power curve, the optimal crystal pulling speed curve, and the optimal diameter curve can be digitally filtered and smoothed to improve the reliability of each data in the second dataset.

[0074] The single-crystal furnace finishing control method of this invention obtains the historical crystal tail contour diameter at historical finishing times of the single-crystal furnace, and the characteristic process parameters related to the historical crystal tail contour diameter; constructs a crystal tail contour diameter prediction model based on the historical crystal tail contour diameter and the characteristic process parameters; predicts the real-time crystal tail contour diameter corresponding to the real-time characteristic process parameters; determines the optimal process parameters for real-time finishing of the single-crystal furnace based on the real-time crystal tail contour diameter and the real-time crystal length, and uses the optimal process parameters as the target characteristic process parameters for single-crystal furnace finishing; and obtains a second dataset for single-crystal furnace finishing, wherein the second dataset includes the real-time characteristic process parameters for each finishing time within a preset time period, and the measured crystal length corresponding to each real-time characteristic process parameter. The tail contour diameter is determined, and the measured crystal tail contour diameter is categorized into different classes based on the diameter curve. The average tail loss of each tail contour is calculated, and the class with the smallest tail loss is identified as the optimal tail contour class. The target diameter curve with the optimal tail contour class as the cluster center is designated as the optimal diameter curve. Based on the single-crystal furnace power curve and crystal pulling speed curve corresponding to the optimal tail contour class, the optimal single-crystal furnace power curve and optimal crystal pulling speed curve corresponding to the optimal diameter curve are fitted. The crystal tail contour diameter prediction model is obtained when the difference between the optimal single-crystal furnace power curve, optimal crystal pulling speed curve, and optimal diameter curve during multiple tailing processes is less than a preset difference threshold. This crystal tail contour diameter prediction model is then used as the target crystal tail contour diameter prediction model for single-crystal furnace tailing. Therefore, by using the optimal single-crystal furnace power curve, optimal crystal pulling speed curve, and optimal diameter curve corresponding to real-time characteristic process parameters as a second dataset, the crystal tail contour diameter prediction model is iteratively optimized using this second dataset. This allows for real-time correction of the crystal tail contour diameter, correcting crystal tail shape deviations, and improving the consistency and error correction capability of single-crystal furnace tailing.

[0075] To achieve the above embodiments, the present invention also proposes a single crystal furnace finishing control device.

[0076] Figure 3 This is a schematic diagram of the structure of a single crystal furnace finishing control device provided in an embodiment of the present invention.

[0077] like Figure 3 As shown, the single crystal furnace finishing control device 30 includes: an acquisition module 31, a construction module 32, a prediction module 33, and a determination module 34.

[0078] The acquisition module 31 is used to acquire a first dataset at the historical end of a single crystal furnace, wherein the first dataset includes the historical crystal tail profile diameter and characteristic process parameters related to the historical crystal tail profile diameter.

[0079] Module 32 is used to construct a crystal tail profile diameter prediction model based on the historical crystal tail profile diameter and the characteristic process parameters;

[0080] Prediction module 33 is used to input real-time feature process parameters into the crystal tail profile diameter prediction model to predict the real-time crystal tail profile diameter corresponding to the real-time feature process parameters.

[0081] The determination module 34 is used to determine the optimal process parameters for the real-time closing of the single crystal furnace based on the real-time crystal tail profile diameter and the real-time crystal length, so as to use the optimal process parameters as the target characteristic process parameters for the closing of the single crystal furnace at the next moment.

[0082] Furthermore, in one possible implementation of this invention, the acquisition module 31 is specifically used for:

[0083] Obtain the production data of the single crystal furnace at the end of the historical period within the preset period, and perform abnormal data screening on the production data to obtain the screening production backup data;

[0084] Based on the mapping relationship between the historical crystal tail profile diameter and characteristic process parameters in the production backup data, the first dataset of the historical end-of-life of the single crystal furnace is constructed.

[0085] Furthermore, in one possible implementation of this invention, the construction module 32 is specifically used for:

[0086] Based on the aforementioned characteristic process parameters, the derived diameter of the historical crystal tail profile is calculated;

[0087] The characteristic process parameters and the derived diameter of the historical crystal tail profile are used as input variables, and the historical crystal tail profile diameter is used as the output variable to construct the crystal tail profile diameter prediction model.

[0088] Furthermore, in one possible implementation of this invention, the determining module 34 includes:

[0089] The determining unit is used to end the single crystal furnace when the real-time crystal length is greater than the preset minimum length, the real-time crystal tail profile diameter is less than the preset maximum tail diameter, and the real-time crystal length is greater than the real-time crystal tail profile diameter.

[0090] The solving unit is used to construct a cost function for the constraints of the single crystal furnace heater power and crystal pulling speed when the real-time crystal length is less than or equal to the preset minimum length, or the real-time crystal tail profile diameter is greater than or equal to the preset maximum tail diameter, or the real-time crystal length is less than or equal to the real-time crystal tail profile diameter, so as to solve for the optimal crystal pulling speed and the optimal heater power, and use the optimal crystal pulling speed and the optimal heater power as the target characteristic process parameters when the single crystal furnace is finished at the next moment.

[0091] Furthermore, in one possible implementation of this invention, the solving unit is specifically used for:

[0092] When the real-time crystal length is less than or equal to the preset minimum length, or the real-time crystal tail profile diameter is greater than or equal to the preset maximum tail diameter, or the real-time crystal length is less than or equal to the real-time crystal tail profile diameter, the search range of the optimal crystal pulling speed and the optimal heater power at the next moment of the single crystal furnace's end is determined based on the real-time characteristic process parameters, as well as the characteristic candidate process parameters at the next moment of the single crystal furnace's end.

[0093] The crystal tail profile diameter prediction model is invoked, and the feature candidate process parameters at the next time moment are used as input to obtain the predicted candidate diameter corresponding to the feature candidate process parameters;

[0094] Based on the predicted candidate diameter, search range, and real-time characteristic process parameters, a cost function constraining the single crystal furnace heater power and crystal pulling speed is constructed to solve for the optimal crystal pulling speed and optimal heater power. The optimal crystal pulling speed and optimal heater power are then used as the target characteristic process parameters for the single crystal furnace at the next moment of closure.

[0095] Furthermore, in one possible implementation of this invention, the apparatus further includes:

[0096] The segmentation module is used to obtain a second dataset at the end of the single crystal furnace. The second dataset includes real-time characteristic process parameters at each end within a preset time period, and the measured crystal tail contour diameter corresponding to each real-time characteristic process parameter. The measured crystal tail contour diameter is then divided into different types of tail contours according to the diameter curve.

[0097] The calculation module is used to calculate the average tail loss of each tail contour. The class with the smallest tail loss is the optimal tail contour class, and the target diameter curve with the optimal tail contour class as the cluster center is the optimal diameter curve.

[0098] The fitting module is used to fit the optimal single crystal furnace power curve and the optimal crystal pulling speed curve corresponding to the optimal diameter curve based on the single crystal furnace power curve and the crystal pulling speed curve corresponding to the optimal tail contour class.

[0099] The second optimization module is used to obtain the crystal tail profile diameter prediction model corresponding to the optimal single crystal furnace power curve, optimal crystal pulling speed curve, and optimal diameter curve when the difference is less than a preset difference threshold during multiple finishing processes, and to use the crystal tail profile diameter prediction model as the target crystal tail profile diameter prediction model during single crystal furnace finishing processes.

[0100] The single-crystal furnace finishing control device of this invention acquires the historical crystal tail contour diameter at the historical finishing time of the single-crystal furnace, as well as the characteristic process parameters related to the historical crystal tail contour diameter; based on the historical crystal tail contour diameter and characteristic process parameters, it constructs a crystal tail contour diameter prediction model; predicts the real-time crystal tail contour diameter corresponding to the real-time characteristic process parameters; based on the real-time crystal tail contour diameter and the real-time crystal length, it determines the optimal process parameters for the real-time finishing of the single-crystal furnace, and uses the optimal process parameters as the target characteristic process parameters for the finishing of the single-crystal furnace at the next moment. Thus, based on the real-time crystal tail contour diameter predicted by the crystal tail contour diameter prediction model, the optimal process parameters for the real-time finishing of the single-crystal furnace are accurately determined, realizing the rationality and efficiency of the single-crystal furnace finishing process and improving the finishing success rate.

[0101] To achieve the above embodiments, the present invention also proposes an electronic device, comprising:

[0102] At least one processor; and

[0103] A memory communicatively connected to the at least one processor; wherein,

[0104] The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the aforementioned method.

[0105] To implement the above embodiments, the present invention also proposes a non-transitory computer-readable storage medium storing computer instructions for causing the computer to perform the aforementioned method.

[0106] To implement the above embodiments, the present invention also proposes a computer program product, including a computer program that, when executed by a processor, implements the method described above.

[0107] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0108] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0109] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of the invention pertain.

[0110] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0111] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any of the following techniques known in the art, or a combination thereof: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0112] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0113] Furthermore, the functional units in the various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0114] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A method of controlling the end of a single crystal pulling process, characterized by, The method includes: Obtain the first dataset from the historical end-of-life of the single crystal furnace, wherein the first dataset includes the historical crystal tail profile diameter and the characteristic process parameters related to the historical crystal tail profile diameter; Based on the aforementioned characteristic process parameters, the derived diameter of the historical crystal tail contour is calculated; the characteristic process parameters and the derived diameter of the historical crystal tail contour are used as input variables, and the historical crystal tail contour diameter is used as the output variable to construct a crystal tail contour diameter prediction model. The real-time feature process parameters are input into the crystal tail profile diameter prediction model to predict the real-time crystal tail profile diameter corresponding to the real-time feature process parameters. When the real-time crystal length is greater than the preset minimum length, the real-time crystal tail contour diameter is less than the preset maximum tail diameter, and the real-time crystal length is greater than the real-time crystal tail contour diameter, the single crystal furnace ends, and the real-time feature process parameter corresponding to the real-time crystal tail contour diameter is used as the target feature process parameter for the single crystal furnace ending. When the real-time crystal length is less than or equal to the preset minimum length, or the real-time crystal tail profile diameter is greater than or equal to the preset maximum tail diameter, or the real-time crystal length is less than or equal to the real-time crystal tail profile diameter, the search range of the optimal crystal pulling speed and the optimal heater power at the next moment of the single crystal furnace's end is determined based on the real-time characteristic process parameters, as well as the characteristic candidate process parameters at the next moment of the single crystal furnace's end. The crystal tail profile diameter prediction model is invoked, and the feature candidate process parameters at the next time moment are used as input to obtain the predicted candidate diameter corresponding to the feature candidate process parameters; Based on the predicted candidate diameter, search range, and real-time characteristic process parameters, a cost function constraining the single crystal furnace heater power and crystal pulling speed is constructed to solve for the optimal crystal pulling speed and optimal heater power. The optimal crystal pulling speed and optimal heater power are then used as the target characteristic process parameters for the single crystal furnace at the next moment of closure.

2. The method of claim 1, wherein, The first dataset obtained when the single crystal furnace was completed in history includes: Obtain the production data of the single crystal furnace at the end of the historical period within the preset period, and perform abnormal data screening on the production data to obtain the screened production backup data; Based on the mapping relationship between the historical crystal tail profile diameter and characteristic process parameters in the production backup data, the first dataset of the historical end-of-life of the single crystal furnace is constructed.

3. The method of claim 1, wherein, The method further includes: A second dataset is obtained at the end of the single crystal furnace. The second dataset includes real-time characteristic process parameters at each end within a preset time period, and the measured crystal tail contour diameter corresponding to each real-time characteristic process parameter. The measured crystal tail contour diameter is divided into different types of tail contours according to the diameter curve. Calculate the average tail loss of each tail contour. The class with the smallest tail loss is the optimal tail contour class. The target diameter curve with the optimal tail contour class as the cluster center is the optimal diameter curve. Based on the single crystal furnace power curve and crystal pulling speed curve corresponding to the optimal tail profile type, fit the optimal single crystal furnace power curve and optimal crystal pulling speed curve corresponding to the optimal diameter curve. Obtain the crystal tail profile diameter prediction model corresponding to the optimal single crystal furnace power curve, optimal crystal pulling speed curve, and optimal diameter curve when the difference is less than a preset difference threshold during multiple finishing stages, and use the crystal tail profile diameter prediction model as the target crystal tail profile diameter prediction model during single crystal furnace finishing stages.

4. A device for controlling the end of a single crystal pulling process, characterized by The device includes: The acquisition module is used to acquire the first dataset at the historical end of the single crystal furnace, wherein the first dataset includes the historical crystal tail profile diameter and the characteristic process parameters related to the historical crystal tail profile diameter. A construction module is used to calculate the derived diameter of the historical crystal tail profile based on the characteristic process parameters; the characteristic process parameters and the derived diameter of the historical crystal tail profile are used as input variables, and the historical crystal tail profile diameter is used as output variable, so as to construct a crystal tail profile diameter prediction model. The prediction module is used to input real-time feature process parameters into the crystal tail profile diameter prediction model to predict the real-time crystal tail profile diameter corresponding to the real-time feature process parameters. The determination module is used to determine the end of the single crystal furnace closing process when the real-time crystal length is greater than a preset minimum length, the real-time crystal tail profile diameter is less than a preset maximum tail diameter, and the real-time crystal length is greater than the real-time crystal tail profile diameter, and the real-time characteristic process parameter corresponding to the real-time crystal tail profile diameter is used as the target characteristic process parameter for the single crystal furnace closing process; when the real-time crystal length is less than or equal to the preset minimum length, or the real-time crystal tail profile diameter is greater than or equal to the preset maximum tail diameter, or the real-time crystal length is less than or equal to the real-time crystal tail profile diameter, the next step of the single crystal furnace closing process is determined based on the real-time characteristic process parameter. The search range for the optimal crystal pulling speed and optimal heater power at the end of the current time step is determined, along with the candidate characteristic process parameters for the next moment when the single crystal furnace is finished. The crystal tail profile diameter prediction model is called, and the candidate characteristic process parameters for the next moment are used as input to obtain the predicted candidate diameter corresponding to the candidate characteristic process parameters. Based on the predicted candidate diameter, the search range, and the real-time characteristic process parameters, a cost function constraining the single crystal furnace heater power and crystal pulling speed is constructed to solve for the optimal crystal pulling speed and optimal heater power. The optimal crystal pulling speed and optimal heater power are then used as the target characteristic process parameters for the next moment when the single crystal furnace is finished.

5. An electronic device, comprising: include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-3.

6. A non-transitory computer-readable storage medium having stored thereon computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-3.

7. A computer program product, characterised in that, Includes a computer program that, when executed by a processor, implements the method according to any one of claims 1-3.

Images