Electrode arc striking control system of quartz crucible

Abstract

The invention discloses an electrode arc striking control system of a quartz crucible, and relates to the technical field of electrode arc striking control systems, a pre-defining module, a target heat determination module, a time point definition module and an error acquisition module are used for acquiring an accumulated heating error according to target arc heat in a local duration before a monitoring time point; wherein the accumulated heating error represents the sum of all local heating errors before the monitoring time point, and the local heating error represents the error between the target arc heat and the actual arc heat within the local duration; the error loss construction module is used for constructing a double-error loss function according to the local heating error and the accumulated heating error; the control parameter updating module is used for updating the execution current and the execution voltage by using a gradient descent algorithm so as to minimize the double-error loss function; according to the invention, by introducing the dual-error loss function and the time-varying weight, the system can maintain the global stability of the quartz crucible in the electrode arc striking control process while ensuring the local precision.

Description

Technical Field

[0001] The present invention relates to the technical field of electrode arc ignition control systems, and particularly to an electrode arc ignition control system for a quartz crucible. Background Art

[0002] A quartz crucible is an important device used in modern industry for high-temperature melting, material processing, and other processes. It is commonly used in fields such as metal smelting, semiconductor manufacturing, and optical fiber production, and has characteristics such as high temperature resistance and corrosion resistance. In these applications, the electrode arc ignition process of the quartz crucible plays a crucial role.

[0003] Electrode arc ignition refers to generating an arc between electrodes through current to heat the substances inside the quartz crucible or perform specific process operations. The arc ignition process requires controlling the current and voltage to ensure that the arc temperature is moderate and to avoid adverse effects on the crucible or the production process caused by an overly strong or weak arc. The stability of the arc is crucial for the quality of the arc ignition process. An overly high arc temperature may cause material damage, while an overly low arc temperature may affect the melting or processing effect. Therefore, how to precisely control the current, voltage, and arc temperature during the electrode arc ignition process is the key to improving the usage efficiency and quality of the quartz crucible.

[0004] Although the prior art has been able to achieve a certain degree of electrode arc ignition control, there are still a series of defects. For example, the patent document with the patent publication number CN104174974A discloses a welding machine arc ignition control circuit. By reducing the pulse width of the pulse drive signal during welding machine arc ignition, the impact current during the welding machine arc ignition process can be significantly reduced. However, the above patent document and the prior art generally only focus on the arc error (such as local heating error) within each local time period and make adjustments based on the local error. Although this method can adjust the instantaneous error during the arc process to a certain extent, it ignores the cumulative error, that is, the error during the arc process accumulates over time, resulting in the stability of the entire arc ignition process not being effectively guaranteed. This method fails to comprehensively consider the dynamic changes in the arc ignition process, leading to the gradual amplification of errors in long-term control and being unable to effectively avoid temperature fluctuations or instability. Summary of the Invention

[0005] In view of the deficiencies of the prior art, the present invention provides an electrode arc ignition control system for a quartz crucible, which solves the technical problems proposed in the background art by constructing a double error loss.

[0006] To achieve the above objectives, the present invention is realized through the following technical solutions: An electrode arc ignition control system for a quartz crucible, the control system comprising: A predefined module for predefining the standard arc ignition duration of electrode arc ignition and its total theoretical arc heat; A target heat determination module, configured to determine the target arc heat within a local duration according to the standard arc starting duration and its total theoretical arc heat; A time point definition module, configured to define monitoring time points within the standard arc starting duration; An error acquisition module, configured to acquire the cumulative heating error according to the target arc heat within the local duration before the monitoring time point; wherein, the cumulative heating error is characterized as the sum of all local heating errors before the monitoring time point, and the local heating error characterizes the error between the target arc heat and the actual arc heat within the local duration; An error loss construction module, configured to construct a double error loss function according to the local heating error and the cumulative heating error; A control parameter update module, configured to update the execution current and the execution voltage using the gradient descent algorithm to minimize the double error loss function; A control parameter output module, configured to, when the double error loss function is minimized, input the execution current and the execution voltage at the time of minimization as the optimal control parameters into the electrode arc starting control system to control the electrode arc starting process.

[0007] In some embodiments, the target heat determination module is specifically configured to: Initialize a virtual plane that coincides with the standard arc starting duration; Construct a plurality of equally spaced parallel dividing planes in the vertical direction of the virtual plane; Use the equally spaced parallel dividing planes to cut the standard arc starting duration to obtain a plurality of adjacent local durations; Between adjacent parallel dividing planes, select any local duration arc starting point and select the local duration that includes the arc starting point; Determine the target arc heat within any local duration according to the total theoretical arc heat within the standard arc starting duration.

[0008] In some embodiments, determining the target arc heat within any local duration according to the total theoretical arc heat within the standard arc starting duration includes: S2-3-1. Determine the average heating power within the standard arc starting duration according to the ratio of the total theoretical arc heat to the standard arc starting duration; S2-3-2. Select the cutting points of the adjacent parallel dividing planes and the standard arc starting duration, define the previous cutting point as the start time point, and the subsequent cutting point as the end time point; S2-3-3. Define the duration between the start time point and the end time point as the any local duration; S2-3-4. Calculate the product of the any local duration and the average heating power, and define the product of the local duration and the average heating power as the target arc heat within the any local duration.

[0009] In some of these embodiments, the error acquisition module is specifically configured to: S4-1. Obtain the local heating error within the first local time period; S4-2. Among several adjacent local time periods, select the next local time period at the end time point of the adjacent first local time period; S4-3. Obtain the local heating error within the next local time period; S4-4. Repeatedly execute the acquisition of the local heating error, and perform cumulative summation on the repeatedly executed local heating errors until the cumulative heating error before the monitoring time point is obtained.

[0010] In some of these embodiments, obtaining the local heating error within the first local time period includes: S4-1-1. Obtain the actual arc heat within the first local time period; S4-1-2. Define the difference between the target arc heat and the actual arc heat within the first local time period as the local heating error within the first local time period.

[0011] In some of these embodiments, obtaining the actual arc heat within the first local time period includes: Obtain the predefined mass and specific heat capacity of the electrode arc-starting hot end; Collect the arc starting temperature at the start time point and the arc real-time temperature at the end time point within the first local time period; Calculate the first temperature difference between the arc real-time temperature at the end time point and the arc starting temperature at the start time point; Calculate the product between the first temperature difference and the predefined mass and specific heat capacity of the electrode arc-starting hot end; Define the product between the first temperature difference and the predefined mass and specific heat capacity of the electrode arc-starting hot end as the actual arc heat within the first local time period.

[0012] In some of these embodiments, the expression of the double error loss function is: ; Wherein, represents the double error loss, represents the local heating error, represents the weighting coefficient, indicating the relative importance of the local error and the cumulative error; represents the local error within the i-th local time period, representing the error term calculated within the local time period; represents the time-varying weight, which is used to adjust the weight of the error of each local time period in the double error loss function.

[0013] In some of these embodiments, the control parameter update module is specifically configured to: S6-1. The learning rate at which the preset execution current and execution voltage are updated in gradient descent; S6-2. Calculate the partial derivative gradients of the double error loss function with respect to the current execution current and the current execution voltage; S6-3. Update the current execution current and the current execution voltage according to the learning rate and the partial derivative gradients; S6-4. According to the updated execution current and execution voltage, obtain the local heating error and the cumulative heating error corresponding to the local duration of the updated execution current and execution voltage; S6-5. Calculate the double error loss function according to the local heating error and the cumulative heating error corresponding to the local duration; S6-6. Loop through S6-2 to S6-5 until the double error loss function is minimized.

[0014] In some of these embodiments, the calculation expression for the partial derivative gradients of the current execution current and the current execution voltage is: ; Wherein, represents the double error loss function, I is the current execution current, U is the current execution voltage, N is the number of all local durations before the monitoring time point, represents the time-varying weight, which is used to adjust the influence of errors of different local durations on the loss; represents the local error within the i-th local duration, represents the cumulative error, is the weighting coefficient, which represents the relative importance of the local error and the cumulative error, and respectively represent the partial derivative gradients of the local error with respect to the execution current and the execution voltage, and respectively represent the partial derivative gradients of the cumulative error with respect to the execution current and the execution voltage.

[0015] In some of these embodiments, the update expressions for the execution current and the execution voltage are: ; Wherein, represents the updated execution current, represents the updated execution voltage, represents the current execution current, represents the current execution voltage, represents the learning rate.

[0016] The present invention provides an electrode arc starting control system for a quartz crucible, which has the following beneficial effects: By calculating the error of each local time period, the current and voltage can be adjusted in real time to avoid overheating or too low arc temperature in a certain time period, thus avoiding quality problems caused by excessive temperature rise or fall.

[0017] Furthermore, in the control process, by combining the local error and the cumulative error, the system can not only cope with the changes within a single time period, but also maintain the overall balance of the arc process in the long term. The weighting coefficient λ controls the relative importance of the local error and the cumulative error in the double-error loss function. By adjusting λ, the system can choose to pay more attention to the local error (when λ is larger) to ensure that the error is controlled during the entire arc starting process. Finally, by combining the optimization objectives of the local heating error and the cumulative error, the optimal current and voltage sequences are output, and the global optimal control current and voltage are obtained by calculating the average values of the current and voltage of all time periods, thereby ensuring the long-term stability and accuracy of the arc starting process while minimizing the local error.

[0018] In summary, the present invention effectively combines local and cumulative error control by introducing the mechanism of the double-error loss function and time-varying weights, and realizes precise current and voltage adjustment during the entire arc starting process. Through reasonable error feedback and optimization, the system can maintain the global stability of the quartz crucible during the electrode arc starting control process while ensuring local accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a structural block diagram of an electrode arc starting control system for a quartz crucible according to the present invention; Figure 2 is a control parameter output flow chart of an electrode arc starting control system for a quartz crucible according to the present invention; Figure 3 is a schematic diagram of the acquisition process of all local heating errors of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0021] Please refer to Figures 1 to 2 , Figure 1 which provides a structural block diagram of an electrode arc starting control system for a quartz crucible, Figure 2 which provides a control parameter output flow chart of an electrode arc starting control system for a quartz crucible; wherein, the control system includes: A predefined module for predefined standard arc-starting duration of electrode arc-starting and its total theoretical arc heat; A target heat determination module for determining the target arc heat within a local duration according to the standard arc-starting duration and its total theoretical arc heat; A time point definition module for defining monitoring time points within the standard arc-starting duration; An error acquisition module for acquiring the cumulative heating error according to the target arc heat within the local duration before the monitoring time point; wherein, the cumulative heating error is characterized as the sum of all local heating errors before the monitoring time point, and the local heating error characterizes the error between the target arc heat and the actual arc heat within the local duration; An error loss construction module for constructing a double error loss function according to the local heating error and the cumulative heating error; A control parameter update module for updating the execution current and execution voltage using the gradient descent algorithm to minimize the double error loss function; Specifically, the gradient descent method is applicable to continuous and differentiable loss functions. In this embodiment, the double error loss function includes local error and cumulative error, so the double error loss function is differentiable in the parameter space of current and voltage. It calculates the partial derivatives of the double error loss function with respect to current and voltage, i.e., the gradient, and then updates the current and voltage according to the direction of the gradient.

[0022] A control parameter output module for, when the double error loss function is minimized, taking the execution current and execution voltage at the time of minimization as the optimal control parameters and inputting them into the electrode arc-starting control system to control the electrode arc-starting process.

[0023] Among them, the optimal execution current and optimal execution voltage actually output the current and voltage sequences of all periods at the time of minimum double error loss, and then obtain the globally optimal execution current and execution voltage by averaging the current and voltage sequences of all periods.

[0024] In this embodiment, the control current and voltage are optimized through the double error loss function. The double error loss function takes into account both local error and cumulative error, ensuring that not only the accuracy of the current period is concerned during the optimization process, but also the stability of the entire arc-starting process is considered. By minimizing the double error loss function and using the gradient descent method to update the current and voltage, the optimal execution current and execution voltage at each time point are finally obtained. This step ensures the globally optimal control parameters by averaging the current and voltage sequences of all periods, significantly improving the stability and accuracy of the arc-starting process, avoiding defects caused by unstable current and voltage, and finally optimizing the arc-starting quality of the quartz crucible.

[0025] Specifically in this embodiment, the steps for the target heat determination module to determine the target arc heat within any local duration include: S2-1. Use equally spaced parallel dividing planes to cut the standard arc starting duration to obtain several adjacent local durations; S2-2. Between adjacent parallel dividing planes, select any local arc starting point and select the local duration containing the arc starting point; S2-3. Determine the target arc heat within any local duration according to the total theoretical arc heat within the standard arc starting duration.

[0026] Before S2-1, it further includes: Initialize a virtual plane that coincides with the standard arc starting duration; Construct several equally spaced parallel dividing planes in the vertical direction of the virtual plane.

[0027] In this embodiment, by equally spacing and dividing the standard arc starting duration, the arc starting process is divided into multiple local durations, and the target arc heat is accurately calculated for each local duration. The equal-spacing division and calculation ensure that each stage of the arc starting process can be fully controlled, avoiding the occurrence of unstable current or voltage, ultimately ensuring the arc starting quality and stability, and avoiding the generation of defects such as bubbles or black spots.

[0028] Further, the step S2-3 specifically further includes: S2-3-1. Determine the average heating power within the standard arc starting duration according to the ratio of the total theoretical arc heat to the standard arc starting duration; S2-3-2. Select the cutting points of the adjacent parallel dividing planes and the standard arc starting duration, define the previous cutting point as the start time point, and the subsequent cutting point as the end time point; S2-3-3. Define the duration between the start time point and the end time point as the any local duration; S2-3-4. Calculate the product of the any local duration and the average heating power, and define the product of the local duration and the average heating power as the target arc heat within the any local duration.

[0029] In this embodiment, by calculating the average heating power, the target arc heat is determined for each local duration. Specifically, first, the average heating power within the standard arc starting duration is calculated according to the ratio of the total theoretical arc heat to the standard arc starting duration. Then, the standard arc starting duration is divided into several local durations, and the arc heat within each local duration is calculated, ensuring that the heat distribution and power transfer within each local duration are controlled, thereby avoiding too high or too low arc temperature and being able to significantly optimize the arc starting quality of the quartz crucible.

[0030] Please refer to Figure 3 , Figure 3 which provides a schematic diagram of the acquisition process for all local heating errors. Specifically in this embodiment, the steps for the error acquisition module to obtain all local heating errors before the monitoring time point include: S4-1. Obtain the local heating error within the first local time period; S4-2. Among several adjacent local time periods, select the next local time period after the end time point of the adjacent first local time period; S4-3. Obtain the local heating error within the next local time period; S4-4. Repeat the acquisition of the local heating error and perform cumulative summation on the repeatedly acquired local heating errors until the cumulative heating error before the monitoring time point is obtained.

[0031] In this embodiment, by gradually obtaining the local heating error, the heat error during the arc process is traced and accumulated. By obtaining the local heating error within each local time period and processing multiple local time periods in sequence, the method ensures real-time monitoring of the arc heat deviation at each stage. This step can accurately capture the error at each time period during the arc process and accumulate it, thereby enabling timely discovery and correction of the temperature fluctuations generated during the arc process, avoiding instability during the arc ignition process, and ensuring the arc ignition quality of the quartz crucible.

[0032] Furthermore, the specific steps of step S4-1 further include: S4-1-1. Obtain the actual arc heat within the first local time period; the specific steps of step S4-1-1 further include: S4-1-1-1. Collect the arc starting temperature at the start time point and the arc real-time temperature at the end time point within the first local time period; S4-1-1-2. Calculate the first temperature difference between the arc real-time temperature at the end time point and the arc starting temperature at the start time point; S4-1-1-3. Calculate the product of the first temperature difference and the predefined mass and specific heat capacity of the electrode arc ignition hot end, and define the product of the first temperature difference and the predefined mass and specific heat capacity of the electrode arc ignition hot end as the actual arc heat within the first local time period.

[0033] Before S4-1-1-1, it further includes: obtaining the predefined mass and specific heat capacity of the electrode arc ignition hot end; where the mass and specific heat capacity are both the rated physical properties of the electrode arc ignition hot end and belong to constants. Generally, they are clearly pre-identified in the manufacturing process of the electrode arc ignition hot end.

[0034] In this embodiment, the actual arc heat in the first local time period is obtained through the temperature difference. Specifically, in this embodiment, by collecting the real-time arc temperature between the arc starting point and subsequent time points and calculating the temperature difference, the actual arc heat in each time period can be quantified.

[0035] S4-1-2: Define the difference between the target arc heat and the actual arc heat in the first local time period as the local heating error in the first local time period.

[0036] In this embodiment, the local heating error is defined by calculating the difference between the actual arc heat and the target arc heat. Specifically, by obtaining the actual arc heat in each local time period and comparing it with the predetermined target arc heat, the heat error in the arc starting process can be monitored and quantified in real time. This step can accurately capture any deviation in the arc starting process, timely correct the instability in arc control, and ensure that the arc temperature remains within the ideal range.

[0037] Specifically, the expression of the double error loss function in this embodiment is: ; Where represents the double error loss, represents the local heating error, that is, the difference calculated in each local time period, which reflects the accuracy at the current moment or time period. represents the weighting coefficient, indicating the relative importance of the local error and the cumulative error. Its value range is 0 ≤ λ ≤ 1. When λ approaches 1, the local error has a greater impact on the total loss; when λ approaches 0, the cumulative error has a greater impact. represents the local error in the i-th local time period, indicating the error term calculated in the local time period, usually the arc heat error; represents the time-varying weight, which is used to adjust the weight of the error in each local time period in the double error loss function.

[0038] Furthermore, the time-varying weight can gradually increase or decrease according to the change of time. For example, in the initial stage of the arc starting process, the tolerance for errors may be higher, so smaller weights are given to these stages; in the later stage of arc starting, as the requirement for process stability increases, higher weights are given to these stages. Therefore, the expression of the time-varying weight based on time change can be: ; Where β is a constant and i is the index of the local time period. As i increases, the weight gradually decreases.

[0039] In this embodiment, the current and voltage adjustments during the arc ignition process are optimized through the specific expression of the double-error loss function. Among them, the double-error loss function combines the local heating error and the cumulative error, and controls the relative importance of these two error terms in the double-error loss function through the weighting coefficient λ. The local heating error represents the heat error of the arc within each local time period, while the cumulative error accumulates the errors in all local time periods to control the impact on the overall ignition process.

[0040] Specifically, the local heating error reflects the difference between the actual heat and the theoretical heat of the arc within each local time period. By calculating the error of each local time period, the current and voltage can be adjusted in real time to avoid overheating or too low arc temperature in a certain time period, thus avoiding quality problems (such as bubbles or black spots) caused by excessive temperature rise or fall.

[0041] The cumulative error is the accumulation of the errors of all local time periods, which measures the error accumulation situation in the entire arc ignition process. During the control process, the cumulative error ensures the long-term stability of the ignition process, preventing the arc adjustment from being too urgent or periodically out of balance. By combining the local error and the cumulative error, the system can not only cope with the changes within a single time period, but also maintain the overall balance of the arc process in the long term. The weighting coefficient λ controls the relative importance of the local error and the cumulative error in the double-error loss function. By adjusting λ, the system can choose to pay more attention to the local error (when λ is larger), such as in the initial stage of arc ignition, or pay more attention to the cumulative error (when λ is smaller) to ensure that the error is controlled throughout the ignition process. For example, in the initial stage of arc ignition, the tolerance for error may be relatively high, so the weight of the local error is small, while in the later stage, as the stability requirement of the arc process increases, the system will increase the attention to the cumulative error. The time-varying weight adjusts the weight of the error of each local time period in the loss function according to different stages of the arc ignition process. For example, in the initial stage of arc ignition, due to the large temperature fluctuation of the system and relatively high tolerance, the weight of the local time period is small; while in the later stage, the arc ignition process tends to be stable, and the cumulative effect of the error gradually appears, and the system increases the weight of the later error. This dynamic adjustment mechanism enables the adjustment of the current and voltage to better meet the requirements in the actual arc ignition process and avoids the impact of temperature fluctuation on the overall arc quality.

[0042] Furthermore, in this embodiment, the specific update steps of the control parameter update module include: S6-1. Preset the learning rate for updating the execution current and execution voltage in gradient descent; S6-2. Calculate the partial derivative gradients of the double-error loss function with respect to the current execution current and the current execution voltage; S6-3. Update the current execution current and the current execution voltage according to the learning rate and the partial derivative gradient. S6-4. According to the updated execution current and execution voltage, obtain the local heating error and the cumulative heating error corresponding to the local duration of the updated execution current and execution voltage. S6-5. Calculate the double-error loss function according to the local heating error and the cumulative heating error corresponding to the local duration. S6-6. Loop through S6-2 to S6-5 until the double-error loss function is minimized.

[0043] Specifically, the control parameter update module continuously optimizes the update process of the execution current and the execution voltage through the gradient descent algorithm. In each iteration step, according to the partial derivative gradient of the double-error loss function with respect to the current execution current and execution voltage, combined with the preset learning rate, the values of the execution current and the execution voltage are adjusted in real time, so that they gradually approach the optimal control parameters within the local duration, and are dynamically corrected during the process of minimizing the double-error loss function, ensuring the control of the arc heat distribution within the standard arc starting duration, so that the changes of the execution current and the execution voltage meet the requirements of the target arc heat, and reducing the control deviation.

[0044] Further, in this embodiment, the calculation expression of the partial derivative gradient of the current execution current and the current execution voltage is: ; Where represents the double-error loss function, I is the current execution current, U is the current execution voltage, N is the number of all local durations before the monitoring time point, represents the time-varying weight, which is used to adjust the influence of errors in different local durations on the loss; represents the local error within the i-th local duration, represents the cumulative error, is the weighting coefficient, representing the relative importance of the local error and the cumulative error, and respectively represent the partial derivative gradients of the local error with respect to the execution current and the execution voltage, and respectively represent the partial derivative gradients of the cumulative error with respect to the execution current and the execution voltage.

[0045] Specifically, the calculation method of the partial derivative gradient calculates the partial derivatives of the double-error loss function with respect to the current execution current and execution voltage respectively, adjusts the influence of different local duration errors on the loss based on the time-varying weight, and comprehensively considers the relative importance of the local error and the cumulative error by combining the weighting coefficient. Within all local duration ranges before the monitoring time point, the gradient influence of the local error and the cumulative error on the execution current and execution voltage is calculated, so as to ensure the effective reduction of error accumulation during the update process of the control parameters and improve the update accuracy of the execution current and execution voltage.

[0046] Further, in this embodiment, the update expressions for the execution current and execution voltage are as follows: ; Wherein, represents the updated execution current, represents the updated execution voltage, represents the current execution current, represents the current execution voltage, represents the learning rate.

[0047] Specifically, the update process of the execution current and execution voltage calculates the partial derivative gradients of the double-error loss function with respect to the current execution current and execution voltage, and combines the preset learning rate to adjust the values of the execution current and execution voltage in the step size of the learning rate, so that they converge towards the optimal control parameters. In each iteration process, the current execution current and execution voltage are updated according to the learning rate multiplied by the corresponding gradient value, thereby gradually reducing the value of the double-error loss function and achieving the minimization of the local heating error and the cumulative heating error.

[0048] Finally, by combining the optimization objectives of the local heating error and the cumulative error and using the gradient descent algorithm to update the current and voltage, it can be ensured that the adjustment of the current and voltage in each time period can minimize the loss function, thereby achieving more precise arc control. The system outputs the optimal current and voltage sequences, and calculates the global optimal control current and voltage through the average values of the current and voltage in all time periods, thereby ensuring the long-term stability and accuracy of the arc ignition process while minimizing the local error.

[0049] In summary, the present invention effectively combines local and cumulative error control by introducing the mechanism of the double-error loss function and the time-varying weight, and realizes precise current and voltage adjustment during the entire arc ignition process. Through reasonable error feedback and optimization, the system can maintain the global stability of the quartz crucible during the electrode arc ignition control process while ensuring local accuracy.

[0050] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (such as infrared, wireless, microwave, etc.).

[0051] The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that contains one or more collections of available media. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVD ), or semiconductor media. The semiconductor media can be a solid-state drive.

[0052] In several embodiments provided in the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division of an underwater terrain change analysis system and method for waterways. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of devices or units can be in electrical, mechanical, or other forms.

[0053] The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed in the present application can easily think of changes or substitutions, which should all be covered by the protection scope of the present application.

Claims

1. An electrode arc control system for a quartz crucible, characterized in that: include: A predefined module is used to predefine the standard arcing time of the electrode arcing and the theoretical total heat of the arc; A target heat determination module is used to determine the target arc heat within a local time length according to the standard arc ignition time length and the theoretical total heat of the arc; The time point definition module is used to define the monitoring time point within the standard arc ignition time; An error acquisition module is used to obtain a cumulative heating error according to a target arc heat in a local time period before a monitoring time point; wherein the cumulative heating error is characterized as the sum of all local heating errors before the monitoring time point, and the local heating error represents the error between the target arc heat and the actual arc heat in a local time period; An error loss construction module, used for constructing a dual error loss function according to a local heating error and a cumulative heating error; A control parameter updating module for updating the execution current and the execution voltage using a gradient descent algorithm to minimize a double error loss function; The control parameter output module is used to input the execution current and execution voltage at the time of minimization as optimal control parameters into the electrode arc ignition control system when the double error loss function is minimized, so as to control the electrode arc ignition process.

2. The electrode arc striking control system of a quartz crucible according to claim 1, characterized in that: The target heat determination module is specifically used for: Initializing a virtual plane that coincides with the standard arc striking time; Constructing a number of parallel dividing surfaces with equal spacing in a direction perpendicular to the virtual plane; The standard arc striking time is cut using parallel dividing planes with equal spacing to obtain a number of adjacent local time lengths; Select any local duration between adjacent parallel split surfaces; According to the theoretical total arc heat within the standard arc ignition time, the target arc heat within any local time is determined.

3. The electrode arc striking control system of a quartz crucible according to claim 2, characterized in that: The target heat determination module is specifically used for: According to the ratio of the theoretical total heat of the arc to the standard arc striking time, the average heating power within the standard arc striking time is determined; Select the cutting points of adjacent parallel dividing surfaces and standard arc striking time, define the front cutting point as the starting time point, and the rear cutting point as the ending time point; Define the duration between the start time point and the end time point as any of the local durations; The product of any local duration and the average heating power is calculated, and the product is defined as the target arc heat in any local duration.

4. The electrode arc striking control system of a quartz crucible according to claim 1, characterized in that: The error acquisition module is specifically used for: Obtaining a local heating error within a first local time length; wherein the previous cutting point of the first local time length is the electrode arc starting point; Among several adjacent partial durations, selecting a next partial duration adjacent to the end time point of the first partial duration; Obtaining a local heating error within a next local time period; The acquisition of the local heating error is repeatedly performed, and the repeated local heating errors are accumulated and summed until the accumulated heating error before the monitoring time point is obtained.

5. The electrode arc striking control system of a quartz crucible according to claim 4, characterized in that: The error acquisition module is specifically used for: Acquiring actual arc heat during the first local time period; The difference between the target arc heat and the actual arc heat in the first local time period is defined as the local heating error in the first local time period.

6. The electrode arc striking control system of a quartz crucible according to claim 5, characterized in that: The error acquisition module is specifically used for: Obtain the predefined mass and specific heat capacity of the arc-starting hot end of the electrode; Collecting the arc starting temperature at the start time point and the arc real-time temperature at the end time point within the first local time period; Calculate a first temperature difference between the arc real-time temperature at the end time point and the arc starting temperature at the start time point; The product of the first temperature difference and the predefined mass and specific heat capacity of the electrode arc ignition hot end is calculated, and the product of the first temperature difference and the predefined mass and specific heat capacity of the electrode arc ignition hot end is defined as the actual arc heat in the first local time length.

7. The electrode arc striking control system of a quartz crucible according to claim 1, characterized in that: The expression of the double error loss function is: ； in, represents the double error loss, Represents the local heating error, represents the weighting coefficient, which indicates the relative importance of local error and cumulative error; represents the local error in the i-th local time period, represents the error term calculated in the local time period; It represents a time-varying weight, which is used to adjust the weight of the error of each local duration in the double error loss function.

8. The electrode arc striking control system of a quartz crucible according to claim 1, characterized in that: The updating step of the control parameter updating module comprises: S6-1, preset the learning rate for updating the execution current and execution voltage in the gradient descent; S6-2, calculating the partial derivative gradient of the dual error loss function with respect to the current execution current and the current execution voltage; S6-3, updating the current execution current and the current execution voltage according to the learning rate and the partial derivative gradient; S6-4, according to the updated execution current and execution voltage, obtaining the local heating error and the cumulative heating error of the local time corresponding to the updated execution current and execution voltage; S6-5, calculating a double error loss function according to the local heating error and the cumulative heating error corresponding to the local time length; S6-6. Loop through S6-2 to S6-5 until the double error loss function is minimized.

9. The electrode arc striking control system of a quartz crucible according to claim 1, characterized in that: The calculation expressions of the partial derivative gradients of the current execution current and the current execution voltage are: ； in, represents the double error loss function, I is the current execution current, U is the current execution voltage, N is the number of all local durations before the monitoring time point, Represents the time-varying weight, which is used to adjust the impact of different local time errors on the loss; represents the local error in the i-th local duration, represents the cumulative error, is the weighting coefficient, which indicates the relative importance of local error and cumulative error. and They represent the partial derivative gradients of the local error to the execution current and the execution voltage, respectively. and They represent the partial derivative gradients of the accumulated error with respect to the execution current and the execution voltage, respectively.

10. The electrode arc striking control system of a quartz crucible according to claim 1, characterized in that: The update expressions of the execution current and the execution voltage are: ； in, represents the updated execution current, Indicates the updated execution voltage, Indicates the current execution current. Indicates the current execution voltage. Represents the learning rate.

Images