Method for temperature control of a shaft furnace body

Abstract

This invention discloses a temperature control method for a vertical furnace body. The method relates to the field of equipment control technology and includes the following steps: real-time temperature control; feature extraction and model training; degradation state assessment; temperature rationality judgment and alarm. This invention achieves precise temperature control by dividing the temperature zone into modules, and establishes a time-series degradation model by collecting temperature and heating element data to assess the element state and correct temperature measurements. Finally, it performs rationality judgment and alarms to ensure stable process temperature, improve the temperature control accuracy of the vertical furnace body, and solve the problem of low temperature control accuracy in existing vertical furnace bodies.

Description

Technical Field

[0001] This invention relates to the field of equipment control technology, and in particular to a method for temperature control of a vertical furnace body. Background Technology

[0002] In semiconductor chip manufacturing, vertical furnaces are a critical piece of equipment, and their temperature control accuracy directly affects process quality and equipment stability. Currently, the typical furnace temperature control scheme for semiconductor equipment involves detecting the furnace temperature using a temperature sensor during each control cycle and adjusting the furnace heating power based on the detected values ​​to stabilize the furnace temperature at a preset target temperature. However, this method has several problems: First, abnormal fluctuations in the furnace temperature sampling values ​​can lead to unstable furnace temperature control, affecting process results; second, uneven temperature distribution within the furnace can cause some areas to be too hot or too cold, affecting process uniformity.

[0003] To address these issues, existing technologies have employed several improved methods, such as temperature compensation and reverse wafer temperature calibration. However, these methods still have limitations: temperature compensation methods struggle to effectively resolve coupling issues between multiple temperature zones, and setting compensation values ​​requires extensive experimentation and adjustment by experienced engineers; while reverse wafer temperature calibration methods rely heavily on process data, and the process results are not easily tested in real time, making timely compensation difficult. Furthermore, in some specialized process equipment, such as physical vapor deposition (PVD) equipment, the reactive gases easily carry away a significant amount of heat, leading to a drop in the internal chamber temperature, which affects the accuracy of process temperature control and wafer throughput efficiency.

[0004] Therefore, there is an urgent need for a method that can achieve precise temperature control and improve the uniformity and stability of furnace temperature. This method should be able to monitor furnace temperature in real time, respond quickly to temperature changes, and independently control different areas according to their temperature requirements, thereby improving process quality and equipment reliability. Simultaneously, it is also necessary to consider how to achieve precise temperature control of critical areas of the furnace without affecting the temperature measurement of the silicon wafers themselves, to ensure the stability and consistency of the process.

[0005] For example, Chinese invention patent CN115951740A discloses a heating furnace temperature control system and control method, including: a power supply circuit, a temperature acquisition circuit, a MOS drive circuit, and a vacuum pump control circuit. The power supply circuit, temperature acquisition circuit, MOS drive circuit, and vacuum pump control circuit are all connected to the MCU main control circuit. The temperature acquisition circuit is connected to the MCU main control circuit and is used to acquire the temperature information of multiple thermal resistors in the vacuum heating furnace and transmit the temperature information of the multiple thermal resistors to the MCU main control circuit in sequence. The MCU main control circuit is connected to the MOS drive circuit and the vacuum pump control circuit. The MCU main control circuit is used to process the temperature information to obtain a heating PWM signal. The MOS drive circuit controls the heating platform in the heating furnace to switch on and off according to the heating PWM signal.

[0006] For example, the Chinese invention patent with publication number CN119781544A discloses a method for controlling the temperature of a refractory furnace, which includes: acquiring a BP neural network constructed with refractory furnace data as input and PID configuration parameters as output; solving the hidden layer weights, initial bias values, and hyperparameters of the BP neural network using a differential evolution algorithm; wherein, during the solution process, the scaling ratio of the mutation process in the differential evolution algorithm changes linearly with a negative correlation with the number of iterations; assigning the solved hidden layer weights, initial bias values, and hyperparameters to the BP neural network and training the BP neural network; inputting real-time acquired refractory furnace data into the trained BP neural network to obtain real-time PID configuration parameters; inputting the desired temperature into the PID controller and configuring the real-time PID configuration parameters into the PID controller to achieve temperature control of the refractory furnace.

[0007] The above-mentioned technology has at least the following technical problems:

[0008] First, existing temperature control methods did not establish a multi-temperature zone independent and coordinated temperature control model at the initial design stage. They failed to achieve precise isolation and balance of heat transfer in each temperature zone through technologies such as zone sensing and dynamic feedback adjustment. This resulted in mutual heat interference and coupling effects between temperature zones, which were difficult to resolve, ultimately leading to uneven furnace temperature distribution and affecting process uniformity.

[0009] Secondly, the core problem with temperature compensation methods is the lack of standardized and automated compensation mechanisms. No algorithm system has been developed that can automatically calculate compensation parameters based on real-time temperature data. Instead, experienced engineers can only rely on repeated trial and error adjustments based on their subjective experience. This not only fails to guarantee compensation accuracy but also makes it difficult to develop personalized compensation schemes for the differentiated temperature needs of different regions, thus making it impossible to achieve independent temperature control.

[0010] Furthermore, the existing temperature control system lacks a clock calibration and data synchronization architecture. Each temperature control module uses an independent operating sequence and lacks technical support for real-time data interaction and parameter calibration. This results in asynchronous acquisition, transmission and execution of temperature parameters in each module, making it impossible to form a unified temperature control benchmark and ultimately disrupting the uniform temperature distribution inside the furnace.

[0011] Finally, the existing heating system lacks advanced intelligent algorithms such as adaptive control and fuzzy control, and it has not built a mechanism for data sharing and collaborative scheduling between modules. As a result, the heating system cannot dynamically optimize the temperature control strategy according to real-time temperature changes, and the heating modules cannot achieve efficient collaboration. It is difficult to achieve the goal of uniform temperature control and precise adjustment in the working area, resulting in low temperature control accuracy of the vertical furnace body. Summary of the Invention

[0012] This application provides a temperature control method for a vertical furnace body, which solves the problem of low temperature control accuracy of the vertical furnace body in the prior art and improves the temperature control accuracy of the vertical furnace body.

[0013] On the one hand, a temperature control method for a vertical furnace body is provided, including the following steps: dividing the furnace body into multiple independent temperature zone modules along the axial direction, collecting temperature data of each temperature zone module in real time, and determining whether the temperature data is greater than the process temperature set value. If so, the cooling unit is adjusted; if not, the heating unit is adjusted. Based on the temperature data and the energizing time and current data of the heating elements in the furnace body, preprocessing is performed to extract effective feature values. Based on the effective feature values, a time series degradation model is trained and established. According to the time series degradation model, the current degradation state of the heating elements in each temperature zone module is evaluated to obtain the temperature measurement value of the processed temperature zone module. It is determined whether the deviation between the temperature measurement value of the processed temperature zone module and the preset measurement temperature is within the set range. If so, the temperature measurement result of the temperature zone module is determined to be reasonable; if not, an alarm is triggered and the temperature measurement is repeated.

[0014] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0015] 1. By dividing the furnace tube into independent temperature zone modules along the axis and constructing dedicated control loops, the furnace tube is divided into multiple independent temperature zone modules, each with its own independent control system. This effectively reduces the mutual influence between modules, improving system reliability and flexibility. Precise control avoids temperature unevenness, significantly improving the accuracy and stability of furnace temperature control and fundamentally solving the industry pain point of uneven axial temperature in traditional vertical furnaces. Each temperature zone is equipped with an independent heating unit, cooling unit, and temperature sensor, forming a closed-loop control logic of "acquisition-judgment-adjustment." This design allows each temperature zone to flexibly set its target temperature according to process requirements, improving the gradient control accuracy of the upper and lower temperature zones and reducing temperature deviation compared to traditional single-loop control. Simultaneously, the independent loop design avoids the impact of fluctuations in a single temperature zone on the overall furnace cavity. Combined with the inherent heat dissipation advantages of the vertical structure, energy consumption is further reduced, significantly improving the accuracy and stability of furnace temperature control.

[0016] 2. By constructing a feature sequence dataset through a time-series degradation model, integrating temperature data, heating element energizing time, and current data, and training with a supervised learning algorithm, key parameters such as remaining lifespan percentage and performance degradation index can be accurately output. This data-driven health assessment method breaks away from the traditional passive "post-failure maintenance" model. Regularly synchronizing data between modules and using time-series algorithms to control heating elements to reach the optimal process temperature setpoint for each region ensures consistent temperature parameters across all modules. This time synchronization mechanism effectively solves the problem of uneven temperature distribution within the furnace, ensuring uniform temperature distribution, improving process stability and consistency, and avoiding waste caused by premature element replacement. This increases the utilization rate of heating element lifespan and further enhances process stability and consistency.

[0017] 3. To address the temperature measurement deviation caused by heating element aging, the solution incorporates a dynamic compensation mechanism based on the degradation state. The system automatically generates a temperature measurement compensation coefficient based on the performance degradation level determined by the degradation model, correcting the raw data collected by the sensors. Employing an advanced heating system and inter-module coordination and communication system, and utilizing algorithms such as PID control, fuzzy control, neural network control, and model predictive control, the accuracy and response speed of temperature control are improved. Simultaneously, an adaptive control algorithm automatically adjusts control parameters according to actual operating conditions, enhancing the adaptability and robustness of the control system. This advanced heating system can rapidly respond to temperature changes, effectively solving the problem of temperature drop in the internal chambers of specialized process equipment and improving wafer production efficiency.

[0018] 4. A dual mechanism of "degradation compensation correction and deviation range verification" addresses the process risks caused by temperature data distortion. In the first layer of protection, the temperature value compensated by the degradation model has eliminated interference from component aging. In the second layer, the system compares the processed data with the preset measurement temperature; it is only considered valid if the deviation is within the set range; otherwise, an alarm is immediately triggered and the measurement is repeated. This design effectively identifies abnormal situations such as sensor failure and signal interference, improving the reliability of the single data acquisition mode. Combined with real-time data stream processing technology, the alarm response delay can be controlled to the millisecond level, providing crucial support for process safety in high-end manufacturing such as semiconductors and aerospace materials. Temperature sensors are placed at key locations on the module, employing both thermocouple and photoelectric non-contact temperature sensors to achieve multi-point temperature measurement, improving the reliability and accuracy of the measurement. This multi-point temperature measurement mechanism, combined with data analysis algorithms, effectively processes the measured temperature data, improving the rationality of the measured temperature and ensuring the accuracy of the temperature range measurement. Through the design of an inter-module communication protocol, data exchange and collaborative control between modules are achieved. A master-slave control mode is adopted, with the master controller coordinating the work of each slave controller to achieve overall temperature control. This communication and control mechanism effectively solves the problem of interference between the quartz tube and the furnace body temperature field, ensuring the stability and reliability of temperature control. Attached Figure Description

[0019] Figure 1 A flowchart illustrating a temperature control method for a vertical furnace body provided in this application embodiment. Detailed Implementation

[0020] This application provides a temperature control method for a vertical furnace body, which solves the problem of low temperature control accuracy in the prior art. It achieves precise temperature control by dividing the temperature zone into modules, and establishes a time series degradation model by collecting temperature and heating element data to evaluate the element status and correct the temperature measurement value. Finally, it makes a rationality judgment and alarm to ensure stable process temperature and improve the temperature control accuracy of the vertical furnace body.

[0021] The technical solution in this application embodiment aims to address the aforementioned problem of low temperature control accuracy in vertical furnace bodies. The overall approach is as follows:

[0022] By dividing the furnace body into multiple independent temperature zone modules along the axial direction, the temperature data of each temperature zone module is collected in real time, and it is determined whether the temperature data is greater than the process temperature set value. If so, the cooling unit is adjusted; if not, the heating unit is adjusted. Based on the temperature data and the energizing time and current data of the heating elements in the furnace body, preprocessing is performed to extract effective feature values. Based on the effective feature values, a time series degradation model is trained and established. According to the time series degradation model, the current degradation state of the heating elements in each temperature zone module is evaluated to obtain the temperature measurement value of the processed temperature zone module. It is determined whether the deviation between the temperature measurement value of the processed temperature zone module and the preset measurement temperature is within the set range. If so, the temperature measurement result of the temperature zone module is determined to be reasonable; if not, an alarm is triggered and the temperature measurement is repeated, thereby improving the temperature control accuracy of the vertical furnace body.

[0023] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0024] like Figure 1 The diagram shown is a flowchart of a temperature control method for a vertical furnace body provided in an embodiment of this application. The method includes the following steps:

[0025] The first step of a temperature control method for a vertical furnace body is to divide the furnace body into multiple independent temperature zone modules along the axial direction, collect the temperature data of each temperature zone module in real time, and determine whether the temperature data is greater than the process temperature set value. If so, the cooling unit is adjusted; if not, the heating unit is adjusted.

[0026] It should be understood that, based on the axial length of the furnace body and process requirements, the number of temperature zone modules and the axial range of each temperature zone module are determined; based on the axial range of each temperature zone module, the corresponding heating unit, independent cooling unit, and temperature sensor for each temperature zone module are determined; wherein, the heating unit, cooling unit, and temperature sensor of each temperature zone module together constitute an independent temperature control loop; the temperature data of each temperature zone module is collected in real time and compared with the process temperature setpoint to obtain a comparison result; according to the comparison result, when the temperature data is greater than the process temperature setpoint, a cooling control signal is generated to adjust the power of the cooling element; when the temperature data is less than the process temperature setpoint, a heating control signal is generated to adjust the power of the heating element; when the temperature data is equal to the process temperature setpoint, the current power is maintained.

[0027] It should be noted that the temperature regulation component includes a temperature sensor, comparator, controller, relay, and thermoelectric cooler, and each temperature zone module is equipped with a water-cooled temperature control system and an air-cooled temperature control system. The temperature sensor collects the real-time temperature of the corresponding temperature zone module, the comparator compares the real-time temperature with the preset target temperature and outputs a temperature difference signal, the controller receives the temperature difference signal and generates a control signal, which is used to control the on / off state of the relay and the operating power of the thermoelectric cooler. The combined effect of the water-cooled temperature control system and the air-cooled temperature control system is used to achieve temperature regulation of the corresponding temperature zone module.

[0028] In this embodiment, scientifically determining the number of temperature zone modules and the axial range of each module based on the furnace body's axial length and process requirements is the first step in achieving precise temperature control. This partitioning strategy directly determines the configuration of all subsequent control units. Its technical effect lies in fundamentally changing the traditional vertical furnace's coarse control mode, which treats the entire furnace body as one or a few large temperature zones. In existing technologies, due to unreasonable or overly coarse temperature zone partitioning, large axial and radial temperature gradients occur in the furnace body, and the temperature requirements of different areas cannot be independently met, resulting in severe temperature non-uniformity problems, directly affecting the uniformity of the process and the yield of the final product. This invention, through refined partitioning, configures each independent temperature zone module with a dedicated heating unit, an independent cooling unit, and a precise temperature sensor. These three components together constitute a complete and autonomous temperature control loop. This "one-to-one" hardware configuration achieves physical decoupling, ensuring that the temperature regulation behavior (whether heating or cooling) of each temperature zone is highly confined to its own area. This minimizes thermal interference to neighboring temperature zones, thus solving the problems of mutual control constraints, temperature fluctuations, and cascading effects caused by thermal coupling between multiple temperature zones in existing technologies. Based on this, the system collects temperature data from each temperature zone module in real time and precisely compares it with the specific process temperature setpoint for that zone, forming a closed-loop feedback control decision-making basis. According to the comparison results, the control system executes differentiated and refined power regulation strategies: when the detected temperature data is higher than the setpoint, the system immediately generates a cooling control signal to precisely adjust the power of the cooling elements (such as semiconductor refrigerators or water / air cooling systems) in that zone to achieve rapid cooling; conversely, when the temperature data is lower than the setpoint, a heating control signal is generated to precisely adjust the power of the heating elements in that zone to achieve rapid heating; and when the temperature data is exactly equal to the setpoint, the system intelligently maintains the current power state of the heating and cooling elements, avoiding unnecessary energy consumption and temperature fluctuations. This dynamic, bidirectional (heating and cooling) precise adjustment mechanism, based on real-time data comparison, not only overcomes the shortcomings of existing technologies such as slow response of control systems and single adjustment methods (usually only heating, with cooling relying on natural heat dissipation, resulting in low efficiency), achieving "rapid response, precise positioning, and dynamic maintenance" of temperature, but also ensures that the temperature in each key area inside the furnace is highly consistent and stable by strictly stabilizing the temperature at the process setpoint. Ultimately, this complete process, from physical partitioning to independent loops, and then to real-time comparison and dynamic adjustment, works together to significantly improve the uniformity, stability, and control accuracy of the overall furnace temperature. It effectively solves a series of problems caused by poor temperature control in existing technologies, such as poor process quality, low equipment stability, and high energy consumption, providing a solid and reliable technical guarantee for high-precision processes such as semiconductor manufacturing.

[0029] The second step of the temperature control method for a vertical furnace body involves preprocessing the temperature data, as well as the energizing time and current data of the heating elements inside the furnace body, to extract effective feature values; based on the effective feature values, a time series degradation model is trained and established.

[0030] It should be understood that the temperature data, as well as the energizing time and current data of the heating element, are preprocessed to extract effective feature values; based on the effective feature values, a time-series degradation model is trained; according to the time-series degradation model, a relationship model between time, lifespan, and current is established; the effective feature values ​​are arranged in chronological order to construct a feature sequence dataset for model training; using the feature sequence dataset, a pre-set initial degradation model is trained using a supervised learning algorithm, enabling the initial degradation model to learn the degradation pattern of the heating element performance over time, thereby obtaining a trained time-series degradation model; based on the trained time-series degradation model, degradation state parameters characterizing the current health state of the heating element are output, the degradation state parameters including at least the remaining lifespan percentage and the performance degradation index.

[0031] It should be added that the actuator is connected to the heating element of the temperature zone module; the controller is used to receive the temperature signal sent by the temperature sensor and control the heating element of the corresponding temperature zone module to work; the heating elements of multiple temperature zone modules are divided into a first heating element group and a second heating element group connected in series. The controller integrates one-to-one control logic and one-to-many control logic to perform parallel control of the first heating element group and the second heating element group to synchronously execute the preset heating scheme.

[0032] Secondly, while completing the basic control of single-temperature zone heating elements, the controller further integrates one-to-one control logic and one-to-many control logic: For heating elements in the first and second heating element groups that require extremely high temperature accuracy and need to be adjusted individually, a one-to-one control logic is adopted. That is, the controller assigns an independent control channel to each such heating element and adjusts its heating power individually according to the temperature deviation of the corresponding temperature zone to ensure the accurate response of a single heating element. For heating elements with similar temperature requirements that can be adjusted collaboratively, a one-to-many control logic is adopted. The controller sends unified control commands to multiple heating elements synchronously through a control channel to achieve synchronous power adjustment of multiple heating elements. Through the integration of the two control logics, the controller performs parallel control of the first and second heating element groups, so that the two groups of heating elements operate simultaneously according to the preset heating scheme (including parameters such as heating power and heating rate at different process stages), avoiding the problem of control delay or inconsistency between the two groups of heating elements.

[0033] This process ensures that controller commands can be quickly transmitted to the heating elements through direct connection between the actuator and the heating element, reducing signal transmission loss. The grouped series design of the heating elements and the integration of dual control logic solve the contradiction of "low efficiency of single control and poor accuracy of group control" in traditional multi-heating element control. One-to-one control logic ensures the heating accuracy of high-demand temperature zones, while one-to-many control logic improves the efficiency of multi-temperature zone collaborative heating. On the other hand, the parallel control of the two groups of heating elements ensures the synchronous execution of the preset heating scheme, avoiding local temperature deviations in the furnace body caused by inconsistent heating pace, and improving the overall temperature uniformity of the multi-temperature zones.

[0034] In this embodiment, the heating element life prediction and health management method based on a time-series degradation model achieves a fundamental shift from passive and delayed equipment maintenance to proactive and predictive maintenance, thus providing unprecedented assurance for the long-term stability and accuracy of vertical furnace temperature control. Specifically, the method first preprocesses the collected temperature data and the heating element's energizing time and current data. This step effectively extracts valid feature values ​​that truly reflect the heating element's performance status from the raw, potentially noisy and interfering signals through filtering, modal analysis, and time-frequency domain transformation, eliminating irrelevant or interfering factors and laying a solid data foundation for the accuracy of the subsequent model. Subsequently, based on these selected valid feature values, a time-series degradation model is trained. The core technical effect of this process is that the model can deeply learn and master the inherent laws and patterns of the gradual degradation of heating element performance over time. To achieve this, effective feature values ​​are arranged chronologically to construct a feature sequence dataset for model training. This dataset is then used to iteratively train a pre-set initial degradation model using a supervised learning algorithm, ultimately resulting in a time-series degradation model that accurately maps the performance changes of the heating element throughout its entire lifecycle. The most critical technical effect of this trained model is its ability to establish a dynamic, quantitative "time-lifetime-current" relationship model. This model not only reveals the performance of the heating element under specific energizing time and current conditions but also outputs degradation state parameters characterizing its current health status in real time, such as the percentage of remaining service life and the performance degradation index. This contrasts sharply with existing technologies that lack quantitative assessment of the aging state of heating elements, leading to an inability to predict and compensate for the decline in temperature control accuracy over time. Existing technologies often only perform reactive maintenance after heating element failure or significant temperature control deviations, severely impacting production continuity and process quality. This invention, through this predictive model, enables the control system to anticipate the performance degradation trend of the heating element and make proactive adjustments accordingly. For example, when the model predicts that a heating element has a low remaining lifespan or an increased performance degradation index, it can automatically and intelligently adjust the current supplied to that heating element to compensate for the decrease in heating efficiency caused by aging. This ensures that within the same process time, the furnace wire can still generate enough heat to accurately reach the specified process temperature inside the furnace. This proactive compensation mechanism ultimately significantly extends the effective lifespan of the heating element, ensuring that the temperature control accuracy of the vertical furnace remains at its initial high level throughout its entire lifespan. It avoids temperature drift, process instability, and sudden equipment failures caused by element aging, thereby significantly improving the overall reliability of the equipment, process consistency, production continuity, and efficiency. Ultimately, it reduces maintenance costs and increases wafer throughput.

[0035] The third step in the temperature control method for a vertical furnace body is to evaluate the current degradation state of the heating elements in each temperature zone module based on a time series degradation model, so as to obtain the temperature measurement value of the processed temperature zone module.

[0036] It should be understood that, based on the time-series degradation model, the current degradation state of the heating element is evaluated to obtain the processed temperature measurement value of the temperature zone module; based on the deviation between the processed temperature measurement value of the temperature zone module and the process temperature setpoint, the current of the heating element is corrected in a closed loop to obtain the processed temperature measurement value of the temperature zone module; evaluating the current degradation state of the heating element based on the time-series degradation model to obtain the processed temperature measurement value of the temperature zone module includes: acquiring the degradation state parameters output by the time-series degradation model; comparing the degradation state parameters with a preset corresponding threshold to obtain a comparison result; judging the degree of performance degradation of the heating element based on the comparison result to obtain a judgment result of the degree of performance degradation; generating a corresponding temperature measurement compensation coefficient based on the judgment result of the degree of performance degradation; and using the temperature measurement compensation coefficient to correct the original temperature measurement value of the temperature zone module acquired in real time to obtain the processed temperature measurement value of the temperature zone module.

[0037] In this embodiment, firstly, the two key parameters of "heating time" and "corresponding current" after the heating element is powered on to reach the preset process temperature are collected at different time periods. These parameters are then preprocessed, including filtering, modal analysis, and time-frequency domain conversion, and post-processed, including data cleaning and differential processing. The fundamental purpose of this series of complex signal processing is to extract the "effective feature values" that best reflect the health status of the heating element from the raw, noisy operating data. For example, the trend of prolonged heating time directly reflects the decrease in heating efficiency, while abnormal fluctuations or increases in the required current under the same process settings indicate the degradation of the resistance value. The preprocessing and post-processing effectively eliminate interference from factors that have no influence, such as ambient temperature fluctuations and power grid disturbances, ensuring data quality and laying a solid foundation for subsequent accurate modeling.

[0038] It needs to be explained that the process involves collecting time data and corresponding current data of the heating element reaching the preset process temperature after being powered on at different time periods; preprocessing the collected time data and corresponding current data, including filtering, modal analysis, and time-frequency domain conversion, extracting effective feature values ​​and removing feature values ​​without influencing factors; postprocessing the preprocessed time data and corresponding current data, including data cleaning to remove outliers, data transformation, and differential processing of the time series data; and using the postprocessed time data and corresponding current data to train the degradation model throughout its entire life cycle. The degradation model is constructed based on a time series algorithm, establishing quantitative indicators of the impact on the service life of the heating element, and selecting feature value data with strong characterization capabilities.

[0039] Next, rigorously processed time and current data are used to train a degradation model based on time series algorithms such as LSTM (Long Short-Term Memory) throughout its entire lifecycle. The core technical effect is the establishment of a dynamically updated, high-fidelity digital mirror of component performance. The key output of this model is the construction of a "quantitative index of the impact on heating element lifespan." This is no longer a qualitative, empirical judgment such as "heating element aging," but a precise, digital measurement. Furthermore, principal component analysis is used to reduce and fuse the selected strong representational features, effectively reducing data dimensionality and eliminating collinearity between features. This extracts a few "principal component data" that carry the majority of degradation information, greatly optimizing the model's efficiency and robustness.

[0040] It should be added that, after selecting eigenvalue data with strong characterization capabilities, the following steps are also included: normalizing the selected eigenvalue data, using principal component analysis for dimensionality reduction and feature fusion to determine multiple principal component data; predicting the remaining lifetime using the trained degradation model, estimating the degradation model parameters using known degradation data, substituting the parameters to obtain the exponential model and its confidence interval, calculating the remaining lifetime of the heating element based on the preset failure threshold; and establishing a time-life-current correlation model based on the remaining lifetime prediction results to ensure that, within the same process time, the heat generated by the heating element enables the furnace body to reach the specified process temperature.

[0041] Then, based on a well-trained degradation model, remaining service life is predicted. Model parameters are estimated using known historical degradation data, and these parameters are substituted to obtain an exponential model describing the performance degradation trajectory and its confidence interval. The most direct technical effect is the realization of personalized, probabilistic predictive maintenance. The system can accurately predict the remaining service life of components based on their actual performance data and provide the range of uncertainty in the prediction. This allows equipment managers to shift from traditional fixed-cycle maintenance or reactive repairs to proactively planning maintenance windows and accurately procuring spare parts based on accurate risk assessments. This minimizes unplanned downtime, avoids the risk of entire batches of products being scrapped due to sudden component failure, and reduces inventory costs.

[0042] Ultimately, based on the remaining lifetime prediction results, a "time-lifetime-current correlation model" was established, a crucial step in transforming predictive information into advanced control strategies. The core effect of this model is ensuring the long-term stability and consistency of furnace temperature control. It enables the control system to proactively compensate and adjust: for example, when the remaining lifetime prediction of a component indicates a gradual decline in its heating efficiency, the system can actively and adaptively fine-tune the power output or energizing time in the control algorithm, intelligently compensating for heat loss due to performance degradation. This ensures that throughout the entire process cycle, even as component performance slowly degrades, the thermal field inside the furnace can consistently and accurately reach the specified process temperature. This fundamentally solves the process drift problem caused by component aging in traditional control systems, providing core technological support for achieving superior quality goals.

[0043] The fourth step in the temperature control method for a vertical furnace body is to determine whether the deviation between the temperature measurement value of the processed temperature zone module and the preset measurement temperature is within the set range. If so, the temperature measurement result of the temperature zone module is deemed reasonable; otherwise, an alarm is triggered and the temperature measurement is repeated.

[0044] It should be understood that the temperature measurement value of the processed temperature zone module is compared with the preset measurement temperature; it is determined whether the deviation is within the set range; if the deviation is within the set range, the temperature measurement result is deemed valid; if the deviation is not within the set range, an alarm is triggered and the temperature measurement is repeated.

[0045] It should be noted that the key locations of the temperature zone module include the heating element setting area, the cooling element setting area, the temperature zone module inlet, and the temperature zone module outlet. Multiple thermocouple-type temperature sensors are arranged circumferentially around the temperature zone module for multi-point temperature measurement. Based on the blackbody radiation law, they achieve non-contact temperature measurement by receiving infrared energy radiated by an object. When measuring the internal temperature of the furnace, multiple thermocouple-type temperature sensors are arranged circumferentially for multi-point temperature measurement. Data analysis algorithms are used to process the measured temperature data to improve the rationality of the measured temperature, thereby determining the measured temperature of the temperature zone. One photoelectric temperature sensor is arranged circumferentially and serves as a control group for the thermocouple-type temperature sensors.

[0046] In this embodiment, the temperature measurement process for the temperature zone module is based on "multi-point acquisition + dual-sensor verification." First, temperature sensors are precisely deployed at key locations within each independent temperature zone module. Simultaneously, thermocouple-type temperature sensors and photoelectric temperature sensors are configured within each module, constructing a three-dimensional temperature measurement architecture of "key area coverage + dual-type sensor verification." After the process starts, multiple thermocouple-type temperature sensors arranged circumferentially first perform multi-point temperature measurements on the temperature zone module, comprehensively capturing the temperature distribution at different locations within the temperature zone and avoiding the biased temperature data caused by a single measurement point. Then, the raw temperature data collected by multiple thermocouple-type temperature sensors is transmitted to the controller. The built-in data analysis algorithm (such as a weighted average algorithm, assigning higher weights to measurement points in temperature-sensitive areas such as near the heating element) processes the data, eliminating the influence of random errors and accurately determining the final measured temperature of the temperature zone module, ensuring that the temperature data truly reflects the overall temperature state of the temperature zone. While the thermocouple-type temperature sensor completes measurement and data processing, a photoelectric temperature sensor arranged circumferentially receives infrared energy radiated by objects inside the furnace based on the blackbody radiation law to measure temperature, and its measurement result is used as the control group temperature data. The controller further calculates the difference between the temperature zone measured by the thermocouple-type temperature sensor and the control group temperature of the photoelectric temperature sensor, and determines whether the difference is within the preset deviation reference range. On the one hand, the high precision and wide range of the thermocouple-type sensor ensure the basic temperature measurement accuracy, and on the other hand, the fast response of the photoelectric sensor enables real-time verification, effectively avoiding temperature measurement errors caused by single sensor failure or drift, thus improving the reliability of the temperature zone measurement results. At the same time, the application of data analysis algorithms further optimizes the integration effect of multi-point temperature measurement data, making the final output temperature zone measurement temperature more consistent with the actual working conditions, providing accurate data support for subsequent heating element adjustment and refrigeration system control.

[0047] This application provides a temperature control system for a vertical furnace body, comprising: an acquisition module for dividing the furnace body into multiple independent temperature zone modules along the axial direction, acquiring temperature data of each temperature zone module in real time, and determining whether the temperature data is greater than the process temperature set value. If so, the cooling unit is adjusted; otherwise, the heating unit is adjusted. A modeling module is used to preprocess the temperature data and the energizing time and current data of the heating elements within the furnace body to extract effective feature values. Based on the effective feature values, a time-series degradation model is trained and established. An evaluation module is used to evaluate the current degradation state of the heating elements in each temperature zone module according to the time-series degradation model to obtain the processed temperature measurement value of the temperature zone module. A judgment module is used to determine whether the deviation between the processed temperature measurement value of the temperature zone module and the preset measurement temperature is within a set range. If so, the temperature measurement result of the temperature zone module is determined to be reasonable; otherwise, an alarm is triggered and the temperature measurement is repeated.

[0048] It should be added that, during the training phase of the heating system, multiple sets of different heating element current adjustment gradients are preset, and each gradient corresponds to multiple preset current adjustment levels. Temperature change data collected within a preset time interval after the heating element is powered on at each level is recorded to establish an initial mapping relationship dataset between current and temperature rise characteristics. A multi-input multi-output control algorithm is adopted, using the initial mapping relationship dataset between current and temperature rise characteristics as the basic model. The deviation between the real-time temperature of each temperature zone module and the preset target temperature is used as the input. The current adjustment of the heating element is used as the output. The current parameters are iteratively optimized through calculation. After each iteration, the temperature uniformity of the temperature zone is compared until the temperature deviation between each temperature zone module is controlled within the preset temperature deviation reference range. This completes the optimization of the heating element's heat output ratio, ensuring that the heating power of the heating element matches the heat demand of the corresponding temperature zone module.

[0049] In this embodiment, by integrating four core modules, a leapfrog upgrade of vertical furnace temperature control from basic adjustment to predictive intelligent maintenance is achieved. Specifically, the acquisition module constitutes a real-time control closed loop. Through zoned temperature control and rapid power adjustment, it lays the foundation for the stable operation of the entire system with high precision and high response speed, ensuring that the initial setting of the process temperature can be quickly achieved and maintained. The establishment module, based on this, deeply mines multi-dimensional operating data such as the energizing time and current of the heating elements, and uses advanced signal processing methods such as filtering and modal analysis to extract effective feature values ​​characterizing the health status of the elements, such as resistance change trends and heating rates. This allows for the training of a high-precision time-series degradation model, enabling the system to discern the performance degradation patterns of the elements from the data. The evaluation module uses this model to perform online health diagnosis on each heating element, quantifying its degradation degree and intelligently compensating and correcting the original temperature measurements. This effectively eliminates measurement errors caused by element aging, solves the problem of "hidden" drift in process temperature caused by the slow degradation of element performance in traditional control, and improves the accuracy of temperature control from instantaneous reliability to full life-cycle reliability. Ultimately, the judgment module acts as a "quality gatekeeper," verifying the rationality of the corrected temperature and analyzing deviations. Upon detecting any anomalies, it immediately triggers an alarm and initiates a retest process. This not only enables real-time diagnosis and safety protection against abnormal states such as sensor malfunctions and sudden component failures, but also forms a self-optimizing closed loop of "measurement-control-evaluation-verification." Overall, this technical solution deeply integrates real-time control, predictive maintenance, and fault self-diagnosis, ultimately ensuring long-term furnace temperature stability and process consistency while significantly reducing unplanned downtime, extending equipment lifespan, and substantially improving production quality and intelligent management levels.

[0050] An embodiment of the present invention provides a computing device, comprising: one or more processors; and a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement a temperature control method for a vertical furnace body.

[0051] One embodiment of this application also provides a computer-readable storage medium for storing a program, which, when executed by a processor, implements a temperature control method for a vertical furnace body.

[0052] Further explanation is needed regarding the real-time acquisition of actual operating data from the vertical furnace. This data includes furnace internal cavity pressure data, reaction gas flow rate data, external ambient temperature data, and heating element aging parameters. The heating element aging parameters are calculated based on the remaining lifespan data output by the degradation model in the time synchronization and lifespan prediction system. The real-time temperature change curves of each temperature zone module under actual operating conditions are compared with the standard temperature change curves of the corresponding process stages in the ideal temperature control model to obtain the curve temperature deviation value. This deviation value includes the instantaneous temperature deviation at each time point and the average temperature deviation of the vertical furnace. The control parameters of the heating system are adjusted based on the curve temperature deviation value. Specifically, if the curve temperature deviation value is greater than or equal to the curve temperature deviation threshold, the proportional coefficient is increased to improve temperature regulation sensitivity, while the derivative coefficient is increased to suppress temperature overshoot. If the curve temperature deviation value is less than the curve temperature deviation threshold, the proportional coefficient is decreased to avoid temperature fluctuations, while the derivative coefficient is decreased to ensure temperature response speed.

[0053] In this embodiment, eliminating temperature coupling between different temperature zones is achieved through a core approach of "training modeling + real-time adaptation." First, during the heating system training phase, an initial mapping dataset of current-temperature rise characteristics is established. Then, a multi-input multi-output (MIMO) control algorithm is employed, using this dataset as the base model. The real-time temperature deviation between each temperature zone module and the target temperature is used as the input, and the heating element current adjustment is used as the output. Current parameters are optimized through iterative calculations. After each iteration, the temperature uniformity of the temperature zones is compared to optimize the heat output ratio, ensuring that the heating power matches the heat demand of each zone and initially eliminating static coupling interference. In subsequent operation, real-time data are collected on the furnace cavity pressure, reaction gas flow rate, external ambient temperature, and heating element aging parameters calculated based on the remaining lifespan of the degradation model. The real-time temperature change curves of each temperature zone are compared with the standard curves corresponding to the ideal model's process stage to obtain the instantaneous and average curve temperature deviation values. This process, through training modeling and real-time parameter adjustment, reduces inter-temperature coupling interference, improves temperature response speed, ensures long-term stability of multiple temperature zones within the preset temperature range, and guarantees the uniformity of the semiconductor process.

[0054] It should be noted that the communication protocol achieves inter-module collaborative control through a cascade control strategy. The cascade control strategy consists of a main control loop and a secondary control loop. The main control loop uses the overall average temperature of the furnace as the control target, while the secondary control loop uses the real-time temperature of each temperature zone module as the control target. The output signal of the main control loop serves as the setpoint reference for the secondary control loop. Furthermore, the communication protocol has a built-in preset coefficient calculation module. The preset coefficients are based on historical process data of the vertical furnace and are generated through least squares fitting. After each preset batch of processes is completed, historical data is automatically retrieved to update the preset coefficients. Through the cascade strategy and the preset coefficient calculation module working together, the output power of the heating element and the operating power of the cooling element of each temperature zone module are accurately determined to achieve a uniform temperature distribution in the furnace working area.

[0055] In this embodiment, the inter-module communication protocol is designed with "cascade control + dynamic coefficients" as the core to build collaborative control logic. The specific process is as follows: First, a cascade control strategy is embedded in the communication protocol architecture, clearly defining the main control loop and the secondary control loop. The main control loop takes the overall average temperature of the furnace as the core control target, collects the temperature data of all temperature zone modules in real time, calculates the average value, and compares it with the preset target temperature of the furnace. The secondary control loop takes the real-time temperature of a single temperature zone module as the control target. The adjustment signal generated by the main control loop based on the overall temperature deviation is directly used as the set value benchmark of the secondary control loop, ensuring that the temperature adjustment of each temperature zone module always revolves around the overall temperature control requirements of the furnace, avoiding overall temperature imbalance caused by individual adjustment of local temperature zones. At the same time, the communication protocol has a built-in preset coefficient calculation module. This module is based on the historical process data of the vertical furnace, and uses the least squares method to fit the historical data to generate preset coefficients that can match different process scenarios (such as heating power coefficient and cooling power coefficient). It is set that after each preset batch of processes is completed, the protocol automatically calls the latest accumulated historical process data to update the preset coefficients, ensuring that the coefficients always adapt to the actual operating conditions. During protocol operation, the cascade control strategy and the preset coefficient calculation module work together: the main control loop adjusts the setpoint of the secondary control loop based on the overall average temperature deviation, and the secondary control loop, combined with the corresponding coefficients output by the preset coefficient calculation module, accurately calculates the output power of the heating element of each temperature zone module (e.g., adjusting the current based on the coefficient) and the operating power of the cooling element (semiconductor cooler, water-cooled / air-cooled system) (e.g., adjusting the power of the semiconductor cooler and the cooling water flow rate), thus achieving differentiated power adjustment for each temperature zone module. This process solves the problem of temperature control coordination between the "overall and local" through cascade control, avoiding the overall temperature fluctuations caused by the independent control of each temperature zone in traditional independent control. Combined with dynamically updated preset coefficients, this improves the power adjustment accuracy of each temperature zone module.

[0056] It is important to understand that the master-slave control mode coordinates the temperature control of each temperature zone module. The master controller can identify the number n of the vertical furnace heating zones in real time and automatically call the decoupler program and PID controller program stored in the internal memory according to the value of n. This establishes a one-to-one correspondence between the n decouplers and the n PID controllers, so that each of the n heating zones forms an independent temperature control loop. Furthermore, the decouplers use a dynamic decoupling algorithm to decouple each temperature control loop, thereby reducing the mutual influence between the temperature zone modules. The master controller initiates a data synchronization command according to a preset periodic threshold. After receiving the command, each slave controller uploads the temperature control parameters of its own temperature zone module. The master controller performs a consistency check on the uploaded temperature control parameters of its own temperature zone module. If there is a parameter deviation, it sends a parameter correction command to the corresponding slave controller to ensure that the temperature control parameters of each temperature zone module are consistent. Moreover, as the heating elements approach their service life with the increase of usage time, the heating power is adjusted in a timely manner through temperature control parameter synchronization to compensate for the decrease in heating efficiency caused by the aging of the elements and maintain the stability of the vertical furnace body temperature.

[0057] In this embodiment, when coordinating the temperature control of each temperature zone module using a master-slave control mode, the master controller first identifies the number n of the vertical furnace heating zones in real time. After identification, the master controller automatically retrieves the decoupler program and PID (Proportional-Integral-Derivative) regulator program pre-stored in the internal memory. According to the rule of "one heating zone corresponds to one decoupler and one PID regulator", a one-to-one correspondence between n decouplers and n PID regulators is established, so that each heating zone independently forms a closed-loop temperature control loop. At the same time, the decoupler adopts a dynamic decoupling algorithm. By establishing a mathematical model of the temperature coupling of each heating zone, the coupling terms between temperature zones are eliminated by matrix inversion operation, reducing the temperature interference caused by the mutual influence of multiple temperature zones to below 5%, and avoiding the fluctuation effect of temperature adjustment in one temperature zone on other temperature zones. Based on independent temperature control and decoupling, the main controller sends data synchronization commands to each slave controller (each slave controller corresponds to an independent control system connected to a temperature zone module) according to a preset periodic threshold. After receiving the command, each slave controller immediately uploads the temperature control parameters of its controlled temperature zone module (including target temperature, real-time heating power, sensor calibration coefficient, and predicted remaining lifespan of the heating element). After receiving all parameters, the main controller uses a built-in verification algorithm to judge the consistency of the parameters. If the parameters uploaded by a slave controller deviate from the overall parameter benchmark, a parameter correction command is immediately sent to that slave controller to ensure that the temperature control parameters of all temperature zone modules remain consistent. When the heating element approaches its service life due to increased usage time, the main controller obtains the predicted remaining lifespan of each heating element through data synchronization, automatically calculates the heating power compensation value based on the lifespan data, and synchronizes the compensated heating power parameters to the corresponding slave controller. The slave controller then adjusts the operating power of the heating element to compensate for the decrease in heating efficiency caused by element aging. This process, through master-slave collaboration and dynamic decoupling, not only ensures independent and precise temperature control of each temperature zone, but also achieves overall parameter unification, making the temperature uniformity of the furnace body stable across multiple temperature zones. At the same time, it avoids local temperature drops caused by aging of heating elements, significantly improving the operational stability of the vertical furnace.

[0058] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0059] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0060] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0061] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0062] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0063] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for temperature control of a vertical furnace body, characterized in that, Includes the following steps: The furnace body is divided into multiple independent temperature zone modules along the axial direction. The temperature data of each temperature zone module is collected in real time, and it is determined whether the temperature data is greater than the process temperature set value. If so, the cooling unit is adjusted; if not, the heating unit is adjusted. Based on temperature data, as well as the energizing time and current data of the heating elements inside the furnace, preprocessing is performed to extract effective feature values; based on the effective feature values, a time series degradation model is trained and established. Based on the time series degradation model, the current degradation state of the heating element in each temperature zone module is evaluated to obtain the temperature measurement value of the processed temperature zone module; Determine whether the deviation between the temperature measurement value of the processed temperature zone module and the preset measurement temperature is within the set range. If yes, the temperature measurement result of the temperature zone module is deemed reasonable. If no, an alarm is triggered and the temperature measurement is repeated. The furnace body is divided into multiple independent temperature zone modules along the axial direction. Temperature data from each temperature zone module is collected in real time, and it is determined whether the temperature data exceeds the process temperature setpoint. If so, the cooling unit is adjusted; otherwise, the heating unit is adjusted. This includes: Based on the axial length of the furnace body and the process requirements, determine the number of temperature zone modules and the axial range of each temperature zone module; Based on the axial range of each temperature zone module, the heating unit, independent cooling unit, and temperature sensor corresponding to each temperature zone module are determined; wherein, the heating unit, cooling unit, and temperature sensor of each temperature zone module together constitute an independent temperature control loop. Temperature data for each temperature zone module is collected in real time and compared with the process temperature setpoint to obtain the comparison result; Based on the comparison results, when the temperature data is greater than the process temperature setpoint, a cooling control signal is generated to adjust the power of the cooling element. When the temperature data is lower than the process temperature setpoint, a heating control signal is generated to adjust the power of the heating element; Maintain the current power when the temperature data equals the process temperature setpoint. Based on temperature data, as well as the energizing time and current data of the heating elements inside the furnace, preprocessing is performed to extract effective feature values; based on these effective feature values, a time-series degradation model is trained and established, including: Temperature data, as well as the energizing time and current data of the heating element, are preprocessed to extract effective feature values; Based on the effective feature values, train the time series degradation model; Based on the time-series degradation model, a relationship model between time, lifetime, and current is established. Based on the effective feature values, a time series degradation model is trained, including: The effective feature values ​​are arranged in chronological order to construct a feature sequence dataset for model training; Using the aforementioned feature sequence dataset, a pre-defined initial degradation model is trained using a supervised learning algorithm, enabling the initial degradation model to learn the degradation pattern of heating element performance over time, thereby obtaining a trained time-series degradation model. Based on the trained time-series degradation model, a degradation state parameter is output to characterize the current health status of the heating element. The degradation state parameter includes at least the percentage of remaining service life and the performance degradation index. Based on the time-series degradation model, the current degradation state of the heating element in each temperature zone module is evaluated to obtain the temperature measurement values ​​of the processed temperature zone module, including: Based on the time series degradation model, the current degradation state of the heating element is evaluated to obtain the temperature measurement value of the processed temperature zone module; Based on the deviation between the temperature measurement value of the processed temperature zone module and the process temperature setting value, the current of the heating element is corrected in a closed loop to obtain the temperature measurement value of the processed temperature zone module. Based on the time-series degradation model, the current degradation state of the heating element is evaluated to obtain the processed temperature measurement values ​​of the temperature zone module, including: Obtain the degradation state parameters output by the time series degradation model; The degradation state parameters are compared with preset corresponding thresholds to obtain comparison results; the degree of performance degradation of the heating element is determined based on the comparison results to obtain the degree of performance degradation determination results. Based on the judgment result of the performance degradation degree, a corresponding temperature measurement value compensation coefficient is generated; The original temperature measurement value of the temperature zone module is corrected by using the temperature measurement value compensation coefficient to obtain the processed temperature measurement value of the temperature zone module. The system determines whether the deviation between the processed temperature measurement value of the temperature zone module and the preset measurement temperature is within the set range. If yes, the temperature measurement result of the temperature zone module is deemed reasonable; otherwise, an alarm is triggered and the temperature measurement is repeated, including: The temperature measurement value of the processed temperature zone module is compared with the preset measurement temperature; it is determined whether the deviation is within the set range; if the deviation is within the set range, the temperature measurement result is deemed valid; if the deviation is not within the set range, an alarm is triggered and the temperature measurement is repeated.

2. A temperature control system for a vertical furnace body, the system implementing the method as described in claim 1, characterized in that, include: The acquisition module is used to divide the furnace body into multiple independent temperature zone modules along the axial direction, collect the temperature data of each temperature zone module in real time, and determine whether the temperature data is greater than the process temperature set value. If so, the cooling unit is adjusted; if not, the heating unit is adjusted. A module is established to preprocess temperature data and the energizing time and current data of the heating elements inside the furnace to extract effective feature values; based on the effective feature values, a time series degradation model is trained and established. The evaluation module is used to evaluate the current degradation state of the heating element in each temperature zone module according to the time series degradation model, so as to obtain the temperature measurement value of the processed temperature zone module. The judgment module is used to determine whether the deviation between the temperature measurement value of the processed temperature zone module and the preset measurement temperature is within the set range. If yes, the temperature measurement result of the temperature zone module is determined to be reasonable. If no, an alarm is triggered and the temperature measurement is repeated.

3. A computing device, characterized in that, include: One or more processors; A storage device for storing one or more programs that, when executed by one or more processors, cause the one or more processors to perform the method as described in claim 1.

4. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program that, when executed by a processor, implements the method as described in claim 1.

Images