A multi-temperature-zone temperature control system optimization method based on gradient algorithm and segmentation control

Abstract

The application provides a multi-temperature-zone temperature control system optimization method based on a gradient algorithm and segmented control, which establishes a temperature control prediction model for each temperature zone by using a gradient algorithm based on the thermal response characteristics of the multi-temperature-zone, determines a heat transfer path based on the predicted temperature data determined based on the temperature control prediction model, divides the temperature control process into multiple stages based on the thermal response characteristics of the multi-temperature-zone, establishes a corresponding stage temperature control strategy based on the temperature control characteristics of each stage, ensures that the system accuracy reaches the best through the optimization based on the gradient algorithm and the segmented control strategy, optimizes the stage temperature control strategy based on the temperature difference between the real-time temperature data and the predicted temperature data, obtains a target stage temperature control strategy, realizes that the system can feedback and adjust the control parameters in real time, collects the actual temperature data at the end of each stage of the multi-temperature-zone, optimizes the heat transfer path based on the actual temperature data, and improves the energy efficiency of the overall temperature control system through the optimization of the heat transfer path.

Description

Technical Field

[0001] This invention relates to multi-zone temperature control, and particularly to an optimization method for a multi-zone temperature control system based on gradient algorithm and segmented control. Background Technology

[0002] With the widespread application of multi-zone temperature control systems, especially in high-precision industrial fields such as semiconductor manufacturing, lithium battery production, and the assembly of high-efficiency electronic devices, the accuracy and energy efficiency of temperature control systems have become key technical issues. Traditional temperature control methods mostly rely on a single PID control strategy. This method can achieve good results in some simple single-zone systems, but its performance is often unsatisfactory in complex multi-zone systems. In the lithium battery production process, due to the significant differences in temperature requirements among the various zones of a multi-zone furnace, existing control algorithms often cannot accurately achieve the temperature control requirements of each zone, leading to problems such as unstable battery quality and insufficient temperature control accuracy during production.

[0003] Currently, most multi-zone temperature control methods still use PID controllers. These controllers struggle to handle the rapid initial temperature changes and the later, more stable temperature phases, resulting in inconsistent performance across different stages of the temperature control process. This limitation is particularly pronounced in complex temperature control tasks. Furthermore, traditional temperature control systems do not adequately consider the differences in thermal response characteristics of each temperature zone, such as thermal response time, heat capacity, and heat transfer efficiency. This leads to unoptimized heat transfer paths, hindering efficient heat transfer and resulting in energy waste.

[0004] Furthermore, existing temperature control systems often neglect the optimization of heat transfer paths, resulting in uneven heat distribution between temperature zones and energy waste. Especially when temperature control targets are not met, existing technologies cannot adaptively adjust control strategies, thus impacting production efficiency and energy utilization.

[0005] In recent years, research on optimization techniques based on gradient algorithms and machine learning has gradually become a hot topic in the optimization of temperature control systems. By analyzing the thermal response characteristics of each temperature zone and using gradient algorithms to optimize the temperature control path, energy loss can be effectively reduced and heat transfer efficiency improved. Furthermore, introducing self-learning algorithms such as reinforcement learning allows the system to monitor temperature changes in real time and automatically adjust control strategies, thereby improving the adaptive capability of the temperature control system in complex environments. Although existing research has proposed various improvement schemes, such as fuzzy control and neural network control, these methods still have many limitations when dealing with multi-temperature zone systems. While existing distributed control systems can partially solve the single-point-of-failure problem of centralized control methods, they still face challenges in handling the complexity and high-precision requirements of multi-temperature zone systems. Therefore, how to design an efficient temperature control system that can cope with temperature zone differences and dynamically adapt to environmental changes has become a key issue in current technological development. Summary of the Invention

[0006] This invention proposes an optimization method for a multi-temperature zone temperature control system based on gradient algorithm and segmented control. The aim is to improve temperature control accuracy and system energy efficiency by analyzing the thermal response characteristics of each temperature zone and dynamically adjusting the temperature control strategy using optimization algorithms, thereby addressing the shortcomings of existing technologies.

[0007] An optimization method for a multi-temperature zone temperature control system based on gradient algorithm and piecewise control includes:

[0008] S1: Based on the thermal response characteristics of multiple temperature zones, a temperature control prediction model for each temperature zone is established using a gradient algorithm, and the heat transfer path is determined based on the predicted temperature data determined by the temperature control prediction model.

[0009] S2: Based on the thermal response characteristics of multiple temperature zones, the temperature control process is divided into multiple stages, and a corresponding stage temperature control strategy is established based on the temperature control characteristics of each stage.

[0010] S3: Collect real-time temperature data from multiple temperature zones, and optimize the stage temperature control strategy based on the temperature difference between the real-time temperature data and the predicted temperature data to obtain the target stage temperature control strategy.

[0011] S4: Collect actual temperature data at the end of each stage in the multi-temperature zone, and optimize the heat transfer path based on the actual temperature data.

[0012] Preferably, in step S1, based on the thermal response characteristics of multiple temperature zones, a temperature control prediction model for each temperature zone is established using a gradient algorithm, including:

[0013] Obtain thermal response time, heat capacity, and heat input for multiple temperature zones;

[0014] Using the gradient algorithm, a temperature control prediction model for each temperature zone is established according to the following formula;

[0015] ;

[0016] in, Indicates the monitored temperature zone at time temperature, Indicates the monitored temperature zone at time temperature, This indicates the input heat of the monitored temperature zone. This indicates the heat capacity of the monitored temperature range. Indicates the time step.

[0017] Preferably, in step S1, determining the heat transfer path based on the predicted temperature data determined by the temperature control prediction model includes:

[0018] Based on the predicted temperature data of each temperature zone determined by the temperature control prediction model, the required heat for each temperature zone is determined. Based on the required heat for each temperature zone and the location distribution of all temperature zones, the initial heat transfer path is determined.

[0019] The heat flow characteristics and temperature gradient of the temperature zone under the initial heat transfer path are obtained. Based on the heat flow characteristics and the location distribution, the direction of heat flow influence between temperature zones is determined. Based on the temperature gradient, the heat influence value between temperature zones is determined.

[0020] A heat interaction model between all temperature zones is established based on the direction of heat flow and the heat influence value.

[0021] Based on the production characteristics of each temperature zone, determine the accuracy of the temperature requirements for each temperature zone.

[0022] Based on the heat interaction model, the comprehensive heat influence value of each temperature zone under the initial heat transfer path is determined by the other temperature zones. Based on the comprehensive heat influence value of each temperature zone, the actual temperature of each temperature zone is determined, and it is judged whether the actual temperature meets the temperature zone's temperature accuracy requirements.

[0023] If so, ensure that the initial heat transfer path meets the requirements;

[0024] Otherwise, based on the relationship between the actual temperature and the required accuracy, and combined with the heat interaction model, the initial heat transfer path is adjusted until the actual temperature meets the temperature accuracy requirements of the temperature zone, thus obtaining the final heat transfer path.

[0025] Preferably, in step S2, based on the multi-temperature zone thermal response characteristics, the temperature control process is divided into multiple stages, including:

[0026] Based on the thermal response characteristics of multiple temperature zones, the temperature control process is divided into a rapid response stage and a stable regulation stage.

[0027] The rapid response phase occurs when the temperature in the temperature zone changes significantly, and the temperature control responds quickly. The stable adjustment phase occurs when the temperature in the temperature zone changes less significantly, and the temperature control responds precisely.

[0028] Preferably, in step S2, a corresponding stage temperature control strategy is established based on the temperature control characteristics of each stage, including:

[0029] Based on the rapid temperature changes during the fast response phase, a proportional-derivative control strategy is adopted.

[0030] Based on the temperature stabilizing during the steady-state adjustment phase, an integral control strategy is adopted.

[0031] Preferably, establishing a corresponding staged temperature control strategy also includes:

[0032] Based on the temperature control characteristics of multiple temperature zones, a temperature difference threshold is set;

[0033] When the absolute value of the temperature difference between the real-time temperature and the target temperature is greater than the temperature difference threshold, the multi-temperature zone temperature control system is determined to be in the rapid response phase.

[0034] When the absolute value of the temperature difference between the real-time temperature and the target temperature is less than or equal to the temperature difference threshold, the multi-temperature zone temperature control system is determined to be in the stable adjustment phase.

[0035] Preferably, in step S3, real-time temperature data from multiple temperature zones is collected, and the staged temperature control strategy is optimized based on the temperature difference between the real-time temperature data and the predicted temperature data to obtain the target staged temperature control strategy, including:

[0036] Obtain the data relationship between heating power, cooling equipment operating status and temperature differences in multiple temperature zones, and establish a thermal comfort matrix for multiple temperature zones under heating power and cooling equipment operating status. The row attributes of the thermal comfort matrix are each temperature zone, and the column data are temperature, humidity and speed.

[0037] Based on the principle of minimizing temperature differences and combined with data relationships, a main reward and penalty rule is established for the effect of temperature differences on heating power and the working status of cooling equipment.

[0038] Based on thermal comfort standards, a secondary reward and punishment rule is established for the thermal comfort matrix regarding heating power and the working status of cooling equipment.

[0039] A primary reward function is established based on the primary reward and punishment rules, and a secondary reward function is established based on the secondary reward and punishment rules.

[0040] Based on the temperature difference between real-time temperature data and predicted temperature data, and the historical data of the strategy parameters of the stage temperature control strategy, an initial strategy optimization model is established by combining machine learning.

[0041] The primary reward function and the secondary reward function are added to the initial policy optimization model and fused to obtain the target policy optimization model.

[0042] Based on real-time temperature difference data, as well as real-time heating power, real-time cooling equipment operating status, and real-time thermal comfort matrix input into the target strategy optimization model, the target stage temperature control strategy is obtained based on the output results.

[0043] Preferably, a primary reward function and a secondary reward function are fused into the initial policy optimization model to obtain a target policy optimization model, including:

[0044] Set the sovereign weight for the primary reward function and the secondary weight for the secondary reward function, where the sovereign weight is greater than the secondary weight.

[0045] Based on the primary and secondary weights, the primary reward function and the secondary reward function are fused to obtain the target reward function;

[0046] The target reward function is added to the initial policy optimization model to obtain the target policy optimization model.

[0047] Preferably, in step S4, the actual temperature data at the end of each stage of the multi-temperature zone is collected, and the heat transfer path is optimized based on the actual temperature data, including:

[0048] Obtain the temperature difference between the actual temperature data and the target temperature data for each temperature zone;

[0049] Establish numerical correspondences between path length consumption, temperature zone area, temperature heat balance, and temperature difference for each temperature.

[0050] Based on the numerical correspondence, the impact values ​​of path length consumption, temperature zone area, and temperature-heat balance on temperature control are determined respectively.

[0051] Based on the impact of path length consumption, temperature zone area, and temperature-heat balance on temperature control, the optimization coefficient for each temperature zone is determined.

[0052] The heat transfer path is optimized based on the aforementioned optimization coefficients.

[0053] Preferably, optimizing the heat transfer path based on the optimization coefficient includes:

[0054] Based on the optimization coefficient of each temperature zone, the corresponding path part in the heat transfer path is optimized to obtain the intermediate transfer path;

[0055] The connecting portion between adjacent temperature ranges in the intermediate heat transfer path is obtained. Based on the smoothness of the path, the connecting portion is optimized a second time to obtain the latest heat transfer path.

[0056] Compared with the prior art, the present invention has achieved the following beneficial effects:

[0057] By leveraging the thermal response characteristics of multiple temperature zones, a temperature control prediction model for each zone is established using a gradient algorithm. Based on the predicted temperature data determined by the temperature control prediction model, the heat transfer path is determined, initially ensuring the theoretical optimality of the heat path. Based on the thermal response characteristics of multiple temperature zones, the temperature control process is divided into multiple stages, and a corresponding stage temperature control strategy is established based on the temperature control characteristics of each stage. Through optimization based on the gradient algorithm and the segmented control strategy, the temperature of each temperature zone can be precisely adjusted to avoid temperature fluctuations and ensure optimal system accuracy. Real-time temperature data of multiple temperature zones is collected, and the stage temperature control strategy is optimized based on the temperature difference between the real-time temperature data and the predicted temperature data to obtain the target stage temperature control strategy. This enables the system to provide real-time feedback and adjust control parameters, thereby adapting to changes in different environmental conditions and maintaining efficient and stable operation. The actual temperature data at the end of each stage of multiple temperature zones is collected, and the heat transfer path is optimized based on the actual temperature data. By optimizing the heat transfer path, energy waste can be reduced, and heat can be evenly distributed among the temperature zones, improving the overall energy efficiency of the temperature control system.

[0058] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in this application.

[0059] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0060] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0061] Figure 1 This is a flowchart of an optimization method for a multi-temperature zone temperature control system based on gradient algorithm and segmented control in an embodiment of the present invention;

[0062] Figure 2 This is a flowchart illustrating how the temperature control process is divided into multiple stages in an embodiment of the present invention;

[0063] Figure 3 This is a flowchart illustrating the target stage temperature control strategy in an embodiment of the present invention. Detailed Implementation

[0064] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0065] Example 1: This embodiment of the invention provides an optimization method for a multi-temperature zone temperature control system based on gradient algorithm and piecewise control, such as... Figure 1 As shown, it includes:

[0066] S1: Based on the thermal response characteristics of multiple temperature zones, a temperature control prediction model for each temperature zone is established using a gradient algorithm, and the heat transfer path is determined based on the predicted temperature data determined by the temperature control prediction model.

[0067] S2: Based on the thermal response characteristics of multiple temperature zones, the temperature control process is divided into multiple stages, and a corresponding stage temperature control strategy is established based on the temperature control characteristics of each stage.

[0068] S3: Collect real-time temperature data from multiple temperature zones, and optimize the stage temperature control strategy based on the temperature difference between the real-time temperature data and the predicted temperature data to obtain the target stage temperature control strategy.

[0069] S4: Collect actual temperature data at the end of each stage in the multi-temperature zone, and optimize the heat transfer path based on the actual temperature data.

[0070] In this embodiment, thermal response characteristics include, for example, thermal response time, heat capacity, and thermal conductivity.

[0071] In this embodiment, the temperature control process is divided into multiple stages, including a rapid response stage and a stable adjustment stage. In the initial stage of the temperature control process, due to the rapid temperature change, a rapid response to the temperature difference change of the temperature control system is required to reduce temperature deviation. A proportional-derivative (PD) control strategy is used in this stage to cope with rapid temperature changes. When the temperature tends to stabilize, the temperature difference gradually decreases. At this point, a fine-tuning control strategy is needed to maintain system stability and avoid over-adjustment. Therefore, this invention employs an integral control (I) strategy.

[0072] In this embodiment, the optimization of the stage temperature control strategy is determined based on the assumption that the relationship between expected return and actual return tends to stabilize.

[0073] In this embodiment, optimizing the heat transfer path specifically involves dynamically adjusting the configuration of the heat source and the cold source.

[0074] The beneficial effects of the above design scheme are as follows: Based on the thermal response characteristics of multiple temperature zones, a temperature control prediction model for each temperature zone is established using a gradient algorithm. Based on the predicted temperature data determined by the temperature control prediction model, the heat transfer path is determined, initially ensuring the theoretical optimality of the heat path. Based on the thermal response characteristics of multiple temperature zones, the temperature control process is divided into multiple stages, and a corresponding stage temperature control strategy is established based on the temperature control characteristics of each stage. Through gradient algorithm-based optimization and segmented control strategies, the temperature of each temperature zone can be precisely adjusted, avoiding temperature fluctuations and ensuring optimal system accuracy. Real-time temperature data from multiple temperature zones is collected, and based on the temperature difference between the real-time temperature data and the predicted temperature data, the stage temperature control strategy is optimized to obtain the target stage temperature control strategy. This enables the system to provide real-time feedback and adjust control parameters, thereby adapting to changes in different environmental conditions and maintaining efficient and stable operation. Actual temperature data at the end of each stage in multiple temperature zones is collected, and based on the actual temperature data, the heat transfer path is optimized. By optimizing the heat transfer path, energy waste can be reduced, allowing heat to be evenly distributed among the temperature zones, thus improving the overall energy efficiency of the temperature control system.

[0075] Example 2: Based on Example 1, this embodiment of the invention provides an optimization method for a multi-temperature zone temperature control system based on gradient algorithm and piecewise control. In step S1, based on the thermal response characteristics of multiple temperature zones, a temperature control prediction model for each temperature zone is established using a gradient algorithm, including:

[0076] Obtain thermal response time, heat capacity, and heat input for multiple temperature zones;

[0077] Using the gradient algorithm, a temperature control prediction model for each temperature zone is established according to the following formula;

[0078] ;

[0079] in, Indicates the monitored temperature zone at time temperature, Indicates the monitored temperature zone at time +1 temperature, This indicates the input heat of the monitored temperature zone. This indicates the heat capacity of the monitored temperature range. Indicates the time step.

[0080] The beneficial effects of the above design scheme are: by using gradient algorithms to establish a temperature control prediction model for each temperature zone based on the thermal response characteristics of multiple temperature zones, a data foundation is provided for determining the temperature control strategy.

[0081] Example 3: Based on Example 2, this embodiment of the invention provides an optimization method for a multi-temperature zone temperature control system based on gradient algorithm and segmented control. In step S1, the heat transfer path is determined based on the predicted temperature data determined by the temperature control prediction model, including:

[0082] Based on the predicted temperature data of each temperature zone determined by the temperature control prediction model, the required heat for each temperature zone is determined. Based on the required heat for each temperature zone and the location distribution of all temperature zones, the initial heat transfer path is determined.

[0083] The heat flow characteristics and temperature gradient of the temperature zone under the initial heat transfer path are obtained. Based on the heat flow characteristics and the location distribution, the direction of heat flow influence between temperature zones is determined. Based on the temperature gradient, the heat influence value between temperature zones is determined.

[0084] A heat interaction model between all temperature zones is established based on the direction of heat flow and the heat influence value.

[0085] Based on the production characteristics of each temperature zone, determine the accuracy of the temperature requirements for each temperature zone.

[0086] Based on the heat interaction model, the comprehensive heat influence value of each temperature zone under the initial heat transfer path is determined by the other temperature zones. Based on the comprehensive heat influence value of each temperature zone, the actual temperature of each temperature zone is determined, and it is judged whether the actual temperature meets the temperature zone's temperature accuracy requirements.

[0087] If so, ensure that the initial heat transfer path meets the requirements;

[0088] Otherwise, based on the relationship between the actual temperature and the required accuracy, and combined with the heat interaction model, the initial heat transfer path is adjusted until the actual temperature meets the temperature accuracy requirements of the temperature zone, thus obtaining the final heat transfer path.

[0089] In this embodiment, the heat flow characteristics of each temperature zone include the heat flow direction and heat flow velocity within the temperature zone.

[0090] In this embodiment, the heat influence value between temperature zones is the influence of other temperature zones on the heat of the current temperature zone.

[0091] In this embodiment, the greater the temperature gradient in the temperature zone, the faster the heat transfer rate and the smaller the heat impact value.

[0092] In this embodiment, the heat interaction model is specifically based on historical data of the direction and value of heat flow influence, combined with machine learning training, and is used to predict the interaction between different temperature zones under different paths.

[0093] In this embodiment, the required temperature accuracy for each temperature zone is determined. For example, the optimal temperature control for the current temperature zone is 80 degrees Celsius, the maximum threshold is 90, and the minimum threshold is 70. The required accuracy is determined by the difference between the maximum threshold and the minimum threshold. The smaller the difference, the higher the required accuracy.

[0094] In this embodiment, the accuracy requirement of the actual temperature to meet the temperature zone's requirements is specifically as follows: Following the above statement, if the actual temperature is 82 degrees, the actual accuracy corresponding to the actual temperature is calculated by first obtaining the absolute value of the difference between the optimal temperature and the actual temperature, dividing the absolute value of the difference by the difference between the maximum threshold and the minimum threshold to obtain the calculated value, and then subtracting the calculated value from 1 to obtain the actual accuracy. The calculation result is... =90%. If the required accuracy is set to 85%, the current situation meets the required accuracy requirement. If the required accuracy is set to 95%, the current situation does not meet the required accuracy requirement.

[0095] The beneficial effects of the above design scheme are as follows: By considering the location distribution of all temperature zones, the heat flow characteristics and temperature gradient of the temperature zones under the initial heat transfer path, and the production characteristics of each temperature zone, the heat transfer path is established and adjusted. In the process, the required heat for each temperature zone is deduced by predicting temperature data, and the initial path is generated in combination with the physical location distribution, so that the heat distribution is aligned with the actual production needs from the source. By comparing the accuracy of the actual temperature with the required temperature, the initial path is cyclically adjusted to ensure that the heat transfer path always adapts to the real-time temperature control requirements. It is especially suitable for dynamic balance under complex working conditions with multiple temperature zones. The production characteristics of the temperature zones are incorporated into the accuracy requirement analysis, so that the temperature control system is upgraded from a simple "temperature control" to "production process adaptation", improving the compatibility and flexibility of automated production, and ultimately ensuring the theoretical optimality of the heat path.

[0096] Example 4: Based on Example 1, this embodiment of the invention provides an optimization method for a multi-temperature zone temperature control system based on gradient algorithm and piecewise control, such as... Figure 2 As shown, in step S2, based on the multi-temperature zone thermal response characteristics, the temperature control process is divided into multiple stages, including:

[0097] Based on the thermal response characteristics of multiple temperature zones, the temperature control process is divided into a rapid response stage and a stable regulation stage.

[0098] The rapid response phase occurs when the temperature in the temperature zone changes significantly, and the temperature control responds quickly. The stable adjustment phase occurs when the temperature in the temperature zone changes less significantly, and the temperature control responds precisely.

[0099] The beneficial effects of the above design scheme are as follows: by dividing the temperature control process into a rapid response stage and a stable adjustment stage based on the thermal response characteristics of multiple temperature zones, the rapid response stage is when the temperature of the temperature zone changes significantly and the temperature control responds quickly, while the stable adjustment stage is when the temperature of the temperature zone changes slightly and the temperature control responds precisely. By using a customized algorithm, local optimization of temperature control is achieved, ultimately leading to global optimization of temperature control.

[0100] Example 5: Based on Example 4, this embodiment of the invention provides an optimization method for a multi-temperature zone temperature control system based on gradient algorithm and segmented control. The method is characterized in that, in step S2, a corresponding stage temperature control strategy is established based on the temperature control characteristics of each stage, including:

[0101] Based on the rapid temperature changes during the fast response phase, a proportional-derivative control strategy is adopted.

[0102] Based on the temperature stabilizing during the steady-state adjustment phase, an integral control strategy is adopted.

[0103] In this embodiment, during the initial stage of the temperature control process, due to the rapid temperature change, the temperature control system needs to respond quickly to changes in temperature difference to reduce temperature deviation. In this stage, a proportional-derivative (PD) control strategy is used to address the rapid temperature changes. The PD control formula is as follows:

[0104] ;

[0105] in: It is a control signal output (such as the power setting of a heater or cooling device); It's the temperature difference. To set the temperature, Real-time temperature; It is the proportional gain coefficient. It is the differential gain coefficient; It is the rate of change of temperature difference, representing the speed at which temperature changes.

[0106] In this embodiment, as the temperature stabilizes, the temperature difference gradually decreases. At this point, a fine-tuned control strategy is needed to maintain system stability and avoid over-adjustment. Therefore, this invention employs an integral control (I-control) strategy. The integral control formula is as follows:

[0107] ;

[0108] in: It is the integral gain coefficient; This is the integral term of the error, representing the cumulative temperature deviation over time. Integral control eliminates persistent, minor errors in the system by adjusting for the accumulated temperature difference, ensuring that the temperature remains stable near the target value. This control strategy prevents over-adjustment and maintains the accuracy of the temperature control system.

[0109] The beneficial effects of the above design scheme are as follows: by adopting a proportional-derivative control strategy based on the rapid temperature change during the fast response phase, and based on the temperature stabilization during the steady adjustment phase, this strategy is suitable for the initial stage of the system. By reducing temperature error and oscillation through rapid response, and by adopting an integral control strategy, the system can prevent over-adjustment and maintain the accuracy of the temperature control system.

[0110] Example 6: Based on Example 5, this embodiment of the invention provides an optimization method for a multi-temperature zone temperature control system based on gradient algorithm and segmented control, establishing a corresponding staged temperature control strategy, and further including:

[0111] Based on the temperature control characteristics of multiple temperature zones, a temperature difference threshold is set;

[0112] When the absolute value of the temperature difference between the real-time temperature and the target temperature is greater than the temperature difference threshold, the multi-temperature zone temperature control system is determined to be in the rapid response phase.

[0113] When the absolute value of the temperature difference between the real-time temperature and the target temperature is less than or equal to the temperature difference threshold, the multi-temperature zone temperature control system is determined to be in the stable adjustment phase.

[0114] In this embodiment, to achieve dynamic switching of the above strategy, the system sets a temperature difference threshold. ,when At that time, the system enters the rapid response phase; when When the condition is maintained for a certain period of time, the system automatically switches to fine-tuning control. Furthermore, the parameters of the control strategy... , , It can adaptively adjust based on the thermal inertia and conductivity characteristics of each temperature zone. For example, a smaller value can be set for a temperature zone with a larger heat capacity. To prevent overshoot, areas with high heat conduction can be enlarged. To improve response speed.

[0115] In this step, the selection and switching of control strategies are completed by the central control module. This module determines the current stage by analyzing indicators such as temperature change curves and error integral trends in real time, and automatically switches the control mode according to preset logic.

[0116] The beneficial effects of the above design scheme are: when the absolute value of the temperature difference between the real-time temperature and the target temperature is greater than the temperature difference threshold, the multi-temperature zone temperature control system is determined to be in the rapid response phase; when the absolute value of the temperature difference between the real-time temperature and the target temperature is less than or equal to the temperature difference threshold, the multi-temperature zone temperature control system is determined to be in the stable adjustment phase, thereby realizing the dynamic switching of the strategy.

[0117] Example 7: Based on Example 1, this embodiment of the invention provides an optimization method for a multi-temperature zone temperature control system based on gradient algorithm and piecewise control, such as... Figure 3 As shown, in step S3, real-time temperature data from multiple temperature zones is collected. Based on the temperature difference between the real-time temperature data and the predicted temperature data, the stage temperature control strategy is optimized to obtain the target stage temperature control strategy, including:

[0118] Obtain the data relationship between heating power, cooling equipment operating status and temperature differences in multiple temperature zones, and establish a thermal comfort matrix for multiple temperature zones under heating power and cooling equipment operating status. The row attributes of the thermal comfort matrix are each temperature zone, and the column data are temperature, humidity and speed.

[0119] Based on the principle of minimizing temperature differences and combined with data relationships, a main reward and penalty rule is established for the effect of temperature differences on heating power and the working status of cooling equipment.

[0120] Based on thermal comfort standards, a secondary reward and punishment rule is established for the thermal comfort matrix regarding heating power and the working status of cooling equipment.

[0121] A primary reward function is established based on the primary reward and punishment rules, and a secondary reward function is established based on the secondary reward and punishment rules.

[0122] Based on the temperature difference between real-time temperature data and predicted temperature data, and the historical data of the strategy parameters of the stage temperature control strategy, an initial strategy optimization model is established by combining machine learning.

[0123] The primary reward function and the secondary reward function are added to the initial policy optimization model and fused to obtain the target policy optimization model.

[0124] Based on real-time temperature difference data, as well as real-time heating power, real-time cooling equipment operating status, and real-time thermal comfort matrix input into the target strategy optimization model, the target stage temperature control strategy is obtained based on the output results.

[0125] In this embodiment, the thermal comfort standard is set according to the actual situation. The closer the thermal comfort matrix is ​​to the thermal comfort standard, the greater the reward. The further away the thermal comfort matrix is ​​from the thermal comfort standard, the greater the penalty.

[0126] In this embodiment, the smaller the temperature difference between the heating power and the cooling equipment under operating conditions, the greater the corresponding reward; conversely, the greater the temperature difference between the heating power and the cooling equipment under operating conditions, the greater the corresponding penalty.

[0127] The beneficial effects of the above design scheme are as follows: Traditional temperature control systems typically only use temperature as a single control objective, while this scheme establishes a multi-dimensional evaluation system through a thermal comfort matrix (integrating temperature, humidity, and airflow velocity). This ensures that the temperature control strategy not only meets temperature accuracy requirements but also quantifies thermal comfort standards into a reward and penalty mechanism through secondary reward and penalty rules. During the optimization process, the system prioritizes the combination of control parameters that meets both temperature accuracy and thermal comfort range. The main reward and penalty rule is centered on "minimizing temperature difference." By establishing a quantitative relationship between temperature difference and heating power and cooling equipment status, the system can quickly adjust control parameters based on real-time errors. The target model, which integrates the reward function, further transforms comfort and control accuracy into model optimization objectives, enabling the system to automatically find the optimal solution when multiple objectives conflict. This improves robustness in complex scenarios and allows the system to provide real-time feedback and adjust control parameters, thereby adapting to changes in different environmental conditions and maintaining efficient and stable operation.

[0128] Example 8: Based on Example 7, this embodiment of the invention provides an optimization method for a multi-temperature zone temperature control system based on gradient algorithm and piecewise control. The method involves fusing a primary reward function and a secondary reward function into the initial strategy optimization model to obtain a target strategy optimization model, including:

[0129] Set the sovereign weight for the primary reward function and the secondary weight for the secondary reward function, where the sovereign weight is greater than the secondary weight.

[0130] Based on the primary and secondary weights, the primary reward function and the secondary reward function are fused to obtain the target reward function;

[0131] The target reward function is added to the initial policy optimization model to obtain the target policy optimization model.

[0132] The beneficial effects of the above design scheme are as follows: by fusing the primary reward function and the secondary reward function based on the primary weight and the secondary weight, a target reward function is obtained. The target reward function is then added to the initial policy optimization model to obtain the target policy optimization model. The target model that fuses the reward function further transforms comfort and control accuracy into model optimization objectives, enabling the system to automatically find the optimal solution when there are multiple objectives in conflict, thereby improving robustness in complex scenarios.

[0133] Example 9: Based on Example 1, this embodiment of the invention provides an optimization method for a multi-temperature zone temperature control system based on gradient algorithm and segmented control. In step S4, actual temperature data at the end of each stage of the multi-temperature zone is collected, and the heat transfer path is optimized based on the actual temperature data, including:

[0134] Obtain the temperature difference between the actual temperature data and the target temperature data for each temperature zone;

[0135] Establish numerical correspondences between path length consumption, temperature zone area, temperature heat balance, and temperature difference for each temperature.

[0136] Based on the numerical correspondence, the impact values ​​of path length consumption, temperature zone area, and temperature-heat balance on temperature control are determined respectively.

[0137] Based on the impact of path length consumption, temperature zone area, and temperature-heat balance on temperature control, the optimization coefficient for each temperature zone is determined.

[0138] ;

[0139] in, This represents the optimization coefficient for the current temperature range. This indicates the impact of the path length consumption in the current temperature zone on temperature control. This represents the minimum impact of path length consumption on temperature control across the entire temperature range. This indicates the impact of the current temperature zone area on temperature control. This represents the minimum impact of the temperature zone area on temperature control across all temperature zones. This indicates the impact of the current temperature zone's heat balance on temperature control. This represents the minimum impact of a moderate heat balance across all temperature zones on temperature control.

[0140] The heat transfer path is optimized based on the aforementioned optimization coefficients.

[0141] In this embodiment, the larger the optimization coefficient, the greater the corresponding path adjustment range.

[0142] In this embodiment, the greater the impact on temperature control, the greater the corresponding impact value.

[0143] The beneficial effects of the above design scheme are as follows: By establishing a numerical correspondence between path length consumption, temperature zone area, heat balance, and temperature difference, the specific impact of each factor on the temperature control effect can be quantified (for example, for every 1 meter increase in path length, the average temperature difference increases by 0.3℃; for every 10㎡ increase in temperature zone area, the temperature difference fluctuation increases by 0.5℃). This allows the system to accurately identify which temperature zones have large temperature differences due to excessively long paths and which temperature zones require additional heat compensation due to excessively large areas. For temperature zones with high impact values ​​from path length consumption, such as those where long paths lead to heat loss, optimizing the path, such as shortening pipe length or replacing insulation materials, can reduce heat loss during transmission. After path optimization, the temperature difference in such temperature zones can be reduced, and the energy consumption of heating equipment can also be reduced, improving temperature control efficiency and balance.

[0144] Example 10: Based on Example 9, this embodiment of the invention provides an optimization method for a multi-temperature zone temperature control system based on gradient algorithm and piecewise control, which optimizes the heat transfer path based on the optimization coefficient, including:

[0145] Based on the optimization coefficient of each temperature zone, the corresponding path part in the heat transfer path is optimized to obtain the intermediate transfer path;

[0146] The connecting portion between adjacent temperature ranges in the intermediate heat transfer path is obtained. Based on the smoothness of the path, the connecting portion is optimized a second time to obtain the latest heat transfer path.

[0147] The beneficial effects of the above design scheme are as follows: by optimizing the corresponding path part in the heat transfer path based on the optimization coefficient of each temperature zone, an intermediate transfer path is obtained. The connecting part between adjacent temperature zones in the intermediate transfer path is obtained. Based on the smoothness of the path, the connecting part is optimized again to obtain the latest heat transfer path. This ensures the smoothness of the path connection between each temperature zone after the path optimization and guarantees the feasibility of the latest heat transfer path.

[0148] In one embodiment, this embodiment provides a specific application in a semiconductor wafer fabrication furnace, namely, an optimization method for a multi-temperature zone temperature control system based on gradient algorithms and segmented control, which specifically includes the following steps:

[0149] A multi-temperature zone dynamic temperature control model is established, with the goal of minimizing temperature fluctuations and energy loss. Real-time data-driven parameter optimization is performed by collecting temperature data in real time and inputting it into the reinforcement learning model. Cross-temperature zone heat co-optimization is also implemented, calculating the optimal heat transfer path between the temperature zones to address the thermal coupling effect between the deposition zone and the annealing zone.

[0150] This method significantly improves temperature control accuracy and system energy efficiency by analyzing the differences in thermal characteristics of each temperature zone and combining dynamic control strategies with real-time data optimization.

[0151] Semiconductor manufacturing furnaces comprise multiple independent temperature zones, including preheating, deposition, and annealing zones, each with significant differences in heat capacity, heat transfer efficiency, and response time. A dynamic temperature control model is constructed using a gradient algorithm. By analyzing the temperature gradient and target temperature deviation within each zone, the heating power allocation is dynamically optimized. The model aims to minimize temperature fluctuations and energy loss, focusing on ensuring the deposition zone temperature remains stable within an error range of ±0.5℃. High-precision distributed fiber optic temperature sensors (measurement error ±0.1℃) are deployed in the deposition and annealing zones to collect temperature data in real time and input it into the reinforcement learning model. The model optimizes temperature deviation and total energy consumption, updating control parameters every 10 seconds and dynamically adjusting heater power and cooling system speed to ensure rapid system response to environmental changes. The optimal heat transfer path between the deposition and annealing zones is calculated to address the thermal coupling effect. The gradient algorithm analyzes temperature gradients, contact areas, and spacing to dynamically adjust the insulation plate opening, reducing ineffective heat exchange. This embodiment applies the described method to a semiconductor manufacturing furnace to solve the temperature oscillation problem caused by high thermal inertia in high-temperature processes. By optimizing the heat transfer path using a gradient algorithm and combining it with a fuzzy PID composite control strategy, the system can adapt to different wafer sizes and material properties, significantly improving manufacturing yield and energy efficiency. Furthermore, this method can be extended to fields such as photovoltaic cell sintering and precision optical coating, adapting to diverse process requirements by adjusting control parameters.

[0152] 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 this application and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. An optimization method for a multi-temperature zone temperature control system based on gradient algorithm and piecewise control, characterized in that, include: S1: Based on the thermal response characteristics of multiple temperature zones, a temperature control prediction model for each temperature zone is established using a gradient algorithm. Based on the predicted temperature data determined by the temperature control prediction model, the heat transfer path is determined, including: Obtain thermal response time, heat capacity, and heat input for multiple temperature zones; Using the gradient algorithm, a temperature control prediction model for each temperature zone is established according to the following formula; ； in, Indicates the monitored temperature zone at time temperature, Indicates the monitored temperature zone at time +1 temperature, This indicates the input heat of the monitored temperature zone. This indicates the heat capacity of the monitored temperature range. Indicates the time step; S2: Based on the thermal response characteristics of multiple temperature zones, the temperature control process is divided into multiple stages, and a corresponding stage temperature control strategy is established based on the temperature control characteristics of each stage. S3: Collect real-time temperature data from multiple temperature zones, and optimize the stage temperature control strategy based on the temperature difference between the real-time temperature data and the predicted temperature data to obtain the target stage temperature control strategy. S4: Collect actual temperature data at the end of each stage in the multi-temperature zone, and optimize the heat transfer path based on the actual temperature data, including: Obtain the temperature difference between the actual temperature data and the target temperature data for each temperature zone; Establish numerical correspondences between path length consumption, temperature zone area, temperature heat balance, and temperature difference for each temperature. Based on the numerical correspondence, the impact values ​​of path length consumption, temperature zone area, and temperature-heat balance on temperature control are determined respectively. Based on the impact of path length consumption, temperature zone area, and temperature-heat balance on temperature control, the optimization coefficient for each temperature zone is determined. ； in, This represents the optimization coefficient for the current temperature range. This indicates the impact of the path length consumption in the current temperature zone on temperature control. This represents the minimum impact of path length consumption on temperature control across the entire temperature range. This indicates the impact of the current temperature zone area on temperature control. This represents the minimum impact of the temperature zone area on temperature control across all temperature zones. This indicates the impact of the current temperature zone's heat balance on temperature control. This represents the minimum impact of a moderate heat balance across all temperature zones on temperature control. The heat transfer path is optimized based on the aforementioned optimization coefficients, including: Based on the optimization coefficient of each temperature zone, the corresponding path part in the heat transfer path is optimized to obtain the intermediate transfer path; The connecting portion between adjacent temperature ranges in the intermediate heat transfer path is obtained. Based on the smoothness of the path, the connecting portion is optimized a second time to obtain the latest heat transfer path.

2. The optimization method for a multi-temperature zone temperature control system based on gradient algorithm and piecewise control according to claim 1, characterized in that, In step S1, the heat transfer path is determined based on the predicted temperature data determined by the temperature control prediction model, including: Based on the predicted temperature data of each temperature zone determined by the temperature control prediction model, the required heat for each temperature zone is determined. Based on the required heat for each temperature zone and the location distribution of all temperature zones, the initial heat transfer path is determined. The heat flow characteristics and temperature gradient of the temperature zone under the initial heat transfer path are obtained. Based on the heat flow characteristics and the location distribution, the direction of heat flow influence between temperature zones is determined. Based on the temperature gradient, the heat influence value between temperature zones is determined. A heat interaction model between all temperature zones is established based on the direction of heat flow and the heat influence value. Based on the production characteristics of each temperature zone, determine the accuracy of the temperature requirements for each temperature zone. Based on the heat interaction model, the comprehensive heat influence value of each temperature zone under the initial heat transfer path is determined by the other temperature zones. Based on the comprehensive heat influence value of each temperature zone, the actual temperature of each temperature zone is determined, and it is judged whether the actual temperature meets the temperature zone's temperature accuracy requirements. If so, ensure that the initial heat transfer path meets the requirements; Otherwise, based on the relationship between the actual temperature and the required accuracy, and combined with the heat interaction model, the initial heat transfer path is adjusted until the actual temperature meets the temperature accuracy requirements of the temperature zone, thus obtaining the final heat transfer path.

3. The optimization method for a multi-temperature zone temperature control system based on gradient algorithm and piecewise control according to claim 1, characterized in that, In S2, based on the multi-temperature zone thermal response characteristics, the temperature control process is divided into multiple stages, including: Based on the thermal response characteristics of multiple temperature zones, the temperature control process is divided into a rapid response stage and a stable regulation stage. The rapid response phase occurs when the temperature in the temperature zone changes significantly, and the temperature control responds quickly. The stable adjustment phase occurs when the temperature in the temperature zone changes less significantly, and the temperature control responds precisely.

4. The optimization method for a multi-temperature zone temperature control system based on gradient algorithm and piecewise control according to claim 3, characterized in that, In step S2, a corresponding stage temperature control strategy is established based on the temperature control characteristics of each stage, including: Based on the rapid temperature changes during the fast response phase, a proportional-derivative control strategy is adopted. Based on the temperature stabilizing during the steady-state adjustment phase, an integral control strategy is adopted.

5. The optimization method for a multi-temperature zone temperature control system based on gradient algorithm and piecewise control according to claim 4, characterized in that, Establishing corresponding phased temperature control strategies also includes: Based on the temperature control characteristics of multiple temperature zones, a temperature difference threshold is set; When the absolute value of the temperature difference between the real-time temperature and the target temperature is greater than the temperature difference threshold, the multi-temperature zone temperature control system is determined to be in the rapid response phase. When the absolute value of the temperature difference between the real-time temperature and the target temperature is less than or equal to the temperature difference threshold, the multi-temperature zone temperature control system is determined to be in the stable adjustment phase.

6. The optimization method for a multi-temperature zone temperature control system based on gradient algorithm and piecewise control according to claim 1, characterized in that, In step S3, real-time temperature data from multiple temperature zones is collected. Based on the temperature difference between the real-time temperature data and the predicted temperature data, the stage temperature control strategy is optimized to obtain the target stage temperature control strategy, including: Obtain the data relationship between heating power, cooling equipment operating status and temperature differences in multiple temperature zones, and establish a thermal comfort matrix for multiple temperature zones under heating power and cooling equipment operating status. The row attributes of the thermal comfort matrix are each temperature zone, and the column data are temperature, humidity and speed. Based on the principle of minimizing temperature differences and combined with data relationships, a main reward and penalty rule is established for the effect of temperature differences on heating power and the working status of cooling equipment. Based on thermal comfort standards, a secondary reward and punishment rule is established for the thermal comfort matrix regarding heating power and the working status of cooling equipment. A primary reward function is established based on the primary reward and punishment rules, and a secondary reward function is established based on the secondary reward and punishment rules. Based on the temperature difference between real-time temperature data and predicted temperature data, and the historical data of the strategy parameters of the stage temperature control strategy, an initial strategy optimization model is established by combining machine learning. The primary reward function and the secondary reward function are added to the initial policy optimization model and fused to obtain the target policy optimization model. Based on real-time temperature difference data, as well as real-time heating power, real-time cooling equipment operating status, and real-time thermal comfort matrix input into the target strategy optimization model, the target stage temperature control strategy is obtained based on the output results.

7. The optimization method for a multi-temperature zone temperature control system based on gradient algorithm and piecewise control according to claim 6, characterized in that, By fusing the primary reward function and secondary reward function into the initial policy optimization model, a target policy optimization model is obtained, including: Set the sovereign weight for the primary reward function and the secondary weight for the secondary reward function, where the sovereign weight is greater than the secondary weight. Based on the primary and secondary weights, the primary reward function and the secondary reward function are fused to obtain the target reward function; The target reward function is added to the initial policy optimization model to obtain the target policy optimization model.

Images