A large surface source blackbody multi-channel balanced heating method, device and medium

Abstract

The present invention belongs to the technical field of temperature control of large surface source black bodies, and discloses a method, device and medium for multi-channel balanced heating of large surface source black bodies, including: S1, data acquisition: generating a simulated heating data set of a large surface source black body through a black body simulation system, and building a multi-channel actual temperature control system at the same time to obtain real heating data sets of different channels; S2, problem modeling; S3, balanced control strategy design; S4, model training; S5, verification and adjustment; S6, iterative optimization. The present invention achieves the goal of enhancing the consistency of the heating rate of each channel, while reducing the influence of thermal interference, and improving the overall temperature control system performance of the large surface source black body. It provides a new research direction and application prospect for the development of temperature control technology.

Description

Technical Field

[0001] The invention relates to the technical field of large surface source black body temperature control, and in particular to a large surface source black body multi-channel balanced heating method, device and medium. Background Art

[0002] In the field of blackbody temperature control, large surface source temperature control systems have attracted much attention due to their important applications in infrared imaging, target simulation, thermal testing and other fields. However, in actual operation, the optimization of system performance and balanced control are key challenges to achieve high-precision temperature control. Large surface source temperature control systems usually adopt a multi-channel design, such as 64-channel temperature control (8×8 array), where each channel is heated independently to achieve overall temperature control. However, due to the complexity of thermal interference, channel differences and environmental factors, the heating rates of each channel are difficult to be completely consistent, which puts high demands on the balance and stability of the system. Thermal interference is one of the core factors affecting the performance of the temperature control system. It is mainly manifested in the heat conduction effect between channels and the influence of the external environment on the heating process. The heat conduction effect makes the temperature change of one channel directly or indirectly affect the temperature change of its adjacent channels. For example, when a channel heats up faster, the adjacent channel may heat up faster due to heat transfer, and vice versa. In addition, the path and intensity of heat transfer are restricted by factors such as the physical structure of the channel and the distribution of thermal resistance, resulting in significant non-uniformity in the thermal interference effect. Environmental factors such as wind speed, humidity and ambient temperature fluctuations further increase the uncertainty of system temperature control, making it difficult for a single-channel heating mode to meet the balance requirements of the overall system.

[0003] There is currently no method to suppress this problem. Summary of the invention

[0004] The object of the present invention is to provide a large surface source blackbody multi-channel balanced heating method, device and medium to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A large surface source blackbody multi-channel balanced heating method comprises the following steps:

[0007] S1, data acquisition: Generate a simulated heating data set of a large surface source blackbody through a blackbody simulation system, and build a multi-channel actual temperature control system to obtain real heating data sets of different channels;

[0008] S2, Problem Modeling: Model the large-surface blackbody multi-channel heating balance problem as a multi-agent collaborative optimization problem, define the state, action and reward function of each channel, establish a dynamic association mechanism between channels, and ensure that the multi-channel heating rates tend to be consistent;

[0009] S3, balanced control strategy design: adopt a multi-channel collaborative control strategy based on reinforcement learning, optimize the heating rate deviation between channels by dynamically adjusting the duty cycle of each channel; design a joint reward mechanism to balance the global temperature averaging effect and the local temperature difference minimization goal;

[0010] S4, model training: A multi-channel balanced control model is constructed based on the reinforcement learning DQN model, and then the multi-channel balanced control model is trained using a simulated heating data set; the synchronization of the heating rates between channels is improved through objective function optimization; the performance of the multi-channel balanced control model is then verified on a real heating data set, and the stability and robustness of the multi-channel balanced control model in a complex environment is evaluated;

[0011] S5, verification and adjustment: Use the real heating data set of S1 to verify the multi-channel balanced control model, analyze the temperature equalization effect and the temperature difference change during the heating process, and compare the deviation between the actual and target temperatures; dynamically adjust the strategy and model parameters of the multi-channel balanced control model according to the verification results to improve the accuracy of balanced heating;

[0012] S6, iterative optimization: Based on the verification feedback from simulation and real environment, the iterative process of modeling, training, and verification is repeatedly executed to optimize the multi-channel balanced control model so that it can achieve balanced heating of large surface source black bodies efficiently and accurately under different power outputs.

[0013] Further, the S2 comprises the following steps:

[0014] S2.1, the multi-channel heating balance problem of large surface source blackbody is modeled as a multi-agent collaborative optimization problem. Each heating channel is regarded as an independent agent. The agents collaborate to complete the overall balanced heating task through a dynamic association mechanism. In the state space design, the state of each agent includes the current temperature T i , Heating rate And the current duty cycle DutyCycle i , forming the state vector S i =[T i ,R i ,DutyCycle i ];

[0015] S2.2, the action space is defined as the adjustment of the duty cycle to ensure the flexibility of heating process control;

[0016] S2.3, the reward function is designed jointly, the goal is to optimize the consistency of the global heating rate and minimize the local temperature difference; the heating rate consistency is achieved by Quantification, is the average heating rate of all channels; the temperature accuracy is measured by -|T i -Tgoal ∣Measurement; Among them, T i is the current temperature, T goal is the target temperature;

[0017] The final joint reward function Reward is in the form of:

[0018]

[0019] Among them, α and β are adjustment coefficients;

[0020] By global rate average Transmit information between channels to ensure the computational efficiency of the multi-channel equalization control model.

[0021] Further, the S3 includes:

[0022] S3.1, the equilibrium control strategy is based on the deep Q network algorithm, which uses a neural network to approximate the Q value function to select the optimal action; in the strategy design, the input is the current state S of each channel i =[T i ,R i ,DutyCycle i ], output is the Q value of each possible action; by selecting the action with the largest Q value, the agent dynamically adjusts the duty cycle to optimize the heating process;

[0023] S3.2, Q-value update follows the Bellman equation:

[0024]

[0025] Among them, S i is the current state, S t is the updated state, A t is the current action, A' represents the possible action to be selected in the next state, η is the learning rate, γ is the discount factor, and max represents the maximum value;

[0026] S3.3, balanced control strategy adopts Greedy method.

[0027] Furthermore, in S3.3 Greedy methods include: Randomly select actions to The probability of selecting the action corresponding to the maximum Q value and gradually decreasing , in order to improve the efficiency of strategy utilization.

[0028] Further, the S4 comprises the following steps:

[0029] S4.1, in the training of the simulated heating data set, multi-channel temperature-time data is generated by a blackbody simulation system and used as input of the reinforcement learning environment; during the training process, noise is added to the simulated heating data set;

[0030] The training process starts with initializing the states of all channels. The agent is based on the current state S t And the policy network selects the current action A t , and reward R according to the environment simulation feedback t and the updated state S t+1 ; Through multiple iterations, the target network is optimized to improve the performance of the balanced control strategy; the experience replay mechanism is introduced during the training process to transfer the state transfer data (S t ,A t ,R t ,S t+1 ) are stored in the experience pool and randomly sampled for training to reduce sample correlation; the target network is used to fix the Q value target to improve the stability of training;

[0031] S4.2, in the verification of the real heating data set, the multi-channel balanced control model obtained by simulation training is applied to the actual temperature control system, focusing on evaluating the temperature equalization effect, heating rate consistency and target temperature accuracy; the evaluation indicators include the mean square error of the heating rate and the absolute value deviation of the local temperature difference.

[0032] Furthermore, in S4.1, noise is added according to the following formula:

[0033]

[0034] In the formula, S total represents the total noise, C represents the proportional constant, T center represents the temperature of the central black body, T 1 -T 8 Represents the temperature of the eight black bodies surrounding the central black body.

[0035] The present invention also provides a large surface source black body multi-channel balanced heating device, comprising one or more processors, for implementing the large surface source black body multi-channel balanced heating method as described above.

[0036] The present invention also provides a readable storage medium on which a program is stored. When the program is executed by a processor, the large surface source blackbody multi-channel balanced heating method as described above is implemented.

[0037] Compared with the prior art, the invention has the following beneficial effects: the invention realizes the enhancement of the consistency of the heating rate of each channel, reduces the influence of thermal interference, and improves the overall temperature control system performance of the large surface source blackbody, providing a new research direction and application prospect for the development of temperature control technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 The present invention is a flow chart of a large surface source blackbody multi-channel balanced heating method.

[0039] Figure 2 It is a schematic diagram of the thermal interference effect analysis of a multi-channel balanced control model in a large surface source blackbody multi-channel balanced heating method of the present invention.

[0040] Figure 3 It is a schematic diagram of the difference in heating time of each channel before and after the intervention of the multi-channel balanced control model in a large surface source blackbody multi-channel balanced heating method of the present invention.

[0041] Figure 4 This is a schematic diagram of multi-channel temperature-time real data, showing ten channels.

[0042] Figure 5 It is a structural schematic diagram of a large surface source blackbody multi-channel balanced heating device of the present invention. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] like Figure 1-Figure 4 As shown, a large surface source blackbody multi-channel balanced heating method is implemented by the following steps:

[0045] S1, data acquisition: Generate a simulated heating data set of a large surface source blackbody through a blackbody simulation system, and build a conventional multi-channel actual temperature control system to obtain real heating data sets for different channels. These data sets provide basic support for the training and verification of multi-channel balanced control models. In the simulation stage, the multi-channel heating data generated by the blackbody simulation system can accurately simulate the heating process of the temperature control system. The data contains information such as temperature changes, heating rate, and duty cycle of each channel. The real heating data set comes from the multi-channel actual temperature control system, reflecting the impact of complex factors such as environmental changes and equipment performance on the heating process. The simulation data and real data work together to help the model better adapt to different control environments and improve the robustness and accuracy of the control strategy.

[0046] S2, Problem Modeling: Model the large-surface blackbody multi-channel heating balance problem as a multi-agent collaborative optimization problem, define the state, action and reward function of each channel, establish a dynamic association mechanism between channels, and ensure that the multi-channel heating rates tend to be consistent. Specifically include:

[0047] S2.1, the multi-channel heating balance problem of large surface source blackbody is modeled as a multi-agent collaborative optimization problem. Each heating channel is regarded as an independent agent, which cooperates with other channels through a dynamic association mechanism to ensure the completion of the overall balanced heating task. The state vector of each agent contains the current temperature T i , Heating rate And the current duty cycle DutyCycle i , forming the state vector S i =[T i ,R i ,DutyCycle i ].

[0048] S2.2, in this state space, the action space is defined as the adjustment of the duty cycle, allowing the agent to flexibly control the heating process.

[0049] S2.3, the reward function is designed jointly, the goal is to optimize the consistency of the global heating rate and minimize the local temperature difference; the heating rate consistency is achieved by Quantification, is the average heating rate of all channels; the temperature accuracy is measured by -|T i -T goal ∣Measurement; Among them, T i is the current temperature, T goal is the target temperature.

[0050] The final joint reward function Reward is in the form of:

[0051]

[0052] Where α and β are adjustment coefficients.

[0053] By global rate average Transmit information between channels to ensure the computational efficiency of the multi-channel equalization control model.

[0054] S3, Balanced Control Strategy Design:

[0055] During the heating process, different thermal interference effects may occur, which will affect the stability and accuracy of the temperature control system. The present invention considers three main thermal interference situations, see Figure 2 :

[0056] The first case ( Figure 2 Middle red box): The current channel is subject to heat transfer from all adjacent channels, and the thermal interference effect is the most significant. In this case, due to the heat transfer from multiple adjacent channels, the temperature of the current channel may rise too fast or too slow, thus affecting the balance and accuracy of the entire temperature control system. Therefore, it is necessary to coordinate the heating rate of each channel by dynamically adjusting the duty cycle, reduce thermal interference, and improve the stability of the temperature control system.

[0057] The second case ( Figure 2 Middle green box): Some adjacent channels generate thermal interference to the current channel, while the ambient channel effectively alleviates the conduction of heat, and the thermal interference effect is weakened. Nevertheless, local temperature may still fluctuate, which requires the multi-channel balancing model to flexibly adjust the duty cycle and balance the heating rate to reduce the impact of thermal interference on the performance of the temperature control system.

[0058] The third case ( Figure 2 Middle yellow box): The current channel is only affected by the thermal interference of a small number of heating channels and is more affected by the ambient temperature fluctuation. At this time, the thermal interference effect is the weakest. In this case, the multi-channel control model needs to accurately control the duty cycle to ensure the overall uniformity of the temperature control system and cope with the temperature rise instability caused by environmental fluctuations.

[0059] For these three thermal interference situations, a dynamic control strategy needs to be designed to adjust the duty cycle of each channel and coordinate the heating rate between channels, so as to reduce thermal interference and improve the overall temperature control performance.

[0060] The design of the balanced control strategy specifically includes:

[0061] S3.1, a reinforcement learning method based on the deep Q network (DQN) algorithm is used to design a multi-channel balancing control strategy. DQN approximates the Q value function through a neural network to select the optimal action for each agent (heating channel) in the current state. The input of each agent is its state S i =[T i ,R i ,DutyCycle i ], and the output is the Q value of each possible action. By selecting the action with the largest Q value, the agent dynamically adjusts the duty cycle to optimize the heating process of each channel.

[0062] S3.2, Q-value update follows the Bellman equation:

[0063]

[0064] Among them, S t is the current state, S t is the updated state, A tis the current action, A' represents the possible action to be selected in the next state, η is the learning rate, γ is the discount factor, and max represents the maximum value; the joint control strategy ensures that each agent makes independent decisions and at the same time uses the global rate average Affects action selection. Joint control strategy: Each independent agent in the multi-channel system (such as each heating channel) coordinates through a global mechanism (global rate average) based on independent decision-making to achieve the goal of overall balanced heating.

[0065] S3.3, balanced control strategy adopts Greedy method, with probability Randomly select actions to The probability of selecting the action corresponding to the maximum Q value and gradually decreasing , in order to improve the efficiency of strategy utilization.

[0066] S4, model training: A multi-channel balanced control model is constructed based on the reinforcement learning DQN model. The reinforcement learning DQN model is the carrier of the deep Q network (DQN) algorithm. The multi-channel balanced control model is trained using a simulated heating data set, and the synchronization of the heating rates between channels is improved through objective function optimization; then the performance of the multi-channel balanced control model is verified on a real heating data set, and the stability and robustness of the multi-channel balanced control model in a complex environment are evaluated. Specifically, it includes:

[0067] S4.1, training on the simulated heating data set, using the multi-channel temperature-time data generated by the blackbody simulation system, and using it as the input of the reinforcement learning environment. During the training process, in order to improve the robustness and fidelity of the multi-channel balanced control model, the present invention adds noise to the simulated heating data. The theory of adding noise is as follows:

[0068] In multi-channel equalization control, the noise generated by the black body is mainly related to the temperature difference between eight adjacent positions. Each black body exchanges heat with the adjacent black body through the temperature difference, which may cause the generation of thermal noise. In order to simplify the calculation, it is assumed that the noise source is only related to the temperature difference between adjacent black bodies, and the influence of the environmental module on the noise is also considered.

[0069] For the black body in the center, it is surrounded by 8 adjacent black bodies. The relationship between the noise generated by the black body and the temperature difference between these adjacent black bodies can be derived through the heat conduction model. Thermal noise (also called thermal wave motion or Johnson-Nyquist noise) is caused by temperature fluctuations inside an object and is usually related to the temperature gradient, the thermal conductivity of the object, and the interaction between the surface and adjacent objects. For a black body surrounded by multiple adjacent black bodies, its thermal noise can be expressed as the temperature difference fluctuations between it and the adjacent black bodies. Under steady-state conditions, heat conduction satisfies Fourier's law:

[0070]

[0071] Where Q is the heat flux, k is the thermal conductivity, is the heat transfer area, and A is the temperature gradient. For a black body, the heat flux depends mainly on the temperature difference between it and the adjacent black body, as well as the thermal conductivity between them.

[0072] The temperature of the central blackbody is set to T center , the temperature of the eight black bodies around it is T 1 -T 8 There is a temperature gradient between the central black body and the eight surrounding black bodies, which leads to heat conduction. According to thermal noise theory, the fluctuation of heat flow will lead to the generation of noise. It can be assumed that the noise is proportional to the temperature difference between each pair of adjacent black bodies. For each pair of adjacent black bodies, there is:

[0073]

[0074] Among them, S i Represents the temperature difference △T between the i-th adjacent black bodies i The noise level caused.

[0075] The total noise produced by the central black body can then be accumulated by adding the noise of the 8 neighboring black bodies, giving:

[0076]

[0077] Therefore, the noise is proportional to the sum of the temperature differences between the central black body and the eight adjacent black bodies. In order to more accurately describe the intensity of the noise, a proportional constant C can be introduced to express the relationship between the noise and the temperature difference. The expression is as follows:

[0078]

[0079] according to Add noise.

[0080] The training process starts with initializing the states of all channels. The agent is based on the current state S t And the policy network selects the current action A t , and reward R according to the environment simulation feedback t and the updated state S t+1 ; Through multiple iterations, the target network is optimized to improve the performance of the balanced control strategy; the experience replay mechanism is introduced during the training process to transfer the state transfer data (S t ,A t ,R t ,S t+1 ) are stored in the experience pool and randomly sampled for training to reduce sample correlation; the target network is used to fix the Q value target to improve the stability of training.

[0081] S4.2, in the verification phase of the real heating data set, the multi-channel balanced control model obtained by simulation training is used in the actual temperature control system, focusing on evaluating the temperature equalization effect, heating rate consistency and target temperature accuracy. The evaluation indicators include the mean square error (MSE) of the heating rate and the absolute value deviation of the local temperature difference. By adjusting hyperparameters such as learning rate and discount factor, the model performance is further optimized to ensure robustness under different power output conditions.

[0082] S5, Verification and Adjustment: Verification using multi-channel real data, Figure 4 The real temperature-time data of ten channels are displayed. The temperature equalization effect and the temperature difference change during the heating process are analyzed, and the deviation between the actual temperature and the target temperature is compared. The multi-channel balanced heating control strategy and model parameters are adjusted according to the verification results to optimize the accuracy of balanced heating. In this process, the impact of each thermal interference situation on the system performance is focused on, and the duty cycle is dynamically adjusted to ensure the stability and accuracy of the overall heating process.

[0083] S6, iterative optimization: Based on the verification feedback from simulation and real environment, the iterative process of modeling, training, and verification is repeatedly performed to optimize the multi-channel balanced control model so that it can efficiently and accurately achieve balanced heating of large surface source black bodies under different power outputs. Through continuous adjustment, the performance of the multi-channel control system under different power output and temperature conditions is improved to ensure efficient and accurate heating effects in various environments. The experimental results of the final adjusted multi-channel balanced control model are as follows: Figure 3 As shown, Figure 3 The time difference of each channel with the average heating time as the zero point before and after the intervention of the multi-channel balanced control model is shown. The specific quantitative evaluation results are shown in Table 1, which compares the performance of the large surface source blackbody multi-channel balanced heating method after intervention with the algorithm-free intervention. The large surface source blackbody multi-channel balanced heating method has outstanding performance in single-channel heating rate and multi-channel heating consistency. In terms of single-channel heating rate, the average steady-state heating time of this method is 3601 seconds, which is 15.3% shorter than 4251 seconds without algorithm intervention. In terms of multi-channel heating rate consistency, the standard deviation D (x) of this method is 745, which is significantly lower than 45334 without algorithm intervention, a decrease of 98.4%. In addition, the maximum steady-state heating time of the large surface source blackbody multi-channel balanced heating method is 3648 seconds, which is 25.4% higher than 4891 seconds without algorithm intervention; the minimum steady-state heating time is 3556 seconds, which is 6.2% higher than 3789 seconds without algorithm intervention.

[0084] Table 1

[0085]

[0086] See also Figure 5 The present invention provides a large surface source black body multi-channel balanced heating device, which includes one or more processors for implementing a large surface source black body multi-channel balanced heating method in the above embodiment.

[0087] The embodiment of a large surface source blackbody multi-channel balanced heating device of the present invention can be applied to any device with data processing capability, and the device with data processing capability can be a device or apparatus such as a computer. The device embodiment can be implemented through software, or through hardware or a combination of software and hardware. Taking software implementation as an example, as a device in a logical sense, it is formed by the processor of any device with data processing capability in which it is located reading the corresponding computer program instructions in the non-volatile memory into the memory for execution. From the hardware level, if Figure 5 As shown, it is a hardware structure diagram of any device with data processing capability in which a large surface source blackbody multi-channel balanced heating device of the present invention is located. Figure 5 In addition to the processor, memory, network interface, and non-volatile memory shown, any device with data processing capabilities in the embodiments may also include other hardware, usually based on the actual functions of the device with data processing capabilities, which will not be described in detail.

[0088] The implementation process of the functions and effects of each unit in the above-mentioned device is specifically described in the implementation process of the corresponding steps in the above-mentioned method, and will not be repeated here.

[0089] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0090] The present invention also provides a readable storage medium on which a program is stored. When the program is executed by a processor, a large surface source blackbody multi-channel balanced heating method in the above embodiment is implemented.

[0091] The readable storage medium may be an internal storage unit of any device with data processing capability described in any of the aforementioned embodiments, such as a hard disk or a memory. The readable storage medium may also be an external storage device, such as a plug-in hard disk, a smart media card (SMC), an SD card, a flash card, etc. equipped on the device. Furthermore, the readable storage medium may also include both an internal storage unit of any device with data processing capability and an external storage device. The readable storage medium is used to store the computer program and other programs and data required by any device with data processing capability, and may also be used to temporarily store data that has been output or is to be output.

[0092] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A large surface source blackbody multi-channel balanced heating method, characterized in that: The following steps are involved: S1, data acquisition: Generate a simulated heating data set of a large surface source blackbody through a blackbody simulation system, and build a multi-channel actual temperature control system to obtain real heating data sets of different channels; including the following steps: S2.1, the multi-channel heating balance problem of large surface source blackbody is modeled as a multi-agent collaborative optimization problem. Each heating channel is regarded as an independent agent. The agents collaborate to complete the overall balanced heating task through a dynamic association mechanism. In the state space design, the state of each agent includes the current temperature T i , Heating rate And the current duty cycle DutyCycle i , forming the state vector S i =[T i ,R i ,DutyCycle i ]; S2.2, the action space is defined as the adjustment of the duty cycle to ensure the flexibility of heating process control; S2.3, the reward function is designed jointly, the goal is to optimize the consistency of the global heating rate and minimize the local temperature difference; the heating rate consistency is achieved by Quantification, is the average heating rate of all channels; the temperature accuracy is measured by -|T i -T goal ∣Measurement; Among them, T i is the current temperature, T goal is the target temperature; The final joint reward function Reward is in the form of: ， Among them, α and β are adjustment coefficients; Average heating rate across all channels Transmit information between channels to ensure the computational efficiency of the multi-channel equalization control model; S2, Problem Modeling: Model the large-surface blackbody multi-channel heating balance problem as a multi-agent collaborative optimization problem, define the state, action and reward function of each channel, establish a dynamic association mechanism between channels, and ensure that the multi-channel heating rates tend to be consistent; S3, balanced control strategy design: adopt a multi-channel collaborative control strategy based on reinforcement learning, optimize the heating rate deviation between channels by dynamically adjusting the duty cycle of each channel; design a joint reward mechanism to balance the global temperature averaging effect and the local temperature difference minimization goal; S4, model training: construct a multi-channel balanced control model based on the reinforcement learning DQN model, and then use the simulated heating data set to train the multi-channel balanced control model; improve the synchronization of the heating rate between channels through objective function optimization; then verify the performance of the multi-channel balanced control model on the real heating data set, and evaluate the stability and robustness of the multi-channel balanced control model in complex environments; including the following steps: S4.1, in the training of the simulated heating data set, multi-channel temperature-time data is generated by a blackbody simulation system and used as input of the reinforcement learning environment; during the training process, noise is added to the simulated heating data set; Wherein, noise is added according to the following formula: ， In the formula, S total represents the total noise, C represents the proportional constant, T center represents the temperature of the central black body, T1-T8 represents the temperature of the eight black bodies around the central black body; The training process starts with initializing the states of all channels. The agent is based on the current state S t and the policy network selects the current action A t , and reward R according to the environment simulation feedback t and the updated state S t+1 ; Through multiple iterations, the target network is optimized to improve the performance of the balanced control strategy; the experience replay mechanism is introduced during the training process to transfer the state transfer data (S t ,A t ,R t ,S t+1 ) are stored in the experience pool and randomly sampled for training to reduce sample correlation; the target network is used to fix the Q value target to improve the stability of training; S4.2, in the verification of the real heating data set, the multi-channel balanced control model obtained by simulation training is applied to the actual temperature control system, focusing on evaluating the temperature equalization effect, heating rate consistency and target temperature accuracy; the evaluation indicators include the mean square error of the heating rate and the absolute value deviation of the local temperature difference; S5, verification and adjustment: Use the real heating data set of S1 to verify the multi-channel balanced control model, analyze the temperature equalization effect and the temperature difference change during the heating process, and compare the deviation between the actual and target temperatures; dynamically adjust the strategy and model parameters of the multi-channel balanced control model according to the verification results to improve the accuracy of balanced heating; S6, iterative optimization: Based on the verification feedback from simulation and real environment, the iterative process of modeling, training, and verification is repeatedly executed to optimize the multi-channel balanced control model so that it can achieve balanced heating of large surface source black bodies efficiently and accurately under different power outputs.

2. A large surface source blackbody multi-channel balanced heating method as claimed in claim 1, characterized in that: The S3 includes: S3.1, the equilibrium control strategy is based on the deep Q network algorithm, which uses a neural network to approximate the Q value function to select the optimal action; in the strategy design, the input is the current state S of each channel i =[T i ,R i ,DutyCycle i ], output is the Q value of each possible action; by selecting the action with the largest Q value, the agent dynamically adjusts the duty cycle to optimize the heating process; S3.2, Q-value update follows the Bellman equation: ， Among them, S i is the current state, S t is the updated state, A t is the current action, A' represents the possible action to be selected in the next state, η is the learning rate, γ is the discount factor, and max represents the maximum value; S3.3, balanced control strategy adopts Greedy approach.

3. A large surface source blackbody multi-channel balanced heating method as claimed in claim 2, characterized in that: S3.3 Greedy methods include: Randomly select actions to The probability of selecting the action corresponding to the maximum Q value and gradually decreasing , in order to improve the efficiency of strategy utilization.

4. A large surface source blackbody multi-channel balanced heating device, characterized in that: It comprises one or more processors for implementing a large surface source blackbody multi-channel balanced heating method according to any one of claims 1-3.

5. A readable storage medium, characterized in that: A program is stored thereon, and when the program is executed by a processor, a large surface source blackbody multi-channel balanced heating method according to any one of claims 1 to 3 is implemented.

Images