A Neural Network-Based Adaptive Cooling Control Method and System for Wall Panels

By constructing a neural network-based adaptive cooling control system for wall panels, the problems of insufficient control accuracy and response speed in the liquid film cooling process of wall panels in traditional methods are solved. This achieves high-precision and fast-response cooling control, reduces costs, and improves resource utilization efficiency.

CN123113303APending Publication Date: 2026-11-17CHINA SHIP DEV & DESIGN CENT
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202511309828.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-09-15
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Traditional PID control methods suffer from large overshoot and long settling time during the liquid film cooling process of wall panels, making it difficult to meet the requirements of high precision and fast response control, and are also greatly affected by environmental factors.

Method used

An adaptive cooling control method based on neural networks is adopted. By acquiring cooling characteristics and environmental characteristic parameters in real time, a cooling and heating neural network model is constructed and trained to dynamically optimize the spray cooling process and achieve precise control of contrast.

Benefits of technology

It improves control accuracy and response speed, reduces control costs, avoids control failure caused by sudden environmental changes in traditional methods, has strong adaptability and dynamic optimization capabilities, and saves water resources.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123113303A_ABST
    Figure CN123113303A_ABST
Patent Text Reader

Abstract

This invention discloses a neural network-based adaptive cooling control method and system for wall panels. The method includes: real-time acquisition of cooling characteristic parameters and environmental characteristic parameters of the wall panel under control during the spray cooling process; the spray cooling process of the wall panel under control includes, based on an initial contrast ratio, multiple control loops that, by executing spraying or interrupting spraying, make the real-time contrast ratio between the wall panel under control and the environment approach and continuously remain below a preset target contrast ratio; inputting the cooling characteristic parameters and environmental characteristic parameters into a pre-constructed and trained cooling neural network model and a heating neural network model to obtain the initial contrast ratio prediction value, spraying control time prediction value, and spray interruption control time prediction value for each control loop; and executing the spray cooling process of the wall panel under control. This invention can meet the control requirements of high precision and fast response in the wall panel cooling control process.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the interdisciplinary field of infrared stealth control and large model training, and in particular to a panel adaptive cooling control method and system based on neural networks. Background Technology

[0002] Liquid film cooling of wall panels is a commonly used cooling method in engineering applications. However, the liquid film cooling process involves relatively complex physical processes, making its simulation and control quite difficult. The liquid film cooling process is affected by many factors such as the environment and water spray parameters, and exhibits characteristics such as nonlinearity and time-varying nature.

[0003] To control the wall panel temperature and ensure strong consistency between its surface temperature and the background environment, a cooling water control method with high precision and fast response is needed. Many environmental factors affect the wall panel temperature, such as solar irradiance, ambient temperature, and wind speed. Traditional PID (proportional-integral-derivative) control suffers from large overshoot and long settling time, making it difficult to meet the requirements of high precision and fast response. Summary of the Invention

[0004] The main objective of this invention is to provide a wall panel adaptive cooling control method and system based on neural networks, which realizes dynamic optimization and efficiency improvement of the control process, improves control accuracy, and reduces control costs.

[0005] The technical solution adopted in this invention is: a wall panel adaptive cooling control method based on neural networks, comprising: The cooling characteristic parameters and environmental characteristic parameters of the wall panel to be controlled during the spray cooling process are acquired in real time. The pre-built and trained cooling neural network model and heating neural network model are input to obtain the initial contrast prediction value, spray control time prediction value and spray cut-off control time prediction value of the wall panel to be controlled and the environment during the spray cooling process. The spray cooling process of the wall panel to be controlled includes multiple control cycles in which the real-time contrast between the wall panel to be controlled and the environment approaches a preset target contrast under the control of the spray control time and the spray interruption control time.

[0006] According to the above technical solution, the cooling characteristic parameters include water temperature and water pressure; the environmental characteristic parameters include initial wall temperature, air temperature, wind speed, and solar irradiance.

[0007] According to the above technical solution, the spray control time includes the initial spray time and the continuous spray time; wherein, the initial spray time is used to represent the spray control time of the first control cycle executed in the spray cooling process of the wall panel to be controlled, and the continuous spray time is used to represent the spray control time in the subsequent multiple control cycles.

[0008] According to the above technical solution, both the cooling neural network model and the heating neural network model include an input layer, a hidden layer, and an output layer; The input parameters of the cooling neural network model's input layer include water pressure, water temperature, initial wall temperature, air temperature, wind speed, and solar irradiance; the output parameters of the output layer include initial spraying time, continuous spraying time, and initial contrast. The input parameters of the input layer of the heating neural network model include initial wall temperature, air temperature, wind speed, and solar irradiance; the output parameters of the output layer include the jet interruption control time.

[0009] According to the above technical solution, the method for determining the number of hidden layer nodes in the cooling neural network model and the heating neural network model includes: estimating the range of values ​​for the number of hidden layer nodes using empirical formulas; evaluating the number of hidden layer nodes within the estimated range using root mean square error and correlation coefficient R, and selecting the number of hidden layer nodes corresponding to the lowest root mean square error. The hidden layers of both the cooling and heating neural network models use the Sigmoid activation function, including the tansig function; the output layers use linear activation functions, including the purelin function.

[0010] According to the above technical solution, the training methods for the cooling neural network model and the heating neural network model include: executing the spray cooling process of multiple wall panels, recording the cooling characteristic parameters and environmental characteristic parameters of a single wall panel during the spray cooling process as sample data, and recording the initial contrast between the corresponding wall panel and the environment, the spray control time, and the spray cut-off control time as verification data. Randomly assign network weights and thresholds to the cooling neural network model and the heating neural network model; The sample data is input into the cooling neural network model and the heating neural network model to obtain the prediction results; Calculate the inverse bias based on the prediction results and validation data, and update the network weights and thresholds of the cooling neural network model and the heating neural network model; Training is complete when all sample data has been traversed or the training objective has been reached; the training objective includes meeting a preset error requirement or reaching a preset maximum number of training iterations.

[0011] According to the above technical solution, during the execution of multiple control cycles of the wall panel to be controlled, before the start of each control cycle, the cooling characteristic parameters and environmental characteristic parameters of the wall panel to be controlled during the spray cooling process are re-acquired in real time and input into the cooling neural network model and the heating neural network model to update the initial contrast between the wall panel to be controlled and the environment, the spray control time and the spray interruption control time of the new control cycle.

[0012] Another aspect of the present invention provides a wall panel adaptive cooling control system based on a neural network, comprising a feature parameter acquisition module, a spray cooling process execution module, a wall panel spray cooling model construction module, and a wall panel spray cooling prediction module; wherein, The feature parameter acquisition module is used to acquire the cooling feature parameters and environmental feature parameters of the wall panel to be controlled in real time during the spray cooling process. The spray cooling process execution module is used to execute multiple control cycles to make the real-time contrast between the wall panel to be controlled and the environment approach the preset target contrast under the control of the spray control time and the spray cut-off control time. The wall panel spray cooling model building module is used to build and train the completed cooling neural network model and heating neural network model; The wall panel spray cooling prediction module is used to input the cooling characteristic parameters and environmental characteristic parameters of the wall panel to be controlled during the spray cooling process into the pre-built and trained cooling neural network model and heating neural network model to obtain the initial contrast prediction value between the wall panel to be controlled and the environment, the spray control time prediction value, and the spray cut-off control time prediction value during the spray cooling process of the wall panel to be controlled.

[0013] Another aspect of the present invention provides a computer storage medium storing a computer program executable by a processor, the computer program performing the above-described neural network-based adaptive cooling control method for wall panels.

[0014] Another aspect of the present invention provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described neural network-based adaptive cooling control method for wall panels.

[0015] The beneficial effects of this invention are as follows: This invention provides a neural network-based adaptive cooling control method and system for wall panels. It determines key parameters and contrast through feature selection, constructs a dynamic control flow, and builds and trains cooling and heating neural network models for different operating conditions, ultimately achieving the prediction of cooling parameters for the wall panel to be controlled. This invention dynamically adjusts the spray strategy through neural networks to match real-time environmental changes, avoiding control failures caused by sudden environmental changes in traditional PID control, and possesses strong adaptability and dynamic optimization capabilities. This invention designs the network structure according to different operating conditions, automatically generating spray control curves, replacing the complex process of traditional experience-based debugging. By precisely controlling the spray time and water volume, it avoids water waste caused by excessive spraying, significantly improving control accuracy and operating efficiency, and achieving resource conservation and cost reduction.

[0016] Furthermore, this invention selects the number of hidden layer nodes in the neural network model by minimizing the root mean square error, and combines RMSE and correlation coefficient R to evaluate accuracy, thereby ensuring high precision and strong generalization ability of the model training.

[0017] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart of the wall panel adaptive cooling control method based on neural network according to an embodiment of the present invention; Figure 2 This is the initial control curve diagram in the wall panel adaptive cooling control method based on neural network according to an embodiment of the present invention; Figure 3 This is a structural diagram of the cooling neural network model in the wall panel adaptive cooling control method based on neural network according to an embodiment of the present invention; Figure 4 This is a structural diagram of the heating neural network model in the wall panel adaptive cooling control method based on neural network according to an embodiment of the present invention; Figure 5 This is a flowchart of the training process of the neural network model in the wall panel adaptive cooling control method based on neural network according to an embodiment of the present invention; Figure 6(a) shows water pressure test data for different groups in a specific application of the wall panel adaptive cooling control method based on neural network according to an embodiment of the present invention. Figure 6(b) shows the water temperature experimental data of different groups in a specific application of the wall panel adaptive cooling control method based on neural network according to the embodiment of the present invention; Figure 6(c) shows experimental data of wall surface temperature for different groups in a specific application of the wall panel adaptive cooling control method based on neural network according to an embodiment of the present invention. Figure 6(d) shows the temperature experimental data of different groups in a specific application of the wall panel adaptive cooling control method based on neural network according to the embodiment of the present invention; Figure 6(e) shows the wind speed experimental data of different groups in a specific application of the wall panel adaptive cooling control method based on neural network according to the embodiment of the present invention; Figure 6(f) shows the solar radiation experimental data of different groups in a specific application of the wall panel adaptive cooling control method based on neural network according to the embodiment of the present invention; Figure 7 This is a comparison of RMSE when the wall panel cooling process is set to different numbers of hidden layer nodes in a specific application of the wall panel adaptive cooling control method based on neural networks in the embodiments of the present invention. Figure 8 This is a training regression diagram of the optimal number of hidden layer nodes when different data are used in the wall panel adaptive cooling control method based on neural networks in a specific application of the wall panel cooling process of the present invention. Figure 9 This is a specific application of the wall panel adaptive cooling control method based on neural networks in the embodiments of the present invention, where the wall panel heating process is set to different numbers of hidden layer nodes for RMSE comparison. Figure 10 This is a training regression diagram of the optimal number of hidden layer nodes when different data are used in the wall panel heating process in a specific application of the wall panel adaptive cooling control method based on neural networks in the embodiments of the present invention. Figure 11 This is a specific application of the wall panel adaptive cooling control method based on neural network in the present invention, where the target contrast f1 = 3% is set as the predicted and experimental values ​​of the control curve. Figure 12 The control curve prediction and experimental values ​​are set when the target contrast f1 = 6% in a specific application of the wall panel adaptive cooling control method based on neural network in this embodiment of the invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0021] It should be noted that the illustrations provided in the embodiments of the present invention are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the number, shape and size of the components in actual implementation. In actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0022] In this invention, it should also be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application. Furthermore, the terms "first" and "second" are used only for descriptive and distinguishing purposes and should not be construed as indicating or implying relative importance.

[0023] Example 1 This embodiment provides a wall panel adaptive cooling control method based on neural networks, the process of which is as follows: Figure 1 As shown, the steps include: S1. Real-time acquisition of cooling characteristic parameters and environmental characteristic parameters of the wall panel to be controlled during the spray cooling process. The spray cooling process of the wall panel to be controlled includes multiple control cycles, based on the initial contrast, to make the real-time contrast between the wall panel to be controlled and the environment approach and remain below the preset target contrast by executing spraying or spray interruption.

[0024] By manually selecting features, feature parameters that have a significant impact on the cooling process of the wall panel are extracted.

[0025] Specifically, the characteristic parameters in the wall panel spray cooling process include water temperature T. w Water pressure P.

[0026] Environmental characteristic parameters include initial wall temperature T0 and air temperature T. a Wind speed v, solar irradiance I.

[0027] Contrast ratio f characterizes the relative temperature difference between the wall panel and the background environment, and is defined as: contrast ratio f = (wall panel radiation intensity - ambient radiation intensity) / ambient radiation intensity.

[0028] Specifically, the spray cooling process for the wall panel is as follows: based on the preset target contrast f1 between the wall panel and the environment, combined with the solar irradiance I and water temperature T of the environment where the wall panel is located... w Temperature T a The wall panel temperature field is quickly estimated using parameters such as wind speed v, and the centrifugal pump and other components of the water supply system are activated to start the wall panel liquid film cooling system.

[0029] The initial control curve generated during the spray cooling process of a certain wall panel is as follows: Figure 2As shown, the target contrast ratio f1 is set as the control target. Based on the initial contrast ratio f2 of the wall panel, the spraying is turned on / off according to the initial spraying time t0, spraying control time t1, and continuous spraying time t2 in the control curve, so that the real-time contrast ratio f of the wall panel approaches and remains lower than the target contrast ratio f1 in each control cycle.

[0030] Specifically, before each control cycle begins, the control system regenerates the control parameters spray control time t1 and continuous spray time t2 based on the changes in environmental characteristic parameters.

[0031] S2. Input the cooling characteristic parameters and environmental characteristic parameters collected in step S1 into the pre-built and trained cooling neural network model and heating neural network model to obtain the initial contrast prediction value, spray control time prediction value and spray interruption control time prediction value of each control cycle.

[0032] Furthermore, the spray control time includes the initial spray time and the continuous spray time. The initial spray time represents the spray control time of the first control cycle executed during the spray cooling process of the wall panel to be controlled, while the continuous spray time represents the spray control time in subsequent control cycles.

[0033] Furthermore, both the cooling neural network model and the heating neural network model include an input layer, a hidden layer, and an output layer.

[0034] For the wall panel cooling process, the main factors affecting the initial spraying time t0, continuous spraying time t2, and initial contrast ratio f2 are water pressure P and water temperature T. w Initial wall temperature T0, air temperature T a Using six parameters—wind speed v and solar irradiance I—as input values, and three parameters—initial spraying time t0, continuous spraying time t2, and initial contrast ratio f2—as output values, a cooling neural network model is constructed as follows: Figure 3 As shown.

[0035] For the wall panel heating process, the main factors affecting the spray control time t1 are wall temperature T0 and air temperature T. a Using wind speed v and solar irradiance I as input values ​​and the jet cessation control time t1 as the output value, a heating neural network model is constructed as follows: Figure 4 As shown.

[0036] Specifically, when determining the number of hidden layer nodes in the cooling neural network model and the heating neural network model, an empirical formula is used. Estimate.

[0037] Where m is the number of input layer nodes, n is the number of output layer nodes, and a is a preset adjustment constant with a value range of [1, 10].

[0038] The model training accuracy and performance are evaluated by the root mean square error and the correlation coefficient R based on the estimation results, and the number of hidden layer nodes corresponding to the lowest root mean square error is selected.

[0039] Preferably, to ensure fast convergence and low training error of the neural network model, the input sample values ​​and the output training results are matched as closely as possible. The hidden layers of both the cooling and heating neural network models use the Sigmoid activation function; in this embodiment, the tansig function is selected. The output layer uses a linear activation function; in this embodiment, the purelin function is selected.

[0040] Furthermore, the training methods and procedures for the cooling neural network model and the heating neural network model are as follows: Figure 5 As shown, the process includes the following steps: performing a spray cooling process on multiple wall panels, recording the cooling characteristic parameters and environmental characteristic parameters of a single wall panel as sample data, and recording the initial contrast between the wall panel and the environment, the spray control time, and the spray interruption control time as verification data.

[0041] Randomly assign network weights and thresholds to the cooling neural network model and the heating neural network model.

[0042] The sample data is input into the cooling neural network model and the heating neural network model to obtain the prediction results.

[0043] The reverse bias is calculated based on the prediction results and validation data, and the network weights and thresholds of the cooling neural network model and the heating neural network model are updated.

[0044] Training is complete when all sample data has been traversed or the training objective has been reached. The training objective includes convergence of training results, meeting preset error requirements, or reaching a preset maximum number of training iterations.

[0045] S3. Execute the spray cooling process of the wall panel to be controlled based on the initial contrast prediction value, spray control time prediction value, and spray interruption control time prediction value of each control cycle.

[0046] This application also provides a neural network-based adaptive cooling control system for wall panels, including a feature parameter acquisition module, a spray cooling process execution module, a wall panel spray cooling model construction module, and a wall panel spray cooling prediction module. Among them, The feature parameter acquisition module is used to acquire cooling feature parameters and environmental feature parameters of the wall panel to be controlled in real time during the spray cooling process.

[0047] The spray cooling process execution module is used to execute the spray cooling process of the wall panel to be controlled based on the initial contrast prediction value, spray control time prediction value, and spray cut-off control time prediction value of each control cycle.

[0048] The wall panel spray cooling model building module is used to build and train the completed cooling neural network model and heating neural network model.

[0049] The wall panel spray cooling prediction module is used to input the cooling characteristic parameters and environmental characteristic parameters into a pre-built and trained cooling neural network model and heating neural network model to obtain the initial contrast prediction value, spray control time prediction value and spray interruption control time prediction value for each control cycle.

[0050] This application also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described neural network-based adaptive cooling control method for wall panels.

[0051] This application also provides a computer-readable storage medium, such as flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, server, app store, etc., which stores a computer program, and the program implements corresponding functions when executed by a processor. The computer-readable storage medium of this embodiment implements a neural network-based adaptive cooling control method for wall panels when executed by a processor.

[0052] Example 2 This embodiment provides a specific application of the neural network-based adaptive cooling control method for wall panels described in Embodiment 1.

[0053] This example uses 3680 sets of wall panel cooling process data calculated by interpolation of experimental data and 11 sets of experimental data of wall panel cooling process as data sources. 70% of the data is selected as training sample data for the established cooling neural network model and heating neural network model, 15% of the data is used for verification, and the remaining 15% of the data is used for testing.

[0054] In this embodiment, the wall panel uses a rectangular steel plate with dimensions of 8m × 5m and a thickness of 10mm. Water pressure P and water temperature T are selected. w Wall temperature T0, air temperature T aSix parameters, including wind speed v and solar irradiance I, were used as model input parameters, and 11 sets of experimental data were obtained, as shown in Figures 6(a), 6(b), 6(c), 6(d), 6(e), and 6(f).

[0055] For the cooling process of the wall panel, based on empirical formulas To estimate the number of hidden layer nodes l, where m is the number of nodes in the input layer, n is the number of nodes in the output layer, and a is an adjustment constant with a value range of [1, 10].

[0056] Based on the number of parameters in the cooling process of the wall panel, m=6 and n=3. Therefore, the network model has the best accuracy when l is in the interval [4,13]. The specific parameter settings of the cooling neural network model are shown in Table 1.

[0057]

[0058] Table 1. Parameter Settings for Cooling Neural Network Model Based on the number of parameters in the heating process of the wall panel, m=4 and n=1. Therefore, the network model has the best accuracy when l is in the interval [3,12]. The specific parameter settings of the heating neural network model are shown in Table 2.

[0059]

[0060] Table 2 Parameter Settings for Heating Neural Network Model Correlation analysis was conducted on cooling neural network models and heating neural network models, and the accuracy and performance of their training models were evaluated by the root mean square error (RMSE) and the correlation coefficient R.

[0061] The relationship between the root mean square error (RMSE) of different numbers of hidden layer nodes during the cooling process of the wall panel is as follows: Figure 7 As shown, when the number of nodes in the hidden layer is 12, the lowest RMSE value is 0.00581, which indicates that the deviation between the predicted value and the actual value is the smallest. At this time, the cooling neural network model is the most accurate.

[0062] The optimal number of hidden layer nodes during the cooling process of the wall panel corresponds to the training regression as follows: Figure 8 As shown, when the number of nodes in the hidden layer is 12, Figure 8 In the four sets of fitting experiment results using training data (70%), validation data (15%), test data (15%), and all data (100%), the solid lines are all straight lines with a slope close to 1, showing a high correlation between the predicted data and the training data.

[0063] The relationship between the root mean square error (RMSE) of different numbers of hidden layer nodes during the cooling process of the wall panel is as follows: Figure 9As shown, when the number of nodes in the hidden layer is 11, the lowest RMSE value is 0.0023, which indicates that the deviation between the predicted value and the actual value is the smallest. At this time, the heated neural network model is the most accurate.

[0064] The optimal number of hidden layer nodes during the cooling process of the wall panel corresponds to the training regression as follows: Figure 10 As shown, when the number of nodes in the hidden layer is 11, Figure 10 In the fitting experiment results of the four groups using training data (70%), validation data (15%), test data (15%), and all data (100%) respectively, the solid lines are all straight lines with a slope close to 1, showing the high correlation between the predicted data and the training data.

[0065] By using the trained cooling neural network model and heating neural network model to predict the control curve, and by inputting the characteristic parameters of the cooling process of the wall panel to be controlled and the characteristic parameters of the environment in which the wall panel to be controlled is located, the predicted values ​​of the spray interruption control time and spray control time of multiple control cycles of the wall panel to be controlled can be obtained.

[0066] In this embodiment, the target contrast is set to 3% and 6% respectively, and prediction is performed based on the neural network model that has been trained.

[0067] When the target contrast is selected as f1=3%, the selected test conditions are shown in Table 3:

[0068] Table 3. Test parameters when target contrast f1 = 3%. When the target contrast ratio f1 = 3%, the predicted values ​​of the control curves obtained by the cooling neural network model and the heating neural network model are as follows: Figure 11 As shown, the experimental and predicted values ​​generally show a high degree of consistency. The initial contrast ratio drops rapidly, then fluctuates around the target contrast ratio of 3%. Ultimately, both approaches 3%, indicating that the cooling control strategy effectively guides the wall panel contrast ratio towards the set target. The small fluctuations in the middle stage reflect the dynamic response of the control system to environmental disturbances or differences in the thermal characteristics of the wall panel, demonstrating adaptive closed-loop adjustment.

[0069] When the target contrast is selected as f1=6%, the selected test conditions are shown in Table 4:

[0070] Table 4. Test parameters when target contrast f1 = 6%. When the target contrast f1 = 6%, the predicted and experimental values ​​of the control curve are as follows: Figure 12As shown, the experimental and predicted value curves highly overlap, indicating that the control model accurately predicts the dynamic changes in the wall panel contrast, and the theoretical strategy matches the actual implementation. Ultimately, both approaches the target contrast of 6%, verifying that the cooling control can guide the wall panel contrast to converge to the preset standard, demonstrating the effectiveness of the system's core function. Fluctuations within the 20-40 time period indicate the response of this prediction process to environmental disturbances or dynamic adjustments to control parameters, reflecting the characteristic of adaptive closed-loop control in continuously correcting deviations.

[0071] from Figure 11 , Figure 12 As can be seen, the predicted value of the initial contrast is relatively close to the experimental measurement. During the cooling and heating processes, the control curves and experimental curves show consistent trends, with only slight differences in control time. The error is generally stable within 1-3 minutes. The main reasons for the error are insufficient experimental samples and rapid changes in environmental conditions during the experiment, which increase the uncertainty of the experimental measurements. The error can be further reduced by increasing the number of experimental samples to train the model and achieve more precise control of the wall panel cooling system.

[0072] In summary, this invention provides a neural network-based adaptive cooling control method and system for wall panels, which can meet the control requirements of high precision and fast response in the wall panel cooling control process.

[0073] It should be noted that, depending on the implementation needs, the various steps / components described in this application can be broken down into more steps / components, or two or more steps / components or parts of the operation of steps / components can be combined into new steps / components to achieve the purpose of this invention.

[0074] The order of the steps in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0075] It should be understood that those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A wall panel adaptive cooling control method based on neural networks, characterized in that, include: Real-time acquisition of cooling characteristic parameters and environmental characteristic parameters of the wall panel under control during the spray cooling process; The spray cooling process of the wall panel to be controlled includes multiple control cycles, based on the initial contrast, to make the real-time contrast between the wall panel to be controlled and the environment approach and remain below the preset target contrast by performing spraying or spraying interruption. The cooling characteristic parameters and environmental characteristic parameters are input into the pre-built and trained cooling neural network model and heating neural network model to obtain the initial contrast prediction value, spray control time prediction value and spray interruption control time prediction value of each control cycle. The spray cooling process of the wall panel to be controlled is executed based on the initial contrast prediction value, spray control time prediction value, and spray cut-off control time prediction value of each control cycle.

2. The wall panel adaptive cooling control method based on neural network according to claim 1, characterized in that, The cooling characteristic parameters include water temperature and water pressure; the environmental characteristic parameters include initial wall temperature, air temperature, wind speed, and solar irradiance.

3. The wall panel adaptive cooling control method based on neural network according to claim 1, characterized in that, The spray control time includes the initial spray time and the continuous spray time; wherein, the initial spray time is used to represent the spray control time of the first control cycle executed during the spray cooling process of the wall panel to be controlled, and the continuous spray time is used to represent the spray control time in the subsequent multiple control cycles.

4. The wall panel adaptive cooling control method based on neural network according to claim 3, characterized in that, Both the cooling neural network model and the heating neural network model include an input layer, a hidden layer, and an output layer; The input parameters of the cooling neural network model's input layer include water pressure, water temperature, initial wall temperature, air temperature, wind speed, and solar irradiance; the output parameters of the output layer include initial spraying time, continuous spraying time, and initial contrast. The input parameters of the input layer of the heating neural network model include initial wall temperature, air temperature, wind speed, and solar irradiance; the output parameters of the output layer include the jet interruption control time.

5. The wall panel adaptive cooling control method based on neural network according to claim 4, characterized in that, The methods for determining the number of hidden layer nodes in cooling neural network models and heating neural network models include: estimating the range of values ​​for the number of hidden layer nodes using empirical formulas; evaluating the number of hidden layer nodes within the estimated range using root mean square error and correlation coefficient R, and selecting the number of hidden layer nodes corresponding to the lowest root mean square error. The hidden layers of both the cooling and heating neural network models use the Sigmoid activation function, including the tansig function; the output layers use linear activation functions, including the purelin function.

6. The wall panel adaptive cooling control method based on neural network according to claim 3, characterized in that, The training methods for the cooling neural network model and the heating neural network model include: executing the spray cooling process of multiple wall panels, recording the cooling characteristic parameters and environmental characteristic parameters of a single wall panel during the spray cooling process as sample data, and recording the initial contrast between the corresponding wall panel and the environment, the spray control time and the spray cut-off control time as validation data. Randomly assign network weights and thresholds to the cooling neural network model and the heating neural network model; The sample data is input into the cooling neural network model and the heating neural network model to obtain the prediction results; Calculate the inverse bias based on the prediction results and validation data, and update the network weights and thresholds of the cooling neural network model and the heating neural network model; Training is complete when all sample data has been traversed or the training objective has been reached; the training objective includes meeting a preset error requirement or reaching a preset maximum number of training iterations.

7. The wall panel adaptive cooling control method based on neural network according to claim 3, characterized in that, During the execution of multiple control cycles of the wall panel to be controlled, before the start of each control cycle, the cooling characteristic parameters and environmental characteristic parameters of the wall panel to be controlled during the spray cooling process are re-acquired in real time and input into the cooling neural network model and the heating neural network model to update the initial contrast between the wall panel to be controlled and the environment, the spray control time and the spray interruption control time of the new control cycle.

8. A wall panel adaptive cooling control system based on neural networks, characterized in that, It includes a feature parameter acquisition module, a spray cooling process execution module, a wall panel spray cooling model construction module, and a wall panel spray cooling prediction module; among which, The feature parameter acquisition module is used to acquire the cooling feature parameters and environmental feature parameters of the wall panel to be controlled in real time during the spray cooling process. The spray cooling process execution module is used to execute the spray cooling process of the wall panel to be controlled based on the initial contrast prediction value, spray control time prediction value, and spray cut-off control time prediction value of each control cycle. The wall panel spray cooling model building module is used to build and train the completed cooling neural network model and heating neural network model; The wall panel spray cooling prediction module is used to input the cooling characteristic parameters and environmental characteristic parameters into a pre-built and trained cooling neural network model and heating neural network model to obtain the initial contrast prediction value, spray control time prediction value and spray interruption control time prediction value for each control cycle.

9. A computer storage medium, characterized in that, It contains a computer program that can be executed by a processor, which performs the neural network-based adaptive cooling control method for wall panels as described in any one of claims 1-7.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the neural network-based adaptive cooling control method for wall panels according to any one of claims 1-7.