Neural Network Control Method for Heat Exchange Module

Abstract

This application relates to the field of heat dissipation technology and provides a neural network control method for a heat exchange module. The method includes obtaining information on the heat exchange performance parameters of the heat exchange module in real time; wherein, the information on the heat exchange performance parameters includes the heat exchange performance parameter values of multiple heat exchange performance parameters; performing combination processing on each heat exchange performance parameter in the information on the heat exchange performance parameters based on a heat exchange performance parameter combination algorithm to obtain multiple heat exchange performance parameter combinations; for each of the heat exchange performance parameter combinations, inputting the heat exchange performance parameter combination into a corresponding target control parameter prediction model to obtain a corresponding control parameter; and controlling the operation of the heat exchange module based on each of the control parameters. With this method, by constructing an independent target control parameter prediction model for each control parameter, the mutual interference and error transmission of the control parameters during the prediction process are reduced, ensuring the accuracy and reliability of the control parameters.

Description

Technical Field

[0001] This application relates to the field of heat dissipation technology, and particularly to a neural network control method for a heat exchange module. Background Art

[0002] In modern industrial and energy fields, as an important energy conversion and transfer device, the heat exchange module is widely used in industries such as chemical engineering, petrochemical, electric power, refrigeration, heating, ventilation, and air conditioning. The performance of the heat exchange module directly affects the energy efficiency, stability, and operating cost of the system. Therefore, effective control of the heat exchange module is of great significance.

[0003] With the development of artificial intelligence and machine learning technologies, neural networks have been gradually introduced into the control of heat exchange modules due to their powerful non-linear mapping ability and self-learning characteristics. Neural networks can establish complex relationships between input variables and output control parameters by learning a large amount of historical data, thereby realizing intelligent control of the system.

[0004] However, in the prior art, a single neural network model is usually used to predict multiple control parameters. This method simplifies the design and implementation of the model to a certain extent, but also brings new problems.

[0005] First, there may be couplings and correlations between the control parameters of the heat exchange module, but they also have their own unique influencing factors and dynamic characteristics. Using a single neural network model to predict multiple control parameters simultaneously easily leads to an increase in model complexity and training difficulty, and it is impossible to fully capture the characteristics of each control parameter.

[0006] Secondly, the prediction of multiple control parameters by a single model may produce interference between parameters and an error amplification effect, resulting in inaccurate prediction results. This may cause instability of the system in actual control, reduce the heat exchange efficiency, and even pose potential safety hazards. Summary of the Invention

[0007] This application provides a neural network control method for a heat exchange module to solve the problems raised in the above background art.

[0008] In a first aspect, this application provides a neural network control method for a heat exchange module, including:

[0009] Obtain a training data set, and respectively determine the key heat exchange performance parameter information corresponding to each control parameter of the heat exchange module based on the training data set, and construct the target training data set corresponding to the control parameter based on the key heat exchange performance parameter information and the training data set;

[0010] Train a preset neural network respectively based on the target training datasets corresponding to the respective control parameters to obtain the target control parameter prediction models corresponding to the respective control parameters;

[0011] Generate a heat transfer performance parameter combination algorithm based on the key heat transfer performance parameter information corresponding to the respective control parameters;

[0012] Obtain the heat transfer performance parameter information of the heat exchange module in real time; wherein, the heat transfer performance parameter information includes the heat transfer performance parameter values of multiple heat transfer performance parameters;

[0013] Perform combination processing on the respective heat transfer performance parameters in the heat transfer performance parameter information based on the heat transfer performance parameter combination algorithm to obtain multiple heat transfer performance parameter combinations;

[0014] For each of the heat transfer performance parameter combinations, input the heat transfer performance parameter combination into the corresponding target control parameter prediction model to obtain the corresponding control parameter;

[0015] Control the operation of the heat exchange module based on the respective control parameters.

[0016] In a possible implementation manner, the training dataset includes multiple mapping relationships, the mapping relationship is the mapping relationship between the heat transfer performance parameter information and the control parameter information, the heat transfer performance parameter information includes the heat transfer performance parameter values of multiple heat transfer performance parameters, the control parameter information includes the control parameter values of multiple control parameters, and the determining the key heat transfer performance parameter information corresponding to each control parameter of the heat exchange module based on the training dataset includes:

[0017] For each of the control parameters, extract the control parameter value corresponding to the control parameter in the control parameter information of the training dataset to obtain the control parameter value sequence corresponding to the control parameter, and generate the variance of the control parameter based on the control parameter value sequence;

[0018] For each of the heat transfer performance parameters, extract the heat transfer performance parameter value corresponding to the heat transfer performance parameter in the heat transfer performance parameter information of the training dataset to obtain the heat transfer performance parameter value sequence corresponding to the heat transfer performance parameter, and generate the variance of the heat transfer performance parameter based on the heat transfer performance parameter value sequence;

[0019] For each of the control parameters, determine the key heat transfer performance parameter information corresponding to the control parameter based on the variance corresponding to the control parameter and the variances corresponding to the respective heat transfer performance parameters;

[0020] Wherein, the determining the key heat transfer performance parameter information corresponding to the control parameter based on the variance corresponding to the control parameter and the variances corresponding to the respective heat transfer performance parameters includes:

[0021] For each of the heat transfer performance parameters, the ratio of the variance of the control parameter to the variance of the heat transfer performance parameter is used as the correlation index between the heat transfer performance parameter and the control parameter, and the correlation index is compared with a preset correlation index. If the correlation index is greater than the preset correlation index, it is determined that the heat transfer performance parameter is the key heat transfer performance parameter corresponding to the control parameter; each of the key heat transfer performance parameters constitutes the key heat transfer performance parameter information corresponding to the control parameter.

[0022] In a possible implementation, the training of the preset neural network based on the target training data sets corresponding to the respective control parameters to obtain the target control parameter prediction models corresponding to the respective control parameters includes:

[0023] For each of the control parameters, the target training data set corresponding to the control parameter is divided into a training set, a validation set, and a calibration set;

[0024] The neural network is trained based on the training set to obtain an initial control parameter prediction model;

[0025] Based on a preset genetic optimization algorithm and the validation set, the model parameters of the initial control parameter prediction model are optimized to obtain an intermediate control parameter prediction model corresponding to the control parameter;

[0026] Based on the calibration set, the model parameters of the intermediate control parameter prediction model are calibrated to obtain the target control parameter prediction models corresponding to the respective control parameters.

[0027] In a possible implementation, the optimization of the model parameters of the initial control parameter prediction model based on a preset genetic optimization algorithm and the validation set to obtain an intermediate control parameter prediction model corresponding to the control parameter includes:

[0028] For each initial model parameter value of the initial control parameter prediction model, a first threshold interval corresponding to the initial model parameter value is generated based on a preset first threshold interval length, and a preset number of first parameter values are extracted within the first threshold interval based on a preset first step size to obtain a first parameter set corresponding to the initial model parameter value; wherein, the initial model parameter value is located in the middle of the first threshold interval;

[0029] The Cartesian product operation is performed on each of the first parameter sets to obtain a first-generation population; the first-generation population includes multiple first-generation individuals;

[0030] Based on the validation set, the fitness corresponding to each of the first-generation individuals is generated, and the first-generation individual with the maximum fitness is determined as the first-generation target individual;

[0031] For each first parameter value of the first-generation target individuals, a second threshold interval corresponding to the first parameter value is generated based on a preset second threshold interval length, and a preset number of second parameter values are extracted within the second threshold interval based on a preset second step size to obtain a second parameter set corresponding to the first parameter value; wherein, the first parameter value is located in the middle of the second threshold interval; the second threshold interval length is one-half of the first threshold interval length, and the second step size is one-half of the first step size;

[0032] Perform a Cartesian product operation on each of the second parameter sets to obtain a second-generation population; the second-generation population includes multiple second-generation individuals; iterate the steps after performing the Cartesian product operation on each of the first parameter sets to obtain the first-generation population until a final target individual is obtained, and the fitness corresponding to the final target individual converges;

[0033] Update the model parameters of the initial control parameter prediction model based on the final target individual to obtain the intermediate control parameter prediction model.

[0034] In a possible implementation manner, the generating the fitness corresponding to the first-generation individuals based on the validation set includes:

[0035] For each of the first-generation individuals, update the model parameters of the initial control parameter prediction model based on the first-generation individual to obtain an updated control parameter prediction model corresponding to the first-generation individual;

[0036] Obtain the prediction accuracy rate and the mean absolute error of the updated control parameter prediction model based on the validation set, and determine the product of the prediction accuracy rate and the reciprocal of the mean absolute error as the fitness of the first-generation individual.

[0037] In a possible implementation manner, the correcting the model parameters of the intermediate control parameter prediction model based on the calibration set to obtain a target control parameter prediction model corresponding to each of the control parameters includes:

[0038] For each mapping relationship in the calibration set, input the key thermal performance parameter information of the mapping relationship into the intermediate control parameter prediction model to obtain a control parameter prediction value corresponding to the mapping relationship, calculate a prediction loss value between the control parameter prediction value and the actual control parameter value corresponding to the mapping relationship based on a preset loss function, and perform backpropagation of the prediction loss value in the intermediate control parameter prediction model to obtain a gradient corresponding to the prediction loss value, and optimize the model parameters of the intermediate control parameter prediction model based on the gradient to obtain the target control parameter prediction model.

[0039] The present application provides a neural network control method for a heat exchange module. The method includes: obtaining a training data set, and respectively determining, based on the training data set, the key heat exchange performance parameter information corresponding to each control parameter of the heat exchange module, and constructing, based on the key heat exchange performance parameter information and the training data set, a target training data set corresponding to the control parameter; training a preset neural network respectively based on the target training data sets corresponding to each control parameter to obtain a target control parameter prediction model corresponding to each control parameter; generating a heat exchange performance parameter combination algorithm based on the key heat exchange performance parameter information corresponding to each control parameter; obtaining in real time the heat exchange performance parameter information of the heat exchange module; where the heat exchange performance parameter information includes the heat exchange performance parameter values of a plurality of heat exchange performance parameters; performing a combination process on each heat exchange performance parameter in the heat exchange performance parameter information based on the heat exchange performance parameter combination algorithm to obtain a plurality of heat exchange performance parameter combinations; for each of the heat exchange performance parameter combinations, inputting the heat exchange performance parameter combination into the corresponding target control parameter prediction model to obtain the corresponding control parameter; and controlling the operation of the heat exchange module based on each of the control parameters. On the one hand, by constructing an independent target control parameter prediction model for each control parameter, each model can focus on learning and capturing the unique influencing factors and dynamic characteristics of the control parameter, thereby significantly improving the prediction accuracy, reducing the mutual interference and error transmission of the control parameters in the prediction process, and ensuring the accuracy and reliability of the control parameters. On the other hand, by separately constructing a target control parameter prediction model for each control parameter, the complexity of each model is effectively reduced, the model structure is made more concise, and the training process is more efficient, which helps to improve the response speed and real-time performance of the overall control system. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0041] Figure 1 It is a schematic flow chart of the neural network control method for the heat exchange module provided by the embodiment of the present application;

[0042] Figure 2 It is a schematic block diagram of the structure of the neural network control system for the heat exchange module provided by the embodiment of the present application;

[0043] Figure 3 It is a schematic block diagram of the structure of the terminal device provided by the embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0044] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0045] The flowchart shown in the accompanying drawings is only an example illustration, and does not necessarily include all the contents and operations / steps, nor does it necessarily need to be executed in the described order. For example, some operations / steps can also be decomposed, combined, or partially merged, so the actual execution order may be changed according to the actual situation.

[0046] It should also be understood that the terms used in the specification of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in the specification of the present application and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an", and "the" are intended to include the plural forms.

[0047] It should be further understood that the term "and / or" used in the specification of the present application and the appended claims refers to any combination and all possible combinations of one or more of the related listed items, and includes these combinations.

[0048] The following will describe in detail some embodiments of the present application with reference to the accompanying drawings. Without conflict, the features in the following embodiments and the embodiments can be combined with each other.

[0049] Please refer to Figure 1 , Figure 1 which is a schematic flowchart of the neural network control method for the heat exchange module provided by the embodiments of the present application. As Figure 1 shown, the neural network control method for the heat exchange module provided by the embodiments of the present application includes steps S1 to S5.

[0050] Step S1: Obtain a training data set, and respectively determine the key heat exchange performance parameter information corresponding to each control parameter of the heat exchange module based on the training data set, and construct the target training data set corresponding to the control parameter based on the key heat exchange performance parameter information and the training data set.

[0051] Specifically, the training data set includes multiple mapping relationships, which are the mapping relationships between heat transfer performance parameter information and control parameter information. The heat transfer performance parameter information includes heat transfer performance parameter values of multiple heat transfer performance parameters, and the control parameter information includes control parameter values of multiple control parameters. The training data set is obtained through the dynamic thermodynamic model corresponding to the heat exchange module. The heat transfer performance parameter information includes, but is not limited to, the coolant inlet temperature, coolant outlet temperature, ambient temperature, inner wall temperature, and outer wall temperature. The control parameter information includes, but is not limited to, the coolant flow rate, valve opening, and pressure.

[0052] Among them, determining the key heat transfer performance parameter information corresponding to each control parameter of the heat exchange module based on the training data set includes the following steps:

[0053] For each of the control parameters, extract the control parameter value corresponding to the control parameter from the control parameter information in the training data set to obtain a control parameter value sequence corresponding to the control parameter, and generate the variance of the control parameter based on the control parameter value sequence;

[0054] For each of the heat transfer performance parameters, extract the heat transfer performance parameter value corresponding to the heat transfer performance parameter from the heat transfer performance parameter information in the training data set to obtain a heat transfer performance parameter value sequence corresponding to the heat transfer performance parameter, and generate the variance of the heat transfer performance parameter based on the heat transfer performance parameter value sequence;

[0055] For each of the control parameters, determine the key heat transfer performance parameter information corresponding to the control parameter based on the variance corresponding to the control parameter and the variances corresponding to each of the heat transfer performance parameters;

[0056] Among them, determining the key heat transfer performance parameter information corresponding to the control parameter based on the variance corresponding to the control parameter and the variances corresponding to each of the heat transfer performance parameters includes:

[0057] For each of the heat transfer performance parameters, use the ratio of the variance of the control parameter to the variance of the heat transfer performance parameter as the correlation index between the heat transfer performance parameter and the control parameter, and compare the correlation index with a preset correlation index. If the correlation index is greater than the preset correlation index, determine that the heat transfer performance parameter is the key heat transfer performance parameter corresponding to the control parameter; each of the key heat transfer performance parameters constitutes the key heat transfer performance parameter information corresponding to the control parameter.

[0058] Among them, constructing the target training data set corresponding to the control parameter based on the key heat transfer performance parameter information and the training data set includes the following steps:

[0059] For each heat transfer performance parameter information in the training data set, extract the corresponding key heat transfer performance parameter values of the key heat transfer performance parameter information from the heat transfer performance parameter information, and construct a mapping relationship between the control parameter values in the control parameter information corresponding to the heat transfer performance parameter information and the key heat transfer performance parameter values, so as to obtain the target training data set.

[0060] It can be understood that by determining the key heat transfer performance parameter information corresponding to the control parameter, the above method can reduce the attention to secondary parameters and the complexity of the model during the subsequent process of training the target control parameter prediction model corresponding to the control parameter, which helps to reduce the computational cost of the heat exchange module control method.

[0061] Step S2: Train a preset neural network based on the target training data sets corresponding to the respective control parameters to obtain the target control parameter prediction models corresponding to the respective control parameters.

[0062] Specifically, step S2 includes the following steps:

[0063] For each of the control parameters, divide the target training data set corresponding to the control parameter into a training set, a validation set, and a calibration set;

[0064] Train the neural network based on the training set to obtain an initial control parameter prediction model;

[0065] Optimize the model parameters of the initial control parameter prediction model based on a preset genetic optimization algorithm and the validation set to obtain an intermediate control parameter prediction model corresponding to the control parameter;

[0066] Calibrate the model parameters of the intermediate control parameter prediction model based on the calibration set to obtain the target control parameter prediction models corresponding to the respective control parameters; specifically, for each mapping relationship in the calibration set, input the key heat performance parameter information of the mapping relationship into the intermediate control parameter prediction model to obtain the control parameter prediction value corresponding to the mapping relationship, and calculate the prediction loss value between the control parameter prediction value and the actual control parameter value corresponding to the mapping relationship based on a preset loss function, and backpropagate the prediction loss value in the intermediate control parameter prediction model to obtain the gradient corresponding to the prediction loss value, and optimize the model parameters of the intermediate control parameter prediction model based on the gradient to obtain the target control parameter prediction model.

[0067] Understandably, the method of training a preset neural network using the target training dataset corresponding to the control parameters to obtain the target control parameter prediction models corresponding to the respective control parameters optimizes the model parameters of the initial control parameter prediction model through a preset genetic optimization algorithm and the validation set to obtain the intermediate control parameter prediction model corresponding to the control parameters, and corrects the model parameters of the intermediate control parameter prediction model based on the calibration set to obtain the target control parameter prediction models corresponding to the respective control parameters, achieving multi-dimensional optimization of the model parameters of the initial control parameter prediction model, which helps improve the prediction accuracy of the target control parameter prediction model.

[0068] Among them, the step of optimizing the model parameters of the initial control parameter prediction model through a preset genetic optimization algorithm and the validation set to obtain the intermediate control parameter prediction model corresponding to the control parameters includes the following steps:

[0069] For each initial model parameter value of the initial control parameter prediction model, a first threshold interval corresponding to the initial model parameter value is generated based on a preset first threshold interval length, and a preset number of first parameter values are extracted within the first threshold interval based on a preset first step size to obtain a first parameter set corresponding to the initial model parameter value; wherein, the initial model parameter value is located in the middle of the first threshold interval;

[0070] Perform a Cartesian product operation on each of the first parameter sets to obtain a first-generation population; the first-generation population includes multiple first-generation individuals;

[0071] Generate the fitness corresponding to each first-generation individual based on the validation set, and determine the first-generation individual with the maximum fitness as the first-generation target individual;

[0072] For each first parameter value of the first-generation target individual, a second threshold interval corresponding to the first parameter value is generated based on a preset second threshold interval length, and a preset number of second parameter values are extracted within the second threshold interval based on a preset second step size to obtain a second parameter set corresponding to the first parameter value; wherein, the first parameter value is located in the middle of the second threshold interval; the second threshold interval length is half of the first threshold interval length, and the second step size is half of the first step size;

[0073] Perform a Cartesian product operation on each of the second parameter sets to obtain a second-generation population; the second-generation population includes multiple second-generation individuals; iterate the steps after performing the Cartesian product operation on each of the first parameter sets to obtain the first-generation population until the final target individual is obtained, and the fitness corresponding to the final target individual converges;

[0074] Update the model parameters of the initial control parameter prediction model based on the final target individual to obtain the intermediate control parameter prediction model.

[0075] Understandably, the method of optimizing the model parameters of the initial control parameter prediction model based on the preset genetic optimization algorithm and the validation set ensures the extensiveness of global search by setting a relatively large threshold interval and sampling step size at the initial stage to cover the entire parameter space. In each generation of iteration, a Cartesian product operation is performed on the parameter set to generate a new population, and then the fitness of each individual is calculated according to the validation set, and the individual with the maximum fitness is selected as the target individual. As the iteration progresses, the length of the threshold interval and the step size are continuously halved, gradually converging to the optimal region of parameter optimization. This method greatly improves the efficiency and accuracy of parameter optimization, and avoids the problems of high computational complexity and slow convergence speed that may occur in traditional optimization methods.

[0076] Among them, generating the fitness corresponding to the first-generation individuals based on the validation set includes the following steps:

[0077] For each of the first-generation individuals, update the model parameters of the initial control parameter prediction model based on the first-generation individual to obtain the updated control parameter prediction model corresponding to the first-generation individual;

[0078] Obtain the prediction accuracy rate and mean absolute error of the updated control parameter prediction model based on the validation set, and determine the product of the prediction accuracy rate and the reciprocal of the mean absolute error as the fitness of the first-generation individual. Specifically, for each mapping relationship in the validation set, input the key heat performance parameter information of the mapping relationship into the updated control parameter prediction model to obtain the predicted control parameter value corresponding to the mapping relationship, calculate the absolute value of the difference between the predicted control parameter value and the standard control parameter value corresponding to the mapping relationship, and determine the average value of each of the absolute values as the mean absolute error.

[0079] Understandably, the method of generating the fitness of the first-generation individuals can more comprehensively and truly reflect the overall performance of the model by combining the two indicators of prediction accuracy rate and mean absolute error. This method can effectively avoid the bias caused by simply relying on the accuracy rate or mean absolute error, and improve the accuracy of the optimization process.

[0080] Step S3: Generate a heat transfer performance parameter combination algorithm based on the key heat transfer performance parameter information corresponding to each control parameter.

[0081] Specifically, construct the matching relationship between each control parameter and its corresponding key heat transfer performance parameter information.

[0082] Step S4: Obtain the heat exchange performance parameter information of the heat exchange module in real time; wherein, the heat exchange performance parameter information includes the heat exchange performance parameter values of multiple heat exchange performance parameters.

[0083] Specifically, obtain the heat exchange performance parameter information of the heat exchange module through various sensors arranged at specific positions.

[0084] Step S5: Perform combination processing on each heat exchange performance parameter in the heat exchange performance parameter information based on the heat exchange performance parameter combination algorithm to obtain multiple heat exchange performance parameter combinations.

[0085] Step S6: For each heat exchange performance parameter combination, input the heat exchange performance parameter combination into the corresponding target control parameter prediction model to obtain the corresponding control parameter.

[0086] Step S7: Control the operation of the heat exchange module based on each control parameter.

[0087] For the method provided in this embodiment, on the one hand, by constructing an independent target control parameter prediction model for each control parameter, each model can focus on learning and capturing the unique influencing factors and dynamic characteristics of the control parameter, thereby significantly improving the prediction accuracy, reducing the mutual interference and error transmission of the control parameters during the prediction process, and ensuring the accuracy and reliability of the control parameters. On the other hand, by separately constructing a target control parameter prediction model for each control parameter, the complexity of each model is effectively reduced, the model structure is made more concise, and the training process is more efficient, which helps to improve the response speed and real-time performance of the overall control system.

[0088] Please refer to Figure 2 , Figure 2 which is a schematic block diagram of the neural network control system 100 of the heat exchange module provided by the embodiment of the present application. As Figure 2 shown, the neural network control system 100 of the heat exchange module provided by the embodiment of the present application includes:

[0089] The first acquisition module 110 is used to acquire the training data set, and respectively determine the key heat exchange performance parameter information corresponding to each control parameter of the heat exchange module based on the training data set, and construct the target training data set corresponding to the control parameter based on the key heat exchange performance parameter information and the training data set.

[0090] The training module 120 is used to train the preset neural network respectively based on the target training data sets corresponding to each control parameter to obtain the target control parameter prediction models corresponding to each control parameter;

[0091] The generation module 130 is used to generate a heat exchange performance parameter combination algorithm based on the key heat exchange performance parameter information corresponding to each control parameter.

[0092] A second acquisition module 140, configured to acquire in real time the heat exchange performance parameter information of the heat exchange module; wherein, the heat exchange performance parameter information includes the heat exchange performance parameter values of a plurality of heat exchange performance parameters.

[0093] A combined processing module 150, configured to perform combined processing on each heat exchange performance parameter in the heat exchange performance parameter information based on the heat exchange performance parameter combination algorithm to obtain a plurality of heat exchange performance parameter combinations.

[0094] An input module 160, configured to input each of the heat exchange performance parameter combinations into a corresponding target control parameter prediction model to obtain a corresponding control parameter.

[0095] A control module, configured to control the operation of the heat exchange module based on each of the control parameters.

[0096] It should be noted that those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described system and each module can refer to the process in the embodiment of the neural network control method of the heat exchange module described above, and will not be elaborated here.

[0097] The neural network control system 100 of the heat exchange module provided in the above embodiment can be implemented in the form of a computer program, and the computer program can run on a terminal device 200 as shown in Figure 3 Figure.

[0098] Please refer to Figure 3 , Figure 3 which is a schematic block diagram of the structure of the terminal device 200 provided in the embodiment of the present application. The terminal device 200 includes a processor 201 and a memory 202. The processor 201 and the memory 202 are connected through a device bus 203. Among them, the memory 202 may include a non-volatile storage medium and an internal memory.

[0099] The non-volatile storage medium can store a computer program. The computer program includes program instructions, and when the program instructions are executed by the processor 201, the processor 201 can be made to execute any of the above neural network control methods of the heat exchange module.

[0100] The processor 201 is configured to provide computing and control capabilities to support the operation of the entire terminal device 200.

[0101] The internal memory provides an environment for the operation of the computer program in the non-volatile storage medium. When the computer program is executed by the processor 201, the processor 201 can be made to execute any of the above neural network control methods of the heat exchange module.

[0102] Those skilled in the art can understand,Figure 3 The structure shown is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the terminal device 200 involved in the solution of this application. Specifically, the terminal device 200 may include more or fewer components than those shown in the figure, or combine some components, or have a different component arrangement.

[0103] It should be understood that the processor 201 may be a central processing unit (CPU), and the processor 201 may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0104] Among them, in some embodiments, the processor 201 is used to run a computer program stored in the memory to implement the following steps:

[0105] Obtain a training data set, and respectively determine the key heat transfer performance parameter information corresponding to each control parameter of the heat exchange module based on the training data set, and construct a target training data set corresponding to the control parameter based on the key heat transfer performance parameter information and the training data set;

[0106] Train a preset neural network respectively based on the target training data sets corresponding to each control parameter to obtain a target control parameter prediction model corresponding to each control parameter;

[0107] Generate a heat transfer performance parameter combination algorithm based on the key heat transfer performance parameter information corresponding to each control parameter;

[0108] Obtain the heat transfer performance parameter information of the heat exchange module in real time; wherein, the heat transfer performance parameter information includes the heat transfer performance parameter values of multiple heat transfer performance parameters;

[0109] Perform combination processing on each heat transfer performance parameter in the heat transfer performance parameter information based on the heat transfer performance parameter combination algorithm to obtain multiple heat transfer performance parameter combinations;

[0110] For each heat transfer performance parameter combination, input the heat transfer performance parameter combination into the corresponding target control parameter prediction model to obtain the corresponding control parameter;

[0111] Control the operation of the heat exchange module based on each of the control parameters.

[0112] It should be noted that those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working process of the terminal device 200 described above can refer to the process of the neural network control method of the heat exchange module described above, and will not be elaborated here.

[0113] The embodiment of the present application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by one or more processors, the one or more processors are caused to implement the neural network control method of the heat exchange module provided by the embodiment of the present application.

[0114] Among them, the computer-readable storage medium may be an internal storage unit of the terminal device 200 in the foregoing embodiment, such as the hard disk or memory of the terminal device 200. The computer-readable storage medium may also be an external storage device of the terminal device 200, such as a plug-in hard disk equipped with the terminal device 200, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc.

[0115] The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of various equivalent modifications or substitutions within the technical scope disclosed by the present application, and these modifications or substitutions should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A neural network control method for a heat exchange module, characterized in that, Including: Obtain a training data set, and respectively determine the key heat transfer performance parameter information corresponding to each control parameter of the heat exchange module based on the training data set, and construct the target training data set corresponding to the control parameter based on the key heat transfer performance parameter information and the training data set; Respectively train a preset neural network based on the target training data set corresponding to each control parameter to obtain the target control parameter prediction model corresponding to each control parameter; Generate a heat transfer performance parameter combination algorithm based on the key heat transfer performance parameter information corresponding to each control parameter; Obtain the heat transfer performance parameter information of the heat exchange module in real time; wherein, the heat transfer performance parameter information includes the heat transfer performance parameter values of multiple heat transfer performance parameters; Perform combination processing on each heat transfer performance parameter in the heat transfer performance parameter information based on the heat transfer performance parameter combination algorithm to obtain multiple heat transfer performance parameter combinations; For each heat transfer performance parameter combination, input the heat transfer performance parameter combination into the corresponding target control parameter prediction model to obtain the corresponding control parameter; Control the operation of the heat exchange module based on each control parameter; Among them, the respectively determining the key heat transfer performance parameter information corresponding to each control parameter of the heat exchange module based on the training data set includes: Respectively determine the variance corresponding to each control parameter and the variance corresponding to each heat transfer performance parameter based on the training data set; For each control parameter, determine the key heat transfer performance parameter information corresponding to the control parameter based on the variance corresponding to the control parameter and the variances corresponding to each heat transfer performance parameter; including: For each heat transfer performance parameter, use the ratio of the variance of the control parameter to the variance of the heat transfer performance parameter as the correlation index between the heat transfer performance parameter and the control parameter, and compare the correlation index with a preset correlation index. If the correlation index is greater than the preset correlation index, determine that the heat transfer performance parameter is the key heat transfer performance parameter corresponding to the control parameter; each key heat transfer performance parameter constitutes the key heat transfer performance parameter information corresponding to the control parameter.

2. The neural network control method of the heat exchange module according to claim 1, characterized in that The training data set includes multiple mapping relationships, the mapping relationship is the mapping relationship between the heat transfer performance parameter information and the control parameter information, the heat transfer performance parameter information includes the heat transfer performance parameter values of multiple heat transfer performance parameters, and the control parameter information includes the control parameter values of multiple control parameters. The respectively determining the variance corresponding to each control parameter and the variance corresponding to each heat transfer performance parameter based on the training data set includes: For each control parameter, extract the control parameter value corresponding to the control parameter from each control parameter information in the training data set to obtain the control parameter value sequence corresponding to the control parameter, and generate the variance of the control parameter based on the control parameter value sequence; For each of the heat transfer performance parameters, extract the heat transfer performance parameter values of the heat transfer performance parameter from the heat transfer performance parameter information of the training data set to obtain a sequence of heat transfer performance parameter values corresponding to the heat transfer performance parameter, and generate the variance of the heat transfer performance parameter based on the sequence of heat transfer performance parameter values.

3. The neural network control method of the heat exchange module according to claim 2, wherein The training of the preset neural network based on the target training data sets corresponding to the respective control parameters to obtain the target control parameter prediction models corresponding to the respective control parameters includes: For each of the control parameters, divide the target training data set corresponding to the control parameter into a training set, a validation set, and a calibration set; Train the neural network based on the training set to obtain an initial control parameter prediction model; Optimize the model parameters of the initial control parameter prediction model based on a preset genetic optimization algorithm and the validation set to obtain an intermediate control parameter prediction model corresponding to the control parameter; Calibrate the model parameters of the intermediate control parameter prediction model based on the calibration set to obtain the target control parameter prediction models corresponding to the respective control parameters.

4. The neural network control method of the heat exchange module according to claim 3, wherein The optimizing the model parameters of the initial control parameter prediction model based on a preset genetic optimization algorithm and the validation set to obtain an intermediate control parameter prediction model corresponding to the control parameter includes: For each initial model parameter value of the initial control parameter prediction model, generate a first threshold interval corresponding to the initial model parameter value based on a preset first threshold interval length, and extract a preset number of first parameter values within the first threshold interval based on a preset first step size to obtain a first parameter set corresponding to the initial model parameter value; wherein, the initial model parameter value is located in the middle of the first threshold interval; Perform a Cartesian product operation on each of the first parameter sets to obtain a first-generation population; the first-generation population includes multiple first-generation individuals; Generate the fitness corresponding to each of the first-generation individuals based on the validation set, and determine the first-generation individual with the maximum fitness as the first-generation target individual; For each first parameter value of the first-generation target individual, generate a second threshold interval corresponding to the first parameter value based on a preset second threshold interval length, and extract a preset number of second parameter values within the second threshold interval based on a preset second step size to obtain a second parameter set corresponding to the first parameter value; wherein, the first parameter value is located in the middle of the second threshold interval; the second threshold interval length is one-half of the first threshold interval length, and the second step size is one-half of the first step size; Perform a Cartesian product operation on each of the second parameter sets to obtain a second-generation population; the second-generation population includes multiple second-generation individuals; iterate the steps after obtaining the first-generation population by performing a Cartesian product operation on each of the first parameter sets until the final target individual is obtained, and the fitness corresponding to the final target individual converges; Update the model parameters of the initial control parameter prediction model based on the final target individual to obtain the intermediate control parameter prediction model.

5. The neural network control method of the heat exchange module according to claim 4, characterized in that, Generating the fitness values corresponding to the first-generation individuals based on the validation set respectively includes: For each of the first-generation individuals, updating the model parameters of the initial control parameter prediction model based on the first-generation individual to obtain the updated control parameter prediction model corresponding to the first-generation individual; Obtaining the prediction accuracy and mean absolute error of the updated control parameter prediction model based on the validation set, and determining the product of the prediction accuracy and the reciprocal of the mean absolute error as the fitness value of the first-generation individual.

6. The neural network control method of the heat exchange module according to claim 3, wherein Calibrating the model parameters of the intermediate control parameter prediction model based on the calibration set to obtain the target control parameter prediction models corresponding to the respective control parameters, including: For each mapping relationship in the calibration set, inputting the key thermal performance parameter information of the mapping relationship into the intermediate control parameter prediction model to obtain the predicted control parameter value corresponding to the mapping relationship, calculating the prediction loss value between the predicted control parameter value and the actual control parameter value corresponding to the mapping relationship based on a preset loss function, and performing backpropagation of the prediction loss value in the intermediate control parameter prediction model to obtain the gradient corresponding to the prediction loss value, and optimizing the model parameters of the intermediate control parameter prediction model based on the gradient to obtain the target control parameter prediction model.

Images