Multi-cavity coordinated control method, device and electronic equipment of microchannel heat exchange module

Abstract

The present application relates to the technical field of heat exchange modules, and discloses a multi-cavity cooperative control method, device and electronic equipment of a microchannel heat exchange module, the method comprising: performing numerical simulation calculations on multiple heat exchange cavities of the microchannel heat exchange module to obtain a training data matrix; performing coefficient calculations and heat flow coupling relationship analysis between cavities to generate a heat flow distribution mathematical model; substituting heat flow distribution data, temperature distribution data and flow distribution data into a non-homogeneous heat conduction equation to construct a multi-objective optimization function; performing chaotic mapping processing and multi-objective cooperative differential evolution to obtain an optimal control solution set; performing Q-learning state value function solution to generate cooperative control instructions; inputting the cooperative control instructions into an execution unit to perform real-time adjustment and online update of the inlet temperature parameters and flow parameters of each heat exchange cavity to obtain a target heat flow distribution model. In this way, information interaction and optimization coordination between each heat exchange cavity are realized, and the heat exchange efficiency is significantly improved.

Description

Technical Field

[0001] The present application relates to the technical field of heat exchange modules, and in particular to a multi-cavity coordinated control method, device and electronic equipment of a microchannel heat exchange module. Background Art

[0002] As the performance requirements for heat exchange equipment continue to increase, traditional single-cavity heat exchange modules can no longer meet application requirements such as high heat flux density and rapid response. Although multi-cavity microchannel heat exchange modules can significantly improve heat exchange efficiency, their complex internal structure leads to a strong heat flux coupling effect between cavities.

[0003] At present, the control method for multi-cavity microchannel heat exchange modules mainly adopts independent PID control strategy. There is a lack of coordination mechanism between the cavities, which makes it difficult to cope with the dynamic heat flux distribution changes under complex working conditions. At the same time, due to the strong nonlinear characteristics and multivariable coupling characteristics of the heat exchange process, traditional control methods are difficult to accurately predict and adjust the heat transfer performance of each cavity, resulting in low overall heat transfer efficiency of the system. In addition, most of the existing heat flux distribution prediction models are based on simplified assumptions and do not fully consider the heat flux coupling relationship between cavities, resulting in insufficient model accuracy. Summary of the invention

[0004] The present application provides a multi-cavity collaborative control method, device and electronic equipment for a microchannel heat exchange module, thereby realizing information interaction and optimized coordination between the heat exchange cavities, and significantly improving the heat exchange efficiency.

[0005] In a first aspect, the present application provides a multi-cavity coordinated control method for a microchannel heat exchange module, the multi-cavity coordinated control method for a microchannel heat exchange module comprising:

[0006] Numerical simulation calculations are performed on multiple heat exchange cavities of the microchannel heat exchange module to obtain temperature distribution data, flow distribution data, and pressure distribution data of each heat exchange cavity to obtain a training data matrix;

[0007] Inputting the training data matrix into the multi-layer block separation model to calculate the temperature field coefficient, flow field coefficient and pressure field coefficient and analyze the thermal flow coupling relationship between cavities to generate a heat flow distribution mathematical model;

[0008] According to the heat flux distribution mathematical model, the heat flux distribution data, temperature distribution data and flow distribution data of each heat exchange cavity are substituted into the non-homogeneous heat conduction equation to construct a multi-objective optimization function;

[0009] Based on the multi-objective optimization function, chaotic mapping processing and multi-objective collaborative differential evolution are performed to obtain an optimal control solution set;

[0010] Performing Markov jump system modeling and Q-learning state value function solution on the optimal control scheme set to generate collaborative control instructions;

[0011] The collaborative control instruction is input into the execution unit, the inlet temperature parameters and flow parameters of each heat exchange cavity are adjusted in real time, and the heat flux distribution mathematical model is updated online to obtain a target heat flux distribution model.

[0012] A second aspect of the present application provides a multi-cavity coordinated control device for a microchannel heat exchange module, the multi-cavity coordinated control device for a microchannel heat exchange module comprising:

[0013] An acquisition module is used to perform numerical simulation calculations on multiple heat exchange cavities of the microchannel heat exchange module, obtain temperature distribution data, flow distribution data and pressure distribution data of each heat exchange cavity, and obtain a training data matrix;

[0014] A calculation module, used for inputting the training data matrix into a multi-layer block separation model to calculate the temperature field coefficient, flow field coefficient and pressure field coefficient and analyze the thermal flow coupling relationship between cavities, and generate a heat flow distribution mathematical model;

[0015] A construction module is used to substitute the heat flux distribution data, temperature distribution data and flow distribution data of each heat exchange cavity into the non-homogeneous heat conduction equation according to the heat flux distribution mathematical model to construct a multi-objective optimization function;

[0016] A processing module, used for performing chaos mapping processing and multi-objective collaborative differential evolution based on the multi-objective optimization function to obtain an optimal control solution set;

[0017] A solution module, used for performing Markov jump system modeling and Q-learning state value function solution on the optimal control scheme set to generate cooperative control instructions;

[0018] The updating module is used to input the collaborative control instruction into the execution unit, adjust the inlet temperature parameters and flow parameters of each heat exchange cavity in real time, and update the heat flux distribution mathematical model online to obtain the target heat flux distribution model.

[0019] The third aspect of the present application provides an electronic device, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the electronic device executes the multi-cavity coordinated control method of the above-mentioned microchannel heat exchange module.

[0020] Compared with the prior art, the present application has the following beneficial effects: the temperature field, flow field and pressure field are separately modeled through a multi-layer block algorithm, and a high-precision mathematical model of heat flux distribution is established in combination with heat-flux coupling analysis, which solves the problem of insufficient prediction accuracy caused by the traditional model not considering the coupling between fields; chaos mapping and multi-objective collaborative differential evolution algorithm are used to optimize the control parameters, which overcomes the defect that the traditional optimization algorithm is prone to fall into local optimality and obtains a better set of control schemes; the Markov jump system and Q-learning algorithm are introduced to construct an intelligent control strategy, which effectively handles the nonlinear characteristics and uncertainties in the heat exchange process and ensures the stability of the control system; a model correction method based on an online update mechanism is designed, and the heat flux distribution model is dynamically corrected through real-time response data, so that the control strategy can adapt to changes in operating conditions; through the multi-cavity collaborative control mechanism, information interaction and optimization coordination between the heat exchange cavities are realized, which significantly improves the overall heat exchange efficiency and system performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0022] The structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification so as to facilitate understanding and reading by persons familiar with this technology. They are not used to limit the conditions under which the present invention can be implemented, and therefore have no substantive technical significance. Any structural modification, change in proportion or adjustment of size, without affecting the effects and purposes that can be achieved by the present invention, should still fall within the scope of the technical contents disclosed by the present invention.

[0023] Figure 1 It is a schematic flow chart of a multi-cavity coordinated control method of a microchannel heat exchange module provided by an embodiment of the present invention;

[0024] Figure 2 It is a schematic block diagram of the structure of a multi-cavity coordinated control device of a microchannel heat exchange module provided by an embodiment of the present invention;

[0025] Figure 3 It is a schematic block diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0028] It should also be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0029] It should be further understood that the term "and / or" used in the specification and appended claims of this application refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations. Figure 1 , an embodiment of the multi-cavity coordinated control method of the microchannel heat exchange module in the embodiment of the present application includes:

[0030] Step 100, performing numerical simulation calculations on multiple heat exchange cavities of the microchannel heat exchange module, obtaining temperature distribution data, flow distribution data, and pressure distribution data of each heat exchange cavity, and obtaining a training data matrix;

[0031] It is understandable that the execution subject of the present application can be a multi-cavity coordinated control device of a microchannel heat exchange module, or a terminal or a server, which is not limited here. The present application embodiment is described by taking a server as the execution subject as an example.

[0032] Specifically, the microchannel heat exchange module is divided into cavities to determine the N heat exchange cavities included in the model. In this process, the basis for cavity division is reasonably selected, such as geometric characteristics, heat transfer performance distribution, and the complexity of the fluid flow path. Temperature boundary conditions and pressure boundary conditions are set for each cavity to ensure that the initial heat exchange cavity model can truly reflect the thermodynamic characteristics in actual operation. Based on the initial model, meshing is performed. The fluid calculation domain uses a hexahedral grid, which can better handle the flow field distribution in the flow channel, while the solid calculation domain uses a tetrahedral grid to improve the calculation accuracy of complex geometric areas. Through meshing, a computational grid model for solving is generated. After the computational grid model is generated, the necessary physical parameters are set to improve the model. The k-epsilon turbulence model parameters are specified for the model to describe the turbulent characteristics of the fluid, and parameters such as the convective heat transfer coefficient, thermal conductivity, and specific heat capacity are set to accurately describe the physical process of heat transfer and heat conduction. The fluid-solid coupling boundary setting operation is performed at the interface between the fluid and the solid to generate a computational physics model with clear physical meaning. Based on the computational physics model, pressure-velocity coupling calculation is performed to solve the momentum conservation equation and continuity equation in fluid flow. Through steady-state iterative solution, the temperature field, flow field and pressure field data of each grid node are gradually converged, which constitute the preliminary original data set. The original data set is grouped and classified into temperature field data, flow field data and pressure field data. The correlation between these data is analyzed using the Pearson correlation coefficient to obtain the correlation matrix describing the correlation of different field data. Based on the data correlation matrix, the heat flow coupling analysis is performed between the cavities of the microchannel heat exchange module. The heat transfer relationship between different cavities is calculated by the thermal conductivity method to quantify the degree of heat exchange between different cavities. At the same time, in order to determine the temperature boundary conditions and flow boundary conditions between cavities, the boundary element method is used, which has significant advantages in dealing with complex boundary problems. The heat flow coupling relationship between cavities is matrix spliced ​​with the original data set to obtain the training data matrix. The matrix integrates multiple data such as heat flow, temperature, pressure and flow, while retaining the coupling characteristics between cavities.

[0033] Step 200, input the training data matrix into the multi-layer block separation model to calculate the temperature field coefficient, flow field coefficient and pressure field coefficient and analyze the thermal flow coupling relationship between cavities to generate a heat flow distribution mathematical model;

[0034] Specifically, the training data matrix is ​​input into the multi-layer block separation model, which performs generalized separable characteristic analysis operations on the temperature field data, flow field data and pressure field data by block processing. During the analysis process, the model uses the mathematical principle of separable characteristics to decompose the complex associations of different field data into independent feature bases, and obtains the basic separation matrix for subsequent calculations. Based on the basic separation matrix, the temperature field coefficient, flow field coefficient and pressure field coefficient are calculated. These field coefficients respectively reflect the coupling relationship and influence degree of temperature, flow and pressure between different cavities in the microchannel heat exchange module. By comprehensively calculating the interaction and correlation between these coefficients, the weight relationship between the field coefficients is established, and the field coefficient weight matrix is ​​generated. Feature analysis is performed based on the field coefficient weight matrix. During the feature analysis process, the field feature recognition matrix describing the characteristics of each field is generated by processing the high-dimensional data of the matrix. The field feature recognition matrix can capture the characteristic patterns of heat flow, temperature and pressure distribution between different cavities in the microchannel heat exchange module. Using the field feature recognition matrix, through the heat flow coupling modeling operation, a group of heat flow coupling equations describing the heat flow distribution relationship between cavities is established. By solving the heat flow coupling equations, the heat flow distribution coefficient between the cavities is obtained, and the heat flow distribution coefficient matrix is ​​constructed to describe the heat transfer relationship and distribution state between the cavities in the microchannel heat exchange module. The heat flow distribution coefficient matrix is ​​energy-balanced. Through the energy balance verification operation, it is verified whether the model conforms to the first law of thermodynamics, ensures the balance between the input heat and the output heat, and generates the heat balance constraint condition. The heat balance constraint condition is combined with the heat flow distribution coefficient matrix to integrate into a unified heat flow distribution calculation model. The heat flow distribution calculation model is verified to ensure the rigor of its mathematical logic and the degree of consistency with the actual operating conditions, and the heat flow distribution mathematical model is obtained.

[0035] Step 300: According to the heat flux distribution mathematical model, the heat flux distribution data, temperature distribution data and flow distribution data of each heat exchange cavity are substituted into the non-homogeneous heat conduction equation to construct a multi-objective optimization function;

[0036] It should be noted that, according to the mathematical model of heat flux distribution, the non-homogeneous heat conduction equation of each heat exchange cavity is modeled to form a basic heat transfer equation, which describes the heat transfer process inside and between heat exchange cavities, and covers the non-uniform heat source, heat conduction characteristics and fluid heat transfer conditions of each cavity. On the basis of the basic heat transfer equation, the temperature distribution data and flow distribution data of each heat exchange cavity are input. In order to perform numerical solution, the basic heat transfer equation is discretized by the finite difference method. The finite difference method converts the continuous partial differential equation into a discrete form, so that the complex heat transfer process can be approximately solved by numerical calculation. Through discretization, the discretized heat transfer equation obtained can describe the distribution state and transfer path of heat between discrete nodes in the heat exchange cavity. Substitute the discretized heat transfer equation into the energy conservation equation to ensure that the heat exchange process conforms to the first law of thermodynamics and generate heat conservation constraints. The constraint expression is used to verify the balance relationship between input energy, transfer energy and output energy. On the basis of the heat conservation constraint, the efficiency objective function is constructed, the goal is to maximize the effective utilization rate of heat and the overall heat exchange efficiency, and provide the core direction for the optimization process. In order to evaluate the performance of the efficiency objective function, a posteriori Cramer-Rao lower bound is calculated for it. Cramer-Rao lower bound is a statistical tool for quantifying the accuracy of estimation. By calculating this lower bound, a performance index function is obtained, which reflects the actual optimization potential of the objective function. The performance index function is weightedly combined with the heat transfer efficiency function to generate a comprehensive evaluation function, so that efficiency optimization and performance reliability are comprehensively considered in the evaluation. Based on the comprehensive evaluation function, the limitations of actual engineering conditions are considered. Pressure loss constraints and heat transfer area constraints are added to the comprehensive evaluation function to ensure that the optimization model not only pursues high efficiency, but also meets the physical limitations of system operation and equipment design conditions. The pressure loss constraint limits the resistance loss of fluid flow in the system, while the heat transfer area constraint limits the physical size and available area range of the heat exchanger. The constrained optimization model is normalized. By mapping the objective function and constraint conditions into a dimensionless range, the influence of different variable scale differences on the optimization results is eliminated, and a multi-objective optimization function is obtained.

[0037] Step 400: Performing chaos mapping processing and multi-objective collaborative differential evolution based on a multi-objective optimization function to obtain an optimal control solution set;

[0038] Specifically, the Logistic chaotic mapping operation is performed on the multi-objective optimization function to generate an initial population. Logistic chaotic mapping is a chaotic system that generates random distribution samples with good coverage through its highly nonlinear properties. In this process, the generated initial population contains M D-dimensional random initial individuals, where M represents the population size and D represents the dimension of the decision variable. The initialization process provides a diverse solution space for optimization. The objective function is calculated for each individual in the initialization population, and these solutions are classified by the non-dominated sorting method. Non-dominated sorting can effectively distinguish the pros and cons of decomposition and retain the Pareto optimal solution when optimizing multiple objectives at the same time. For the sorted non-dominated solutions, they are archived and stored to obtain the initial Pareto solution set, which includes the current optimal non-dominated solution. Based on the Pareto solution set, the population is mutated using the differential mutation strategy. Differential mutation explores new solution spaces by generating mutation vectors, where the generation formula of mutation vectors is Vi = Xr1 + F(Xr2 - Xr3). Among them, Vi is the mutant individual, Xr1, Xr2, and Xr3 are different parent individuals randomly selected from the population, and F is a scaling factor used to adjust the amplitude of the mutation. The characteristic of the differential mutation strategy is that it can efficiently explore complex multi-objective solution spaces and avoid falling into local optimal solutions. After the mutant population is generated, it is crossovered with the parent population according to the binomial crossover probability to generate a test population. The crossover operation enhances the diversity of the population by combining the genetic information of the mutant individual and the parent individual. At the same time, to ensure the feasibility of the solution, the test population is subjected to boundary constraints, and the individuals that exceed the constraint range are adjusted to the legal range to generate a feasible solution population. A dual-population co-evolution operation is performed on the feasible solution population, and the population is divided into an exploration population and a development population, and information is exchanged between the two populations. The exploration population is responsible for expanding the search range of the solution space, while the development population focuses on local search to refine the quality of the solution. Through the synergy of the two populations, a balance can be achieved between global and local searches, effectively improving the evolutionary efficiency of the population. After the co-evolution is completed, the co-evolution population is merged with the parent population, and the merged population is sorted and screened using the fast non-dominated sorting algorithm and crowding calculation. Fast non-dominated sorting can efficiently handle the solution sorting problem in multi-objective optimization problems, while crowding calculation is used to evaluate the diversity of solutions to ensure that the final screened population has both excellent performance and is evenly distributed in the solution space. Through this screening process, the target population is generated. The elite retention strategy is implemented on the target population, and the current optimal individuals are directly retained in the next generation population to avoid the loss of optimal solutions due to evolution. At the same time, the external archive set is updated to record the optimal solution set obtained during the evolution process. On this basis, the convergence judgment is made based on the evolving population. When the evolutionary generation reaches the set maximum value or the solution set of the population converges to a stable state, the iteration is stopped, and the optimal control solution set is finally output.

[0039] Step 500: Perform Markov jump system modeling and Q-learning state value function solution on the optimal control solution set to generate collaborative control instructions;

[0040] Specifically, according to the optimal control scheme set, the operating characteristics of the microchannel heat exchange module are modeled in state space. In the modeling process, the temperature distribution, flow distribution and pressure distribution are set as the state variables of the system, while the inlet temperature and flow of the heat exchange cavity are defined as control variables to construct the state space equation that describes the dynamic behavior of the system. This equation fully reflects the heat flow coupling and fluid dynamics characteristics inside and between cavities. The state space equation is processed in piecewise homogenization. The entire heat exchange process is divided into multiple sub-intervals, and a continuous residence time parameter is introduced in each sub-interval to capture the time-varying characteristics of the system state in different regions. In this way, the state space equation is converted into a piecewise state equation group, and the partitioning of the complex system is realized, so that the model can more accurately describe the local state characteristics. The piecewise state equation group is converted into a linear matrix inequality form. The complex nonlinear problem is linearized to facilitate the solution in combination with the Q-learning algorithm. On this basis, the state value function Q(s,a) is constructed, where s represents the current state of the system and a represents the control action. The state value function model is used to quantify the value of a certain control action taken by the system under a specific state, providing a direct basis for optimizing the control strategy. In order to realize the calculation of the state value function, the kernel matrix estimation operation is performed on the model. The quadratic programming method is used to solve the minimization problem min||AX-b||, where A is the state transfer matrix, X is the kernel matrix to be solved, and b is the observation vector, and the kernel matrix parameters are obtained. This parameter describes the probability of system state transition and the influence of control variables on state changes. Based on the kernel matrix parameters, the Q-learning value iteration equation is constructed, and the ε-greedy strategy is used to optimize the value iteration process. The ε-greedy strategy is an effective strategy optimization method. By retaining a certain exploration ability while searching for the best solution, the algorithm avoids falling into the local optimum, thereby improving the overall performance of the strategy optimization model. The optimization process of the value iteration equation generates a strategy optimization model, which can provide specific operational guidance for the coordinated control of the heat exchange cavity. After the strategy optimization is completed, the generated model is analyzed for closed-loop stability. By strictly verifying the stability of the closed-loop system, the stability constraint conditions are obtained to ensure that the control strategy is reliable and robust in practical applications. The stability constraints and the strategy optimization model are combined to form complete heat exchange cavity coordinated control instructions. These instructions include temperature adjustment instructions and flow adjustment instructions, which can dynamically adjust the inlet temperature and flow parameters of each cavity, so that the system can achieve efficient and stable operation while meeting the heat exchange requirements.

[0041] Step 600: Input the coordinated control instruction into the execution unit, adjust the inlet temperature parameters and flow parameters of each heat exchange cavity in real time, and update the heat flux distribution mathematical model online to obtain the target heat flux distribution model.

[0042] Specifically, the collaborative control instructions are parsed into a temperature adjustment instruction sequence and a flow adjustment instruction sequence, and these instruction sequences are divided according to the control cycle to generate segmented control instructions. The generation of segmented control instructions helps to achieve refined control, so that the adjustment action in each cycle can better meet the dynamic requirements of the real-time operation of the system. The generated segmented control instructions are input into the execution unit to drive the inlet temperature control valve and flow control valve of each heat exchange cavity, and the real-time control parameters are formed by accurately controlling the valve opening. The real-time control parameters directly act on the inlet conditions of the heat exchange cavity to change its temperature and flow distribution. In this process, the temperature response data, flow response data and pressure response data of each heat exchange cavity are collected at the same time to form a cavity response data set that fully reflects the dynamic behavior of the cavity. The collected cavity response data set is filtered and normalized to eliminate noise interference and standardize the data. Data filtering uses appropriate filtering algorithms, such as Kalman filtering or sliding mean filtering, to eliminate sensor noise and environmental interference; normalization processing adjusts the data range to adapt it to the input requirements of the heat flow distribution mathematical model to obtain standardized response data. The standardized response data is substituted into the mathematical model of heat flux distribution, and the model parameters are identified online based on the real-time operating status. Online identification dynamically adjusts the coefficient parameters of the model so that it can more accurately reflect the actual operating conditions and obtain the model correction parameters. Based on the model correction parameters, the coefficient matrix and coupling matrix of the mathematical model of heat flux distribution are updated. The update operation uses the real-time collected data to continuously correct the core parameters in the mathematical model that describe the relationship between heat flux, temperature and flow distribution, so that the model can reflect the coupling strength and interaction behavior between each cavity in real time. The updated mathematical model of heat flux distribution is more in line with the actual operating status. Through the above dynamic control and online update process, the target heat flux distribution model is obtained. This model not only includes the feedback effect of real-time control instructions, but also realizes rapid adaptation to the dynamic operating environment through online correction, thereby ensuring the optimal heat flux distribution and heat exchange efficiency between the cavities of the microchannel heat exchange module.

[0043] In the embodiments of the present application, the temperature field, flow field and pressure field are separately modeled by a multi-layer block algorithm, and a high-precision mathematical model of heat flux distribution is established in combination with heat-flux coupling analysis, which solves the problem of insufficient prediction accuracy caused by the traditional model not considering the coupling between fields; chaos mapping and multi-objective collaborative differential evolution algorithm are used to optimize the control parameters, which overcomes the defect that the traditional optimization algorithm is prone to fall into local optimality and obtains a better set of control schemes; Markov jump system and Q-learning algorithm are introduced to construct an intelligent control strategy, which effectively handles the nonlinear characteristics and uncertainties in the heat exchange process and ensures the stability of the control system; a model correction method based on an online update mechanism is designed, and the heat flux distribution model is dynamically corrected through real-time response data, so that the control strategy can adapt to changes in operating conditions; through the multi-cavity collaborative control mechanism, information interaction and optimization coordination between the heat exchange cavities are realized, which significantly improves the overall heat exchange efficiency and system performance.

[0044] In a specific embodiment, the process of executing step 100 may specifically include the following steps:

[0045] The microchannel heat exchange module is divided into cavities to obtain N heat exchange cavities, and temperature boundary conditions and pressure boundary conditions are set for the inlet and outlet of each heat exchange cavity to obtain an initial heat exchange cavity model;

[0046] Perform meshing on the initial heat exchange cavity model, wherein hexahedral mesh parameters are set for the fluid calculation domain and tetrahedral mesh parameters are set for the solid calculation domain, to obtain a calculation mesh model;

[0047] The k-epsilon turbulence model parameters, convection heat transfer coefficient parameters, thermal conductivity parameters and specific heat capacity parameters are set for the computational grid model, and the fluid-solid coupling boundary setting operation is performed to obtain the computational physical model;

[0048] Based on the computational physics model, pressure-velocity coupling calculation is performed, and steady-state iterative solution operations of temperature field, flow field and pressure field are performed on each grid node to obtain the original data set;

[0049] The original data set is grouped to obtain temperature field data, flow field data and pressure field data, and the Pearson correlation coefficient is used to calculate the correlation between different field data to obtain the data correlation matrix;

[0050] Based on the data correlation matrix, the heat-flow coupling analysis is performed, the heat transfer coefficient between each heat exchange cavity is calculated by the thermal conductivity method, and the temperature boundary conditions and flow boundary conditions between cavities are determined by the boundary element method to obtain the heat-flow coupling relationship between cavities;

[0051] The thermal-fluid coupling relationship between cavities and the original data set are subjected to matrix concatenation operations to obtain a training data matrix.

[0052] Specifically, the microchannel heat exchange module is divided into cavities according to the geometric structure of the module and the fluid flow characteristics. An independent heat exchange cavity. The division of each cavity should be based on its actual fluid flow path, heat exchange surface area and heat flux distribution to ensure that the divided model can reflect the actual heat exchange process. After the cavity division is completed, set the inlet and outlet boundary conditions for each heat exchange cavity, including temperature boundary conditions and pressure boundary conditions. Meshing is performed on each initial heat exchange cavity model. The meshing of the fluid calculation domain uses hexahedral mesh parameters. This mesh structure has high computational efficiency and accuracy and is suitable for fluid flow analysis; while the solid calculation domain uses tetrahedral mesh parameters. This unstructured grid can better adapt to complex geometric shapes. In meshing, the size h of each grid unit is set to ensure that the resolution is high enough to capture subtle heat transfer and flow characteristics. After meshing, a computational mesh model is generated to provide geometric and topological information for subsequent physical model construction. Based on the computational mesh model, the physical parameters of fluid dynamics and heat transfer characteristics are set. Adopt - The turbulence model describes the turbulent behavior of the fluid, where represents the turbulent kinetic energy, represents the turbulent kinetic energy dissipation rate. These two variables satisfy the turbulent transport equation:

[0053] ;

[0054] ;

[0055] in is the density, is the velocity field, is the dynamic viscosity coefficient, is the turbulent viscosity coefficient, and is the model constant, is the turbulence production term, The model constant of the turbulence production term is used to control the conversion rate of turbulent kinetic energy into turbulent dissipation, is the model constant for the turbulence dissipation term, which is used to control the decay of the turbulence dissipation rate. By setting these parameters, we can ensure that the model can accurately describe the turbulence characteristics of the fluid. Set the convective heat transfer coefficient for the solid part , thermal conductivity and specific heat capacity , through the fluid-solid coupling boundary condition, the heat transfer behavior of the fluid calculation domain and the solid calculation domain are connected to form a complete computational physics model. Based on the above model, pressure-velocity coupling calculation is performed to solve the momentum conservation and continuity equations in the fluid flow. For each grid node, the temperature field is calculated using numerical methods. , Flow Field and pressure field Perform steady-state iterative solution to generate a complete raw data set to describe the heat transfer and flow characteristics inside the cavity. Group the raw data set and organize it into temperature field data , flow field data and pressure field data In order to analyze the relationship between various types of data, the Pearson correlation coefficient was used. Calculate the correlation of different field data. The calculation formula of the correlation coefficient is:

[0056] ;

[0057] in and Represent the values ​​of two sets of data respectively. and By calculating the correlation, the data correlation matrix is ​​obtained to reflect the coupling relationship between the temperature field, flow field and pressure field. Based on the data correlation matrix, the thermal-fluid coupling analysis is performed. The thermal conductivity method is used to calculate the heat transfer coefficient between the cavities. , the formula is:

[0058] ;

[0059] in Indicates that from the cavity To the cavity At the same time, the boundary element method is used to determine the temperature boundary conditions and flow boundary conditions between the cavities to complete the modeling of the heat flow coupling relationship between the cavities. The heat flow coupling relationship between the cavities is matrix-joined with the original data set to form a training data matrix. This matrix contains the temperature, flow, pressure and heat flow coupling information of all heat exchange cavities. For example, if a heat exchange module contains 5 cavities, the training data matrix is ​​as follows:

[0060] ;

[0061] Each row represents the state of a cavity and its heat flow interaction with other cavities. Through the above steps, a comprehensive numerical model is constructed.

[0062] In a specific embodiment, the process of executing step 200 may specifically include the following steps:

[0063] The training data matrix is ​​input into the multi-layer block separation model, and the generalized separable characteristic analysis operation is performed on the temperature field data, the flow field data and the pressure field data to obtain the basic separation matrix;

[0064] The temperature field coefficient, flow field coefficient and pressure field coefficient are calculated based on the basic separation matrix, and the weight relationship between the temperature field coefficient, flow field coefficient and pressure field coefficient is established to obtain the field coefficient weight matrix;

[0065] Performing feature analysis based on the field coefficient weight matrix to obtain a field feature identification matrix, and performing thermal-fluid coupling modeling operations on the field feature identification matrix to obtain a thermal-fluid coupling equation group;

[0066] Based on the heat flow coupling equations, the heat flow distribution coefficient between the cavities is solved to obtain the heat flow distribution coefficient matrix, and the energy balance check operation is performed on the heat flow distribution coefficient matrix to obtain the heat balance constraint condition;

[0067] The heat balance constraint condition is combined with the heat flux distribution coefficient matrix to obtain a heat flux distribution calculation model, and a verification operation is performed on the heat flux distribution calculation model to output a heat flux distribution mathematical model.

[0068] Specifically, the training data matrix is ​​input into the multi-layer block separation model. The model uses the generalized separable characteristic analysis operation to convert the temperature field data , flow field data and pressure field data Decompose and extract the independent features and interactions of different physical fields. In the generalized separable characteristic analysis, the training data matrix is ​​represented as:

[0069] ;

[0070] in, , , are temperature, velocity and pressure data respectively, Indicates the number of cavities, Represents the grid node number. Through the separable characteristic analysis, the above matrix is ​​decomposed into a series of basic separation matrices:

[0071] ;

[0072] in , are the basic separation matrices for temperature field, flow field and pressure field respectively. Based on the basic separation matrix, the temperature field coefficient is calculated , flow field coefficient and the pressure field coefficient These coefficients describe the distribution characteristics of each physical field inside the heat exchange cavity. For example, the temperature field coefficient Calculated by the following formula:

[0073] ;

[0074] in , , The temperature field is , , The gradient of the direction. Similarly, the flow field coefficient and the pressure field coefficient Through the gradient calculation of velocity field and pressure field, and by analyzing the coupling relationship between these coefficients, the weight matrix of temperature field coefficient, flow field coefficient and pressure field coefficient is established:

[0075] ;

[0076] in Display field Hechang The coupling weights describe the mutual influence between different fields. Based on the field coefficient weight matrix, feature analysis is performed to extract key physical features and generate a field feature identification matrix. The feature identification matrix is ​​used to capture the main thermal flow behavior of the system, and its construction formula is:

[0077] ;

[0078] The matrix Contains comprehensive heat flow characteristic information. Perform heat flow coupling modeling operations on the field characteristic identification matrix, and use the thermal conductivity method and boundary conditions to construct the heat flow coupling equations between cavities. The basic form of the coupling equation is:

[0079] ;

[0080] in For cavity To the cavity The heat flow rate, is the thermal conductivity, and is the temperature of the corresponding cavity. By solving the heat flow coupling equations, the heat flow distribution coefficient matrix between cavities is obtained:

[0081]

[0082] This matrix reflects the heat exchange relationship between the cavities in the system. In order to ensure the physical consistency of the calculation, the heat flow distribution coefficient matrix is ​​checked for energy balance. According to the principle of energy conservation, the input heat of the system Should be equal to the heat output and store heat The sum of:

[0083] ;

[0084] in Represent the input, output and stored heat respectively. By verifying the above relationship, the heat balance constraint condition is obtained. The heat balance constraint condition is combined with the heat flow distribution coefficient matrix to construct a complete heat flow distribution calculation model:

[0085] ;

[0086] in is the heat flux distribution calculation model, is the constraint weight coefficient, is the constraint matrix. The heat flux distribution calculation model is verified by numerical solution and experimental comparison to ensure that the model has high accuracy and robustness, and finally the heat flux distribution mathematical model is output.

[0087] In a specific embodiment, the process of executing step 300 may specifically include the following steps:

[0088] According to the mathematical model of heat flux distribution, the non-homogeneous heat conduction equation of each heat exchange cavity is modeled to obtain the basic heat transfer equation;

[0089] Input the temperature distribution data and flow distribution data into the basic heat transfer equation, use the finite difference method to perform discretization, and obtain the discretized heat transfer equation;

[0090] Substitute the discretized heat transfer equation into the energy conservation equation to obtain the heat conservation constraint, and construct the efficiency objective function based on the heat conservation constraint;

[0091] The efficiency objective function is calculated by performing a posteriori Cramer-Rao lower bound calculation to obtain a performance index function, and the performance index function is weightedly combined with the heat transfer efficiency function to obtain a comprehensive evaluation function;

[0092] Pressure loss constraints and heat exchange area constraints are added to the comprehensive evaluation function to obtain a constrained optimization model, and the constrained optimization model is normalized to obtain a multi-objective optimization function.

[0093] Specifically, based on the mathematical model of heat flux distribution, the heat transfer characteristics of each heat exchange cavity are modeled to form a non-homogeneous heat conduction equation. Assume that the temperature distribution in a heat exchange cavity is , its nonhomogeneous heat conduction equation is expressed as:

[0094] ;

[0095] in is the fluid density, is the specific heat capacity, is the thermal conductivity, is the volume heat source term. This equation describes the heat transfer process inside the cavity, where the left side is the time variation term of heat, and the right side is the heat conduction term and the internal heat source term. In the above basic heat transfer equation, input the actual temperature distribution data and traffic distribution data , and discretize it using the finite difference method. The finite difference method converts the continuous partial differential equation into a discrete form for numerical solution. For example, for the temperature field in three-dimensional space, the discretized equation is expressed as:

[0096] ;

[0097] in Represents the time step and spatial nodes The temperature value, is the time step, are the discrete step lengths of space respectively. Substitute the discretized heat transfer equation into the energy conservation equation to construct the heat balance relationship of the system. For each cavity, the energy conservation equation is expressed as:

[0098] ;

[0099] in and are the heat fluxes entering and leaving the cavity, respectively, and V is the cavity volume. This constraint ensures the balance between the input heat, output heat, and internal storage heat of the system. Based on the heat conservation equation, an efficiency objective function is constructed to quantify the heat transfer efficiency of the system. The efficiency objective function is in the form of:

[0100] ;

[0101] in For effective use of heat, is the total heat input to the system. This objective function is used to maximize the efficiency of heat utilization. To evaluate the optimization performance of the objective function, the efficiency objective function is subjected to the a posteriori Cramer-Rao lower bound (CRLB) calculation to obtain the performance index function. CRLB is a statistical tool used to evaluate the lower bound of the estimated variance, and its calculation formula is:

[0102] ;

[0103] in is the Fisher information, defined as:

[0104] I ;

[0105] represents the parameter estimates, is the likelihood function. Through CRLB calculation, we get the performance index function , which is used to measure the robustness of the optimization results. The performance index function and the heat transfer efficiency function are weighted and combined to generate a comprehensive evaluation function, which is in the form of:

[0106] ;

[0107] in and are weight coefficients, reflecting the relative importance of efficiency and performance indicators. Based on the comprehensive evaluation function, pressure loss constraints and heat exchange area constraints are added to construct a constrained optimization model. The pressure loss constraint is expressed as:

[0108] ;

[0109] in is the actual pressure loss of the system, is the maximum allowable pressure loss; the heat exchange area constraint can be expressed as:

[0110] ;

[0111] in is the heat exchange area actually used, is the available heat transfer area. The constrained optimization model is normalized to eliminate the influence of the variable scale and obtain the multi-objective optimization function. The normalization process is completed by the following formula:

[0112] ;

[0113] in and are the minimum and maximum values ​​of the objective function respectively. The normalized multi-objective optimization function is used to optimize the heat transfer performance of the cavity and meet the system constraints.

[0114] In a specific embodiment, the process of executing step 400 may specifically include the following steps:

[0115] Perform Logistic chaos mapping operation on the multi-objective optimization function to generate M D-dimensional random initial individuals, where M represents the population size and D represents the decision variable dimension, and obtain the initialized population;

[0116] Calculate the objective function and perform non-dominated sorting on the initialized population, archive and store the non-dominated solutions, and obtain the initial Pareto solution set;

[0117] Based on the initial Pareto solution set, a mutation vector is constructed, and a differential mutation operation Vi = Xr1 + F(Xr2 - Xr3) is performed on the parent individuals, where Vi is the mutant individual, Xr1, Xr2, and Xr3 are randomly selected different parent individuals, and F is the scaling factor to obtain the mutant population;

[0118] The mutant population and the parent population are crossovered according to the binomial crossover probability to generate a test population, and the test population is subjected to boundary constraints to obtain a feasible solution population;

[0119] Perform a dual-population co-evolution operation on the feasible solution population, divide the population into an exploration population and an exploitation population, and perform information exchange to obtain a co-evolution population;

[0120] The co-evolution population is merged with the parent population, and the merged population is sorted and screened using the fast non-dominated sorting algorithm and crowding calculation to obtain the target population;

[0121] The elite retention strategy is implemented on the target population, and the current optimal individual is directly introduced into the next generation population. At the same time, the external archive set is updated to obtain the evolved population, and convergence judgment is performed based on the evolved population. When the evolutionary generation reaches the set value or the population converges, the iteration is stopped and the optimal control solution set is output.

[0122] Specifically, the Logistic chaos mapping operation is performed on the multi-objective optimization function to generate the initial population. Logistic mapping is a chaotic dynamics model, which is expressed as:

[0123] ;

[0124] in is the current value of the chaotic sequence, is the control parameter (value To ensure chaos). Through this mapping, random values ​​as the initial solution of the population, represents the population size, represents the dimension of the decision variable. Each individual is represented as a Dimensional vector:

[0125] ;

[0126] These individuals constitute the initial population, ensuring that the population has good diversity and global exploration capabilities. The objective function is calculated for the initial population to evaluate the performance of each individual on the optimization objective. For example, for a dual-objective optimization problem, the objective function is recorded as and . According to the objective function values ​​of the individuals, the non-dominated sorting method is used to sort the population. The rule of non-dominated sorting is that if an individual is better than another individual in all objectives, it is called "dominated". Non-dominated solutions constitute the initial Pareto solution set, which are regarded as "equilibrium solutions" in multi-objective optimization problems and are used to guide subsequent optimization. Construct a mutation vector based on the initial Pareto solution set. Differential mutation is a method of generating new solutions through differences between individuals, and its expression is:

[0127] ;

[0128] in It is a mutant individual. , , are three different individuals randomly selected from the population, and Is the basic individual, is the scaling factor (usually taken between ). New solution generated by mutation It can open up new areas in the search space and improve the exploration ability of the population. The mutant population and the parent population are crossovered according to the binomial probability. Perform a crossover operation to generate a test population. The crossover operation further enhances the diversity of the population by mixing the genetic information of the mutant individuals and the parent individuals. The crossover formula is:

[0129] ;

[0130] in The experimental population is The individual in The value of the dimension, rand (0,1) is a random number. After the crossover is completed, the boundary constraint processing is performed on the test population to ensure that the generated individual solutions are within the range of the variables (for example, for the variables Range , the values ​​outside the range are clipped to the boundary value). After the experimental population is processed by boundary constraints, it enters the dual-population co-evolution stage. The population is divided into an exploration population and a development population, which are responsible for global search and local optimization respectively. The exploration population finds new optimal solutions by expanding the search range, while the development population performs local refinement search based on existing solutions. Through the information interaction between the two populations, it is possible to take into account both global optimization capability and local optimization efficiency to form a co-evolutionary population. The co-evolutionary population is merged with the parent population, and the merged population is sorted and screened by the fast non-dominated sorting algorithm and crowding calculation. The fast non-dominated sorting algorithm can efficiently identify non-dominated solutions in the population, and the crowding calculation is used to measure the uniformity of the distribution of solutions on the Pareto front. Finally, the target population is screened, and these individuals represent the best results of the current optimization. The elite retention strategy is implemented on the target population, and the current best individuals are directly retained to the next generation population to avoid the loss of the optimal solution during the iteration process. At the same time, the external archive set is updated to record the optimal Pareto solution set in all iterations. This strategy ensures that the algorithm keeps approaching the real Pareto frontier during the entire optimization process. Convergence is judged based on the evolving population. By comparing the degree of change of solutions between populations, if the diversity of solutions is reduced to the set threshold or the maximum evolutionary generation is reached, the iteration is stopped and the optimal control solution set is output.

[0131] In a specific embodiment, the process of executing step 500 may specifically include the following steps:

[0132] Based on the optimal control solution set, the microchannel heat exchange module is modeled in state space, the temperature distribution, flow distribution and pressure distribution are set as state variables, the inlet temperature and flow of the heat exchange cavity are set as control variables, and the state space equation is obtained;

[0133] The state space equation is subjected to piecewise homogenization, the heat transfer process is divided into multiple sub-intervals, and a residence time parameter is introduced into each sub-interval to obtain a piecewise state equation group;

[0134] The piecewise state equations are transformed into linear matrix inequality form, and the Q-learning state value function Q(s,a) is constructed, where s is the system state and a is the control action, and the state value function model is obtained;

[0135] Perform kernel matrix estimation operation on the state value function model, and use quadratic programming method to solve the minimization problem min||AX-b||, where A is the state transfer matrix, X is the kernel matrix to be solved, and b is the observation vector, to obtain the kernel matrix parameters;

[0136] The Q-learning value iteration equation is constructed based on the kernel matrix parameters, and the value iteration equation is optimized by ε-greedy strategy to obtain the strategy optimization model.

[0137] A closed-loop stability analysis is performed based on the strategy optimization model to obtain stability constraints, and the stability constraints and the strategy optimization model are combined to generate collaborative control instructions for the heat exchange cavity. The collaborative control instructions include temperature adjustment instructions and flow rate adjustment instructions.

[0138] Specifically, the state space modeling of the microchannel heat exchange module is performed based on the optimal control solution set. Traffic distribution and pressure distribution Defined as the state variable of the system , expressed as:

[0139] ;

[0140] At the same time, the inlet temperature of the heat exchange cavity and import flows Set as control variable ,Right now:

[0141] ;

[0142] Based on these variables, the state space equations are established to describe the dynamic behavior of the system:

[0143] ;

[0144] in is the state transition matrix, is the control gain matrix, is the system noise vector, which is used to describe the influence of external disturbances. In order to simplify the model and enhance its adaptability, the state space equation is piecewise homogenized. The heat transfer process is divided into multiple subintervals, for example, based on the fluid flow path or temperature gradient distribution, and the system is assumed to have homogeneous characteristics in each subinterval. The duration parameter is introduced in each subinterval , which represents the average time the fluid stays in this interval. Through this treatment, the state equation is decomposed into a set of piecewise state equations:

[0145] ;

[0146] in , , Subintervals The state, control variables and noise in the system are transformed into the linear matrix inequality (LMI) form so as to be combined with the reinforcement learning algorithm. The general form of the linear matrix inequality is:

[0147] ;

[0148] in is the symmetric positive definite matrix to be solved, is the cost weight matrix. By expressing the dynamic characteristics of the state space model in LMI form, the Q-learning state value function is constructed :

[0149] ;

[0150] in is the reward function, Discount factor ), which represents the weight of future rewards. Through this function, the system is quantified in state Select Control Action In order to solve the state value function, the kernel matrix estimation operation is performed on it and converted into a quadratic programming problem. Assume that the state transfer matrix is , the kernel matrix is , the observation vector is , the problem is of the form:

[0151] ;

[0152] in represents the Euclidean norm. By solving this minimization problem, we get the kernel matrix parameters , which is used to describe the mapping relationship between state and control variables. Based on the kernel matrix parameters, the Q-learning value iteration equation is constructed:

[0153] ;

[0154] in is the learning rate, Indicates the maximum expected value of the next state. To improve the efficiency and stability of the algorithm, the value iteration equation is executed - Greedy strategy optimization. Under this strategy, the algorithm uses probability Select the optimal control action with probability Randomly explore other actions to achieve a balance between the optimal solution and exploration ability. Perform closed-loop stability analysis based on the policy optimization model. Verify the stability of the system by using the optimized control strategy and generate stability constraints. Stability constraints are expressed using Lyapunov functions, for example:

[0155] ;

[0156] in is a positive definite matrix, is the gradient of the Lyapunov function. This constraint ensures that the system remains stable under optimal control. The stability constraint is combined with the strategy optimization model to generate coordinated control instructions for the heat exchange cavity. These instructions include temperature control instructions and flow control instructions, such as:

[0157] ;

[0158] in and is the control function optimized based on Q-learning strategy.

[0159] In a specific embodiment, the process of executing step 600 may specifically include the following steps:

[0160] The coordinated control instruction is parsed into a temperature adjustment instruction sequence and a flow adjustment instruction sequence, and the temperature adjustment instruction sequence and the flow adjustment instruction sequence are divided into control cycles to obtain segmented control instructions;

[0161] The segmented control instruction is input into the execution unit, the opening of the inlet temperature regulating valve and the flow regulating valve of each heat exchange cavity is controlled, and the real-time control parameters are obtained. The temperature response data, flow response data and pressure response data of each heat exchange cavity are collected based on the real-time control parameters to obtain a cavity response data set;

[0162] The cavity response data set is filtered and normalized to obtain standardized response data, and the standardized response data is substituted into the heat flux distribution mathematical model for online identification of model parameters to obtain model correction parameters;

[0163] The coefficient matrix and coupling matrix of the heat flux distribution mathematical model are updated based on the model correction parameters to obtain the target heat flux distribution model.

[0164] Specifically, the collaborative control instruction is parsed into a temperature adjustment instruction sequence and a flow adjustment instruction sequence. Assume that the control instruction is ,in is the inlet temperature adjustment value, is the inlet flow adjustment value. The resolved temperature instruction sequence and flow command sequence Corresponding to According to the operation characteristics of the heat exchange cavity, the temperature and flow rate adjustment sequence is divided into control cycles to generate segmented control instructions. Assuming the control cycle is , the control instructions in each cycle are expressed as:

[0165] ;

[0166] in Indicates the current cycle index. These segmented control instructions provide precise adjustment targets for subsequent execution units. The segmented control instructions are input into the execution unit to control the opening of the inlet temperature control valve and flow control valve of each heat exchange cavity. Assume that the opening of the temperature control valve is , and with The opening of the flow control valve is proportional to , and with The calculation formula of the opening of the regulating valve is:

[0167] ;

[0168] in , is the allowable temperature adjustment range, , The allowable flow adjustment range. After adjusting the valve opening in real time, collect the temperature response data of each heat exchange cavity. , Traffic response data and pressure response data , these data constitute the cavity response dataset:

[0169] Response Dataset ;

[0170] After collecting the cavity response data, the data set is preprocessed, including data filtering and normalization operations. Data filtering eliminates noise interference through the Kalman filter algorithm, and its state update formula is:

[0171] ;

[0172] in is the current state estimate, is the measured value, is the observation matrix, is the Kalman gain. Normalization scales the data to a standard range, for example:

[0173] ;

[0174] The standardized response data after filtering and normalization is substituted into the heat flux distribution mathematical model to identify the model parameters online. Assume that the coefficient matrix of the heat flux distribution mathematical model is And the coupling matrix is , the identification problem is expressed as minimizing the following error function:

[0175] ;

[0176] in is the response data vector, is the temperature data matrix, is the flow data matrix. The identification problem is solved by the least squares method to obtain the model correction parameters and . Using the correction parameters and , update the coefficient matrix and coupling matrix of the heat flux distribution mathematical model:

[0177] ;

[0178] The updated model can more accurately describe the current heat flux distribution state, thereby generating a target heat flux distribution model.

[0179] The above describes the multi-cavity coordinated control method of the microchannel heat exchange module in the embodiment of the present application. The following describes the multi-cavity coordinated control device 10 of the microchannel heat exchange module in the embodiment of the present application. Figure 2 In one embodiment of the present application, a multi-cavity coordinated control device 10 of a microchannel heat exchange module includes:

[0180] An acquisition module 11 is used to perform numerical simulation calculations on multiple heat exchange cavities of the microchannel heat exchange module, obtain temperature distribution data, flow distribution data and pressure distribution data of each heat exchange cavity, and obtain a training data matrix;

[0181] The calculation module 12 is used to input the training data matrix into the multi-layer block separation model to calculate the temperature field coefficient, flow field coefficient and pressure field coefficient and analyze the thermal flow coupling relationship between cavities, and generate a heat flow distribution mathematical model;

[0182] A construction module 13 is used to substitute the heat flow distribution data, temperature distribution data and flow distribution data of each heat exchange cavity into the non-homogeneous heat conduction equation according to the heat flow distribution mathematical model to construct a multi-objective optimization function;

[0183] A processing module 14 is used to perform chaos mapping processing and multi-objective collaborative differential evolution based on a multi-objective optimization function to obtain an optimal control solution set;

[0184] A solution module 15 is used to perform Markov jump system modeling and Q-learning state value function solution on the optimal control solution set to generate cooperative control instructions;

[0185] The updating module 16 is used to input the coordinated control instruction into the execution unit, adjust the inlet temperature parameters and flow parameters of each heat exchange cavity in real time, and update the heat flux distribution mathematical model online to obtain the target heat flux distribution model.

[0186] Through the collaborative cooperation of the above-mentioned components, the temperature field, flow field and pressure field are separately modeled through a multi-layer block algorithm, and combined with the heat-flux coupling analysis, a high-precision mathematical model of heat flux distribution is established, which solves the problem of insufficient prediction accuracy caused by the traditional model not considering the coupling between fields; chaos mapping and multi-objective collaborative differential evolution algorithm are used to optimize the control parameters, which overcomes the defect that the traditional optimization algorithm is prone to fall into the local optimum and obtains a better set of control schemes; the Markov jump system and Q-learning algorithm are introduced to construct an intelligent control strategy, which effectively handles the nonlinear characteristics and uncertainties in the heat exchange process and ensures the stability of the control system; a model correction method based on the online update mechanism is designed, and the heat flux distribution model is dynamically corrected through real-time response data, so that the control strategy can adapt to changes in operating conditions; through the multi-cavity collaborative control mechanism, information interaction and optimization coordination between the heat exchange cavities are realized, which significantly improves the overall heat exchange efficiency and system performance.

[0187] See also Figure 3 , Figure 3 This is a schematic block diagram of the structure of an electronic device 300 provided in an embodiment of the present application. The electronic device 300 includes a processor 301 and a memory 302. The processor 301 and the memory 302 are connected via a device bus 303, wherein the memory 302 may include a non-volatile storage medium and an internal memory.

[0188] 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 301, the processor 301 can execute any of the above-mentioned multi-cavity coordinated control methods of the microchannel heat exchange module.

[0189] The processor 301 is used to provide computing and control capabilities to support the operation of the entire electronic device 300 .

[0190] 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 301, the processor 301 can execute any of the above-mentioned multi-cavity coordinated control methods of the microchannel heat exchange module.

[0191] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a partial structure related to the present application scheme, and does not constitute a limitation on the electronic device 300 involved in the present application scheme. The specific electronic device 300 may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0192] It should be understood that the processor 301 may be a central processing unit (CPU), and the processor 301 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) 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.

[0193] It should be noted that technicians in the relevant field can clearly understand that for the convenience and simplicity of description, the specific working process of the electronic device 300 described above can refer to the corresponding process of the multi-cavity collaborative control method of the aforementioned microchannel heat exchange module, and will not be repeated here.

[0194] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0195] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable an electronic device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.

[0196] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A multi-cavity coordinated control method for a microchannel heat exchange module, characterized in that: The method comprises: Numerical simulation calculations are performed on multiple heat exchange cavities of the microchannel heat exchange module to obtain temperature distribution data, flow distribution data, and pressure distribution data of each heat exchange cavity to obtain a training data matrix; The training data matrix is ​​input into the multi-layer block separation model to calculate the temperature field coefficient, flow field coefficient and pressure field coefficient and analyze the thermal-fluid coupling relationship between cavities to generate a mathematical model of thermal flux distribution; specifically comprising: inputting the training data matrix into the multi-layer block separation model, performing a generalized separable characteristic analysis operation on the temperature field data, flow field data and pressure field data to obtain a basic separation matrix; calculating the temperature field coefficient, flow field coefficient and pressure field coefficient based on the basic separation matrix, and establishing a weight relationship between the temperature field coefficient, the flow field coefficient and the pressure field coefficient to obtain a field coefficient weight matrix ; Based on the field coefficient weight matrix, feature analysis is performed to obtain a field feature identification matrix, and a heat-flow coupling modeling operation is performed on the field feature identification matrix to obtain a heat-flow coupling equation group; based on the heat-flow coupling equation group, the heat flow distribution coefficient between the cavities is solved to obtain a heat flow distribution coefficient matrix, and an energy balance verification operation is performed on the heat flow distribution coefficient matrix to obtain a heat balance constraint condition; the heat balance constraint condition is combined with the heat flow distribution coefficient matrix to obtain a heat flow distribution calculation model, and a verification operation is performed on the heat flow distribution calculation model to output a heat flow distribution mathematical model; According to the heat flux distribution mathematical model, the heat flux distribution data, temperature distribution data and flow distribution data of each heat exchange cavity are substituted into the non-homogeneous heat conduction equation to construct a multi-objective optimization function; specifically, the method comprises: according to the heat flux distribution mathematical model, the non-homogeneous heat conduction equation of each heat exchange cavity is modeled to obtain a basic heat transfer equation; the temperature distribution data and flow distribution data are input into the basic heat transfer equation, and the finite difference method is used for discretization to obtain a discretized heat transfer equation; the discretized heat transfer equation is substituted into the energy conservation equation to obtain a heat conservation constraint, and an efficiency objective function is constructed based on the heat conservation constraint; a posteriori Cramer-Rao lower bound calculation is performed on the efficiency objective function to obtain a performance index function, and the performance index function is weightedly combined with the heat exchange efficiency function to obtain a comprehensive evaluation function; a pressure loss constraint and a heat exchange area constraint are added to the comprehensive evaluation function to obtain a constrained optimization model, and the constrained optimization model is normalized to obtain a multi-objective optimization function; Based on the multi-objective optimization function, chaotic mapping processing and multi-objective collaborative differential evolution are performed to obtain an optimal control solution set; Performing Markov jump system modeling and Q-learning state value function solution on the optimal control scheme set to generate collaborative control instructions; The collaborative control instruction is input into the execution unit, the inlet temperature parameters and flow parameters of each heat exchange cavity are adjusted in real time, and the heat flux distribution mathematical model is updated online to obtain a target heat flux distribution model.

2. The multi-cavity coordinated control method of the microchannel heat exchange module according to claim 1, characterized in that: The numerical simulation calculation is performed on multiple heat exchange cavities of the microchannel heat exchange module to obtain temperature distribution data, flow distribution data and pressure distribution data of each heat exchange cavity, and obtain a training data matrix, including: The microchannel heat exchange module is divided into cavities to obtain N heat exchange cavities, and temperature boundary conditions and pressure boundary conditions are set for the inlet and outlet of each heat exchange cavity to obtain an initial heat exchange cavity model; Performing meshing on the initial heat exchange cavity model, wherein hexahedral mesh parameters are set for the fluid calculation domain and tetrahedral mesh parameters are set for the solid calculation domain, to obtain a calculation mesh model; Setting k-epsilon turbulence model parameters, convection heat transfer coefficient parameters, thermal conductivity parameters and specific heat capacity parameters for the computational grid model, performing fluid-solid coupling boundary setting operations, and obtaining a computational physical model; Based on the computational physics model, pressure-velocity coupling calculation is performed, and steady-state iterative solution operations of temperature field, flow field and pressure field are performed on each grid node to obtain an original data set; The original data set is grouped to obtain temperature field data, flow field data and pressure field data, and the Pearson correlation coefficient is used to calculate the correlation between different field data to obtain a data correlation matrix; Based on the data association matrix, a heat-flow coupling analysis is performed, a heat transfer coefficient between each heat exchange cavity is calculated by using a thermal conductivity method, and a temperature boundary condition and a flow boundary condition between cavities are determined by using a boundary element method to obtain a heat-flow coupling relationship between cavities; A matrix concatenation operation is performed on the heat flow coupling relationship between the cavities and the original data set to obtain a training data matrix.

3. The multi-cavity coordinated control method of the microchannel heat exchange module according to claim 1, characterized in that: The method of performing chaotic mapping processing and multi-objective collaborative differential evolution based on the multi-objective optimization function to obtain an optimal control solution set includes: Performing a Logistic chaotic mapping operation on the multi-objective optimization function to generate M D-dimensional random initial individuals, where M represents the population size and D represents the decision variable dimension, to obtain an initialized population; Calculate the objective function and perform non-dominated sorting on the initialized population, archive and store the non-dominated solutions, and obtain an initial Pareto solution set; A mutation vector is constructed based on the initial Pareto solution set, and a differential mutation operation Vi = Xr1 + F(Xr2 - Xr3) is performed on the parent individuals, where Vi is the mutation individual, Xr1, Xr2, and Xr3 are randomly selected different parent individuals, and F is a scaling factor, to obtain a mutation population; Performing a crossover operation on the variant population and the parent population according to the binomial crossover probability to generate a test population, and performing boundary constraint processing on the test population to obtain a feasible solution population; Performing a dual-population co-evolution operation on the feasible solution population, dividing the population into an exploration population and an exploitation population, and performing information interaction to obtain a co-evolution population; Merging the co-evolving population with the parent population, and sorting and screening the merged population using a fast non-dominated sorting algorithm and crowding calculation to obtain a target population; An elite retention strategy is implemented on the target population, the current optimal individual is directly introduced into the next generation population, and the external archive set is updated at the same time to obtain an evolved population. Convergence judgment is performed based on the evolved population. When the evolutionary generation reaches a set value or the population converges, the iteration is stopped and the optimal control solution set is output.

4. The multi-cavity coordinated control method of the microchannel heat exchange module according to claim 3, characterized in that: The step of performing Markov jump system modeling and Q-learning state value function solving on the optimal control scheme set to generate collaborative control instructions includes: Based on the optimal control solution set, the microchannel heat exchange module is modeled in state space, the temperature distribution, flow distribution and pressure distribution are set as state variables, the inlet temperature and flow of the heat exchange cavity are set as control variables, and the state space equation is obtained; Performing piecewise homogenization processing on the state space equation, dividing the heat transfer process into a plurality of sub-intervals, and introducing a continuous residence time parameter in each sub-interval to obtain a piecewise state equation group; The piecewise state equation group is converted into a linear matrix inequality form, and a Q-learning state value function Q(s,a) is constructed, where s is the system state and a is the control action, to obtain a state value function model; Performing a kernel matrix estimation operation on the state value function model, using a quadratic programming method to solve the minimization problem min||AX-b||, where A is the state transfer matrix, X is the kernel matrix to be solved, and b is the observation vector, to obtain the kernel matrix parameters; Constructing a Q-learning value iteration equation based on the kernel matrix parameters, and performing ε-greedy strategy optimization on the value iteration equation to obtain a strategy optimization model; A closed-loop stability analysis is performed based on the strategy optimization model to obtain stability constraints, and the stability constraints and the strategy optimization model are combined to generate collaborative control instructions for the heat exchange cavity, wherein the collaborative control instructions include temperature adjustment instructions and flow rate adjustment instructions.

5. The multi-cavity coordinated control method of the microchannel heat exchange module according to claim 4, characterized in that: The collaborative control instruction is input into the execution unit, the inlet temperature parameters and flow parameters of each heat exchange cavity are adjusted in real time, and the heat flux distribution mathematical model is updated online to obtain the target heat flux distribution model, including: Parsing the collaborative control instruction into a temperature adjustment instruction sequence and a flow adjustment instruction sequence, and dividing the temperature adjustment instruction sequence and the flow adjustment instruction sequence into control cycles to obtain segmented control instructions; The segmented control instruction is input into the execution unit, the opening of the inlet temperature regulating valve and the flow regulating valve of each heat exchange cavity is controlled to obtain real-time control parameters, and the temperature response data, flow response data and pressure response data of each heat exchange cavity are collected based on the real-time control parameters to obtain a cavity response data set; Performing data filtering and normalization processing on the cavity response data set to obtain standardized response data, and substituting the standardized response data into the heat flux distribution mathematical model to perform online identification of model parameters to obtain model correction parameters; Based on the model correction parameters, the coefficient matrix and coupling matrix of the heat flux distribution mathematical model are updated to obtain a target heat flux distribution model.

6. A multi-cavity coordinated control device for a microchannel heat exchange module, characterized in that: A method for cooperatively controlling a multi-cavity microchannel heat exchange module according to any one of claims 1 to 5, wherein the multi-cavity cooperatively controlling device of the microchannel heat exchange module comprises: An acquisition module is used to perform numerical simulation calculations on multiple heat exchange cavities of the microchannel heat exchange module, obtain temperature distribution data, flow distribution data and pressure distribution data of each heat exchange cavity, and obtain a training data matrix; A calculation module, used for inputting the training data matrix into a multi-layer block separation model to calculate the temperature field coefficient, flow field coefficient and pressure field coefficient and analyze the thermal flow coupling relationship between cavities, and generate a heat flow distribution mathematical model; A construction module is used to substitute the heat flux distribution data, temperature distribution data and flow distribution data of each heat exchange cavity into the non-homogeneous heat conduction equation according to the heat flux distribution mathematical model to construct a multi-objective optimization function; A processing module, used for performing chaos mapping processing and multi-objective collaborative differential evolution based on the multi-objective optimization function to obtain an optimal control solution set; A solution module, used for performing Markov jump system modeling and Q-learning state value function solution on the optimal control scheme set to generate cooperative control instructions; The updating module is used to input the collaborative control instruction into the execution unit, adjust the inlet temperature parameters and flow parameters of each heat exchange cavity in real time, and update the heat flux distribution mathematical model online to obtain the target heat flux distribution model.

7. An electronic device, characterized in that: The electronic device comprises: a memory and at least one processor, wherein instructions are stored in the memory; The at least one processor calls the instructions in the memory so that the electronic device executes the multi-cavity coordinated control method of the microchannel heat exchange module according to any one of claims 1-5.

Images