Thermal management system optimization method and apparatus

Abstract

The application provides a kind of heat management system optimization method and device, it is related to data processing technical field, comprising: based on the working condition parameter of heat management system, initial flow vector for fluid driving element and constraint condition are determined;Wherein, constraint condition includes heat exchange constraint and flow boundary constraint, heat exchange constraint is characterized in that the total heat exchange amount in working cycle is equal to target heat exchange amount;With initial flow vector as iteration initial value, iteratively execute marginal cost coordination operation until the update flow vector obtained by current iteration meets the preset convergence condition, and the update flow vector that meets the preset convergence condition is determined as optimal flow sequence;Input flow vector is initial flow vector at first iteration, and is the update flow vector obtained by previous iteration at subsequent iteration;Control instruction is generated based on optimal flow sequence, and the operating state of fluid driving element is adjusted based on control instruction, to reduce the operating energy consumption of heat management system.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and apparatus for optimizing a thermal management system. Background Technology

[0002] Thermal management systems are core supporting components for high-power load equipment, aerospace equipment, new energy equipment, and high-efficiency power systems. In complex application scenarios such as power batteries and aero engines, the dynamic operation of thermal management systems generally exhibits typical characteristics such as strong nonlinearity, multi-physics coupling, heat transfer delay, temporal state correlation, and multiple constraint boundary limitations. Within a complete operating cycle, the thermal management system needs to achieve optimal control of pump power, energy consumption, or total system weight while meeting constraints such as maximum temperature limits, total heat exchange requirements, and flow rate limits. This is a typical high-dimensional, dynamic, constrained nonlinear optimization problem.

[0003] The results of dynamic optimization can often be used for the optimized control of thermal management systems. The optimized time-varying flow sequence can serve as feedforward commands for the control loop. During actual operation, the control loop adjusts the operating state of the fluid drive components based on the optimized trajectory, thereby reducing energy consumption. However, the control loop has strict requirements for real-time performance; the optimization process must be completed within a finite time. If the solution time is too long, it becomes difficult to support finer time discretization, thus limiting further improvements in the energy efficiency of the thermal management system.

[0004] In related technologies, the dynamic operation optimization of thermal management systems commonly employs general optimization methods such as sequential quadratic programming, interior-point methods, projected gradient methods, penalty function methods, and augmented Lagrange multiplier methods. These optimization methods are designed for general optimization scenarios. However, in high-dimensional discrete-time scenarios, the decision variables are highly dimensional, the number of iterations is large, and the computational load is high, often making it difficult to meet the millisecond to second-level real-time optimization requirements of airborne and vehicle-mounted control systems. Furthermore, these optimization methods are prone to slow convergence, insufficient optimization accuracy, and poor robustness when dealing with characteristics such as heat transfer delays, temporal coupling, and nonlinear heat transfer.

[0005] Therefore, it is evident that the dynamic operation optimization methods for thermal management systems in related technologies struggle to balance high optimization accuracy and high computational efficiency within a limited computational time, thus hindering the real-time dynamic optimization and control capabilities of thermal management systems. Consequently, how to quickly and accurately solve the high-dimensional dynamic operation optimization problem of thermal management systems while satisfying multiple constraints has become a pressing technical problem in this field. Summary of the Invention

[0006] This invention provides a method and apparatus for optimizing a thermal management system, which addresses the problem that existing dynamic operation optimization methods for thermal management systems struggle to balance high optimization accuracy and high computational efficiency within a limited computation time, thus restricting the real-time dynamic optimization and control capabilities of the thermal management system.

[0007] This invention provides a method for optimizing a thermal management system, comprising: Based on the operating parameters of the thermal management system, the initial flow vector and constraints for the fluid drive component are determined; wherein, the constraints include heat transfer constraints and flow boundary constraints, and the heat transfer constraints characterize that the total heat transfer within the working cycle is equal to the target heat transfer. Using the initial flow vector as the initial value for iteration, the marginal cost coordination operation is performed iteratively until the updated flow vector obtained in the current iteration satisfies the preset convergence condition, and the updated flow vector that satisfies the preset convergence condition is determined as the optimal flow sequence. The marginal cost coordination operation includes: calculating the marginal cost corresponding to each discrete time step for the input flow vector in this iteration, coordinating and optimizing the input flow vector based on the marginal cost, and generating an updated flow vector for this iteration by combining the heat transfer constraint and the flow boundary constraint; the marginal cost represents the incremental rate of instantaneous energy consumption relative to the total heat transfer; the input flow vector is the initial flow vector in the first iteration and the updated flow vector obtained in the previous iteration in subsequent iterations; Control commands are generated based on the optimal flow sequence, and the operating state of the fluid drive component is adjusted based on the control commands to reduce the operating energy consumption of the thermal management system.

[0008] Optionally, based on the operating parameters of the thermal management system, the initial flow vector and constraints for the fluid drive component are determined, including: Based on the operating parameters, a dynamic model of the thermal management system is constructed using the dynamic heat flow method; wherein, the dynamic model has a delayed thermodynamic potential characterizing the delay of heat transfer in the pipeline, and the delay time corresponding to the delayed thermodynamic potential changes with the flow velocity. Based on the dynamic model, the optimization objective is to minimize the total energy consumption of the fluid drive during the working cycle, and the initial flow vector is generated.

[0009] Optionally, calculating the marginal cost corresponding to each discrete time step includes: Calculate the instantaneous energy consumption corresponding to each of the discrete time steps; wherein the instantaneous energy consumption is proportional to the drag coefficient of the fluid drive component and the cube of the flow rate within the corresponding discrete time step; Based on the dynamic model, the heat transfer gradient of the total heat transfer relative to the flow rate within each discrete time step is calculated; The marginal cost is determined by the ratio of the partial derivative of the instantaneous energy consumption with respect to the corresponding flow rate to the heat exchange gradient.

[0010] Optionally, the input flow vector is coordinated and optimized based on the marginal cost, and an updated flow vector for this iteration is generated by combining the heat transfer constraint and the flow boundary constraint, including: Based on the consistency assessment results of the marginal cost, variable release or variable freeze operations are performed on the variables in the input flow vector to obtain the set of free variables and the corresponding average marginal cost. The iterative search direction is determined based on the difference between the marginal cost and the average marginal cost of each variable in the set of free variables. Perform a flow scaling operation that satisfies the heat transfer constraint along the iterative search direction, determine the iteration step size, and generate the updated flow vector for this iteration.

[0011] Optionally, based on the consistency assessment result of the marginal cost, variable release or variable freeze operations are performed on the variables in the input flow vector to obtain the set of free variables and the corresponding average marginal cost, including: If the marginal cost of a frozen variable that is at the lower limit of the flow boundary constraint is less than the difference between the average marginal cost and the preset tolerance, the corresponding frozen variable is released to the set of free variables. If the marginal cost of a frozen variable at the upper limit of the flow boundary constraint is greater than the sum of the average marginal cost and the preset tolerance, the corresponding frozen variable is released to the set of free variables. In response to the variable release operation, the average marginal cost is dynamically updated.

[0012] Optionally, a flow scaling operation satisfying the heat transfer constraint is performed along the iterative search direction to determine the iteration step size and generate the updated flow vector for this iteration, including: Solve for the scaling factor used to make the scaled flow vector satisfy the heat transfer constraint; Candidate flow vectors are constructed using the scaling factor, the initial exploration step size, and the iterative search direction; If the decrease in total energy consumption corresponding to the candidate flow vector relative to the total energy consumption of the previous iteration is greater than a preset threshold, the candidate flow vector is determined as the updated flow vector for this iteration.

[0013] Optionally, the determination of the preset convergence condition includes: The maximum absolute deviation between the marginal cost of each variable in the set of free variables and the average marginal cost is determined to obtain the marginal cost consistency error. If the marginal cost consistency error is less than the preset convergence threshold, the preset convergence condition is determined to be met.

[0014] Optionally, the working cycle includes a heat generation stage and a heat dissipation stage; During the heat generation stage, the heat source generates heat with a preset thermal power, and the constraint conditions also include that the maximum temperature of the thermal management system is lower than a preset safe temperature. During the heat dissipation phase, the heat source is turned off; The initial flow vector corresponds to each discrete time step of the heat dissipation stage, and the flow rate within each discrete time step of the heat generation stage remains at the upper limit of the flow boundary constraint.

[0015] Optionally, the thermal management system includes a pump, a heat source, a phase change thermal storage unit, and a radiator, wherein the fluid drive is the pump, and the operating energy consumption of the fluid drive is the pump power. Adjusting the operating state of the fluid drive based on the control command includes adjusting the pump speed or the valve opening in the thermal management system based on the control command.

[0016] The present invention also provides a thermal management system optimization device, comprising: The determination module is used to determine the initial flow vector and constraints for the fluid drive component based on the operating parameters of the thermal management system; wherein the constraints include heat transfer constraints and flow boundary constraints, and the heat transfer constraints represent that the total heat transfer during the working cycle is equal to the target heat transfer. The update module is used to iteratively perform marginal cost coordination operations with the initial flow vector as the initial value of the iteration until the updated flow vector obtained in the current iteration satisfies the preset convergence condition, and the updated flow vector that satisfies the preset convergence condition is determined as the optimal flow sequence. The marginal cost coordination operation includes: calculating the marginal cost corresponding to each discrete time step for the input flow vector in this iteration, coordinating and optimizing the input flow vector based on the marginal cost, and generating an updated flow vector for this iteration by combining the heat transfer constraint and the flow boundary constraint; the marginal cost represents the incremental rate of instantaneous energy consumption relative to the total heat transfer; the input flow vector is the initial flow vector in the first iteration and the updated flow vector obtained in the previous iteration in subsequent iterations; An optimization module is used to generate control commands based on the optimal flow sequence and adjust the operating state of the fluid drive component based on the control commands to reduce the operating energy consumption of the thermal management system.

[0017] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the thermal management system optimization method as described above.

[0018] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the thermal management system optimization method as described above.

[0019] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the thermal management system optimization method as described above.

[0020] This invention provides a method and apparatus for optimizing a thermal management system. By acquiring the operating parameters of the thermal management system, the initial flow vector and constraints for the fluid drive component are determined. Using the initial flow vector as the initial value for iteration, a marginal cost coordination operation is iteratively performed. In each iteration, the marginal cost corresponding to each discrete time step is calculated. Based on the marginal cost, the input flow vector is coordinated and optimized. Then, combined with heat transfer constraints and flow boundary constraints, an updated flow vector is generated until a preset convergence condition is met to obtain the optimal flow sequence. Finally, control commands are generated based on the optimal flow sequence to adjust the operating state of the fluid drive component. This transforms the original high-dimensional, constrained multivariate dynamic optimization problem into an optimization process focused on marginal cost coordination. While strictly satisfying the total heat transfer constraints and flow boundary constraints, the iterative computation is significantly reduced, thereby greatly improving the solution speed while ensuring optimization accuracy. This method can meet the dynamic optimization and control requirements of airborne and vehicle-mounted thermal management systems with strong real-time requirements, and effectively supports more refined time discretization to further improve the energy efficiency of the thermal management system. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0022] Figure 1 This is a flowchart illustrating the thermal management system optimization method provided by the present invention; Figure 2 This is a diagram of the thermal management system architecture provided in an embodiment of the present invention; Figure 3 This is the dynamic heat flow model of the thermal management system provided in the embodiments of the present invention; Figure 4 A flowchart of the calculation method provided by this invention; Figure 5 A schematic diagram of the structure of the thermal management system optimization device provided by the present invention; Figure 6 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0024] The thermal management system optimization method provided in this embodiment of the invention can be executed by a controller, processor, control system, embedded optimization module, or real-time computing unit integrated in an airborne or vehicle-mounted platform.

[0025] For ease of explanation, the following embodiments are described using the controller of the thermal management system as the execution subject. Those skilled in the art will understand that the technical solution can be implemented by replacing the execution subject with a processor or other computing unit.

[0026] Figure 1 This is a flowchart illustrating the thermal management system optimization method provided by the present invention, as shown below. Figure 1 As shown, the method includes the following: Step 110: Based on the operating parameters of the thermal management system, determine the initial flow vector and constraints for the fluid drive component; wherein, the constraints include heat transfer constraints and flow boundary constraints, and the heat transfer constraints characterize that the total heat transfer within the working cycle is equal to the target heat transfer. Step 120: Using the initial flow vector as the initial value for iteration, perform the marginal cost coordination operation iteratively until the updated flow vector obtained in the current iteration satisfies the preset convergence condition, and determine the updated flow vector that satisfies the preset convergence condition as the optimal flow sequence. The marginal cost coordination operation includes: calculating the marginal cost corresponding to each discrete time step for the input flow vector in this iteration, coordinating and optimizing the input flow vector based on the marginal cost, and generating an updated flow vector for this iteration by combining the heat transfer constraint and the flow boundary constraint; the marginal cost represents the incremental rate of instantaneous energy consumption relative to the total heat transfer; the input flow vector is the initial flow vector in the first iteration and the updated flow vector obtained in the previous iteration in subsequent iterations; Step 130: Generate control commands based on the optimal flow sequence, and adjust the operating state of the fluid drive unit based on the control commands to reduce the operating energy consumption of the thermal management system.

[0027] In this application, the controller first acquires the operating parameters of the thermal management system. These operating parameters refer to a set of data describing the input and boundary conditions of the thermal management system during operation, and may include inlet temperature and flow rate on the environmental side, heat generation power of the heat source, physical property data of the working fluid, and energy consumption characteristic parameters of the fluid drive components, etc.

[0028] Operating parameters can be obtained by reading pre-configured parameters from a host computer or storage unit, by real-time acquisition through temperature and flow sensors deployed in the thermal management system, or by combining real-time acquired values ​​with pre-stored calibration values ​​and then correcting them.

[0029] Fluid drive components are parts that provide power for the flow of working fluid in a thermal management system. In different application scenarios, they can be pumps, fans, or other drive elements.

[0030] Based on the acquired operating parameters, the controller determines the initial flow vector and constraints for the fluid drive components.

[0031] The initial flow vector refers to a set of values ​​representing the flow rates of the working fluid driven by the fluid actuator at each discrete time step after discretizing the working cycle over time, arranged sequentially. The flow components corresponding to each discrete time step are the decision variables to be determined, denoted as... ; Where n is the number of discrete time steps. Constraints are used to define the feasible range of the flow vector, including both heat transfer constraints and flow boundary constraints, and can be expressed as follows: ; In the formula, Q is the heat transfer function, representing the total amount of heat dissipated to the environment during the working cycle. To exchange heat for the target and Let be the lower and upper bounds of the flow rate at each discrete time step, respectively, and satisfy 0 < 1. < The heat transfer constraint characterizes the total heat transfer within the working cycle as equal to the target heat transfer. The purpose of determining this constraint is to ensure that the heat generated by the heat source can be completely processed.

[0032] The flow boundary constraint indicates that the flow rate at each discrete time step is neither lower than the lower bound nor higher than the upper bound. The purpose of this constraint is to ensure that the flow rate conforms to the physical operating capabilities of the fluid drive components and the pipeline. The optimization objective is to minimize the operating energy consumption of the fluid drive components within their operating cycle, which can be written as: ; In the formula P Let be the energy consumption function. Pi For the first i Instantaneous energy consumption over a discrete time step Δt Let be the time interval. To ensure that the flow vector satisfies the flow boundary constraints, a clipping mapping is defined as . ; Mapping can confine any flow component to a lower and upper bound. When determining the initial flow vector, a set of initial values ​​can be chosen. Then, the scalar equation is solved using the bisection method. ; Get scaling factor β(0) and order ; This ensures that the initial flow vector satisfies the flow boundary constraints while the corresponding total heat exchange equals the target heat exchange.

[0033] Then, the controller uses the initial flow vector as the initial value for iteration and iteratively performs marginal cost coordination operations until the updated flow vector obtained in the current iteration satisfies the preset convergence condition, and determines the updated flow vector that satisfies the preset convergence condition as the optimal flow sequence.

[0034] An iterative optimization approach is adopted to gradually improve traffic allocation. Each iteration is adjusted based on the results of the previous iteration, so that the traffic vector gradually moves towards a direction with lower energy consumption while still satisfying the constraints.

[0035] The marginal cost reconciliation operation is the core processing step in this embodiment. The following operations are performed on the input flow vector for this iteration.

[0036] The controller first calculates the marginal cost for each discrete time step. Marginal cost represents the incremental rate of change of instantaneous energy consumption relative to total heat exchange, and is defined as follows: Marginal cost refers to the instantaneous energy consumption change caused by a unit change in heat exchange at a specific discrete time step. Since the impact of flow rate changes on energy consumption and heat exchange varies at different discrete time steps, marginal cost can be used to measure at which time steps it is more cost-effective to increase the flow rate and at which time steps it is more reasonable to decrease the flow rate.

[0037] The controller then coordinates and optimizes the input flow vector based on the calculated marginal cost, and then generates the updated flow vector for this iteration by combining heat transfer constraints and flow boundary constraints.

[0038] When the marginal costs of different discrete time steps are inconsistent, shifting the flow from the time step with higher marginal cost to the time step with lower marginal cost can reduce the total energy consumption while keeping the total heat exchange constant.

[0039] By redistributing traffic in this way, the marginal costs of each time step are gradually made to converge, thereby approaching the optimal traffic allocation in terms of energy consumption.

[0040] During the generation and updating of the flow vector, the controller adjusts the flow rate in conjunction with heat transfer constraints so that the total heat transfer corresponding to the adjusted flow vector is still equal to the target heat transfer. At the same time, it limits the flow rate that exceeds the upper and lower limits in conjunction with the flow boundary constraints, so that each flow component falls back to the feasible range.

[0041] It should be noted that the input flow vector is the initial flow vector in the first iteration, and the updated flow vector obtained from the previous iteration in subsequent iterations.

[0042] The input for each round of marginal cost coordination operation is the output of the previous round, thus forming a smooth flow of data between iterations and ensuring that the optimization process progresses round by round.

[0043] The preset convergence condition is used to determine whether the iteration can be terminated. It means that the optimization process is considered converged when the marginal costs at each time step tend to be consistent and the improvement from continued iteration is sufficiently small. When the updated flow vector meets the preset convergence condition, the controller determines this updated flow vector as the optimal flow sequence; when the updated flow vector does not meet the preset convergence condition, the controller uses this updated flow vector as the input flow vector for the next iteration and continues to perform marginal cost coordination operations, repeating this cycle until the preset convergence condition is met.

[0044] Finally, the controller generates control commands based on the obtained optimal flow sequence and adjusts the operating state of the fluid drive components accordingly to reduce the energy consumption of the thermal management system. The optimal flow sequence provides the appropriate flow rate value at each discrete time step, and the controller generates corresponding control commands based on this. These control commands can be used as feedforward commands.

[0045] The method of adjusting the operating state of the fluid drive based on control commands can be to adjust the output power of the fluid drive or to adjust the flow regulation elements in the relevant flow path, so that the actual flow rate of the working fluid driven by the fluid drive tracks the optimal flow sequence, thereby reducing the operating energy consumption while meeting the heat exchange requirements and operating constraints.

[0046] As can be seen from the above processing, this embodiment combines marginal cost coordination optimization with the actual operation and control of the thermal management system, transforming the high-dimensional constrained flow optimization problem into a stepwise approximation of marginal cost consistency, and using the optimization results for the actual adjustment of the driving components. This scheme significantly improves computational efficiency while maintaining optimization accuracy, and can meet the needs of dynamic optimization and control of the thermal management system in strong real-time scenarios.

[0047] Optionally, calculating the marginal cost corresponding to each discrete time step includes: Calculate the instantaneous energy consumption corresponding to each of the discrete time steps; wherein the instantaneous energy consumption is proportional to the drag coefficient of the fluid drive component and the cube of the flow rate within the corresponding discrete time step; Based on the dynamic model, the heat transfer gradient of the total heat transfer relative to the flow rate within each discrete time step is calculated; The marginal cost is determined by the ratio of the partial derivative of the instantaneous energy consumption with respect to the corresponding flow rate to the heat exchange gradient.

[0048] In this application, the controller first calculates the instantaneous energy consumption corresponding to each discrete time step. In this embodiment, the instantaneous energy consumption is proportional to the drag coefficient of the fluid drive component and the cube of the flow rate within the corresponding discrete time step. The energy consumption target can be specifically written as: ; In the formula, c is the drag coefficient, which characterizes the magnitude of the resistance encountered by the working medium when it flows in the flow path. The larger the flow rate and the larger the drag coefficient, the higher the instantaneous energy consumption required to maintain that flow rate. Since instantaneous energy consumption increases cubically with flow rate, rationally allocating the flow rate between different time steps is of great significance for reducing total energy consumption.

[0049] The controller then calculates the heat transfer gradient of the total heat transfer relative to the flow rate at each discrete time step based on the dynamic model. The heat transfer gradient is denoted as the partial derivative of the function Q with respect to each component direction. ; It reflects the change in total heat exchange during the working cycle when the flow rate is changed slightly at a certain discrete time step.

[0050] Because dynamic models incorporate delay characteristics and time-series coupling, changes in flow rate at a given time step not only affect heat transfer at the current moment but also influence heat transfer at subsequent moments through delayed thermodynamic potential. Therefore, the heat transfer gradient can be comprehensively calculated using the dynamic model to account for this influence. In practical calculations, the heat transfer gradient can be obtained using forward differencing. ; In the formula hFor a relatively small disturbance, For the first i Unit vectors in each direction.

[0051] The controller then determines the marginal cost as the ratio of the partial derivative of instantaneous energy consumption with respect to the corresponding flow rate to the heat transfer gradient. ; Partial derivative of instantaneous energy consumption with respect to the corresponding flow rate The heat transfer gradient characterizes the change in energy consumption caused by a change in unit flow rate. The change in heat exchange caused by a unit change in flow rate is used to characterize the change in heat exchange. By comparing the two, we can obtain the change in energy consumption corresponding to the unit change in heat exchange.

[0052] When instantaneous energy consumption is proportional to the cube of flow rate, the partial derivative of instantaneous energy consumption with respect to flow rate is related to the square of flow rate, thus enabling the marginal cost to be directly calculated from the flow rate and heat exchange gradient.

[0053] This embodiment combines instantaneous energy consumption, resistance coefficient, flow rate cube relationship, and heat transfer gradient calculated by dynamic model to determine marginal cost, so that marginal cost can accurately reflect the actual relationship between energy consumption and heat transfer at each time step.

[0054] Optionally, the input flow vector is coordinated and optimized based on the marginal cost, and an updated flow vector for this iteration is generated by combining the heat transfer constraint and the flow boundary constraint, including: Based on the consistency assessment results of the marginal cost, variable release or variable freeze operations are performed on the variables in the input flow vector to obtain the set of free variables and the corresponding average marginal cost. The iterative search direction is determined based on the difference between the marginal cost and the average marginal cost of each variable in the set of free variables. Perform a flow scaling operation that satisfies the heat transfer constraint along the iterative search direction, determine the iteration step size, and generate the updated flow vector for this iteration.

[0055] In this application, the controller first performs variable release or variable freeze operations on the variables in the input flow vector based on the consistency assessment results of the marginal cost, so as to obtain the set of free variables and the corresponding average marginal cost.

[0056] At point At this point, the free variable index set is defined as the set of components whose values ​​fall between the upper and lower bounds. ; The set of frozen variable indicators is the complement of the set of free variables. ; Variable freezing refers to temporarily fixing flow components that are at the upper or lower limits of flow boundary constraints, preventing them from participating in the current round of flow adjustment; variable releasing refers to bringing previously frozen flow components back into the adjustable range. For a set of free variables, its average marginal cost is defined as the average of the marginal costs of all components in the set. ; Average marginal cost is used as a benchmark to measure the marginal cost of each flow component.

[0057] The controller then determines the iterative search direction based on the difference between the marginal cost and the average marginal cost of each variable in the set of free variables. For each component belonging to the set of free variables, the difference in marginal cost is first calculated. Then calculate the flow increment. Based on this, a preliminary direction vector is defined, with the flow increments taken for components belonging to the set of free variables and the zeros taken for components belonging to the set of frozen variables. To ensure that the total heat exchange remains constant during flow rate adjustments, the projection coefficient is calculated. And construct the final search direction. The iterative search direction is used to indicate whether each flow component should be increased or decreased in this iteration and the relative magnitude of the adjustment.

[0058] For flow components with marginal costs higher than average marginal costs, appropriately reducing their flow rate is beneficial to reducing total energy consumption; for flow components with marginal costs lower than average marginal costs, appropriately increasing their flow rate is beneficial to reducing total energy consumption while maintaining total heat exchange.

[0059] The controller then performs a flow scaling operation that satisfies the heat transfer constraints along the iterative search direction, determines the iteration step size, and generates the updated flow vector for this iteration.

[0060] Flow scaling refers to adjusting the flow rate along the search direction while appropriately scaling the overall flow rate so that the total heat transfer corresponding to the adjusted flow rate vector still satisfies the heat transfer constraint. The iteration step size controls the overall magnitude of the flow rate adjustment in this round. Too large a step size may cause the optimum to be missed, while too small a step size will result in slow convergence. Therefore, it is necessary to select an appropriate iteration step size.

[0061] This embodiment distinguishes adjustable flow components by releasing and freezing variables, constructs a search direction based on the difference between marginal cost and average marginal cost, and generates an updated flow vector by combining flow scaling that satisfies heat transfer constraints, so that each iteration effectively reduces energy consumption while maintaining constraints.

[0062] Optionally, based on the consistency assessment result of the marginal cost, variable release or variable freeze operations are performed on the variables in the input flow vector to obtain the set of free variables and the corresponding average marginal cost, including: If the marginal cost of a frozen variable that is at the lower limit of the flow boundary constraint is less than the difference between the average marginal cost and the preset tolerance, the corresponding frozen variable is released to the set of free variables. If the marginal cost of a frozen variable at the upper limit of the flow boundary constraint is greater than the sum of the average marginal cost and the preset tolerance, the corresponding frozen variable is released to the set of free variables. In response to the variable release operation, the average marginal cost is dynamically updated.

[0063] In this application, when the controller performs a variable release operation, it makes the following judgment for the flow components at the boundary.

[0064] When the frozen variables at the lower bound of the flow boundary constraint satisfy Furthermore, its marginal cost is significantly low, that is... When this happens, the controller releases the corresponding frozen variables into the set of free variables.

[0065] This situation illustrates that although the flow component is already at its lower limit, increasing it can help reduce total energy consumption. When the frozen variable at the upper limit of the flow boundary constraint satisfies... Moreover, its marginal cost is significantly high, that is When this happens, the controller releases the corresponding frozen variables into the set of free variables.

[0066] This situation illustrates that although the flow component is already at its upper limit, reducing the flow component can help reduce the total energy consumption.

[0067] In the formula The preset tolerance is a margin value used to avoid frequent release and freezing of marginal costs near the baseline. Introducing this margin helps to improve the stability of the iteration process.

[0068] The controller dynamically updates the average marginal cost in response to variable release operations. Whenever a frozen variable is released into the set of free variables, the composition of the free variable set changes, and the corresponding average marginal cost changes. This also changes accordingly, so the average marginal cost needs to be recalculated in the form of the average value given in the above embodiments to keep it consistent with the current set of free variables.

[0069] Dynamic updates ensure that subsequent release decisions and search direction calculations are based on the latest benchmark.

[0070] This embodiment introduces a preset tolerance and combines it with the marginal cost of the boundary flow component to make a release judgment, and then dynamically updates the average marginal cost, making the variable release and freezing process more stable and reasonable, which is conducive to the smooth convergence of the iteration process.

[0071] Optionally, a flow scaling operation satisfying the heat transfer constraint is performed along the iterative search direction to determine the iteration step size and generate the updated flow vector for this iteration, including: Solve for the scaling factor used to make the scaled flow vector satisfy the heat transfer constraint; Candidate flow vectors are constructed using the scaling factor, the initial exploration step size, and the iterative search direction; If the decrease in total energy consumption corresponding to the candidate flow vector relative to the total energy consumption of the previous iteration is greater than a preset threshold, the candidate flow vector is determined as the updated flow vector for this iteration.

[0072] In this application, the controller first solves for the scaling factor used to make the scaled flow vector satisfy the heat transfer constraint.

[0073] Let the iteration step size α Take as the initial exploration step size α (0) Solving for the scaling factor using the bisection method scalar equations The scaling factor is the proportional value used to scale the flow vector adjusted along the search direction. Since the total heat exchange may deviate from the target heat exchange after adjusting the flow along the search direction, a scaling factor is applied to the overall flow to make the total heat exchange corresponding to the scaled flow vector equal to the target heat exchange again.

[0074] The controller then uses the obtained scaling factor, initial exploration step size, and iterative search direction to construct candidate flow vectors, and determines whether the decrease in total energy consumption corresponding to the candidate flow vectors relative to the total energy consumption of the previous iteration meets the requirements, i.e., it determines... Whether it is true or not. In the formula... A preset threshold is used to define the acceptable level of energy efficiency improvement.

[0075] When the above equation holds true, it indicates that adjusting by the current step size has resulted in sufficient energy savings. The controller accepts the current step size and records it as... α (k) Then, generate the updated flow vector for this iteration using the following formula. When the above formula does not hold, the controller reduces the detection step size by setting α=ρ·α and then recalculates the scaling factor and evaluates again until an updated flow vector that meets the energy consumption reduction requirement is obtained, where ρ is an attenuation coefficient less than 1.

[0076] By judging the reduction in energy consumption, the iteration step size can be adaptively matched to the optimization requirements of the current iteration point.

[0077] This embodiment ensures that the heat transfer constraints are still met after the flow rate is adjusted by solving the scaling factor, and adaptively determines the iteration step size by evaluating the energy consumption reduction, so that each iteration can achieve stable energy consumption improvement while meeting the constraints.

[0078] Optionally, the determination of the preset convergence condition includes: The maximum absolute deviation between the marginal cost of each variable in the set of free variables and the average marginal cost is determined to obtain the marginal cost consistency error. If the marginal cost consistency error is less than the preset convergence threshold, the preset convergence condition is determined to be met.

[0079] In this application, the controller determines the maximum absolute deviation between the marginal cost and the average marginal cost of each variable in the set of free variables, thereby obtaining the marginal cost consistency error. Marginal cost consistency error is used to measure the degree of convergence between the marginal costs of the current flow components.

[0080] When the marginal cost of each flow component is closer to the average marginal cost, the maximum absolute deviation is smaller, indicating that the flow allocation is closer to the energy-optimal state.

[0081] The controller satisfies the condition when the marginal cost consistency error is less than the preset convergence threshold. Determine if the preset convergence condition is met, and output the optimal solution. In the formula The preset convergence threshold is a threshold value used to define whether the degree of convergence of marginal costs is sufficient.

[0082] When the marginal cost consistency error is not less than the preset convergence threshold, the controller instructs... k=k+1 The updated flow vector is then used as the input for the next iteration to continue the marginal cost reconciliation operation.

[0083] This embodiment uses the maximum absolute deviation between the marginal cost of the free variable set and the average marginal cost as the convergence criterion, so that the convergence judgment directly reflects the optimization goal of marginal cost consistency, which is conducive to terminating the iteration in a timely manner while ensuring the optimization accuracy.

[0084] Optionally, the working cycle includes a heat generation stage and a heat dissipation stage; During the heat generation stage, the heat source generates heat with a preset thermal power, and the constraint conditions also include that the maximum temperature of the thermal management system is lower than a preset safe temperature. During the heat dissipation phase, the heat source is turned off; The initial flow vector corresponds to each discrete time step of the heat dissipation stage, and the flow rate within each discrete time step of the heat generation stage remains at the upper limit of the flow boundary constraint.

[0085] In this embodiment, the working cycle includes a heat generation phase and a heat dissipation phase. The heat generation phase refers to the period when the heat source is in operation and inputs heat into the thermal management system, while the heat dissipation phase refers to the period when the heat source stops operating and the thermal management system dissipates the previously accumulated heat.

[0086] During the heat generation phase, the heat source generates heat at a preset thermal power. To ensure that the system temperature remains within a safe range, the constraint also includes that the maximum temperature of the thermal management system is lower than the preset safe temperature.

[0087] The preset safe temperature is an upper limit set to protect system components and ensure operational safety. The introduction of the maximum temperature constraint requires that the flow rate during the heat generation phase be maintained at a level sufficient to suppress temperature rise. During the heat dissipation phase, the heat source is shut off, and the thermal management system primarily dissipates the previously stored heat.

[0088] Furthermore, the initial flow vector corresponds to each discrete time step in the heat dissipation stage, and the flow rate within each discrete time step in the heat generation stage remains at the upper limit Ui of the flow boundary constraint.

[0089] The heat generation stage requires a high flow rate to meet the maximum temperature constraint. Setting this flow rate as an upper limit ensures temperature safety and concentrates the optimization variables on the heat dissipation stage, thus reducing the dimensionality of the optimization problem. Based on this, the controller mainly optimizes the flow rate in the heat dissipation stage by coordinating marginal costs.

[0090] This embodiment distinguishes between the heat generation stage and the heat dissipation stage, fixes the flow rate of the heat generation stage to an upper limit, and optimizes only the flow rate of the heat dissipation stage. This reduces the number of decision variables while ensuring temperature safety, which helps to improve optimization efficiency.

[0091] Optionally, the thermal management system includes a pump, a heat source, a phase change thermal storage unit, and a radiator, wherein the fluid drive is the pump, and the operating energy consumption of the fluid drive is the pump power. Adjusting the operating state of the fluid drive based on the control command includes adjusting the pump speed or the valve opening in the thermal management system based on the control command.

[0092] Figure 2 This is a diagram of the thermal management system architecture provided in an embodiment of the present invention. Figure 3 This is the dynamic heat flow model of the thermal management system provided in the embodiments of the present invention.

[0093] like Figure 2As shown, the thermal management system in this embodiment includes a pump, a heat source, a phase change thermal storage unit, and a radiator. The pump, as a fluid drive, drives the cooling medium to circulate within the system; the heat source generates heat during operation; the phase change thermal storage unit stores a portion of the heat; and the radiator dissipates the heat to the environment. In this system, the fluid drive is the pump, and the operating energy consumption of the fluid drive is the pump's pumping power.

[0094] In this scenario, the system operates on a continuous cycle. At the beginning of the cycle, the heat source starts and generates heat with a certain thermal power. The pump drives the cooling medium to transport the heat, with a portion stored in the phase change thermal storage unit and the remainder dissipated to the environment through the radiator. At the end of the cycle, the heat source shuts off, and the pump drives the cooling medium to dissipate the heat stored in the phase change thermal storage unit to the environment through the radiator. When the cooling medium flow rate is discretized over time, if the time step is a small fixed interval, the dimensionality of the decision variables is high, constituting a high-dimensional constrained optimization problem. The controller, following the processing procedure of the aforementioned embodiment, performs marginal cost coordination optimization on the cooling medium flow rate sequence to minimize the pump power target. To optimize the objective, under the premise that the total heat exchange equals the target heat exchange and the upper and lower limits of the flow rate are satisfied, the optimal flow rate sequence that minimizes the pump work is obtained. .

[0095] like Figure 3 As shown, when constructing the dynamic model of this system using the dynamic heat flow method, the phase change thermal storage unit can be equivalently represented as a counter-current heat exchanger with an infinite heat flow on one side. The delayed effect in the pipes and heat exchanger channels is characterized by the delayed thermodynamic potential, and the delay time corresponding to the delayed thermodynamic potential is calculated in conjunction with the flow rate change. With the help of this dynamic model, the controller can calculate the heat transfer gradient of the total heat transfer with the flow rate at each time step by forward difference, and determine the marginal cost at each time step by combining the partial derivative of instantaneous energy consumption with respect to flow rate.

[0096] After obtaining the optimal flow sequence, the controller adjusts the operating state of the fluid drive components based on control commands. Specifically, this can involve adjusting the pump speed or the valve opening in the thermal management system. Adjusting the pump speed changes the flow rate of the cooling medium driven by the pump, while adjusting the valve opening changes the flow distribution in each flow path. This allows the actual operating flow rate to track the optimal flow sequence, thus reducing pump power. The time-varying flow value obtained from the optimal flow sequence can be used as feedforward commands for control. In actual operation, the control system adjusts the pump speed or valve opening based on this optimal trajectory.

[0097] To illustrate the effectiveness of this scheme, the optimality and computational efficiency of the proposed optimization algorithm based on the equal incremental criterion are verified by comparing it with generalized optimization algorithms. The algorithms compared include sequential quadratic programming, interior point method, projected gradient method, penalty function method, and augmented Lagrange multiplier method. All optimization calculations were performed on the same computer, and the algorithm was implemented using Python 3.11.4.

[0098] The optimized operating parameters are shown in Table 1. The optimization objective is to minimize the total pump work, and the constraints include the total heat exchange being equal to the target value (1 MJ) and the upper and lower limits of the flow rate (minimum flow rate 0.2 kg / s, maximum flow rate 1 kg / s).

[0099] Table 1 Operating Parameter Information of Thermal Management System

[0100] More specifically, sequential quadratic programming performs best in both optimality and computational speed. The proposed algorithm based on the equal incremental criterion achieves an optimal pumping power similar to that of sequential quadratic programming, but with significantly reduced computation time, requiring only 1.79 seconds, a 91.49% reduction compared to the 21.03 seconds of sequential quadratic programming. The optimization results of the interior point method, projected gradient method, penalty function method, and augmented Lagrange multiplier method are slightly worse, and their computation time is longer.

[0101] Both algorithms yielded total heat exchange that strictly satisfied the constraint condition (1 MJ), with a relative deviation of 0.08% for total pump work, a maximum relative deviation of 4.11% for flow rate at each time step, and flow rate deviations of less than 1% at other time steps. This indicates that the proposed algorithm achieves an order-of-magnitude improvement in computational efficiency while obtaining an optimal solution similar to that of sequential quadratic programming.

[0102] In one alternative embodiment, Figure 4 A flowchart illustrating the calculation method provided by this invention. Figure 4 As shown, the processing logic of the thermal management system optimization in this embodiment during the iterative optimization process can be described as follows.

[0103] The process begins at the initial stage, where the controller first performs initialization processing to determine the initial flow vector and algorithm parameters for the fluid drive components, and uses the initial flow vector as the input flow vector for the first iteration.

[0104] The controller then enters the iterative phase of marginal cost coordination. In each iteration, the controller first performs gradient calculation to obtain the heat transfer gradient of the total heat transfer within the working cycle relative to the flow rate in each discrete time step, and determines the marginal cost corresponding to each discrete time step accordingly. The controller then executes variable release and freeze, adjusting the flow components at the upper or lower limit between freeze and release based on the relationship between the marginal cost of each flow component and the average marginal cost of the free variable set, thus obtaining the current free variable set.

[0105] After obtaining the set of free variables, the controller determines the iteration direction and iteration step size. It constructs the iteration search direction based on the difference between the marginal cost and the average marginal cost of each flow component within the free variable set, and determines the iteration step size for this round by combining heat transfer constraints and energy consumption reduction requirements. The controller then updates the iteration points, adjusting and scaling the flow along the iteration search direction and according to the iteration step size to generate the updated flow vector for this iteration.

[0106] The controller then enters the convergence judgment stage to determine whether the updated flow vector meets the preset convergence conditions. If the judgment result is negative, meaning the preset convergence conditions have not yet been met, the process returns to the gradient calculation stage along the branch extending laterally from the diamond-shaped judgment box, using the updated flow vector as the input flow vector for the next iteration to continue iterative optimization. If the judgment result is positive, meaning the preset convergence conditions have been met, the process enters the termination stage along the path extending from the bottom of the diamond-shaped judgment box. The controller determines the current updated flow vector as the optimal flow sequence and outputs it, terminating the iteration process.

[0107] pass Figure 4 The processing flow shown in this embodiment organizes initialization, gradient calculation, variable release and freezing, determination of iteration direction and step size, iteration point update and convergence judgment into a closed-loop iterative process that advances round by round. This allows the flow allocation to gradually approach the energy-optimal state where the marginal cost of each time step is consistent, under the premise of satisfying the constraints, thereby improving the computational efficiency while ensuring the optimization accuracy.

[0108] Figure 5 A schematic diagram of the structure of the thermal management system optimization device provided by the present invention is shown below. Figure 5 As shown, it includes: The determination module 510 is used to determine the initial flow vector and constraints for the fluid drive component based on the operating parameters of the thermal management system; wherein, the constraints include heat transfer constraints and flow boundary constraints, and the heat transfer constraints characterize that the total heat transfer in the working cycle is equal to the target heat transfer. The update module 520 is used to iteratively perform marginal cost coordination operation with the initial flow vector as the initial value of the iteration until the updated flow vector obtained in the current iteration satisfies the preset convergence condition, and the updated flow vector that satisfies the preset convergence condition is determined as the optimal flow sequence. The marginal cost coordination operation includes: calculating the marginal cost corresponding to each discrete time step for the input flow vector in this iteration, coordinating and optimizing the input flow vector based on the marginal cost, and generating an updated flow vector for this iteration by combining the heat transfer constraint and the flow boundary constraint; the marginal cost represents the incremental rate of instantaneous energy consumption relative to the total heat transfer; the input flow vector is the initial flow vector in the first iteration and the updated flow vector obtained in the previous iteration in subsequent iterations; The optimization module 530 is used to generate control commands based on the optimal flow sequence and adjust the operating state of the fluid drive component based on the control commands to reduce the operating energy consumption of the thermal management system.

[0109] Figure 6 This is a schematic diagram of the structure of the electronic device provided by the present invention, such as... Figure 6 As shown, the electronic device may include a processor 610, a communications interface 620, a memory 630, and a communication bus 640, wherein the processor 610, communications interface 620, and memory 630 communicate with each other via the communication bus 640. The processor 610 can call logical instructions in the memory 630 to execute a thermal management system optimization method. This method includes: determining an initial flow vector and constraints for the fluid drive component based on the operating parameters of the thermal management system; wherein the constraints include heat transfer constraints and flow boundary constraints, and the heat transfer constraints characterize the total heat transfer within the operating cycle as equal to the target heat transfer. Using the initial flow vector as the initial value for iteration, the marginal cost coordination operation is performed iteratively until the updated flow vector obtained in the current iteration satisfies the preset convergence condition, and the updated flow vector that satisfies the preset convergence condition is determined as the optimal flow sequence. The marginal cost coordination operation includes: calculating the marginal cost corresponding to each discrete time step for the input flow vector in this iteration, coordinating and optimizing the input flow vector based on the marginal cost, and generating an updated flow vector for this iteration by combining the heat transfer constraint and the flow boundary constraint; the marginal cost represents the incremental rate of instantaneous energy consumption relative to the total heat transfer; the input flow vector is the initial flow vector in the first iteration and the updated flow vector obtained in the previous iteration in subsequent iterations; Control commands are generated based on the optimal flow sequence, and the operating state of the fluid drive component is adjusted based on the control commands to reduce the operating energy consumption of the thermal management system.

[0110] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0111] On the other hand, the present invention also provides a computer program product, the computer program product including a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the thermal management system optimization method provided by the above methods. The method includes: determining an initial flow vector and constraints for the fluid drive component based on the operating parameters of the thermal management system; wherein the constraints include heat transfer constraints and flow boundary constraints, and the heat transfer constraints characterize the total heat transfer within the working cycle as equal to the target heat transfer. Using the initial flow vector as the initial value for iteration, the marginal cost coordination operation is performed iteratively until the updated flow vector obtained in the current iteration satisfies the preset convergence condition, and the updated flow vector that satisfies the preset convergence condition is determined as the optimal flow sequence. The marginal cost coordination operation includes: calculating the marginal cost corresponding to each discrete time step for the input flow vector in this iteration, coordinating and optimizing the input flow vector based on the marginal cost, and generating an updated flow vector for this iteration by combining the heat transfer constraint and the flow boundary constraint; the marginal cost represents the incremental rate of instantaneous energy consumption relative to the total heat transfer; the input flow vector is the initial flow vector in the first iteration and the updated flow vector obtained in the previous iteration in subsequent iterations; Control commands are generated based on the optimal flow sequence, and the operating state of the fluid drive component is adjusted based on the control commands to reduce the operating energy consumption of the thermal management system.

[0112] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the thermal management system optimization method provided by the above methods. The method includes: determining an initial flow vector and constraints for a fluid drive component based on the operating parameters of the thermal management system; wherein the constraints include heat transfer constraints and flow boundary constraints, and the heat transfer constraints characterize the total heat transfer within the working cycle as equal to the target heat transfer. Using the initial flow vector as the initial value for iteration, the marginal cost coordination operation is performed iteratively until the updated flow vector obtained in the current iteration satisfies the preset convergence condition, and the updated flow vector that satisfies the preset convergence condition is determined as the optimal flow sequence. The marginal cost coordination operation includes: calculating the marginal cost corresponding to each discrete time step for the input flow vector in this iteration, coordinating and optimizing the input flow vector based on the marginal cost, and generating an updated flow vector for this iteration by combining the heat transfer constraint and the flow boundary constraint; the marginal cost represents the incremental rate of instantaneous energy consumption relative to the total heat transfer; the input flow vector is the initial flow vector in the first iteration and the updated flow vector obtained in the previous iteration in subsequent iterations; Control commands are generated based on the optimal flow sequence, and the operating state of the fluid drive component is adjusted based on the control commands to reduce the operating energy consumption of the thermal management system.

[0113] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0114] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0115] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for optimizing a thermal management system, characterized in that, include: Based on the operating parameters of the thermal management system, the initial flow vector and constraints for the fluid drive component are determined; wherein, the constraints include heat transfer constraints and flow boundary constraints, and the heat transfer constraints characterize that the total heat transfer within the working cycle is equal to the target heat transfer. Using the initial flow vector as the initial value for iteration, the marginal cost coordination operation is performed iteratively until the updated flow vector obtained in the current iteration satisfies the preset convergence condition, and the updated flow vector that satisfies the preset convergence condition is determined as the optimal flow sequence. The marginal cost coordination operation includes: calculating the marginal cost corresponding to each discrete time step for the input flow vector in this iteration, coordinating and optimizing the input flow vector based on the marginal cost, and generating an updated flow vector for this iteration by combining the heat transfer constraint and the flow boundary constraint; the marginal cost represents the incremental rate of instantaneous energy consumption relative to the total heat transfer; the input flow vector is the initial flow vector in the first iteration and the updated flow vector obtained in the previous iteration in subsequent iterations; Control commands are generated based on the optimal flow sequence, and the operating state of the fluid drive component is adjusted based on the control commands to reduce the operating energy consumption of the thermal management system.

2. The thermal management system optimization method according to claim 1, characterized in that, Based on the operating parameters of the thermal management system, the initial flow vector and constraints for the fluid drive components are determined, including: Based on the operating parameters, a dynamic model of the thermal management system is constructed using the dynamic heat flow method; wherein, the dynamic model has a delayed thermodynamic potential characterizing the delay of heat transfer in the pipeline, and the delay time corresponding to the delayed thermodynamic potential changes with the flow velocity. Based on the dynamic model, the optimization objective is to minimize the total energy consumption of the fluid drive during the working cycle, and the initial flow vector is generated.

3. The thermal management system optimization method according to claim 2, characterized in that, The calculation of the marginal cost corresponding to each discrete time step includes: Calculate the instantaneous energy consumption corresponding to each of the discrete time steps; wherein the instantaneous energy consumption is proportional to the drag coefficient of the fluid drive component and the cube of the flow rate within the corresponding discrete time step; Based on the dynamic model, the heat transfer gradient of the total heat transfer relative to the flow rate within each discrete time step is calculated; The marginal cost is determined by the ratio of the partial derivative of the instantaneous energy consumption with respect to the corresponding flow rate to the heat exchange gradient.

4. The thermal management system optimization method according to claim 2, characterized in that, Based on the marginal cost, the input flow vector is coordinated and optimized. Combining the heat transfer constraint and the flow boundary constraint, the updated flow vector for this iteration is generated, including: Based on the consistency assessment results of the marginal cost, variable release or variable freeze operations are performed on the variables in the input flow vector to obtain the set of free variables and the corresponding average marginal cost. The iterative search direction is determined based on the difference between the marginal cost and the average marginal cost of each variable in the set of free variables. Perform a flow scaling operation that satisfies the heat transfer constraint along the iterative search direction, determine the iteration step size, and generate the updated flow vector for this iteration.

5. The thermal management system optimization method according to claim 4, characterized in that, Based on the consistency assessment results of the marginal cost, variable release or variable freezing operations are performed on the variables in the input flow vector to obtain the set of free variables and the corresponding average marginal cost, including: If the marginal cost of a frozen variable that is at the lower limit of the flow boundary constraint is less than the difference between the average marginal cost and the preset tolerance, the corresponding frozen variable is released to the set of free variables. If the marginal cost of a frozen variable at the upper limit of the flow boundary constraint is greater than the sum of the average marginal cost and the preset tolerance, the corresponding frozen variable is released to the set of free variables. In response to the variable release operation, the average marginal cost is dynamically updated.

6. The thermal management system optimization method according to claim 4, characterized in that, Perform a flow scaling operation that satisfies the heat transfer constraint along the iterative search direction, determine the iteration step size, and generate the updated flow vector for this iteration, including: Solve for the scaling factor used to make the scaled flow vector satisfy the heat transfer constraint; Candidate flow vectors are constructed using the scaling factor, the initial exploration step size, and the iterative search direction; If the decrease in total energy consumption corresponding to the candidate flow vector relative to the total energy consumption of the previous iteration is greater than a preset threshold, the candidate flow vector is determined as the updated flow vector for this iteration.

7. The thermal management system optimization method according to claim 4, characterized in that, The determination of the preset convergence condition includes: The maximum absolute deviation between the marginal cost of each variable in the set of free variables and the average marginal cost is determined to obtain the marginal cost consistency error. If the marginal cost consistency error is less than the preset convergence threshold, the preset convergence condition is determined to be met.

8. The thermal management system optimization method according to claim 1, characterized in that, The working cycle includes a heat generation phase and a heat dissipation phase; During the heat generation stage, the heat source generates heat with a preset thermal power, and the constraint conditions also include that the maximum temperature of the thermal management system is lower than a preset safe temperature. During the heat dissipation phase, the heat source is turned off; The initial flow vector corresponds to each discrete time step of the heat dissipation stage, and the flow rate within each discrete time step of the heat generation stage remains at the upper limit of the flow boundary constraint.

9. The thermal management system optimization method according to claim 1, characterized in that, The thermal management system includes: a pump, a heat source, a phase change thermal storage unit, and a radiator. The fluid drive is the pump, and the operating energy consumption of the fluid drive is the pump power. Adjusting the operating state of the fluid drive based on the control command includes adjusting the pump speed or the valve opening in the thermal management system based on the control command.

10. A thermal management system optimization device, characterized in that, include: The determination module is used to determine the initial flow vector and constraints for the fluid drive component based on the operating parameters of the thermal management system; wherein the constraints include heat transfer constraints and flow boundary constraints, and the heat transfer constraints represent that the total heat transfer during the working cycle is equal to the target heat transfer. The update module is used to iteratively perform marginal cost coordination operations with the initial flow vector as the initial value of the iteration until the updated flow vector obtained in the current iteration satisfies the preset convergence condition, and the updated flow vector that satisfies the preset convergence condition is determined as the optimal flow sequence. The marginal cost coordination operation includes: calculating the marginal cost corresponding to each discrete time step for the input flow vector in this iteration, coordinating and optimizing the input flow vector based on the marginal cost, and generating an updated flow vector for this iteration by combining the heat transfer constraint and the flow boundary constraint; the marginal cost represents the incremental rate of instantaneous energy consumption relative to the total heat transfer; the input flow vector is the initial flow vector in the first iteration and the updated flow vector obtained in the previous iteration in subsequent iterations; An optimization module is used to generate control commands based on the optimal flow sequence and adjust the operating state of the fluid drive component based on the control commands to reduce the operating energy consumption of the thermal management system.

Images