A multi-objective collaborative optimization method for new energy vehicle thermal management system

Abstract

A new energy vehicle thermal management system multi-objective collaborative optimization method, comprising the following steps: constructing a digital twin model for the thermal management system for simulation, collecting dynamic operating condition parameters during the simulation process; based on the dynamic operating condition parameters, performing thermodynamic analysis on the thermal management system; based on the data obtained from the thermodynamic analysis, constructing target functions corresponding to dynamic mixing efficiency, system total cost rate and environmental impact rate respectively; optimizing the target functions with the optimization objectives of maximizing the dynamic mixing efficiency of the thermal management system, minimizing the system total cost rate and minimizing the environmental impact rate. This method breaks through the static constraint limitations of traditional thermodynamic optimization, solves the energy efficiency balance problem under complex operating conditions such as electric compressor power consumption, battery thermal management conflict, multi-heat source collaborative control, and is suitable for vehicle thermal management system design, control strategy optimization and performance verification of pure electric or hybrid electric vehicles.

Description

Technical Field

[0001] This invention relates to the field of thermal management technology for new energy vehicles, specifically a multi-objective collaborative optimization method for thermal management systems of new energy vehicles. Background Technology

[0002] The thermal management system of new energy vehicles is a highly integrated and complex system with multiple heat sources coupled together. It needs to coordinate the management of the differentiated thermal demands of subsystems such as batteries, motors, and electronic controls under dynamic operating conditions. Traditional optimization methods have several limitations: First, single-objective optimization relying on indicators such as COP or heating / cooling capacity cannot quantify environmental costs and economic indicators. Research and analysis conclude that such optimization methods targeting system performance and energy efficiency will increase the total life cycle cost of the thermal management system by 19%-23%, making such optimization methods difficult to apply in engineering. Second, the use of fixed thermodynamic constraints (such as an evaporation temperature setting range of 5-15℃) makes it difficult to adapt to extreme environmental temperature fluctuations (-30℃-50℃) and transient changes in vehicle load, resulting in a high energy efficiency degradation rate after system optimization. Third, the core advantage of multi-objective optimization is that it can comprehensively evaluate and utilize multiple aspects of the system's performance, avoiding the limitations and biases of traditional single-objective optimization methods in certain aspects. However, general-purpose multi-objective algorithms (such as NSGA-II) have drawbacks when dealing with high-dimensional nonlinear problems, such as uneven distribution of Pareto solution sets and high constraint conflict rate, which can easily lead to invalid solution sets. Finally, traditional independent optimization strategies for each subsystem can easily cause temperature control conflicts between the battery and the thermal management of the passenger compartment, making it difficult to meet the usage requirements of each temperature control object. Summary of the Invention

[0003] The purpose of this invention is to propose a multi-objective collaborative optimization method for the thermal management system of new energy vehicles, which breaks through the static constraints of traditional thermodynamic optimization and solves the energy efficiency balance problem under complex operating conditions such as electric compressor power consumption, battery thermal management conflict, and multi-heat source collaborative control. It is applicable to the design, control strategy optimization and performance verification of the whole vehicle thermal management system of pure electric or hybrid vehicles.

[0004] The technical solution adopted in this invention is: a multi-objective collaborative optimization method for a thermal management system of new energy vehicles, comprising the following steps: A digital twin model of the thermal management system is constructed for simulation, and dynamic operating parameters are collected during the simulation process; Based on the dynamic operating parameters, a thermodynamic efficiency analysis is performed on the thermal management system; based on the data obtained from the thermodynamic efficiency analysis, objective functions corresponding to the dynamic mixing efficiency, the total system cost rate, and the environmental impact rate are constructed respectively. The objective function is optimized with the goals of maximizing the dynamic mixing efficiency, minimizing the total system cost rate, and minimizing the environmental impact rate of the thermal management system; and a composite performance evaluation index model is established to comprehensively evaluate the optimization results. A global sensitivity analysis is performed on the adjustable parameters of the thermal management system to identify the decision variables in the thermal management system that affect the optimization objective, and the constraints on the decision variables are determined. Based on the objective function, decision variables, and constraints, the output value of the composite performance evaluation index model is used as the fitness value. A multi-objective optimization algorithm is used to optimize the decision variables, and the optimal operating parameters of each mode of the thermal management system under a wide temperature range are output using the superior-inferior solution distance method.

[0005] As a preferred embodiment, a thermodynamic analysis of the thermal management system is performed, specifically including: A dual-mode thermodynamic cycle model of the thermal management system under multi-source coupling conditions is established, and thermal analysis data is calculated based on thermodynamic thermal equilibrium theory. The thermal analysis data includes product thermal efficiency, fuel thermal efficiency, thermal damage, conventional thermal efficiency, and advanced thermal efficiency of each component of the thermal management system.

[0006] As a preferred embodiment, the components of the thermal management system include an in-cabin heat exchanger, an out-of-cabin heat exchanger, a power battery heat exchanger, a throttle valve, and an electric compressor.

[0007] As a preferred option, the dynamic mixing efficiency is equivalent to the traditional mixing efficiency, the advanced mixing efficiency, and... Based on the weighted sum of dynamic weight coefficients, each of the dynamic weight coefficients is generated by an LSTM neural network according to the correlation between vehicle speed, ambient temperature and battery SOC.

[0008] As a preferred option, the model for dynamic mixing efficiency is: in, Indicates the dynamic mixing ratio; This indicates traditional efficiency. , This indicates that the system has effectively output the product. This indicates the amount of fuel consumed by the system. Indicates advanced efficiency. , This indicates the unavoidable loss of the system; This indicates that the system can avoid endogenous fuel generation. Indicates the energy efficiency ratio; , , Let α and β represent the dynamic weighting coefficients, and α + β + γ = 1.

[0009] As the preferred option, the composite performance evaluation index model is as follows: in, This is a composite performance evaluation value; For dynamic mixing efficiency; The total cost rate, , This represents the total investment cost of all components in the system. This is the sum of the losses of all components in the system; For environmental impact rate, , The environmental impact rate of each component of the system; This represents the environmental impact rate of losses for each component of the system.

[0010] As a preferred embodiment, the digital twin model integrates an intelligent distribution valve, which dynamically regulates the flow between the passenger compartment and the battery thermal management branch.

[0011] As a preferred embodiment, the decision variables include at least one of the following: condenser saturation temperature, evaporator saturation temperature, evaporator superheat, condenser subcooling, and compressor efficiency.

[0012] As a preferred option, the multi-objective optimization algorithm is the improved NSGA-II algorithm, which uses dynamic crossover probability in its crossover process: P c =0.8-0.2×(t / G max ) Among them, P c G represents the crossover probability; t represents the current iteration number; G max This represents the maximum number of generations.

[0013] As a preferred approach, when determining the constraints of the decision variables, the constraints are encoded as penalty functions: P(x) = exp(10 × |T actual -T target |) Where P(x) is the penalty function; T actual Indicates the actual temperature of the object being controlled; T target This indicates the target temperature of the object being controlled.

[0014] Compared with the prior art, the beneficial effects of the present invention are: This invention innovatively proposes a dynamic weighted thermodynamic constraint mechanism, optimizing the traditional evaluation and calculation model based on COP and efficiency into a hybrid efficiency model that incorporates dynamic weight coefficients and traditional efficiency, advanced efficiency, and COP. By establishing the relationship between the α-β-γ three-weight coefficients and vehicle operating parameters, dynamic adaptive adjustment of the optimization is achieved, enabling the system to achieve a comprehensive improvement in operating efficiency, system cost, and environmental friendliness under a wide temperature range of -40-60℃. An innovative SEI composite performance evaluation system is constructed, integrating three-dimensional thermodynamic optimization objectives—system hybrid efficiency, life-cycle cost rate, and environmental impact rate—for the first time, quantifying the overall system performance through a nonlinear model. An improved multi-objective collaborative optimization architecture is developed, integrating the adaptive NSGA-II algorithm to reduce invalid solution sets and improve system optimization efficiency. Compared to traditional NSGA-II multi-objective optimization, the Pareto solution set coverage is doubled. This provides an optimization paradigm for the thermal management of new energy vehicles that combines thermodynamic perfection, economic rationality, and environmental friendliness. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of the architecture of the new energy thermal management system involved in this invention; Figure 2 This is a schematic diagram of the AMEsim simulation model of the present invention; Figure 3 This is a schematic diagram of the thermodynamic cycle TS of the present invention; Figure 4 This is a schematic diagram of the improved NSGA-Ⅱ algorithm of the present invention; Figure 5 This is a flowchart illustrating the improved NSGA-II algorithm of the present invention. Figure 6 This is a flowchart illustrating a specific implementation of the present invention; Figure 7 This is a schematic diagram of the calculation results of 100 iterations of single-objective iterative optimization of the hybrid efficiency of the present invention; Figure 8 This is a schematic diagram illustrating the calculation results of the total cost rate of this invention after 100 iterations of single-objective iterative optimization. Figure 9 This is a schematic diagram of the calculation results of the single-objective iterative optimization of the environmental impact rate of the present invention after 100 iterations. Figure 10 For the multi-objective optimization cooling mode of the present invention and Pareto frontier; Figure 11 For the multi-objective optimization heating mode of this invention and Pareto frontier; Figure 12 For the multi-objective optimization cooling mode of the present invention and Pareto frontier; Figure 13 For the multi-objective optimization heating mode of this invention and Pareto frontier; Figure 14 For the multi-objective optimization cooling mode of the present invention and Pareto frontier; Figure 15 For the multi-objective optimization heating mode of this invention and Pareto frontier; Figure 16 This is a schematic diagram of the loss decomposition of system components according to the present invention. Detailed Implementation

[0017] The present invention will now be described in detail through exemplary embodiments. However, it should be understood that, without further description, elements, structures, and features in one embodiment may be advantageously incorporated into other embodiments.

[0018] It should be noted that, unless otherwise defined, the technical or scientific terms used herein should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "a," "an," or "the," etc., used in the specification and claims of this patent application do not express a limitation on quantity, but rather indicate the presence of at least one; the terms "first," "second," and "third," as used herein, should not be considered as a limitation on the order of components, but are merely for distinguishing different components; the terms "comprising," "including," etc., indicate that the elements or objects preceding "comprising" or "including" encompass the elements or objects listed following "comprising" or "including" and their equivalents, but do not exclude other elements or objects having the same function.

[0019] To more clearly describe the multi-objective collaborative optimization method of the new energy vehicle thermal management system, combined with the appendix... Figure 1-16 This embodiment is described as follows: like Figure 1-16As shown, a multi-objective collaborative optimization method for a new energy vehicle thermal management system is presented. The components of the new energy vehicle thermal management system include: an internal heat exchanger and air conditioning unit (HVAC), internal heat exchangers (internal evaporator HEx1, internal condenser HEx2), an external heat exchanger HEx, a battery cooling plate, a solenoid valve SV, an electronic expansion valve EXV, a liquid vapor separator, a throttle valve, and an electric compressor. For detailed architecture, see [link to relevant documentation]. Figure 1 The specific method for establishing the mechanism is as follows: Step 1: Constructing the system's thermodynamic cycle. A digital twin model of the thermal management system is built for simulation. Dynamic operating parameters are collected during the simulation, specifically including: Based on the architecture of the thermal management system for new energy vehicles, a digital twin simulation model integrating intelligent distribution valves was built on the AMEsim platform. Figure 2 This valve enables dynamic flow control between the passenger compartment and the battery thermal management branch (e.g., a flow distribution ratio of 1:0.8 to 1:1.2 when the SOC is >80% in cooling mode). The architecture also illustrates the signal transmission channel, the electricity transmission channel, the mechanical connection channel, the heat transmission channel, the refrigerant transfer channel, the control strategy, the DC / DC converter, the electric motor, and the battery.

[0020] The system defines an extreme operating condition test matrix covering low-temperature cold start (-40℃, SOC < 20%) and high-temperature fast charging (60℃, 2C charging rate). It collects 32 dynamic parameters for pure electric vehicles under WLTC conditions, including compressor power, working fluid phase change point temperature, and heat exchanger pressure drop. Simultaneously, it records key operating condition variables in real time, such as vehicle speed (0-120km / h), ambient temperature and humidity, and battery SOC (20%-100%). Based on this data, a dual-mode thermodynamic cycle model / temperature-entropy diagram TS (cooling / heating) is established. Figure 3 ), annotate the flow distribution characteristics of key state points, 1-9 correspond to Figure 1 In the system architecture diagram, operating state points 1-5 correspond to the refrigerant-side heat transfer process, and 6-9 correspond to the air-side heat transfer process. The intelligent distribution valve improves heat exchange efficiency by 18.2% under -40℃ conditions, while simultaneously constructing a three-dimensional dynamic parameter space based on vehicle speed, ambient temperature, and state of charge (SOC). This step lays the foundation for the subsequent hybrid efficiency model ε. hybrid =α·ε ex +β·ε enProvide a data basis for the dynamic adjustment of the weight coefficient of +γ·COP, and synchronously collect the total life cycle cost rate C tot And the original data of Btot to support the calculation of the SEI composite efficiency index in step 2. Break through the limitations of the working condition adaptability of the traditional fixed constraint model.

[0021] Step 2: Based on the dynamic working condition parameters, conduct a thermodynamic exergy analysis of the thermal management system; according to the data obtained from the thermodynamic exergy analysis, construct objective functions corresponding to the dynamic mixing efficiency, system total cost rate, and exergy environmental impact rate, including the following steps: 1. Establish a system thermodynamic analysis model: The system working fluid can also be called the exergy flow, and the exergy value calculation formula for each thermodynamic state point of the system is as follows: (1) In the formula, is the exergy value of the thermodynamic state point; is the mass flow rate of the system working fluid; is the specific exergy value of the exergy flow; is the specific enthalpy of the exergy flow per unit mass flow; is the specific enthalpy of the exergy flow per unit mass flow under the environmental state; is the environmental temperature; is the specific entropy of the exergy flow per unit mass flow; is the specific entropy of the exergy flow per unit mass flow under the environmental state.

[0022] The physical exergy based on unit mass can be further divided into thermal exergy and mechanical exergy. The former is mainly determined by temperature, and the latter is mainly determined by pressure. Then, the expression of the specific exergy value of the system working fluid is shown in Equation (2): (2) In the formula, , are the specific exergies of thermal exergy and mechanical exergy respectively; represents the specific enthalpy of the exergy flow per unit mass flow at state point , represents the specific entropy of the exergy flow per unit mass flow at state point . State point is defined as the state point of the exergy flow under the given pressure and the reference temperature ; represents that the pressure remains unchanged, which is the constraint condition for calculating the thermodynamic exergy; represents that the environmental temperature remains unchanged, which is the constraint condition for calculating the mechanical exergy; represents the specific enthalpy of the exergy flow per unit mass flow under the environmental state; It represents the specific entropy of a unit mass flow rate under environmental conditions.

[0023] According to the first and second laws of thermodynamics, ∠ can be explained by the concept of "fuel-product", and the equilibrium equations for ∠ can be established as shown in equations (3) and (4): System components: (3) In the formula, For components The amount of fuel consumed; For components The effective output of the product can also be referred to as the "product output". For components The destruction of heat represents the irreversibility of heat transfer.

[0024] Overall system: (4) In the formula, The fuel consumed by the system; To ensure effective product output from the system; The sum of losses of all components in the system; The loss of a system is a loss that occurs only at the system-wide level. It refers to a loss that is no longer used in the system and only occurs at the system-wide level.

[0025] The loss is decomposed into endogenous and exogenous components, and its expression is as follows: (5) In the formula, It is an inherent loss in the component. It is an external loss in the component. As an unavoidable loss in components, To avoid varying losses, the intrinsic varying losses associated with component k are... This refers to the efficiency loss of component k when all other components operate ideally, but component k operates at its true efficiency. Endogenous efficiency loss is only related to the inefficiency of component k itself. Exogenous efficiency loss... It is the remaining part of the total loss after deducting the intrinsic loss of component k. The exogenous loss of component k is caused by the inefficiency of other components in the system.

[0026] To reflect the impact of technological limitations on the system, the loss can be decomposed into avoidable and unavoidable parts, and its expression is: (6) (7) In the formula, As an unavoidable loss in components, To avoid energy losses, improving system energy efficiency requires minimizing avoidable energy losses in components. By breaking down the energy losses of each system component into unavoidable and avoidable parts, the energy efficiency of each component can be assessed more accurately. It is calculated by formula (6) in the inevitable cycle of ITMS. The system components are working under optimal operating conditions. The system is running in an inevitable cycle state. The optimal conditions of each component reflect its maximum potential for improvement.

[0027] To obtain a more in-depth thermodynamic analysis, the two pyrolysis methods are combined, and the varnish loss is further divided into four parts: avoidable endogenous varnish loss, avoidable exogenous varnish loss, unavoidable endogenous varnish loss, and unavoidable exogenous varnish loss. The decomposition diagram is shown below. Figure 16 As shown; Its calculation expression is: (8) (9) In the formula, Due to unavoidable endogenous losses, kW; For unavoidable exogenous losses, kW. To avoid endogenous losses, kW; To avoid exogenous losses, kW. It can be calculated using formula (9), but it cannot be reduced due to technical limitations on component k. The loss that cannot be reduced due to technical limitations of other components can be calculated through unavoidable loss decomposition. It can be calculated by decomposing endogenous losses. By improving the efficiency of the components under consideration, this part of the losses can be reduced. It can be calculated by decomposing exogenous losses, and this part of the losses can be reduced by improving the efficiency of the remaining components or optimizing the system structure.

[0028] From the above models, we can derive models for the traditional and advanced efficiency of the system, as shown in equations (10) and (11): Traditional system efficiency : (10) System Advanced Efficiency : (11) In the formula, This represents the unavoidable loss of the system, which can be derived from Equation 7; The system's avoidable endogenous fuel is represented by Equation 8.

[0029] Therefore, based on the thermodynamic analysis model, an innovative model for dynamic mixing efficiency is proposed. Specifically, as shown in Equation 12: (12) The dynamic weighting coefficients α, β, and γ are generated in real time using an LSTM neural network. Input parameters include vehicle speed (0-120 km / h), ambient temperature (-40~60℃), and battery SOC (20%-100%), satisfying the constraint α+β+γ=1, with each coefficient dynamically adjusted within the range of 0.2-0.5. This model overcomes the limitations of traditional single-dimensional efficiency evaluation, improving the thermodynamic characterization accuracy by 18.7% in WLTC cycle testing.

[0030] System cost analysis model: Investment costs of each component in the system Mainly composed of the annualized investment cost of components With operation and maintenance costs The composition can be represented as: (13) The cost balance equation for system components is shown in Equation 14, which states that the effective output cost of a component is the sum of the input fuel cost and investment cost. The cost calculation formula for each flow is shown in Equation 15, and the cost calculation for component failure can be referenced in Equation 16. (14) (15) (16) In the formula, , , These are the unit product cost, unit fuel cost, and unit flow cost of system components, respectively. , , These are the product value, fuel value, and flow value of the component, respectively; The unit cost of flow. The total cost of the flow is the unit cost multiplied by the total flow.

[0031] System environmental impact model: The environmental impact balance equations for the components in the system are shown in equations (17) and (18): (17) (18) In the formula, , , The values ​​are the unit product environmental impact, unit fuel environmental impact, and unit flow environmental impact of thermal management system components, respectively, in mPts / kJ; The environmental impact rate of each component in the system is expressed in mPts / h. For the unit environmental impact of flow, The total environmental impact of flow is expressed as environmental impact per unit flow * total flow.

[0032] The formula for calculating the environmental impact rate of component k is: (19) System optimization objective function: Based on the above thermodynamic analysis model, with the optimization objectives of maximizing the system's dynamic mixing efficiency, minimizing the system's total cost rate, and minimizing the environmental impact rate, the objective function is established as follows: Dynamic mixing efficiency f1: (20) In the formula, W com E is the input power to the compressor. P,hex1,2 Product output for crew compartment heat exchanger; E P,chiller The product output for the battery heat exchanger.

[0033] Total cost rate f2: (twenty one) In the formula, Represents the total cost ratio; ∑ The total investment cost of all components of the system; ∑ This is the sum of the losses of each component in the system.

[0034] Environmental impact rate f3: (twenty two) In the formula, Indicates the environmental impact rate; ∑ The environmental impact rate of each component of the system; ∑ This represents the environmental impact rate of losses for each component of the system.

[0035] The unknown variables in equations (20) to (22) can all be calculated based on thermodynamic cycle data using thermodynamic analysis, cost analysis, and environmental impact analysis models.

[0036] A composite performance evaluation index, SEI (System Effectiveness Index), is constructed concurrently. The key parameter system comprises three dimensions: 1. Thermodynamic dimension: Inheriting and expanding traditional thermodynamic analysis theory, it innovatively couples traditional system efficiency, advanced system efficiency, and COP to comprehensively quantify energy utilization efficiency; 2. Economic dimension: Optimizing the cost model architecture, based on thermodynamic analysis methods, it integrates initial investment costs and operating costs, addressing the problem that traditional economic models cannot fully cover the entire life cycle cost; 3. Environmental dimension: Upgrading the environmental impact model, improving upon the limitations of traditional models that only consider direct emissions. The specific SEI composite performance evaluation index model is shown in the following formula: (twenty three) The establishment of the SEI composite performance evaluation index gives the above evaluation scheme significant advantages over traditional performance evaluation schemes. Firstly, in terms of energy efficiency characterization, it expands from a single dimension to a three-dimensional fusion, and the real-time calculation is improved by dynamic weight adjustment, enhancing operational adaptability by 35%, resulting in a more comprehensive energy loss assessment. Secondly, the cost evaluation dimension expands from initial investment cost to total life-cycle cost. Thirdly, the environmental dimension expands from direct emissions to total life-cycle emissions. The SEI system can reduce the dimensionality of a three-objective optimization problem to maximizing a single index, shortening multi-objective decision-making time by 41.2%, while simultaneously addressing industry pain points in the background technology.

[0037] Step 3: Establish a sensitivity analysis model to perform a global sensitivity analysis on the adjustable parameters of the thermal management system, and screen out the decision variables in the thermal management system that affect the optimization objective, specifically including: The global sensitivity analysis method (Sobol index method) explores the influence relationship between system inputs and outputs from the perspective of variance. Specific steps include decomposing the system model into a multi-parameter form consisting of individual parameters and their combinations, and calculating the contribution of each individual input parameter and parameter combination to the total output variance.

[0038] Total variance: (twenty four) Conditional variance: (25) Among them, the thermal management system model under study is considered to be represented as y represents the system output. The difference between the total variance and the conditional variance represents the variance of variable x. j The magnitude of the influence on the output of system y is shown in formula (26). If the obtained mean is small, it indicates the opposite of the influence on variable x. j It has a significant impact on the output of system y.

[0039] (26) Formula (27) can be obtained from mathematical statistics. (27) In the Sobol index method, the ratio of the two is defined as a first-order sensitivity index, also known as the main effect index: (28) If all input variables are divided into x j and Two categories, Indicates division by x j Considering all other variables as a whole, the total effect index is defined in the Sobol index method as follows: (29) For the k-th input variable, Let Sx be the sensitivity index of the k-th input variable; in the Sobol index method, multiple input variables are analyzed, such as Sx. j， x k When considering the overall impact on the system output, this sensitivity index is shown in formula (30), and is also called the second-order interaction effect index. .

[0040] (30) Step 4: Based on the results of the global sensitivity analysis, determine the decision variables in the thermal management system design, and define the value range and constraints of these variables, specifically including: In this study of the mechanism, the following five decision variables were selected for analysis: Condenser saturation temperature Evaporator saturation temperature Evaporator superheat condenser subcooling Compressor efficiency η com In engineering applications that solve multi-objective optimization problems, constraints are typically imposed on the trade-offs of decision variables. These constraints usually stem from theoretical feasibility and engineering application requirements. For example, the permissible water flow velocity on the tube side of a shell-and-tube heat exchanger should be maintained between 1 and 3 m / s to prevent fouling and corrosion; for evaporators and condensers, it is widely recommended in engineering applications that the ratio of tube length to shell diameter should be between 5 and 15. In the design of thermal management systems, the recommended values ​​for superheat and subcooling are within 10°C. Because there are multiple pipe channels in the evaporators and condensers of a thermal management system, and multiple heat exchange processes occur, the highest temperature of the cold flow must always be lower than the lowest temperature of the hot flow to prevent temperature crossover during heat exchange. Table 1 shows the constraints on the decision variables in a thermal management system. Where T0 is the ambient temperature; ΔT... cond,min and ΔT BEVap,min These are the minimum temperature differences between the highest temperature of the cold fluid and the lowest temperature of the hot fluid in the condenser and evaporator, respectively; T cond,air,in and T cond,air,in These are the air inlet temperatures for the condenser and evaporator, respectively.

[0041] Step 5: Based on the objective function, decision variables, and constraints, using the output value of the composite performance evaluation index model as the fitness value, a multi-objective optimization algorithm is employed to optimize the decision variables. The optimal operating parameters for each mode of the thermal management system under a wide temperature range are then output using the superior-inferior solution distance method. Specifically, this includes: The multi-objective optimization algorithm employs the NSGA-II algorithm, which identifies non-dominated individuals through techniques such as non-dominated sorting and crossover mutation, and then introduces a shared function in the decision vector space to gradually solve the optimization problem. In 2000, this algorithm was further improved by introducing an elite retention strategy. Based on the established system model, optimization objectives, parameters to be optimized, and constraints, this mechanism designs a multi-objective optimization algorithm for a thermal management system based on NSGA-II. This algorithm evaluates the merits of the calculation results based on the Pareto dominance principle, without requiring the specification of weight coefficients for each objective. Related schematic diagrams are shown in [the diagram]. Figure 4 .

[0042] NSGA-II ensures population diversity by calculating crowding distance between individuals. The main process includes initial sorting, selection, crossover, mutation, and screening. Figure 5 As shown, data is first selected and initialized. Let t=0, and the population P is initialized. t The first generation population N is produced, and the population 2N is obtained through evolutionary iteration. The objective function values ​​are compared to generate a new parent population G = G + 1. The maximum population P is obtained through the three basic operations of selection, crossover, and mutation in a genetic algorithm. t,max Output the result if t < G max G maxThis indicates the maximum number of iterations; then proceed to step 2 until the maximum number of iterations is reached. In the calculation, the evolutionary population is set to 100, the maximum number of generations is 50, the mutation probability is 0.1, and the crossover probability is 0.9.

[0043] Develop an adaptive crossover probability module P c = 0.8 - 0.2×(t / G max ), where P c represents the crossover probability, which determines the likelihood of two parent individuals generating new offspring through gene crossover in a genetic algorithm. The dynamic crossover probability design described above allows the crossover probability to decrease linearly with the number of iterations (gradually decreasing from an initial 0.8), which helps maintain population diversity in the early stages of the algorithm and tends to retain superior individuals in later stages; t represents the current iteration number, indicating which generation the algorithm has reached; G max This represents the maximum number of generations, and is one of the termination conditions for the algorithm to stop running.

[0044] Simultaneously, the adaptive constraint conditions defined in step 4 are encoded as a penalty function P(x) = exp(10×|T actual -T target |), T actual Indicates the actual temperature of the object being controlled; T target | represents the target temperature of the temperature control object; the function of this penalty function is to calculate the absolute value of the difference between the actual temperature and the target temperature and amplify it exponentially to penalize solutions that deviate from the target temperature, thereby forcing the solutions generated by the optimization algorithm to meet the thermodynamic constraints, effectively solving the problem of low optimization efficiency caused by traditional constraints and poor handling.

[0045] For the multi-objective optimization of the thermal management system, taking the optimization objective in step 4 as an example, by adjusting the controllable decision variable parameters of the system, a balance point is found among the conflicting objectives, so that it reaches a relatively optimal state under specific constraints, thus completing the multi-objective optimization modeling of the thermal management system of new energy vehicles based on thermodynamic optimal constraints.

[0046] Based on the above invention, taking a typical new energy vehicle thermal management system as an example, a thermodynamic analysis is conducted on its integrated cooling and heating modes. Using the three optimization objectives of efficiency, total cost rate, and total environmental impact rate from the thermodynamic analysis, the system's multi-objective optimization is completed. The specific steps are as follows: Step S1: Establish the AMEsim simulation model of the thermal management system. Based on the cooling and heating operating modes of the thermal management system, establish its thermodynamic cycle one by one and construct the system thermodynamic cycle diagram. For example, the specific parameters of the cooling cycle are shown in Table 1, and the cycle data are shown in Table 2. Table 1 System Refrigeration Cycle Parameters Table 2 Operating data of the system refrigeration cycle Step S2: Under the determined operating condition mode, according to Step 2 in the invention content, complete the thermodynamic analysis, cost analysis, and environmental impact analysis of the system, and establish the system's optimization objective function, as follows: Based on the above thermodynamic analysis model, with the system's efficiency (maximization), total system cost rate (minimization), and environmental impact rate (minimization) as optimization objectives, an optimization objective function is established as follows: Dynamic mixing efficiency f1: Total cost rate f2: Environmental impact rate f3: Based on thermodynamic cycle data, the unknown variables in the formula are calculated using thermodynamic analysis, cost analysis, and environmental impact analysis models, thereby obtaining the system's real-time efficiency, total cost rate, and environmental impact rate.

[0047] Step S3: Using steps 3-4 of the invention description, determine the decision variables in the system design and define the value range and constraints of these variables. In this study of the mechanism, the following five decision variables are selected for analysis: Condenser saturation temperature Evaporator saturation temperature Evaporator superheat condenser subcooling Compressor efficiency η com The numerical range of the constraints is as follows: Table 3 Constraints on decision variables in a thermal management system Step S4: Using step 5 in the invention description, establish a multi-objective optimization model for the system. Specifically: After determining the objective function, decision variables, and constraints, a multi-objective optimization algorithm for the thermal management system based on NSGA-II is designed. NSGA-II ensures population diversity by calculating the crowding degree between individuals. The main process is initialization sorting, selection, crossover, mutation, and screening. First, data is selected and initialized, t=0, and the population Pt is initialized. The first generation population N is produced, and the population 2N is obtained through evolutionary iteration. The objective function value is compared, and a new parent population G=G+1 is generated. The maximum population Pt,max is obtained through the three basic operations of selection, crossover, and mutation of the genetic algorithm, and the result is output. If t<Gmax, then go to step 4 until the maximum number of iterations is reached. In this patent, 100 iterations are usually used.

[0048] Step S5: Using steps 2 and 5 in the invention, based on the three set optimization objectives, conduct a single-objective optimization study of the system one by one, and determine the interaction relationship between the optimization objectives and their impact on the system, as follows: The calculation results of system efficiency through 100 iterations of single-objective iterative optimization are as follows: Figure 7 As shown, with the increase of the number of iterations, the system efficiency value shows a trend of first gradually increasing and then stabilizing, ultimately achieving the highest single-objective optimization value. Through optimizing the single-objective solution for efficiency, the system energy efficiency is significantly improved. To achieve optimal efficiency, component improvements are needed, which in turn increases the system's cost rate and environmental impact rate. Similar conclusions are drawn from the single-objective optimization of the system's total cost rate and total environmental impact rate, as detailed below. Figure 8 , 9 As shown in Table 4, the final decision variables of the three different single-objective optimization schemes are as follows.

[0049] Table 4. Output results of decision variables under different single-objective optimization schemes Step 6: Based on Step 5 of this invention, complete the multi-objective optimization of the system to obtain the improved logic of the system. Specifically, the large set of optimal solutions obtained from the multi-objective optimization is evaluated and ranked using the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). A weighting factor (usually based on experience or the importance of each objective) is assigned to each objective and / or decision to select a single final solution. The Pareto solution is used to form the optimal solution front. The multi-objective optimization results of the thermal management system are as follows: Figure 10-15 As shown.

[0050] Table 5 Optimization Results of Cooling Mode Table 6 Optimization Results of Heating Mode Tables 5 and 6 show the extreme values ​​of the system optimization objective function under the two modes, respectively. It can be seen that single-objective optimization has its own disadvantages, while in multi-objective optimization scenarios, these objectives are considered simultaneously. Therefore, a trade-off is made between solutions for two conflicting objectives, providing an optimization solution that falls between the extreme values ​​produced by single-objective methods.

[0051] The parts not described in detail in the above embodiments are existing technologies.

[0052] It should be noted that although the present invention has been described through the above embodiments, the present invention may have many other embodiments. Without departing from the spirit and scope of the present invention, those skilled in the art can obviously make various corresponding changes and modifications to the present invention, but all such changes and modifications should fall within the scope of protection of the appended claims and their equivalents.

Claims

1. A multi-objective collaborative optimization method for a thermal management system of a new energy vehicle, characterized in that, Includes the following steps: A digital twin model of the thermal management system is constructed for simulation, and dynamic operating parameters are collected during the simulation process; Based on the dynamic operating parameters, a thermodynamic efficiency analysis is performed on the thermal management system; based on the data obtained from the thermodynamic efficiency analysis, objective functions corresponding to the dynamic mixing efficiency, the total system cost rate, and the environmental impact rate are constructed respectively. The objective function is optimized with the goals of maximizing the dynamic mixing efficiency, minimizing the total system cost rate, and minimizing the environmental impact rate of the thermal management system; and a composite performance evaluation index model is established to comprehensively evaluate the optimization results. A global sensitivity analysis is performed on the adjustable parameters of the thermal management system to identify the decision variables in the thermal management system that affect the optimization objective, and the constraints on the decision variables are determined. Based on the objective function, decision variables, and constraints, the output value of the composite performance evaluation index model is used as the fitness value. A multi-objective optimization algorithm is used to optimize the decision variables, and the optimal operating parameters of each mode of the thermal management system under a wide temperature range are output using the superior-inferior solution distance method.

2. The multi-objective collaborative optimization method for a new energy vehicle thermal management system according to claim 1, characterized in that: A thermodynamic analysis of the thermal management system is performed, specifically including: A dual-mode thermodynamic cycle model of the thermal management system under multi-source coupling conditions is established, and thermal analysis data is calculated based on thermodynamic thermal equilibrium theory. The thermal analysis data includes product thermal efficiency, fuel thermal efficiency, thermal damage, conventional thermal efficiency, and advanced thermal efficiency of each component of the thermal management system.

3. The multi-objective collaborative optimization method for a new energy vehicle thermal management system according to claim 2, characterized in that: The components of the thermal management system include an in-cabin heat exchanger, an out-of-cabin heat exchanger, a power battery heat exchanger, a throttle valve, and an electric compressor.

4. The multi-objective collaborative optimization method for a thermal management system of a new energy vehicle according to claim 1, characterized in that: Dynamic hybrid efficiency is categorized into traditional efficiency, advanced efficiency, and... Based on the weighted sum of dynamic weight coefficients, each of the dynamic weight coefficients is generated by an LSTM neural network according to the correlation between vehicle speed, ambient temperature and battery SOC.

5. The multi-objective collaborative optimization method for a thermal management system of a new energy vehicle according to claim 4, characterized in that: The model for dynamic mixing efficiency is as follows: in, Indicates the dynamic mixing ratio; This indicates traditional efficiency. , This indicates that the system has effectively output the product. This indicates the amount of fuel consumed by the system. Indicates advanced efficiency. , This indicates the unavoidable loss of the system; This indicates that the system can avoid endogenous fuel generation. Indicates the energy efficiency ratio; , , Let α and β represent the dynamic weighting coefficients, and α + β + γ = 1.

6. The multi-objective collaborative optimization method for a thermal management system of a new energy vehicle according to claim 1, characterized in that: The composite performance evaluation index model is as follows: in, This is a composite performance evaluation value; For dynamic mixing efficiency; The total cost rate, , This represents the total investment cost of all components in the system. This is the sum of the losses of all components in the system; For environmental impact rate, , The environmental impact rate of each component of the system; This represents the environmental impact rate of losses for each component of the system.

7. The multi-objective collaborative optimization method for a thermal management system of a new energy vehicle according to claim 1, characterized in that: The digital twin model integrates an intelligent distribution valve, which dynamically regulates the flow between the passenger compartment and the battery thermal management branch.

8. The multi-objective collaborative optimization method for a thermal management system of a new energy vehicle according to claim 1, characterized in that: The decision variables include at least one of the following: condenser saturation temperature, evaporator saturation temperature, evaporator superheat, condenser subcooling, and compressor efficiency.

9. The multi-objective collaborative optimization method for a thermal management system of a new energy vehicle according to claim 1, characterized in that, The multi-objective optimization algorithm is an improved version of the NSGA-II algorithm, which uses dynamic crossover probabilities in its crossover process: P c =0.8-0.2×(t / G max ) Among them, P c G represents the crossover probability; t represents the current iteration number; G max This represents the maximum number of generations.

10. A multi-objective collaborative optimization method for a thermal management system of a new energy vehicle according to claim 1, characterized in that, When determining the constraints of the decision variables, the constraints are encoded as penalty functions: P(x)=exp(10×|T actual -T target |) Where P(x) is the penalty function; T actual Indicates the actual temperature of the object being controlled; T target This indicates the target temperature of the object being controlled.

Images