A global sensitivity data-driven vehicle thermal management system modeling method and system

Abstract

The application discloses a kind of global sensitivity data-driven vehicle thermal management system modeling method and system. Modeling method includes the following steps: building the component physical model of all components in vehicle thermal management system;Each of the component physical model is extracted according to the following process respectively, and the obtained component physical model of all components in vehicle thermal management system is spliced and integrated to obtain a system-level model;Combined with the imported system-level model dataset and the obtained sensitive parameter combination, iterative calculation is carried out until the target function deviation between the system-level model output result and the measured value is up to standard, and the final sensitive parameter combination is output. The present application solves the problem of low efficiency and precision dependence on artificial experience in traditional modeling by the core logic of component-level sensitive parameter optimization and system-level integration iteration.

Description

Technical Field

[0001] This invention belongs to the field of vehicle thermal management technology, and relates to a modeling method and system for a vehicle thermal management system based on global sensitivity data. Background Technology

[0002] Thermal Management Systems (TMS) play a crucial role in ensuring cabin comfort and vehicle safety, attracting increasing attention from the industry, especially for new energy vehicles. Finding the best balance between energy consumption, comfort, and safety has become a research hotspot, and efficiently and quickly developing a thermal management system will be a core competitive advantage for companies. Modeling and simulation are vital for guiding vehicle development and testing. They can accurately simulate the performance of thermal management systems under various operating conditions, understanding changes in temperature distribution, heat transfer, and cooling efficiency under different conditions. This allows designers to optimize for specific application scenarios, thereby improving overall vehicle performance. Furthermore, it enables the comparison and evaluation of multiple thermal management system solutions during the design phase, significantly reducing the cost and time of later actual testing.

[0003] Modeling has a significant impact on the reliability and accuracy of simulation results. Currently, modeling is generally based on physical models and involves manual adjustments and optimizations based on engineers' personal experience, which is time-consuming and highly subjective. There is an urgent need for an efficient modeling method that integrates physical models with data. Summary of the Invention

[0004] The purpose of this invention is to provide a modeling method and system for vehicle thermal management systems based on global sensitivity data, so as to achieve accurate system modeling based on test data and improve the modeling and simulation accuracy of thermal management systems.

[0005] To achieve the objectives of this invention, the technical solution provided by this invention is as follows:

[0006] First Invention

[0007] This application provides a modeling method for a vehicle thermal management system based on global sensitivity data, including the following steps:

[0008] Step S1: Build physical models of all components in the vehicle thermal management system;

[0009] Step S2: For each component's physical model, sensitive parameter combinations are extracted according to the following process, as follows:

[0010] Step S2.1: Extract the key parameters of the component physical model and set the numerical range of the key parameters;

[0011] Step S2.2: Import the component-level model dataset;

[0012] Step S2.3: Based on the key parameters, the key parameter matrix is ​​randomly generated using the Monte Carlo sampling method;

[0013] Step S2.4: Input the component-level model dataset and the key parameter matrix into the component physical model, and calculate the model output results;

[0014] Step S2.5: Based on the model output, calculate the first-order and total-order exponents of the key parameters with respect to the model output;

[0015] Step S2.6: Based on the magnitude of the first-order exponent and the total-order exponent, extract three to five of the most sensitive parameters for each model output result as a sensitive parameter combination;

[0016] Step S2.7: Establish the objective function between the model output and the measured values;

[0017] Step S2.8: Determine if there is a sensitive parameter combination that meets the target value requirement of the objective function; if so, directly output the corresponding component physical model and sensitive parameter combination; if not, select a sensitive parameter combination that is closest to the target value requirement, perform iterative calculation on the sensitive parameter combination, adjust the value of the sensitive parameter combination until the deviation of the objective function meets the standard, and output the corresponding component physical model and sensitive parameter combination.

[0018] Step S3: Integrate the component physical models of all components in the vehicle thermal management system obtained in Step S2 to obtain a system-level model; combine the imported system-level model dataset and the sensitive parameter combination obtained in Step S2 to perform iterative calculations until the objective function deviation between the system-level model output and the measured value meets the standard, and output the final sensitive parameter combination.

[0019] Furthermore, in step S1, the vehicle thermal management system is a heat pump air conditioning system, whose components include: a compressor, a liquid-cooled condenser (LCC), a condenser, a chiller, an evaporator, an electronic expansion valve, and a battery pack.

[0020] Furthermore, in step S2.1, the process of extracting key parameters from the component physical model is as follows: sequentially determine whether the parameters of the component physical model are structural parameters; if so, discard them; if not, retain them as key parameters.

[0021] Further, step S2.3 includes the following:

[0022] Step 2.3.1: The key parameters of the component physical model are... n (x1, x2, x3, …, x) n The sample size is... mTwo matrices, A and B, are randomly generated using Monte Carlo sampling:

[0023] ;

[0024] ;

[0025] Step 2.3.2: Construct the key parameter matrix AB i Matrix, where i=1~ n AB i The matrix is ​​formed by replacing the i-th column of matrix A with the i-th column of matrix B.

[0026] Furthermore, in step S2.5, the first-order exponent of the key parameters on the model output is calculated. S i Sum of total order index S Ti As shown in the following formula:

[0027] ;

[0028] ;

[0029] in: V For variance; y i Output the results for the model; x i These are key parameters for the physical model of the component.

[0030] Further, in step S2.7, the objective function between the model output and the measured values ​​is established as follows:

[0031] ;

[0032] Where y=f(x1, x2, x3, …, x n )= y i ; y a This is the actual value. deltay is the objective value of the objective function.

[0033] Furthermore, in step S2.8, when the target value is within 5%, the target value requirement of the objective function is met.

[0034] Further, in step S2.8, the values ​​of the sensitivity parameter combination are adjusted using the following formula:

[0035] ;

[0036] In the formula: xt These are the adjusted parameter values. x t-1 The parameter values ​​are those that bring the objective function's target value closest to 5%. k This is an adjustment factor.

[0037] Second aspect

[0038] This application provides a vehicle thermal management system modeling system based on global sensitivity data, including the following units: a component physical model building unit, a component physical model sensitive parameter combination extraction unit, and an integration unit;

[0039] The component physical model building unit is used to build the component physical models of all components in the vehicle thermal management system.

[0040] The component physical model sensitive parameter combination extraction unit is used to extract sensitive parameter combinations for each component physical model according to the following process:

[0041] Step S2.1: Extract the key parameters of the component physical model and set the numerical range of the key parameters;

[0042] Step S2.2: Import the component-level model dataset;

[0043] Step S2.3: Based on the key parameters, the key parameter matrix is ​​randomly generated using the Monte Carlo sampling method;

[0044] Step S2.4: Input the component-level model dataset and the key parameter matrix into the component physical model, and calculate the model output results;

[0045] Step S2.5: Based on the model output, calculate the first-order and total-order exponents of the key parameters with respect to the model output;

[0046] Step S2.6: Based on the magnitude of the first-order exponent and the total-order exponent, extract three to five of the most sensitive parameters for each model output result as a sensitive parameter combination;

[0047] Step S2.7: Establish the objective function between the model output and the measured values;

[0048] Step S2.8: Determine if there is a sensitive parameter combination that meets the target value requirement of the objective function; if so, directly output the corresponding component physical model and sensitive parameter combination; if not, select a sensitive parameter combination that is closest to the target value requirement, perform iterative calculation on the sensitive parameter combination, adjust the value of the sensitive parameter combination until the deviation of the objective function meets the standard, and output the corresponding component physical model and sensitive parameter combination.

[0049] The integration unit is used to stitch together and integrate the component physical models of all components in the obtained vehicle thermal management system to obtain a system-level model; combined with the imported system-level model dataset and the obtained sensitive parameter combination, iterative calculation is performed until the objective function deviation between the system-level model output result and the measured value meets the standard, and the final sensitive parameter combination is output.

[0050] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0051] The solution of this invention effectively solves the problems of low efficiency and reliance on human experience in traditional modeling by using the core logic of "component-level sensitive parameter optimization and system-level integrated iteration". Its specific beneficial effects include the following:

[0052] (1) This invention uses global sensitivity analysis (first-order exponent + total-order exponent) to screen key parameters, avoid interference from irrelevant parameters, and focus on core influencing factors; establishes objective functions for model output and measured values, and achieves adaptive correction through parameter iterative adjustment to ensure that the model output is highly consistent with the actual working conditions; iterative verification is carried out again in the system-level integration stage to further eliminate errors after component splicing and improve the overall model accuracy.

[0053] (2) The present invention automatically identifies 3-5 of the most sensitive parameters, eliminating the need to optimize all parameters one by one, thus greatly reducing the amount of calculation and modeling cycle; it replaces the traditional mode of "relying on engineers' experience to manually adjust parameters", avoiding the complex and repetitive work of repeated trial and error, and reducing labor costs; it can complete the simulation evaluation of multiple schemes in advance during the design stage, reducing the number of real vehicle tests in the later stage, shortening the development cycle and saving testing costs.

[0054] (3) This invention outputs the sensitive parameter combination of each component, providing a clear direction for the performance optimization of the vehicle thermal management system; the solution is applicable to heat pump air conditioning systems of various models such as pure electric and hybrid vehicles, with strong compatibility and high reusability; the automated parameter adjustment logic can be adapted to different working condition data, providing a technical basis for subsequent model upgrades and expansions. Attached Figure Description

[0055] Figure 1 A schematic diagram of the modeling method for a vehicle thermal management system based on global sensitivity data provided in an embodiment of the present invention;

[0056] Figure 2 This is a schematic diagram of the system structure of a vehicle thermal management system provided by the present invention. Detailed Implementation

[0057] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0058] It should be noted that the proposed solution is applicable to the modeling and development of thermal management systems for various types of vehicles, including pure electric vehicles and hybrid electric vehicles. The proposed solution divides modeling into two levels: component models and system models.

[0059] like Figure 1 As shown, this embodiment provides a modeling method for a vehicle thermal management system based on global sensitivity data, including the following steps:

[0060] Step S1: Build physical models of all components in the vehicle thermal management system;

[0061] Step S2: For each component's physical model, sensitive parameter combinations are extracted according to the following process, as follows:

[0062] Step S2.1: Extract the key parameters of the component physical model and set the numerical range of the key parameters;

[0063] Step S2.2: Import the component-level model dataset;

[0064] Step S2.3: Based on the key parameters, the key parameter matrix is ​​randomly generated using the Monte Carlo sampling method;

[0065] Step S2.4: Input the component-level model dataset and the key parameter matrix into the component physical model, and calculate the model output results;

[0066] Step S2.5: Based on the model output, calculate the first-order and total-order exponents of the key parameters with respect to the model output;

[0067] Step S2.6: Based on the magnitude of the first-order exponent and the total-order exponent, extract three to five of the most sensitive parameters for each model output result as a sensitive parameter combination;

[0068] Step S2.7: Establish the objective function between the model output and the measured values;

[0069] Step S2.8: Determine if there is a sensitive parameter combination that meets the target value requirement of the objective function; if so, directly output the corresponding component physical model and sensitive parameter combination; if not, select a sensitive parameter combination that is closest to the target value requirement, perform iterative calculation on the sensitive parameter combination, adjust the value of the sensitive parameter combination until the deviation of the objective function meets the standard, and output the corresponding component physical model and sensitive parameter combination.

[0070] Step S3: Integrate the component physical models of all components in the vehicle thermal management system obtained in Step S2 to obtain a system-level model; combine the imported system-level model dataset and the sensitive parameter combination obtained in Step S2 to perform iterative calculations until the objective function deviation between the system-level model output and the measured value meets the standard, and output the final sensitive parameter combination.

[0071] Accordingly, the present invention also provides a vehicle thermal management system modeling system based on global sensitivity data, comprising the following units: a component physical model building unit, a component physical model sensitive parameter combination extraction unit, and an integration unit;

[0072] The component physical model building unit is used to build the component physical models of all components in the vehicle thermal management system.

[0073] The component physical model sensitive parameter combination extraction unit is used to extract sensitive parameter combinations for each component physical model according to the following process:

[0074] Step S2.1: Extract the key parameters of the component physical model and set the numerical range of the key parameters;

[0075] Step S2.2: Import the component-level model dataset;

[0076] Step S2.3: Based on the key parameters, the key parameter matrix is ​​randomly generated using the Monte Carlo sampling method;

[0077] Step S2.4: Input the component-level model dataset and the key parameter matrix into the component physical model, and calculate the model output results;

[0078] Step S2.5: Based on the model output, calculate the first-order and total-order exponents of the key parameters with respect to the model output;

[0079] Step S2.6: Based on the magnitude of the first-order exponent and the total-order exponent, extract three to five of the most sensitive parameters for each model output result as a sensitive parameter combination;

[0080] Step S2.7: Establish the objective function between the model output and the measured values;

[0081] Step S2.8: Determine if there is a sensitive parameter combination that meets the target value requirement of the objective function; if so, directly output the corresponding component physical model and sensitive parameter combination; if not, select a sensitive parameter combination that is closest to the target value requirement, perform iterative calculation on the sensitive parameter combination, adjust the value of the sensitive parameter combination until the deviation of the objective function meets the standard, and output the corresponding component physical model and sensitive parameter combination.

[0082] The integration unit is used to stitch together and integrate the component physical models of all components in the obtained vehicle thermal management system to obtain a system-level model; combined with the imported system-level model dataset and the obtained sensitive parameter combination, iterative calculation is performed until the objective function deviation between the system-level model output result and the measured value meets the standard, and the final sensitive parameter combination is output.

[0083] like Figure 2 The diagram shown is a schematic of the system structure of a vehicle thermal management system provided by the present invention. The vehicle thermal management system is a heat pump air conditioning system, and its components include: a compressor, a liquid-cooled condenser (LCC), a condenser, a chiller, an evaporator, an electronic expansion valve, a battery pack, a water pump, etc.

[0084] Taking a heat pump air conditioning system as an example, the modeling process is as follows:

[0085] First: Build physical models of all components of the heat pump air conditioning system as constraints for subsequent processes;

[0086] Taking the identification and optimization of condenser component modeling parameters as an example, the component physical model is calculated based on the input refrigerant mass flow rate / pressure / enthalpy, air velocity / pressure / temperature / humidity, etc., to obtain refrigerant outlet pressure / enthalpy / temperature, air outlet temperature, wall temperature, refrigerant-side heat transfer, air-side heat transfer, etc. The specific formulas are as follows:

[0087] (1) Calculation of wall temperature:

[0088] ;

[0089] in: T wall The wall temperature; Q ref For the refrigeration side heat exchange (heat exchange between the refrigerant and the wall); Q air This refers to air-side heat exchange (heat exchange between air and the wall). m wall For wall surface quality; Cp wall Specific heat of the wall surface; t It is a time variable.

[0090] (2) The heat exchange on the refrigerant side of the condenser is divided into single-phase heat exchange (subcooled zone, superheated zone) and two-phase heat exchange (two-phase zone). The calculation formula for the heat exchange is the same, and the total heat exchange is equal to the sum of the heat exchange in the subcooled zone, superheated zone, and two-phase zone:

[0091] ;

[0092] ;

[0093] ;

[0094] ;

[0095] in: Q subcool For heat exchange in the supercooled zone; Q two-phase For heat exchange in the two-phase region; Q supheat For heat exchange in the overheated area; T subcool,ref This refers to the refrigerant temperature in the subcooled zone. h subcool,ref The convective heat transfer coefficient between the refrigerant and the wall in the subcooled zone; S subcool,ref This refers to the heat exchange area between the refrigerant and the wall surface in the subcooled zone. T two-phase,ref This refers to the refrigerant temperature in the subcooled zone. h two-phase,ref The convective heat transfer coefficient between the refrigerant and the wall in the subcooled zone; S two-phase,ref This refers to the heat exchange area between the refrigerant and the wall surface in the subcooled zone. T supheat,ref This refers to the refrigerant temperature in the subcooled zone. h supheat,ref The convective heat transfer coefficient between the refrigerant and the wall in the subcooled zone; S supheat,ref This refers to the heat exchange area between the refrigerant and the wall surface in the subcooled zone. K ref_heat This is the correction factor for the heat exchange on the refrigerant side.

[0096] (3) Heat transfer coefficient in single-phase region (subcooled region, superheated region):

[0097] ;

[0098] ;

[0099] in: ξ The coefficient of friction; Re It is the Reynolds number; Pr It is a Prandtl number; λThe thermal conductivity of the refrigerant; D For water conservancy diameter; ε The absolute coefficient of friction inside the microchannel;

[0100] (4) Heat transfer coefficient in the two-phase region:

[0101] ;

[0102] in: h LO The convective heat transfer coefficient when only liquid is flowing; x Dryness; ρ l The density of the liquid phase; ρ g This represents the gas phase density.

[0103] (5) The total pressure drop on the refrigerant side of the condenser is equal to the sum of the pressure drops in the subcooled zone, the superheated zone, and the two-phase zone:

[0104] ;

[0105] (6) Pressure drop in single-phase regions (subcooled region, superheated region):

[0106] ;

[0107] in, z For length; The average specific volume; G For mass flow rate.

[0108] (7) Pressure drop in the two-phase region:

[0109] ;

[0110] ;

[0111] in: A Calculation of pressure drop in saturated liquid phase; B This is for calculating the pressure drop in the saturated gas phase.

[0112] In summary, the refrigerant outlet temperature and pressure are obtained based on the total heat exchange, total pressure drop, refrigerant inlet temperature, and refrigerant inlet pressure on the refrigerant side. The refrigerant outlet enthalpy value is then found based on the refrigerant outlet temperature, outlet pressure, and pressure-enthalpy diagram.

[0113] Air-side heat exchange calculation:

[0114] ;

[0115] in: T air This refers to the air inlet temperature. Twall The wall temperature; h air The convective heat transfer coefficient between air and the wall; A chan_air This refers to the heat exchange area between the air and the wall. For fin surface efficiency; K air_heat This is the air-side heat transfer correction factor.

[0116] The air outlet temperature is calculated based on the air-side heat exchange, air-side inlet temperature, air specific heat, and air mass.

[0117] Second: Based on the component physical model, key parameters in the component physical model are extracted, and structural parameters are removed. At the same time, the numerical range of key parameters is set, and a multi-dimensional parameter numerical boundary space is constructed, as shown in Table 1.

[0118] Table 1

[0119] ;

[0120] Third: Import the component-level model dataset. If the data input is not standardized, it needs to be processed first. The inlet includes refrigerant-side flow rate / pressure / enthalpy and air-side air velocity / temperature / humidity. The outlet includes heat exchange, refrigerant outlet temperature, refrigerant-side pressure drop, air outlet temperature, and air-side pressure drop. The data should be as comprehensive as possible (containing more than 5 sets of data), and the test boundary coverage should be broad to ensure the accuracy and reliability of sensitivity parameters and subsequent calibration and optimization results. If the refrigerant inlet conditions are not inlet pressure and enthalpy, but inlet temperature and inlet pressure, the data needs to be preliminarily processed. The inlet enthalpy can be found from the pressure-enthalpy diagram first to ensure the standardization of data input.

[0121] Fourth: Based on key parameters, a key parameter matrix is ​​randomly generated using the Monte Carlo sampling method; including the following:

[0122] (1) The key parameters of the component physical model are: n (x1, x2, x3, …, x) n The sample size is... m Two matrices, A and B, are randomly generated using Monte Carlo sampling:

[0123] ;

[0124] ;

[0125] (2) Construct the key parameter matrix AB i Matrix, where i=1~ n AB iThe matrix is ​​formed by replacing the i-th column of matrix A with the i-th column of matrix B.

[0126] Fifth: Input the component-level model dataset and the key parameter matrix into the component physical model, and calculate the model output results;

[0127] Sixth: Based on the model output, calculate the first-order and total-order exponents of the key parameters with respect to the model output; as shown in the following formula:

[0128] ;

[0129] ;

[0130] in: V For variance; y i Output the results for the model; x i This is a key parameter in the physical model of the component. A larger exponent value indicates a greater impact on the model output value, making it a key indicator to focus on during system modeling.

[0131] Seventh: Based on the magnitude of the first-order exponent and the total-order exponent, extract three to five of the most sensitive parameters for each model output as a sensitive parameter combination; other parameters are no longer considered during modeling optimization, saving modeling time.

[0132] Eighth: Establish the objective function between the model output and the measured values; each model output should have a corresponding objective function; as follows:

[0133] ;

[0134] Where y=f(x1, x2, x3, …, x n )= y i ; y a This is the actual value. deltay The objective value of the objective function;

[0135] Ninth: Determine if there is a sensitive parameter combination that meets the target value requirement of the objective function; if so, directly output the corresponding component physical model and sensitive parameter combination; if not, select a sensitive parameter combination that is closest to the target value requirement, perform iterative calculation on the sensitive parameter combination, adjust the value of the sensitive parameter combination until the deviation of the objective function meets the standard, and output the corresponding component physical model and sensitive parameter combination.

[0136] The target value requirement refers to the fact that when the target value is within 5%, the sensitive parameter combination closest to the target value refers to the parameter combination whose target value is closest to 5%.

[0137] The values ​​of the sensitivity parameter combination are adjusted using the following formula:

[0138] ;

[0139] In the formula: x t These are the adjusted parameter values. x t-1 The parameter values ​​are those that bring the objective function's target value closest to 5%. k This is an adjustment factor.

[0140] Tenth: Building a system-level model of the heat pump air conditioning system;

[0141] Currently used heat pump air conditioning systems typically possess multiple functions, including both passenger compartment cooling and battery cooling / heating. Cooling-side components include compressors, liquid-cooled condensers (LCCs), condensers, evaporators, electronic expansion valves, and battery coolers, while coolant-side components include battery pack coolant channels and water pumps. The physical models of these components are integrated and combined, and signal transmission calculations are performed according to a unified naming convention. Specifically, the flow distribution calculation for the battery cooler and evaporator branches requires two key nodes: node 1 and node 2. Node 1 calculates the distribution of total flow, pressure, and enthalpy, while node 2 summarizes and calculates the flow, pressure, and enthalpy of both branches.

[0142] The calculation process for node 1 is as follows: Inputs include condenser outlet flow rate, outlet pressure, outlet enthalpy, battery cooler outlet pressure, evaporator outlet pressure, and the opening degrees of the two electronic expansion valves; outputs include battery cooler side inlet flow rate, inlet pressure, inlet enthalpy, and evaporator side inlet flow rate, inlet pressure, and inlet enthalpy. First, the value of k1 (0~1) is calculated based on the difference between the evaporator outlet pressure and the battery cooler outlet pressure. When the difference is greater than 0, the value of k1 gradually increases by 0.001 per step; when the difference is less than 0, the value of k1 gradually decreases by 0.001 per step to ensure the stability of the calculation process. Second, the evaporator side flow rate is calculated based on k1, the inlet flow rate, and the opening degree of the battery cooler side electronic expansion valve. When the opening degree of the battery cooler side electronic expansion valve is greater than 0, the evaporator side flow rate is the product of k1 and the inlet flow rate; otherwise, it is the inlet flow rate. Next, the battery cooler side flow rate is calculated based on k1, inlet flow rate, and the opening degree of the evaporator-side electronic expansion valve. When the opening degree of the evaporator-side electronic expansion valve is greater than 0, the battery cooler side flow rate is the product of (1-k1) and the inlet flow rate; otherwise, it is the inlet flow rate. Finally, the battery cooler side inlet pressure is the condenser outlet pressure, and the inlet enthalpy is the condenser outlet enthalpy. The evaporator side inlet pressure is the condenser outlet pressure, and the inlet enthalpy is the condenser outlet enthalpy.

[0143] The calculation process for node 2 is as follows: Inputs are the battery cooler outlet flow rate, outlet pressure, and outlet enthalpy; outputs are the total flow rate, compressor inlet pressure, and inlet enthalpy. First, the total flow rate is the sum of the flow rates of the two branches. Second, the compressor inlet pressure is the ratio of the sum of (evaporator outlet pressure * evaporator-side flow rate) and (battery cooler outlet pressure * battery cooler-side flow rate) to the total flow rate. Third, the compressor inlet enthalpy is the ratio of the sum of (evaporator outlet enthalpy * evaporator-side flow rate) and (battery cooler outlet enthalpy * battery cooler-side flow rate) to the total flow rate.

[0144] Eleventh: System-level model parameter optimization

[0145] After system-level model integration, system-level model parameters are optimized and adjusted. Referring to the component physical model parameter optimization method, the initial values ​​of system-level model parameters are the final determined values ​​of component physical model parameters.

[0146] Specifically, by combining the imported system-level model dataset and the obtained combination of sensitive parameters, iterative calculations are performed until the deviation of the objective function between the output of the system-level model and the measured value meets the standard, and the final combination of sensitive parameters is output.

[0147] Finally, it should be noted that the above embodiments are merely illustrative and explanatory of the present invention, and are not intended to limit the present invention to the scope of the described embodiments. Furthermore, those skilled in the art will understand that the present invention is not limited to the above embodiments, and many more variations and modifications can be made based on the teachings of the present invention, all of which fall within the scope of protection claimed by the present invention.

Claims

1. A globally sensitivity data-driven based vehicle thermal management system modeling method, characterized in that, Comprising the following steps: Step S1: building component physical models of all components in the vehicle thermal management system; Step S2: extracting sensitive parameter combinations for each component physical model according to the following process, specifically as follows: Step S2.1: extracting key parameters of the component physical model, while setting the numerical range of the key parameters; Step S2.2: importing component-level model data sets; Step S2.3: based on the key parameters, randomly generating a key parameter matrix using the Monte Carlo sampling method; Step S2.4: inputting the component-level model data sets and the key parameter matrix into the component physical model to calculate the model output results; Step S2.5: based on the model output results, calculating the first-order index and total-order index of the key parameters on the model output results; Step S2.6: according to the size of the first-order index and the total-order index value, for each model output result, extracting three to five most sensitive parameters as the sensitive parameter combination; Step S2.7: establishing a target function between the model output results and the measured values; Step S2.8: determining whether there is a sensitive parameter combination that meets the target value requirement of the target function; if so, directly output the corresponding component physical model and sensitive parameter combination; if not, select a sensitive parameter combination closest to the target value requirement, and iteratively calculate the sensitive parameter combination to adjust the value of the sensitive parameter combination until the target function deviation meets the standard, and output the corresponding component physical model and sensitive parameter combination; Step S3: integrating the component physical models of all components in the vehicle thermal management system obtained in step S2 to obtain a system-level model; combining the imported system-level model data sets and the sensitive parameter combinations obtained in step S2 to iteratively calculate until the target function deviation between the system-level model output results and the measured values meets the standard, and output the final sensitive parameter combination.

2. The global sensitivity data-driven vehicle thermal management system modeling method of claim 1, wherein, In step S1, the vehicle thermal management system is a heat pump air conditioning system, and the components thereof include: a compressor, a liquid-cooled condenser LCC, a condenser, a Chiller, an evaporator, an electronic expansion valve, and a battery pack.

3. The global sensitivity data-driven vehicle thermal management system modeling method of claim 1, wherein, In step S2.1, the process of extracting key parameters of the component physical model is as follows: sequentially determining whether the parameters of the component physical model are structure parameters, if so, eliminating; if not, retaining as key parameters.

4. The global sensitivity data-driven vehicle thermal management system modeling method of claim 1, wherein, The step S2.3 comprises the following: Step 2.3.1: Key parameters of the component physical model are n (x1, x2, x3, …, x n ), and the number of samples is m two matrices A and B are randomly formed by Monte Carlo sampling: ； ； Step 2.3.2: Constructing the key parameter matrix AB i matrix, where i = 1 n ; AB i The matrix is formed by replacing the i-th column of the A matrix with the i-th column of the B matrix.

5. The globally sensitivity data-driven based vehicle thermal management system modeling method of claim 4, wherein, In step S2.5, the first-order exponential of the key parameters on the model output result is calculated S i and the total order exponential S Ti as follows: ； ； wherein: V is the variance; y i is the model output result; x i is a key parameter of the component physical model.

6. The global sensitivity data-driven vehicle thermal management system modeling method of claim 5, wherein, In step S2.7, the target function between the model output results and the measured values is established as follows: ； where y = f(x1, x2, x3, …, x n )= y i ; y a is the actual value, deltay is the target value of the objective function.

7. The globally sensitivity data-driven based vehicle thermal management system modeling method of claim 6, wherein, In step S2.8, when the target value is within 5%, the target value requirement of the target function is met.

8. The globally sensitivity data-driven based vehicle thermal management system modeling method of claim 7, wherein, In step S2.8, the value of the sensitive parameter combination is adjusted by the following formula: ； In the formula: x t is the adjusted parameter value, x t-1 is the parameter value when the target value of the target function is closest to 5%, k is the adjustment factor.

9. A globally sensitivity data driven based vehicle thermal management system modeling system, characterized in that, Comprising the following units: a component physical model building unit, a component physical model sensitive parameter combination extraction unit, and an integration unit; The component physical model building unit is configured to build component physical models of all components in the vehicle thermal management system; The component physical model sensitive parameter combination extraction unit is configured to extract sensitive parameter combinations for each component physical model according to the following process, specifically as follows: Step S2.1: extract key parameters of the component physical model, and set a numerical range for the key parameters; Step S2.2: import the component-level model data set; Step S2.3: based on the key parameters, randomly generate a key parameter matrix by using the Monte Carlo sampling method; Step S2.4: input the component-level model data set and the key parameter matrix into the component physical model, and calculate the model output result; Step S2.5: based on the model output result, calculate the first-order index and the total-order index of the key parameters on the model output result; Step S2.6: according to the values of the first-order index and the total-order index, for each model output result, extract three to five most sensitive parameters as a sensitive parameter combination; Step S2.7: establish a target function between the model output result and the measured value; Step S2.8: determine whether there is a sensitive parameter combination that meets the target value requirement of the target function; if yes, directly output the corresponding component physical model and the sensitive parameter combination; if no, select a sensitive parameter combination that is closest to the target value requirement, and iteratively calculate the sensitive parameter combination to adjust the value of the sensitive parameter combination until the deviation of the target function reaches the standard, and output the corresponding component physical model and the sensitive parameter combination; The integration unit is configured to splice and integrate the obtained component physical models of all components in the vehicle thermal management system to obtain a system-level model; and iteratively calculate the obtained sensitive parameter combination in combination with the imported system-level model data set until the deviation of the target function between the system-level model output result and the measured value reaches the standard, and output the final sensitive parameter combination.

Images