Methods, devices and electronic equipment for evaluating aerodynamic noise of vehicle air conditioning system ducts

Abstract

This invention discloses a method, device, and electronic equipment for evaluating aerodynamic noise in vehicle air conditioning system ducts. The method includes: establishing a vehicle CFD model based on duct data, air outlet grille data, and passenger compartment data of the vehicle air conditioning system; simulating the flow field of the vehicle CFD model based on preset operating conditions, preset constant calculation parameters, and preset unsteady calculation parameters of the vehicle air conditioning system to obtain sound pressure data at preset monitoring points inside the vehicle; converting the sound pressure data at the preset monitoring points inside the vehicle into aerodynamic noise audio data, and determining the objective parameters and subjective evaluation scores of the aerodynamic noise audio data; training a neural network model based on the objective parameters and subjective evaluation scores to obtain a target neural network model, and evaluating the aerodynamic noise based on the target neural network model. This method can expose NVH risks of ducts in advance, effectively guide the NVH performance development of vehicle air conditioning system ducts, and help shorten the development cycle and reduce development costs.

Description

Technical Field

[0001] This invention relates to the field of duct aerodynamic noise evaluation technology, and in particular to a method, device and electronic equipment for evaluating duct aerodynamic noise in a vehicle air conditioning system. Background Technology

[0002] With the rapid development of science and technology, the technology of vehicle air conditioning systems has gradually improved, providing users with an increasingly convenient and comfortable vehicle experience. However, while providing a comfortable temperature environment for the driver and passengers, vehicle air conditioning systems also increase noise interference inside the vehicle, affecting the acoustic comfort of the in-vehicle environment. The noise generated by vehicle air conditioning systems includes mechanical noise, electromagnetic noise, and aerodynamic noise. When the vehicle air conditioning system is running at a low setting, mechanical and electromagnetic noise dominate; when it is running at a high setting, aerodynamic noise dominates.

[0003] Considering that aerodynamic noise interferes with users more than mechanical and electromagnetic noise, and that aerodynamic noise is highly correlated with the ductwork of vehicle air conditioning systems, current technologies involve mounting vehicle air conditioning system duct samples on a vehicle or fixing them to a test bench for aerodynamic noise testing. While this provides some understanding of the aerodynamic noise generated by the vehicle air conditioning system ducts, the exposure of duct NVH problems is usually delayed, hindering predictive management of duct NVH risks. Furthermore, obtaining accurate test results typically requires repeated testing and evaluation of the duct samples, leading to high development costs, time, and manpower for duct NVH performance, which does not meet the requirements for efficient and high-quality product performance development. Summary of the Invention

[0004] This invention aims to at least partially solve one of the technical problems in related technologies. Therefore, one objective of this invention is to propose a method for evaluating the aerodynamic noise of ductwork in vehicle air conditioning systems. This method can expose NVH risks of ductwork in advance, effectively guide the development of NVH performance of vehicle air conditioning system ductwork, and help shorten the development cycle and reduce development costs.

[0005] The second objective of this invention is to provide an electronic device.

[0006] The third objective of this invention is to provide a device for evaluating the aerodynamic noise of a vehicle air conditioning system duct.

[0007] To achieve the above objectives, a first aspect of the present invention provides a method for evaluating the aerodynamic noise of a vehicle air conditioning system duct, the method comprising:

[0008] A vehicle CFD model is established based on the duct data, air outlet grille data, and passenger compartment data of the vehicle's air conditioning system.

[0009] The flow field of the vehicle CFD model is simulated based on the preset operating conditions, preset constant calculation parameters and preset unsteady calculation parameters of the vehicle air conditioning system to obtain the sound pressure data of preset monitoring points inside the vehicle.

[0010] The sound pressure data from the preset monitoring points inside the vehicle are converted into aerodynamic noise audio data, and the objective parameters and subjective evaluation scores of the aerodynamic noise audio data are determined.

[0011] The neural network model is trained based on the objective parameters and the subjective evaluation scores to obtain the target neural network model, and the aerodynamic noise is evaluated based on the target neural network model.

[0012] According to the vehicle air conditioning system duct aerodynamic noise evaluation method of the present invention, the duct does not require vehicle-mounted testing or bench testing. It can accurately obtain sound pressure data of preset monitoring points inside the vehicle based on the vehicle CFD model, and convert the obtained sound pressure data into aerodynamic noise audio data. Based on the objective parameters and subjective evaluation scores of the aerodynamic noise audio data, a target neural network model is determined. Thus, aerodynamic noise can be directly evaluated based on the target neural network model, which is beneficial for early exposure of NVH risks of the duct. Furthermore, the subjective evaluation results output by the target neural network model can effectively guide the NVH performance development of the vehicle air conditioning system duct, which helps to shorten the development cycle and reduce development costs.

[0013] In some embodiments of the present invention, establishing a vehicle CFD model based on the duct data, air outlet grille data, and passenger compartment data of the vehicle air conditioning system includes: establishing a geometric model based on the duct data, air outlet grille data, and passenger compartment data of the vehicle air conditioning system; processing the geometric model to obtain a boundary geometric model corresponding to the CFD calculation domain of aerodynamic noise; and meshing the geometric model based on the boundary geometric model to obtain the vehicle CFD model. Simulating the flow field of the vehicle CFD model based on preset operating conditions, preset constant calculation parameters, and preset unsteady calculation parameters of the vehicle air conditioning system includes: performing steady-state calculations on the vehicle CFD model based on the preset operating conditions and preset constant calculation parameters of the vehicle air conditioning system to obtain initial sound pressure data; and performing unsteady calculations on the vehicle CFD model based on the initial sound pressure data and preset unsteady calculation parameters to obtain sound pressure data at preset monitoring points inside the vehicle.

[0014] In some embodiments of the present invention, the step of converting the sound pressure data of the preset monitoring points inside the vehicle into aerodynamic noise audio data includes: determining the sampling frequency according to the preset unsteady calculation parameters; and sampling the sound pressure data of the preset monitoring points inside the vehicle according to the sampling frequency to obtain the aerodynamic noise audio data.

[0015] In some embodiments of the present invention, determining the objective parameters of the aerodynamic noise audio data includes: performing a fast Fourier transform on the aerodynamic noise audio data to obtain an energy spectrum; and determining the energy of a first frequency range, the energy of a second frequency range, the energy of a third frequency range, and the A-weighted sound pressure level based on the energy spectrum, wherein the minimum value of the second frequency range is greater than or equal to the maximum value of the first frequency range, and the maximum value of the second frequency range is less than or equal to the minimum value of the third frequency range.

[0016] In some embodiments of the present invention, the step of training a neural network model based on the objective parameters and the subjective evaluation scores includes: allocating the objective parameters and the subjective evaluation scores into training samples, test samples, and prediction samples according to a preset ratio; training a neural network model based on the training samples, and obtaining a neural network model to be optimized when the error is less than or equal to a first threshold; optimizing the neural network model to be optimized based on the test samples and a genetic algorithm to obtain the target neural network model; performing subjective evaluation prediction on the objective parameters in the prediction samples based on the target neural network model, and determining the effectiveness of the target neural network model based on the error between the first prediction score of the subjective evaluation prediction and the subjective evaluation score in the prediction samples.

[0017] In some embodiments of the present invention, the preset ratio is 6:2:2.

[0018] In some embodiments of the present invention, optimizing the neural network model to be optimized based on the test samples and the genetic algorithm includes: inputting objective parameters from the test samples into the neural network model to be optimized to obtain a second prediction score; calculating the mean squared error between the second prediction score and the subjective evaluation score from the test samples, and using the mean squared error as the fitness value of the genetic algorithm; and optimizing the neural network model to be optimized using the genetic algorithm based on the fitness value.

[0019] In some embodiments of the present invention, the genetic algorithm has 50 generations, the crossover probability is 0.7, and the mutation probability is 0.1.

[0020] To achieve the above objectives, a second aspect of the present invention provides an electronic device, the electronic device including a memory, a processor, and a vehicle air conditioning system duct aerodynamic noise evaluation program stored in the memory and executable on the processor. When the processor executes the duct aerodynamic noise evaluation program, it implements the vehicle air conditioning system duct aerodynamic noise evaluation method of any of the above embodiments.

[0021] According to the electronic device of the present invention, the duct does not require vehicle testing or bench testing. It can accurately obtain sound pressure data of preset monitoring points inside the vehicle based on the vehicle CFD model, and convert the obtained sound pressure data into aerodynamic noise audio data. Based on the objective parameters and subjective evaluation scores of the aerodynamic noise audio data, the target neural network model is determined. Thus, the aerodynamic noise can be directly evaluated based on the target neural network model, which is beneficial for early exposure of NVH risks of the duct. Furthermore, the subjective evaluation results output by the target neural network model can effectively guide the development of NVH performance of the vehicle air conditioning system duct, which is beneficial for shortening the development cycle and reducing development costs.

[0022] To achieve the above objectives, a third aspect of the present invention provides an aerodynamic noise evaluation device for a vehicle air conditioning system duct. The device includes: a modeling module for establishing a vehicle CFD model based on duct data, air outlet grille data, and passenger compartment data of the vehicle air conditioning system; a simulation module for simulating the flow field of the vehicle CFD model based on preset operating conditions, preset constant calculation parameters, and preset unsteady calculation parameters of the vehicle air conditioning system to obtain sound pressure data at preset monitoring points inside the vehicle; a determination module for converting the sound pressure data at the preset monitoring points inside the vehicle into aerodynamic noise audio data and determining the objective parameters and subjective evaluation scores of the aerodynamic noise audio data; and a training module for training a neural network model based on the objective parameters and the subjective evaluation scores to obtain a target neural network model, and evaluating aerodynamic noise based on the target neural network model.

[0023] According to the vehicle air conditioning system duct aerodynamic noise evaluation device of the present invention, the duct does not require vehicle-mounted testing or bench testing. It can accurately obtain sound pressure data of preset monitoring points inside the vehicle based on the vehicle CFD model, and convert the obtained sound pressure data into aerodynamic noise audio data. Based on the objective parameters and subjective evaluation scores of the aerodynamic noise audio data, a target neural network model is determined. Thus, aerodynamic noise can be directly evaluated based on the target neural network model, which is beneficial for early exposure of NVH risks of the duct. Furthermore, the subjective evaluation results output by the target neural network model can effectively guide the NVH performance development of the vehicle air conditioning system duct, which helps to shorten the development cycle and reduce development costs.

[0024] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0025] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, wherein:

[0026] Figure 1 This is a flowchart illustrating a method for evaluating aerodynamic noise in ducts according to an embodiment of the present invention.

[0027] Figure 2 This is a flowchart illustrating a method for evaluating aerodynamic noise in ducts according to another embodiment of the present invention;

[0028] Figure 3 This is a schematic diagram of the boundary geometry model in a duct aerodynamic noise evaluation method according to an embodiment of the present invention;

[0029] Figure 4 This is a flowchart illustrating a method for evaluating aerodynamic noise in ducts according to another embodiment of the present invention;

[0030] Figure 5 This is a flowchart illustrating a method for evaluating aerodynamic noise in ducts according to another embodiment of the present invention;

[0031] Figure 6 This is a flowchart illustrating a method for evaluating aerodynamic noise in ducts according to another embodiment of the present invention;

[0032] Figure 7 This is a schematic diagram of a scenario involving a subjective evaluation control panel in a duct aerodynamic noise evaluation method according to an embodiment of the present invention.

[0033] Figure 8 This is a flowchart illustrating a method for evaluating aerodynamic noise in ducts according to another embodiment of the present invention;

[0034] Figure 9 This is a schematic diagram of an objective parameter information panel in a duct aerodynamic noise evaluation method according to an embodiment of the present invention;

[0035] Figure 10 This is a flowchart illustrating a method for evaluating aerodynamic noise in ducts according to another embodiment of the present invention;

[0036] Figure 11 This is a schematic diagram of the topology of a neural network model in a duct aerodynamic noise evaluation method according to an embodiment of the present invention;

[0037] Figure 12This is a schematic diagram comparing the first predicted score and the actual subjective evaluation score in a duct aerodynamic noise evaluation method according to an embodiment of the present invention.

[0038] Figure 13 This is a flowchart illustrating a method for evaluating aerodynamic noise in ducts according to another embodiment of the present invention;

[0039] Figure 14 This is a flowchart illustrating a method for evaluating aerodynamic noise in ducts according to another embodiment of the present invention;

[0040] Figure 15 This is a schematic diagram showing the change of the minimum mean square error value of the genetic algorithm in a duct aerodynamic noise evaluation method according to an embodiment of the present invention.

[0041] Figure 16 This is a schematic diagram comparing the output value of the target neural network model with the subjective evaluation score in a duct aerodynamic noise evaluation method according to an embodiment of the present invention;

[0042] Figure 17 This is a structural block diagram of an electronic device according to an embodiment of the present invention;

[0043] Figure 18 This is a structural block diagram of a duct aerodynamic noise evaluation device according to an embodiment of the present invention. Detailed Implementation

[0044] Embodiments of the present invention are described in detail below. Examples of these embodiments are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0045] To clearly illustrate the method, apparatus, and electronic equipment for evaluating the aerodynamic noise of the vehicle air conditioning system according to embodiments of the present invention, the following is in conjunction with... Figure 1 The flowchart illustrating the method for evaluating aerodynamic noise in ductwork is described below. Figure 1 As shown, the method for evaluating the aerodynamic noise of a vehicle air conditioning system duct in this embodiment of the invention includes the following steps:

[0046] S11: Establish a vehicle CFD model based on the vehicle's air conditioning system duct data, air outlet grille data, and passenger compartment data.

[0047] S13: Simulate the flow field of the vehicle CFD model based on the preset operating conditions, preset constant calculation parameters and preset unsteady calculation parameters of the vehicle air conditioning system to obtain the sound pressure data of preset monitoring points inside the vehicle.

[0048] S15: Convert the sound pressure data of the preset monitoring points inside the vehicle into aerodynamic noise audio data, and determine the objective parameters and subjective evaluation scores of the aerodynamic noise audio data;

[0049] S17: Train a neural network model based on objective parameters and subjective evaluation scores to obtain a target neural network model, and evaluate aerodynamic noise based on the target neural network model.

[0050] According to the vehicle air conditioning system duct aerodynamic noise evaluation method of the present invention, the duct does not require vehicle-mounted testing or bench testing. It can accurately obtain sound pressure data of preset monitoring points inside the vehicle based on the vehicle CFD model, and convert the obtained sound pressure data into aerodynamic noise audio data. Based on the objective parameters and subjective evaluation scores of the aerodynamic noise audio data, a target neural network model is determined. Thus, aerodynamic noise can be directly evaluated based on the target neural network model, which is beneficial for early exposure of NVH risks of the duct. Furthermore, the subjective evaluation results output by the target neural network model can effectively guide the NVH performance development of the vehicle air conditioning system duct, which helps to shorten the development cycle and reduce development costs.

[0051] It is understandable that since the air ducts of a vehicle's air conditioning system face directly towards the passengers and driver, the aerodynamic noise inside the vehicle mainly originates from the air duct outlets during operation. In related technologies, the control of aerodynamic noise in vehicle air conditioning systems primarily focuses on the individual unit and component-level air ducts, i.e., identifying and controlling NVH risks solely through individual duct CFD simulations and experiments. The simulation control indicators for individual ducts in related technologies include the pressure loss P at the duct inlet and outlet. lose and outlet flow velocity V out Pressure loss P lose The calculation formula is: P lose =P in -P out , where P in P indicates the total pressure at the duct inlet. out This represents the total pressure at the duct outlet. The pressure and velocity values ​​are obtained from the flow field results calculated by CFD for the duct, and then determined based on the pressure loss P at the duct inlet and outlet. lose and outlet flow velocity V out The magnitude of the value determines the intensity of aerodynamic noise. That is, the pressure loss P at the inlet and outlet of the duct. lose and outlet flow velocity V out The larger the value, the stronger the aerodynamic noise; the pressure loss P at the inlet and outlet of the duct. lose and outlet flow velocity V out The smaller the value, the weaker the aerodynamic noise.

[0052] In other words, the relevant technologies are limited to the prediction and evaluation of aerodynamic noise at the component level of ducts, and cannot achieve the prediction and evaluation of aerodynamic noise at the vehicle level. Furthermore, the relevant technologies can only indirectly predict the aerodynamic noise generated by the duct through the pressure loss at the duct inlet and outlet and the outlet flow velocity, and cannot directly quantitatively predict and evaluate the objective parameters of aerodynamic noise. In addition, since the relevant technologies only predict and evaluate the aerodynamic noise of ducts at the component level, after the ducts are installed in the vehicle, the actual aerodynamic noise generated may deviate significantly from the prediction results depending on the actual vehicle environment, which cannot guarantee the NVH performance of the ducts during actual use and affects the user experience.

[0053] In the duct aerodynamic noise evaluation method of this invention, a vehicle-level CFD model is pre-constructed. By simulating the vehicle-level CFD model, relatively accurate and interference-free sound pressure data at preset monitoring points inside the vehicle, corresponding to the overall vehicle environment, can be directly obtained. The sound pressure data is converted into aerodynamic noise audio data, facilitating direct and intuitive evaluation of aerodynamic noise and calculation of its objective parameters. Furthermore, a target neural network model is trained based on the objective parameters and subjective evaluation scores. This allows for a more accurate and efficient evaluation of the aerodynamic noise generated by the duct inside the vehicle, effectively guiding the development of aerodynamic noise control for vehicle air conditioning systems.

[0054] Specifically, the duct data, air vent grille data, and passenger compartment data of the vehicle's air conditioning system can be pre-generated data. The vehicle CFD model can be built directly based on the pre-generated data, which helps to improve development efficiency and ensure development quality.

[0055] The duct data of the vehicle air conditioning system can include duct data of various duct schemes, the air outlet grille data can include air outlet grille data of various air outlet grille schemes, and the passenger compartment data can include passenger compartment data of various vehicle models. Based on the duct data, air outlet grille data, and passenger compartment data of the vehicle air conditioning system, multiple vehicle CFD models can be established. Then, based on the preset operating conditions, preset constant calculation parameters, and preset unsteady calculation parameters of the vehicle air conditioning system, the flow field of each vehicle CFD model can be simulated, and multiple sets of sound pressure data of preset monitoring points inside the vehicle can be obtained.

[0056] Furthermore, the sound pressure data from each set of preset monitoring points inside the vehicle are converted to obtain multiple sets of aerodynamic noise audio data. Objective parameters and subjective evaluation scores are calculated for each set of aerodynamic noise audio data. The obtained objective parameters are used as input to the neural network model, and the obtained subjective evaluation scores are used as output to train the neural network model. This results in a target neural network model that can effectively replace the tedious manual evaluation work in the later stages of developing the aerodynamic noise performance of the vehicle's air conditioning system ducts. When the objective parameters of unknown aerodynamic noise audio data are input into the target neural network model, the model can automatically, accurately, and efficiently output the user's objective perception corresponding to the unknown aerodynamic noise audio data.

[0057] Please see Figure 2 In some embodiments of the present invention, step S11 includes:

[0058] S111: Establish a geometric model based on the vehicle's air conditioning system duct data, air outlet grille data, and passenger compartment data;

[0059] S113: Process the geometric model to obtain the boundary geometric model corresponding to the CFD computational domain of aerodynamic noise;

[0060] S115: Mesh the geometric model based on the boundary geometric model to obtain the vehicle CFD model.

[0061] In this way, by establishing a vehicle CFD model, it is possible to simulate the aerodynamic noise of the entire vehicle duct without testing on a test bench, and it is also convenient for subsequent whole vehicle testing and optimization, shortening the development cycle and reducing development costs.

[0062] Specifically, the duct data for the vehicle's air conditioning system may include one or more of the following: duct shape, duct size, and duct installation location within the vehicle. The air vent grille corresponds to the grille installed at the air outlet of the vehicle's air conditioning system duct. Air vent grille data may include one or more of the following: air vent grille shape, air vent grille size, and air vent grille installation location within the vehicle. Passenger compartment data may include one or more of the following: seat shape, seat size, number of seats, seat installation location within the passenger compartment, passenger compartment shape, and passenger compartment size. The duct data, air vent grille data, and passenger compartment data for the vehicle's air conditioning system can be pre-generated data. Directly building a geometric model based on this pre-generated data improves development efficiency and ensures development quality.

[0063] Considering that the mesh settings of the boundary region of the geometric model and the CFD computation domain of the geometric model are usually different when performing mesh generation, the boundary geometric model corresponding to the CFD computation domain of aerodynamic noise is pre-determined by processing the geometric model. After determining the boundary geometric model, the boundary region and the CFD computation domain in the geometric model are meshed according to the boundary geometric model to obtain the CFD model.

[0064] In one example, taking the roof duct of the air conditioning system of a large SUV as an example, the resulting boundary geometry model is as follows: Figure 3 As shown in the figure, the boundary geometry model 100 includes the crew compartment 12, the air duct 14, and the air outlet grille 16.

[0065] In some embodiments, prior to step S115, the method further includes determining mesh generation parameters. Mesh generation parameters may include volume mesh type, number of boundary layer mesh layers, growth rate, total thickness, and basic size range of the volume mesh. After determining the mesh generation parameters, the boundary region and CFD computational domain in the geometric model are meshed according to the mesh generation parameters and the boundary geometric model, respectively.

[0066] In one example, the mesh division parameters for the passenger compartment, air duct, air outlet grille, air duct grille wall area, and passenger compartment wall area are set as shown in Table 1.

[0067] Table 1

[0068] Control Area Mesh generation parameters Crew cabin Tetrahedral, three prism layers, growth rate 1.1, total thickness 10.0mm, 3.0-15.0mm air duct Tetrahedral, three prism layers, growth rate 1.1, total thickness 1.0mm / 2.0mm, 1.0-3.0mm Air vent grille Tetrahedral, three prismatic layers, growth rate 1.1, total thickness 0.5mm, 0.5-1.0mm Duct grille wall area Tetrahedral, three prism layers, growth rate 1.1, total thickness 0.5mm, <0.5mm Crew cabin wall area Tetrahedral, three prism layers, growth rate 1.1, total thickness 3.0 mm, <10.0 mm

[0069] Please see Figure 4 In some embodiments of the present invention, step S13 includes:

[0070] S131: Perform steady-state calculations on the vehicle CFD model based on the preset operating conditions and preset constant calculation parameters of the vehicle air conditioning system to obtain initial sound pressure data.

[0071] S133: Perform unsteady calculations on the vehicle CFD model based on the initial sound pressure data and preset unsteady calculation parameters to obtain the sound pressure data of preset monitoring points inside the vehicle.

[0072] In this way, the sound pressure data corresponding to the aerodynamic noise of the whole vehicle duct can be obtained more accurately.

[0073] Specifically, the preset operating conditions of the vehicle air conditioning system may include the system's operating level. Preset constant calculation parameters may include a first turbulence model, a first iteration number, boundary conditions, and a Y+ value. Boundary conditions may include inlet flow boundary, outlet pressure boundary, and wall boundary. The inlet flow boundary can be understood as the gas flow rate corresponding to the duct outlet. The inlet flow boundary can be determined based on the vehicle air conditioning system's operating level. It can be understood that there is a correspondence between the operating level and the gas flow rate; after determining the vehicle air conditioning system's operating level, the gas flow rate can be determined based on this correspondence, and thus the inlet flow boundary can be determined based on the gas flow rate. The gas flow rate can be either volumetric flow rate or mass flow rate; gas mass flow rate and gas volumetric flow rate can be converted to each other based on gas density. The outlet pressure boundary can be understood as the pressure corresponding to the passenger compartment pressure relief valve outlet. The wall boundary may include a non-slip wall. In one embodiment, the areas in the CFD model other than the duct outlet and the passenger compartment pressure relief valve outlet are set as non-slip wall boundaries.

[0074] Preset unsteady calculation parameters may include the second turbulence model, acoustic model, time step, second iteration number, and solution time. The sound pressure data at the preset monitoring points inside the vehicle are time-domain sound pressures, which can reflect the change of sound pressure at the preset monitoring points inside the vehicle over time.

[0075] In one example, the preset constant calculation parameters are set as shown in Table 2, and the preset unsteady calculation parameters are set as shown in Table 3. It can be understood that during the steady-state calculation process, the first iteration number of the preset constant calculation parameters is a variable value, related to the first turbulence model, and the maximum value of the first iteration number can be set to 2000.

[0076] Table 2

[0077] Settings Setting value First Turbulence Model Low Re number SST-K-Omega Import flow boundary Mass flow rate corresponding to the setting of the vehicle's air conditioning system wall boundary No-slip wall, wall function + low Reynolds number model Export pressure boundary Atmospheric pressure Y+ value 1-20

[0078] Table 3

[0079]

[0080]

[0081] In some embodiments of the present invention, the preset monitoring points inside the vehicle include one or more of the following: the right ear area of ​​the driver, the right ear area of ​​the middle left passenger, the right ear area of ​​the rear left passenger, and the air vent grille area.

[0082] In this way, sound pressure data at different locations can be obtained according to actual needs, thereby determining the aerodynamic noise audio data at different locations.

[0083] Specifically, one area can be selected as the preset monitoring point inside the vehicle, or two or more areas can be selected as preset monitoring points inside the vehicle; there is no limitation here. When the vehicle CFD model includes seats and air vent grilles, the driver's right ear area, the right ear area of ​​the middle left passenger, and the right ear area of ​​the rear left passenger can be determined based on the position of the seat headrest, and the air vent grille area can be determined based on the position and size of the air vent grille.

[0084] In one example, the area 10cm to the right of the driver's seat headrest is designated as the driver's right ear area, the area 10cm to the right of the middle row left seat headrest is designated as the middle left row passenger's right ear area, and the area 10cm to the right of the rear row left seat headrest is designated as the rear left row passenger's right ear area.

[0085] Please see Figure 5 In some embodiments of the present invention, step S15 includes:

[0086] S151: Determine the sampling frequency based on preset unsteady calculation parameters;

[0087] S152: Sample the sound pressure data of preset monitoring points inside the vehicle according to the sampling frequency to obtain aerodynamic noise audio data.

[0088] Thus, compared to related technologies that indirectly predict the aerodynamic noise generated by the duct based on the pressure loss at the inlet and outlet of the duct and the outlet flow velocity, the present invention can more intuitively perceive and evaluate the aerodynamic noise generated by the duct of the vehicle air conditioning system inside the vehicle.

[0089] Specifically, the sampling frequency is kept consistent with the time step of the preset unsteady calculation parameters, thus ensuring that the collected aerodynamic noise audio data fully retains the detailed information in the original signal. In some embodiments, the aerodynamic noise audio data is in .wav audio format.

[0090] In one example, the sampling frequency is 10. -4 The aerodynamic noise audio data obtained by sampling sound pressure data shows a peak accuracy of over 90% and an A-weighted total sound pressure level accuracy of over 90% within 2000Hz, compared to aerodynamic noise-related data collected by a noise sensor. This indicates that the aerodynamic noise audio data obtained by this method is accurate and reliable. It can be understood that, according to the Nyquist sampling theorem, when the sampling frequency is 10... -4 At time s, signals with frequencies up to 5000Hz can be acquired.

[0091] Please see Figure 6 In some embodiments of the present invention, step S15 includes:

[0092] S153: Based on the same scoring mechanism, obtain scores from multiple users for aerodynamic noise audio data;

[0093] S154: Calculate the average score of multiple users' ratings of the aerodynamic noise audio data, and use the average score as the subjective evaluation score of the aerodynamic noise audio data.

[0094] In this way, the bias of personal evaluation is effectively overcome, and the subjective evaluation score more objectively and truthfully reflects the user's subjective feelings about the current aerodynamic noise audio data.

[0095] In some embodiments, the rating mechanism uses a 10-point scale, where a higher score indicates a better subjective experience and a lower score indicates a worse subjective experience.

[0096] The more users participate in the subjective evaluation, the more objective and truthful the final subjective evaluation score will be. In some embodiments, the subjective evaluation score of the aerodynamic noise audio data is determined based on the average score of at least six users' ratings of the aerodynamic noise audio data.

[0097] In some embodiments, step S153 further includes: providing a subjective evaluation control panel, receiving a first input command input by a user on the subjective evaluation control panel, and determining the user's rating of the aerodynamic noise audio data based on the received first input command.

[0098] Please combine Figure 7In one example, the subjective evaluation control panel includes a playback control area, a sample selection area, an evaluator registration area, a booming sound scoring input area, a whooshing sound scoring input area, a hissing sound scoring input area, and a comprehensive score display area. Playback control commands input in the playback control area can be used to play, pause, continue, switch to the next, or switch to the previous aerodynamic noise audio. Sample selection commands input in the sample selection area can be used to select the aerodynamic noise audio to be played. Different users can be distinguished based on unique personal identification information input in the evaluator registration area; this personal identification information may include one or more of numbers, letters, symbols, and Chinese characters. The booming sound scoring input area may include a first anchor point, a first progress bar, and a first score display box. Each position on the first progress bar corresponds to a score, and the first displayed value of the first score display box can be determined based on the stationary position of the first anchor point on the first progress bar. The whooshing sound scoring input area may include a second anchor point, a second progress bar, and a second score display box. Each position on the second progress bar corresponds to a score, and the second displayed value of the second score display box can be determined based on the stationary position of the second anchor point on the second progress bar. The hissing sound scoring input area may include a third anchor point, a third progress bar, and a third score display box. Each position on the third progress bar corresponds to a score. The third displayed value of the third score display box can be determined based on the stationary position of the third anchor point on the third progress bar. The comprehensive scoring display area may include a fourth anchor point, a fourth progress bar, and a fourth score display box. After determining the first, second, and third displayed values, the average of the first, second, and third displayed values ​​is calculated, and the calculation result is used as the current user's score for the current aerodynamic noise audio data. Simultaneously, the score is displayed in the fourth score display box, and the fourth anchor point is controlled to remain stationary at the position on the fourth progress bar corresponding to the score.

[0099] Please see Figure 8 In some embodiments of the present invention, step S15 includes:

[0100] S155: Perform a fast Fourier transform on the aerodynamic noise audio data to obtain the energy spectrum;

[0101] S156: Determine the energy of the first frequency range, the energy of the second frequency range, the energy of the third frequency range, and the A-weighted sound pressure level based on the energy spectrum, wherein the minimum value of the second frequency range is greater than or equal to the maximum value of the first frequency range, and the maximum value of the second frequency range is less than or equal to the minimum value of the third frequency range.

[0102] In this way, four objective parameters of the aerodynamic noise audio data can be determined.

[0103] Specifically, when performing the Fast Fourier Transform, the Hanning window can be selected as the calculation window function, the analysis time can be set to 0.2048s, and a total of 20 segments can be analyzed to obtain the root mean square value as the spectral output of the FFT, so as to obtain the FFT energy spectrum.

[0104] Objective parameters may include the energy in a first frequency range, the energy in a second frequency range, the energy in a third frequency range, and the A-weighted sound pressure level. In some embodiments, the first frequency range is 20-200Hz, the second frequency range is 200-500Hz, and the third frequency range is 500-2000Hz. It can be understood that the human ear perceives aerodynamic noise in the 20-200Hz range as a "rumbling" sound, aerodynamic noise in the 200-500Hz range as a "whooshing" sound, and aerodynamic noise in the 500-2000Hz range as a "hissing" sound. This establishes a correlation between objective parameters and subjective evaluation scores.

[0105] In some implementations, the energy of the first frequency range, the energy of the second frequency range, the energy of the third frequency range, and the A-weighted sound pressure level can be determined using Matlab software based on the FFT energy spectrum.

[0106] In some embodiments, step S15 further includes: providing an objective parameter information panel, receiving a second input command from the user on the objective parameter information panel, and performing a fast Fourier transform on the aerodynamic noise audio data according to the received second input command; displaying the energy spectrum, the energy of the first frequency range, the energy of the second frequency range, the energy of the third frequency range, and the A-weighted sound pressure level. Please refer to... Figure 9 In one example, the objective parameter information panel includes an energy spectrum calculation parameter setting area, an objective parameter calculation result display area, a time domain data display area for the current aerodynamic noise audio data, and an energy spectrum display area for the current aerodynamic noise audio data.

[0107] Please see Figure 10 In some embodiments of the present invention, step S17 includes:

[0108] S171: According to a preset ratio, objective parameters and subjective evaluation scores are allocated to training samples, test samples, and prediction samples;

[0109] S173: Train the neural network model based on the training samples, and obtain the neural network model to be optimized when the error is less than or equal to the first threshold.

[0110] S175: Optimize the neural network model to be optimized based on the test samples and the genetic algorithm to obtain the target neural network model;

[0111] S177: Subjectively evaluate and predict the objective parameters in the prediction sample based on the target neural network model, and determine the effectiveness of the target neural network model based on the error between the first prediction score of the subjective evaluation and the subjective evaluation score in the prediction sample.

[0112] In this way, a target neural network model is obtained to evaluate newly generated aerodynamic noise audio data, replacing the manual evaluation of newly generated aerodynamic noise audio data, while ensuring that the evaluation results are true and objective.

[0113] Specifically, in some embodiments of the present invention, a preset ratio of 6:2:2 is used, meaning that 6 parts of the obtained objective parameters and subjective evaluation scores are used as training samples, 2 parts are used as test samples, and 2 parts are used as prediction samples. This ensures that the obtained target neural network model is realistic and reliable. In one example, if a total of 40 aerodynamic noise audio data points are obtained, then the objective parameters and subjective evaluation scores of 24 of these aerodynamic noise audio data points are used as training samples, the objective parameters and subjective evaluation scores of 8 of these aerodynamic noise audio data points are used as test samples, and the objective parameters and subjective evaluation scores of 8 of these aerodynamic noise audio data points are used as prediction samples.

[0114] Neural network models can include back propagation (BP) neural networks. A BP neural network can include an input layer, hidden layers, and an output layer. The relationship between the number of neurons in the hidden layer (n2) and the number of neurons in the input layer (n1) can be expressed by the formula: n2 = 2n1 + 1. In one example, when the objective parameters are 20-200Hz energy, 200-500Hz energy, 500-2000Hz energy, and A-weighted sound pressure level, the topology of the BP neural network is as follows: Figure 11 As shown, the topology of the BP neural network consists of 4 input layer neurons, 9 hidden layer neurons, and 1 output layer neuron.

[0115] In some implementations, the first threshold is 0.01, meaning that the neural network model to be optimized is obtained when the error between the output obtained by the neural network model based on the objective parameters in the training samples and the subjective evaluation scores in the training samples is less than or equal to 0.01.

[0116] In one example, please combine Figure 12 , Figure 12The figure shows a comparison between the first predicted score obtained by subjectively evaluating the objective parameters in the predicted sample based on the target neural network model and the actual subjective evaluation score in the predicted sample. It can be seen from the figure that the target neural network model has high prediction accuracy, and the error range between it and the actual subjective evaluation score in the predicted sample is within 5%. That is, the effectiveness of the target neural network model is 5%. The target neural network model of this method has good generalization prediction ability and can be used for subjective evaluation of unknown aerodynamic noise audio data.

[0117] Please see Figure 13 In some embodiments of the present invention, step S175 includes:

[0118] S1751: Input the objective parameters from the test samples into the neural network model to be optimized to obtain the second prediction score;

[0119] S1753: Calculate the mean squared error (MSE) between the second predicted score and the subjective evaluation scores in the test sample, and use the mean squared error as the fitness value of the genetic algorithm;

[0120] S1755: The genetic algorithm is used to optimize the neural network model to be optimized based on the fitness value.

[0121] In this way, a genetic algorithm can be used to optimize the neural network model to be optimized and obtain the target neural network model.

[0122] Specifically, the specific steps for optimizing the neural network model to be optimized are as follows: Figure 14 As shown, S211: Determine the topology of the neural network model; S212: Encode the weights and thresholds of the neural network model to obtain the initial population; S213: Decode to obtain the current weights and thresholds; S214: Assign the current weights and thresholds to the neural network model to be optimized; S215: Test the neural network model to be optimized using test samples; S216: Test the mean squared error (MSE); S217: Calculate the fitness; S218: Select the chromosomes with the highest fitness (top 95%) for replication; S219: Crossover operation; S220: Mutation operation; S221: Generate the next generation population; S222: Determine if the required number of generations is met. If not, return to step S212; if met, proceed to step S223; S223: Decode; S224: Obtain the target weights and thresholds of the neural network model; S225: Obtain the target neural network model.

[0123] In one example, the minimum mean squared error of the subjective evaluation prediction scores of all test samples in each generation of the genetic algorithm is as follows: Figure 15As shown, the target weights and target thresholds of the neural network model are obtained. The output value of the target neural network based on the objective parameters in the training samples is compared with the subjective evaluation scores in the training samples. Figure 16 As shown.

[0124] In some embodiments of the present invention, the number of generations of the genetic algorithm is 50, the crossover probability of the genetic algorithm is 0.7, and the mutation probability of the genetic algorithm is 0.1.

[0125] It should be noted that the specific values ​​mentioned above are only for illustrating the implementation of the present invention in detail, and should not be construed as limiting the present invention. In other examples, implementation methods, or embodiments, other values ​​may be selected according to the present invention, and no specific limitations are made here.

[0126] To implement the above embodiments, this invention also proposes a computer-readable storage medium storing a duct aerodynamic noise evaluation program for a vehicle air conditioning system. When the duct aerodynamic noise evaluation program is executed by a processor, it implements the duct aerodynamic noise evaluation method for a vehicle air conditioning system according to any of the above embodiments.

[0127] According to the computer-readable storage medium of the present invention, the duct does not require vehicle testing or bench testing. It can accurately obtain sound pressure data of preset monitoring points inside the vehicle based on the vehicle CFD model, and convert the obtained sound pressure data into aerodynamic noise audio data. Based on the objective parameters and subjective evaluation scores of the aerodynamic noise audio data, a target neural network model is determined. This allows for direct evaluation of aerodynamic noise based on the target neural network model, which is beneficial for early exposure of NVH risks of the duct. Furthermore, the subjective evaluation results output by the target neural network model can effectively guide the development of NVH performance of the vehicle air conditioning system duct, which helps to shorten the development cycle and reduce development costs.

[0128] It should be noted that the above explanation of the implementation method and beneficial effects of the duct aerodynamic noise evaluation method is also applicable to the computer-readable medium of the present invention, and will not be elaborated in detail here to avoid redundancy.

[0129] To implement the above embodiments, this invention also proposes an electronic device that can implement the duct aerodynamic noise evaluation method of any of the above embodiments. Figure 17 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Figure 17 As shown, the electronic device 300 proposed in this invention includes a memory 32, a processor 34, and a vehicle air conditioning system duct aerodynamic noise evaluation program 36 stored in the memory 32 and capable of running on the processor 34. When the processor 32 executes the duct aerodynamic noise evaluation program 36, it implements the vehicle air conditioning system duct aerodynamic noise evaluation method of any of the above embodiments.

[0130] According to the electronic device 300 of the present invention, the duct does not require vehicle testing or bench testing. It can accurately obtain sound pressure data of preset monitoring points inside the vehicle based on the vehicle CFD model, and convert the obtained sound pressure data into aerodynamic noise audio data. Based on the objective parameters and subjective evaluation scores of the aerodynamic noise audio data, a target neural network model is determined. This allows for direct evaluation of aerodynamic noise based on the target neural network model, which is beneficial for early exposure of NVH risks of the duct. Furthermore, the subjective evaluation results output by the target neural network model can effectively guide the development of NVH performance of the vehicle air conditioning system duct, which helps to shorten the development cycle and reduce development costs.

[0131] Specifically, electronic devices 300 include, but are not limited to, tablet computers, laptops, personal computers, servers, etc.

[0132] It should be noted that the above explanation of the implementation method and beneficial effects of the duct aerodynamic noise evaluation method also applies to the electronic device 300 of the present invention. To avoid redundancy, it will not be elaborated in detail here.

[0133] To achieve the above embodiments, this invention also proposes a duct aerodynamic noise evaluation device for a vehicle air conditioning system, which can implement the duct aerodynamic noise evaluation method of any of the above embodiments. Figure 18 This is a schematic diagram of a duct aerodynamic noise evaluation device according to an embodiment of the present invention. Figure 18 As shown, the aerodynamic noise evaluation device 200 for a vehicle air conditioning system proposed in this invention includes a modeling module 22, a simulation module 24, a determination module 26, and a training module 28. The modeling module 22 is used to establish a vehicle CFD model based on the duct data, air outlet grille data, and passenger compartment data of the vehicle air conditioning system. The simulation module 24 is used to simulate the flow field of the vehicle CFD model based on preset operating conditions, preset constant calculation parameters, and preset unsteady calculation parameters of the vehicle air conditioning system to obtain sound pressure data at preset monitoring points inside the vehicle. The determination module 26 is used to convert the sound pressure data at the preset monitoring points inside the vehicle into aerodynamic noise audio data and determine the objective parameters and subjective evaluation scores of the aerodynamic noise audio data. The training module 28 is used to train a neural network model based on the objective parameters and subjective evaluation scores to obtain a target neural network model, and to evaluate the aerodynamic noise based on the target neural network model.

[0134] According to the embodiment of the present invention, the aerodynamic noise evaluation device 200 for vehicle air conditioning system ducts eliminates the need for vehicle-mounted testing and bench testing of the ducts. It can accurately obtain sound pressure data of preset monitoring points inside the vehicle based on the vehicle CFD model, convert the obtained sound pressure data into aerodynamic noise audio data, and determine the target neural network model based on the objective parameters and subjective evaluation scores of the aerodynamic noise audio data. This allows for direct evaluation of aerodynamic noise based on the target neural network model, which is beneficial for early exposure of NVH risks in the ducts. Furthermore, it can effectively guide the NVH performance development of vehicle air conditioning system ducts based on the subjective evaluation results output by the target neural network model, thereby shortening the development cycle and reducing development costs.

[0135] In some embodiments of the present invention, the modeling module 22 includes a modeling unit, a processing unit, and a meshing unit. The modeling unit is used to establish a geometric model based on the duct data, air outlet grille data, and passenger compartment data of the vehicle's air conditioning system. The processing unit is used to process the geometric model to obtain a boundary geometric model corresponding to the CFD computational domain of aerodynamic noise. The meshing unit is used to mesh the geometric model according to the boundary geometric model to obtain a vehicle CFD model.

[0136] In some embodiments of the present invention, the simulation module 24 includes a steady-state calculation unit and an unsteady-state calculation unit. The steady-state calculation unit is used to perform steady-state calculations on the vehicle CFD model according to the preset operating conditions and preset steady-state calculation parameters of the vehicle air conditioning system to obtain initial sound pressure data. The unsteady-state calculation unit is used to perform unsteady-state calculations on the vehicle CFD model according to the initial sound pressure data and preset unsteady-state calculation parameters to obtain sound pressure data at preset monitoring points inside the vehicle.

[0137] In some embodiments of the present invention, the preset monitoring points inside the vehicle include one or more of the following: the right ear area of ​​the driver, the right ear area of ​​the middle left passenger, the right ear area of ​​the rear left passenger, and the air vent grille area.

[0138] In some embodiments of the present invention, the determining module 26 includes a first determining unit and a sampling unit. The first determining unit is used to determine the sampling frequency according to preset unsteady calculation parameters. The sampling unit is used to sample the sound pressure data of preset monitoring points inside the vehicle according to the sampling frequency to obtain aerodynamic noise audio data.

[0139] In some embodiments of the present invention, the determining module 26 further includes an acquisition unit and a calculation unit. The acquisition unit is used to acquire ratings of the aerodynamic noise audio data from multiple users based on the same scoring mechanism. The calculation unit is used to calculate the average of the ratings of the aerodynamic noise audio data from multiple users, and use the average as the subjective evaluation score of the aerodynamic noise audio data.

[0140] In some embodiments of the present invention, the determining module 26 further includes a transformation unit and a second determining unit. The transformation unit is used to perform a fast Fourier transform on the aerodynamic noise audio data to obtain an energy spectrum. The second determining unit is used to determine the energy of a first frequency range, the energy of a second frequency range, the energy of a third frequency range, and the A-weighted sound pressure level based on the energy spectrum, wherein the minimum value of the second frequency range is greater than or equal to the maximum value of the first frequency range, and the maximum value of the second frequency range is less than or equal to the minimum value of the third frequency range.

[0141] In some embodiments of the present invention, the training module 28 includes an allocation unit, a training unit, an optimization unit, and a prediction unit. The allocation unit is used to allocate objective parameters and subjective evaluation scores into training samples, test samples, and prediction samples according to a preset ratio. The training unit is used to train a neural network model based on the training samples and obtain a neural network model to be optimized when the error is less than or equal to a first threshold. The optimization unit is used to optimize the neural network model to be optimized based on the test samples and a genetic algorithm to obtain a target neural network model. The prediction unit is used to perform subjective evaluation predictions on the objective parameters in the prediction samples based on the target neural network model, and determine the effectiveness of the target neural network model based on the error between the first predicted score of the subjective evaluation and the subjective evaluation score in the prediction samples.

[0142] In some embodiments of the present invention, the preset ratio is 6:2:2.

[0143] In some embodiments of the present invention, the optimization unit includes an input subunit, a computation subunit, and an optimization subunit. The input subunit is used to input objective parameters from the test samples into the neural network model to be optimized to obtain a second predicted score. The computation subunit is used to calculate the mean squared error between the second predicted score and the subjective evaluation scores from the test samples, and uses the mean squared error as the fitness value of the genetic algorithm. The optimization subunit is used to optimize the neural network model to be optimized using the genetic algorithm based on the fitness value.

[0144] In some embodiments of the present invention, the number of generations of the genetic algorithm is 50, the crossover probability of the genetic algorithm is 0.7, and the mutation probability of the genetic algorithm is 0.1.

[0145] It should be noted that the above explanation of the implementation method and beneficial effects of the duct aerodynamic noise evaluation method also applies to the duct aerodynamic noise evaluation device 200 of the present invention. To avoid redundancy, it will not be elaborated in detail here.

[0146] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0147] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0148] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0149] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0150] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0151] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," and "circumferential" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this invention and simplifying the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0152] Furthermore, the terms "first," "second," etc., used in the embodiments of this invention are for descriptive purposes only and should not be construed as indicating or implying relative importance, or implicitly specifying the number of technical features indicated in this embodiment. Therefore, features defined with terms such as "first" and "second" in the embodiments of this invention can explicitly or implicitly indicate that the embodiment includes at least one of those features. In the description of this invention, the word "multiple" means at least two or more, such as two, three, four, etc., unless otherwise explicitly specified in the embodiments.

[0153] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for evaluating the aerodynamic noise of a vehicle air conditioning system duct, characterized in that, include: A vehicle CFD model is established based on the duct data, air outlet grille data, and passenger compartment data of the vehicle's air conditioning system. The flow field of the vehicle CFD model is simulated based on the preset operating conditions, preset constant calculation parameters and preset unsteady calculation parameters of the vehicle air conditioning system to obtain the sound pressure data of preset monitoring points inside the vehicle. The sound pressure data from the preset monitoring points inside the vehicle are converted into aerodynamic noise audio data, and the objective parameters and subjective evaluation scores of the aerodynamic noise audio data are determined. The neural network model is trained based on the objective parameters and the subjective evaluation scores to obtain a target neural network model, and the aerodynamic noise is evaluated based on the target neural network model. The objective parameters for determining the aerodynamic noise audio data include: The aerodynamic noise audio data is subjected to a fast Fourier transform to obtain the energy spectrum; The energy of the first frequency range, the energy of the second frequency range, the energy of the third frequency range, and the A-weighted sound pressure level are determined based on the energy spectrum, wherein the minimum value of the second frequency range is greater than or equal to the maximum value of the first frequency range, and the maximum value of the second frequency range is less than or equal to the minimum value of the third frequency range. The step of training the neural network model based on the objective parameters and the subjective evaluation score includes: According to a preset ratio, the objective parameters and the subjective evaluation scores are allocated into training samples, test samples, and prediction samples; The neural network model is trained based on the training samples, and the neural network model to be optimized is obtained when the error is less than or equal to the first threshold. The neural network model to be optimized is optimized based on the test samples and the genetic algorithm to obtain the target neural network model; The objective parameters in the prediction sample are subjectively evaluated and predicted based on the target neural network model, and the effectiveness of the target neural network model is determined based on the error between the first prediction score of the subjective evaluation and the subjective evaluation score in the prediction sample.

2. The method for evaluating aerodynamic noise in ducts according to claim 1, characterized in that, The process of establishing a vehicle CFD model based on the vehicle's air conditioning system duct data, air vent grille data, and passenger compartment data includes: A geometric model is established based on the duct data of the vehicle air conditioning system, the air outlet grille data, and the passenger compartment data; The geometric model is processed to obtain a boundary geometric model corresponding to the CFD computational domain of aerodynamic noise; The boundary geometry model is divided into meshes to obtain the vehicle CFD model; The simulation of the flow field of the vehicle CFD model based on the preset operating conditions, preset constant calculation parameters, and preset unsteady calculation parameters of the vehicle air conditioning system includes: The vehicle CFD model is subjected to steady-state calculation based on the preset operating conditions of the vehicle air conditioning system and the preset constant calculation parameters to obtain initial sound pressure data. The vehicle CFD model is subjected to unsteady calculations based on the initial sound pressure data and the preset unsteady calculation parameters to obtain the sound pressure data of the preset monitoring points inside the vehicle.

3. The method for evaluating aerodynamic noise in ducts according to claim 1, characterized in that, The step of converting the sound pressure data from the preset monitoring points inside the vehicle into aerodynamic noise audio data includes: The sampling frequency is determined based on the preset unsteady calculation parameters; The sound pressure data at the preset monitoring points inside the vehicle are sampled according to the sampling frequency to obtain the aerodynamic noise audio data.

4. The method for evaluating aerodynamic noise in ducts according to claim 1, characterized in that, The preset ratio is 6:2:

2.

5. The method for evaluating aerodynamic noise in ducts according to claim 1, characterized in that, The optimization of the neural network model to be optimized based on the test samples and the genetic algorithm includes: The objective parameters from the test samples are input into the neural network model to be optimized to obtain a second prediction score; Calculate the mean squared error between the second predicted score and the subjective evaluation score in the test sample, and use the mean squared error as the fitness value of the genetic algorithm; The genetic algorithm is used to optimize the neural network model to be optimized based on the fitness value.

6. The method for evaluating aerodynamic noise in ducts according to claim 5, characterized in that, The genetic algorithm has 50 generations, a crossover probability of 0.7, and a mutation probability of 0.

1.

7. An electronic device, characterized in that, The system includes a memory, a processor, and a duct aerodynamic noise evaluation program for a vehicle air conditioning system stored in the memory and capable of running on the processor. When the processor executes the duct aerodynamic noise evaluation program, it implements the duct aerodynamic noise evaluation method for a vehicle air conditioning system as described in any one of claims 1-6.

8. A device for evaluating the aerodynamic noise of a vehicle air conditioning system duct, characterized in that, include: The modeling module is used to create a vehicle CFD model based on the duct data, air outlet grille data, and passenger compartment data of the vehicle's air conditioning system. The simulation module is used to simulate the flow field of the vehicle CFD model based on the preset operating conditions, preset constant calculation parameters and preset unsteady calculation parameters of the vehicle air conditioning system, so as to obtain the sound pressure data of preset monitoring points inside the vehicle. The determination module is used to convert the sound pressure data of the preset monitoring points inside the vehicle into aerodynamic noise audio data, and to determine the objective parameters and subjective evaluation scores of the aerodynamic noise audio data; The training module is used to train a neural network model based on the objective parameters and the subjective evaluation scores to obtain a target neural network model, and to evaluate aerodynamic noise based on the target neural network model. The objective parameters for determining the aerodynamic noise audio data include: The aerodynamic noise audio data is subjected to a fast Fourier transform to obtain the energy spectrum; The energy of the first frequency range, the energy of the second frequency range, the energy of the third frequency range, and the A-weighted sound pressure level are determined based on the energy spectrum, wherein the minimum value of the second frequency range is greater than or equal to the maximum value of the first frequency range, and the maximum value of the second frequency range is less than or equal to the minimum value of the third frequency range. The step of training the neural network model based on the objective parameters and the subjective evaluation score includes: According to a preset ratio, the objective parameters and the subjective evaluation scores are allocated into training samples, test samples, and prediction samples; The neural network model is trained based on the training samples, and the neural network model to be optimized is obtained when the error is less than or equal to the first threshold. The neural network model to be optimized is optimized based on the test samples and the genetic algorithm to obtain the target neural network model; The objective parameters in the prediction sample are subjectively evaluated and predicted based on the target neural network model, and the effectiveness of the target neural network model is determined based on the error between the first prediction score of the subjective evaluation and the subjective evaluation score in the prediction sample.

Images