CN117313249A - Whole vehicle wind noise voice definition prediction method, equipment and storage medium - Google Patents
Whole vehicle wind noise voice definition prediction method, equipment and storage medium Download PDFInfo
- Publication number
- CN117313249A CN117313249A CN202311615551.4A CN202311615551A CN117313249A CN 117313249 A CN117313249 A CN 117313249A CN 202311615551 A CN202311615551 A CN 202311615551A CN 117313249 A CN117313249 A CN 117313249A
- Authority
- CN
- China
- Prior art keywords
- definition
- vehicle
- prediction
- data
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 71
- 238000009413 insulation Methods 0.000 claims abstract description 89
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 11
- 238000013135 deep learning Methods 0.000 claims description 30
- 230000005540 biological transmission Effects 0.000 claims description 27
- 238000012360 testing method Methods 0.000 claims description 24
- 238000005457 optimization Methods 0.000 claims description 11
- 238000004088 simulation Methods 0.000 claims description 11
- 238000013145 classification model Methods 0.000 claims description 10
- 238000012549 training Methods 0.000 claims description 7
- 238000005070 sampling Methods 0.000 claims description 6
- 230000001052 transient effect Effects 0.000 claims description 3
- 238000011161 development Methods 0.000 abstract description 17
- 238000013461 design Methods 0.000 abstract description 15
- 238000012545 processing Methods 0.000 abstract description 5
- 230000008569 process Effects 0.000 description 15
- 230000035945 sensitivity Effects 0.000 description 12
- 230000006870 function Effects 0.000 description 11
- 238000004364 calculation method Methods 0.000 description 10
- 230000004044 response Effects 0.000 description 10
- 238000010521 absorption reaction Methods 0.000 description 9
- 238000011156 evaluation Methods 0.000 description 7
- 238000004590 computer program Methods 0.000 description 6
- 230000004927 fusion Effects 0.000 description 6
- 238000010276 construction Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 239000012530 fluid Substances 0.000 description 3
- 238000012795 verification Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000008030 elimination Effects 0.000 description 2
- 238000003379 elimination reaction Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000009533 lab test Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000011056 performance test Methods 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 238000012827 research and development Methods 0.000 description 2
- 239000013585 weight reducing agent Substances 0.000 description 2
- 239000012814 acoustic material Substances 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000001364 causal effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000005034 decoration Methods 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 230000008676 import Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/15—Vehicle, aircraft or watercraft design
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M9/00—Aerodynamic testing; Arrangements in or on wind tunnels
- G01M9/02—Wind tunnels
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/10—Noise analysis or noise optimisation
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Geometry (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Computational Mathematics (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Aviation & Aerospace Engineering (AREA)
- Automation & Control Theory (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Fluid Mechanics (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention relates to the technical field of data processing, and discloses a method, equipment and a storage medium for predicting wind noise voice definition of a whole vehicle. According to the method, corresponding first prediction definition, second prediction definition, third prediction definition, fourth prediction definition, fifth prediction definition and definition contribution are determined by acquiring volume data of an in-vehicle sound cavity, external modeling characteristic data, in-vehicle reverberation time data, sound insulation parameter data, external noise source data and air tightness data of a target vehicle type and further respectively combining intelligent prediction models among all pre-constructed design parameters and air noise performance, all prediction definition and definition contribution are fused through an intelligent algorithm, air noise voice definition is obtained, rapid prediction of the overall vehicle air noise voice definition is achieved, development efficiency and accuracy are greatly improved, and development cost is reduced.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a method, equipment and a storage medium for predicting wind noise voice definition of a whole vehicle.
Background
With the development of the automobile industry and the improvement of automobile quality, consumers have higher and higher requirements on automobile comfort, challenges are also presented to the improvement of automobile wind noise level, and the wind noise problem is a high-rise problem which seriously affects the experience of consumers, so that the method has very important significance in the rapid and intelligent control of the whole automobile wind noise performance in the development of automobile types.
The wind noise performance of the whole vehicle is relatively wide in involved aspects, in the current wind noise design and development process in industry, wind noise performance development engineers, modeling teams, vehicle body structure teams, interior and exterior decoration design teams and the like are required to conduct design and verification development for a longer period, a large amount of manpower is required to repeatedly model, calculate and post-process, and the wind noise performance development period is long, the cost is high and the effect is not ideal.
In view of this, the present invention has been made.
Disclosure of Invention
In order to solve the technical problems, the invention provides a method, equipment and a storage medium for predicting the wind noise voice definition of a whole vehicle, which can realize the rapid prediction of the wind noise voice definition of the whole vehicle, further realize the rapid evaluation of the wind noise performance of the vehicle type and solve the problems of long development period, high cost and poor effect of the wind noise performance in the prior art.
The embodiment of the invention provides a method for predicting the wind noise voice definition of a whole vehicle, which comprises the following steps:
acquiring volume data, exterior modeling characteristic data, interior reverberation time data, sound insulation parameter data, external noise source data and air tightness data of an interior acoustic cavity of a target vehicle type;
determining a corresponding first prediction definition based on the in-vehicle acoustic cavity volume data and a pre-built volume linear model, determining a corresponding second prediction definition based on the exterior modeling feature data and a pre-built exterior modeling network model, determining a corresponding third prediction definition based on the in-vehicle reverberation time data and a pre-built reverberation linear model, determining a corresponding fourth prediction definition based on the sound insulation parameter data and a pre-built sound insulation network model, and determining a corresponding fifth prediction definition based on the external noise source data and a pre-built noise source network model;
and determining corresponding definition contribution according to the air tightness data and a pre-constructed air tightness grading model, and analyzing and fusing the first prediction definition, the second prediction definition, the third prediction definition, the fourth prediction definition, the fifth prediction definition and the definition contribution through an intelligent optimization algorithm to obtain the wind noise voice definition of the target vehicle type.
The embodiment of the invention provides electronic equipment, which comprises:
a processor and a memory;
the processor is used for executing the steps of the whole vehicle wind noise voice definition prediction method according to any embodiment by calling the program or the instructions stored in the memory.
The embodiment of the invention provides a computer readable storage medium, which stores a program or instructions for causing a computer to execute the steps of the whole vehicle wind noise voice definition prediction method according to any embodiment.
The embodiment of the invention has the following technical effects:
the method comprises the steps of acquiring volume data, external modeling characteristic data, in-vehicle reverberation time data, sound insulation parameter data, external noise source data and air tightness data of a target vehicle model, further respectively combining a pre-built volume linear model, an external modeling network model, a reverberation linear model, a sound insulation network model, a noise source network model and an air tightness grading model to determine corresponding first prediction definition, second prediction definition, third prediction definition, fourth prediction definition, fifth prediction definition and definition contribution, analyzing and fusing all prediction definition and definition contribution through an intelligent optimization algorithm to obtain wind noise voice definition of the target vehicle model, realizing rapid prediction of wind noise voice definition of the whole vehicle, further realizing rapid evaluation of wind noise performance of different vehicle models, providing support of wind noise performance intelligent design, target evaluation, weight reduction characteristic difference analysis and the like for research and development engineers, leading wind noise performance development and verification work to an import stage, enabling wind noise performance design to be rapidly developed, noise elimination chamber noise performance analysis or noise elimination chamber analysis to be carried out, and wind noise performance test efficiency to be greatly reduced, and wind performance test efficiency can be greatly estimated; in addition, the method combines the volume of the sound cavity, the external modeling characteristic, the sound absorption environment in the vehicle, the sound insulation performance, other external noise sources and the air tightness of the whole vehicle, analyzes the wind noise performance of different vehicle types, and can ensure the analysis accuracy of the wind noise performance.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are needed in the description of the embodiments or the prior art will be briefly described, and it is obvious that the drawings in the description below are some embodiments of the present invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart of a method for predicting the wind noise voice definition of a whole vehicle according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below. It will be apparent that the described embodiments are only some, but not all, embodiments of the invention. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the invention, are within the scope of the invention.
Before the detailed description of the method for predicting the wind noise voice definition of the whole vehicle, which is provided by the embodiment of the invention, the technical problem solved by the method is described. In the prior art, common means and tools for developing the wind noise performance of the whole vehicle are wind tunnel test, fluid simulation analysis (Computational Fluid Dynamics, CFD), anechoic chamber test, reverberation box test and statistical energy simulation analysis (Statistical Energy Analysis, SEA). However, the general laboratory test needs to be completed in the middle and later stages of model design, the model is largely determined, the acoustic material is determined, the whole vehicle performance target is formulated, the changeable optimization scheme is few, the target formulation rationality cannot be judged and verified, and the simulation analysis is supported by a large amount of manpower and computational resources, the cost is high, the result accuracy is low, so that the conventional laboratory test, fluid and statistical energy simulation analysis tool cannot be well involved in wind noise design in the early stage of vehicle model design, meanwhile, the contribution and internal correlation of the related parameters such as external model characteristics and acoustic packages to the wind noise of the whole vehicle in the current industry are not clear, and in the vehicle model development, how the two are well coordinated, the accurate basis is not yet.
Therefore, in order to solve the technical problems, the embodiment of the invention provides an accurate, rapid and intelligent prediction method for the wind noise voice definition of the whole vehicle, which is used for assisting a development engineer in fully considering the internal relations among the external modeling, the path sound insulation, the sound absorption environment in the vehicle and the air tightness of the whole vehicle in a project pre-research stage, and considering various influencing factors in the vehicle type design to develop a low wind noise vehicle type.
The method for predicting the wind noise voice definition of the whole vehicle is mainly suitable for predicting the wind noise voice definition of the vehicle type, so that the wind noise performance of the vehicle type can be evaluated according to the wind noise voice definition obtained through prediction. The method for predicting the wind noise voice definition of the whole vehicle provided by the embodiment of the invention can be executed by electronic equipment integrated in a computer.
Fig. 1 is a flowchart of a method for predicting wind noise and voice definition of a whole vehicle according to an embodiment of the present invention. Referring to fig. 1, the method for predicting the wind noise voice definition of the whole vehicle specifically includes:
s110, acquiring in-vehicle acoustic cavity volume data, external modeling characteristic data, in-vehicle reverberation time data, sound insulation parameter data, external noise source data and air tightness data of a target vehicle type.
In the embodiment of the invention, the wind noise performance of the whole vehicle of a target vehicle type can be decomposed into six parts, namely the volume of a sound cavity, the external modeling characteristic, the sound absorption environment in the vehicle, the sound insulation performance, other external noise sources and the air tightness of the whole vehicle, the influence of each part on the wind noise performance of the whole vehicle is respectively determined, and the influence of all parts on the wind noise performance of the whole vehicle is fused through an optimization algorithm, so that the final wind noise voice definition is obtained. Wind noise speech intelligibility is understood to be the percentage of speech units in which speech can be clearly identified, in AI%, e.g. 66 AI% means that 66% of the speech units can be clearly identified.
Specifically, for a target vehicle model, the relevant data of six parts including the volume of the acoustic cavity, the external modeling characteristic, the sound absorption environment in the vehicle, the sound insulation performance, other external noise sources and the air tightness can be obtained.
The vehicle interior sound cavity volume data of the target vehicle type can comprise a vehicle length, a vehicle width, a vehicle height and a wheel base. The influence of the whole vehicle size such as the vehicle length, the vehicle width, the vehicle height, the wheelbase and the like on wind noise can be practically converted into the influence of the volume of the sound cavity in the vehicle on the wind noise.
The external modeling characteristic data of the target vehicle type can be composed of values of relevant parameters corresponding to external modeling areas such as a front face, a machine cover, a wiper, an A column, a rearview mirror and the like. For example, considering that the area with the greatest influence on the front exhaust noise is the a pillar, the rear view mirror, the triangle cover plate, the external modeling characteristic parameters may include the a pillar inclination angle, the a pillar width, the a pillar and windshield level difference, the a pillar and bright strip level difference, the rear view mirror Y-direction dimension, the rear view mirror Z-direction dimension, the mirror arm windward thickness dimension, the rear view mirror shell and body distance, the rear view mirror shell and X-axis included angle, and the like. Note that X, Y, Z refers to three directions in a vehicle body coordinate system, wherein X is a direction from a vehicle head to a vehicle tail, Y is a direction from a vehicle door to a center line of the vehicle, and Z is a direction from a vehicle bottom to a vehicle roof.
In the embodiment of the invention, besides the influence of an external noise source on the wind noise of the whole vehicle, the sound-absorbing environment in the vehicle has a certain influence on the wind noise of the whole vehicle, and the change of the sound-absorbing environment in the vehicle can be evaluated according to the reverberation time data in the vehicle.
For example, the in-vehicle reverberation time data can be calculated by the formula: t60=kv/a, where T60 is in-vehicle reverberation time data, K is a constant related to spatial humidity, V is the volume of the closed chamber, and a is the total sound absorption amount. Specifically, the in-vehicle reverberation time data under different frequencies can be measured to obtain a reverberation time spectrum curve, and further the in-vehicle reverberation time data corresponding to 1000Hz can be extracted therefrom for subsequent speech intelligibility prediction.
The sound insulation parameter data of the target vehicle type may include sound insulation parameters under respective transmission paths. For example, the sound insulation performance of the whole vehicle can be decomposed into five parts, namely cabin-in-vehicle (cabin-in-vehicle main driving head), floor-in-vehicle (floor-in-vehicle main driving head), door-in-vehicle (door-in-vehicle main driving head), four-wheel-in-vehicle (four-wheel-in-vehicle main driving head) and exhaust-in-vehicle (tail-row-in-vehicle main driving head), so that sound insulation parameters under cabin-in-vehicle transmission paths, floor-in-vehicle transmission paths, door-in-vehicle transmission paths, four-wheel-in-vehicle transmission paths and exhaust-in-vehicle transmission paths can be obtained, and sound insulation parameter data of a target vehicle type can be obtained.
The sound insulation parameter in each transmission path may be an average sound insulation amount, that is, an average value of PBNR (Power-Based Noise Reduction, energy-based sound insulation value) measured in all frequency bands. Such as:
；
wherein n is the number of frequency bands for measuring sound insulation,、/>、……、/>is 1/3 octave frequency band sound insulation, sound insulation parameters under each transmission path, < ->Is the average sound insulation, i.e. the sound insulation parameter under the transmission path.
The external noise source data of the target vehicle type may include, among others, powertrain noise (or cabin noise), four-wheel cavity noise, and exhaust noise. In the embodiment of the invention, the reason for decomposing the external noise source into the power assembly noise, the four-wheel cavity noise and the exhaust noise is as follows: assuming that external noise sources under the working condition of 120km/h are divided into cabin noise (s 1), four-wheel cavity noise (s 2) and exhaust noise (s 3), the wind noise source is sw, and the in-vehicle sound pressure level generated by each independent noise source is SPL by adopting single noise source analysis 1 、SPL 2 、SPL 3 、SPL w . Converting sound pressure level into energy E 1 、E 2 、E 3 、E w Then the total energy in the vehicle is E total =E 1 +E 2 +E 3 +E w The total energy in the vehicle is converted, and the total sound pressure level in the vehicle can be obtained:
；
the statistical energy analysis test can determine that the total sound pressure level in the vehicle has a certain influence on the wind noise performance, and further determine that the noise of the power assembly, the noise of the four-wheel cavity and the exhaust noise influence the wind noise performance of the whole vehicle.
The air tightness data of the target vehicle model can be understood as the air tightness value of the target vehicle model.
S120, determining corresponding first prediction definition based on the volume data of the sound cavity in the vehicle and a pre-built volume linear model, determining corresponding second prediction definition based on the external modeling characteristic data and a pre-built external modeling network model, determining corresponding third prediction definition based on the reverberation time data in the vehicle and a pre-built reverberation linear model, determining corresponding fourth prediction definition based on the sound insulation parameter data and a pre-built sound insulation network model, and determining corresponding fifth prediction definition based on the external noise source data and a pre-built noise source network model.
Specifically, after the volume data, the exterior modeling characteristic data, the interior reverberation time data, the sound insulation parameter data, the external noise source data, and the air tightness data of the interior acoustic cavity of the target vehicle model are obtained, each data may be input to a volume linear model, an exterior modeling network model, a reverberation linear model, a sound insulation network model, a noise source network model, and an air tightness classification model, respectively.
The volume linear model and the reverberation linear model can be linear models obtained through experimental data fitting; the external modeling network model, the sound insulation network model and the noise source network model can be neural network models obtained through analysis training of an intelligent optimization algorithm based on a database containing design parameters and wind noise performance; the gas tightness classification model may be a mathematical classification model defined by experimental data.
It should be noted that, considering that the rule of influence of the external modeling feature difference corresponding to different vehicle types on wind noise is different, for example, the inclination angle of the car a pillar is generally smaller than that of the off-road vehicle, so that the influence of the car a pillar and the off-road vehicle on wind noise performance is different, in order to further improve the evaluation accuracy of the wind noise performance of the whole vehicle, each corresponding model can be respectively constructed for each vehicle type.
Optionally, the method provided by the embodiment of the invention further includes: and respectively constructing a corresponding volume linear model, an external modeling network model, a reverberation linear model, a sound insulation network model, a noise source network model and an air tightness grading model aiming at different vehicle types.
Specifically, for each vehicle category, such as an off-road vehicle, a car, a truck, a semitrailer, or the like, a corresponding volumetric linear model, an external modeling network model, a reverberation linear model, a sound insulation network model, a noise source network model, and an air tightness classification model can be respectively constructed according to the design parameters and the large database of wind noise performance under the vehicle category.
Further, for a target vehicle model, a vehicle category corresponding to the target vehicle model can be determined first, so that wind noise voice definition of the target vehicle model can be predicted according to a volume linear model, an exterior modeling network model, a reverberation linear model, a sound insulation network model, a noise source network model and an air tightness classification model corresponding to the vehicle category. By classifying the vehicle types to construct each model under different vehicle types, the evaluation accuracy of the wind noise performance of the whole vehicle can be further improved.
In the embodiment of the invention, the volume linear model can be a prediction model of the volume of the sound cavity in the vehicle corresponding to the voice definition, the external modeling network model can be a network model of the external modeling characteristic corresponding to the voice definition, the reverberation linear model can be a prediction model of the reverberation time in the vehicle corresponding to the voice definition, the sound insulation network model can be a network model of the sound insulation parameter corresponding to the voice definition, and the noise source network model can be a network model of the noise source corresponding to the voice definition.
In a specific embodiment, the construction of the volumetric linear model comprises the following steps:
step 11, acquiring wind tunnel test data, wherein the wind tunnel test data comprise volume values of acoustic cavities in all vehicles and corresponding definition values;
and step 12, fitting a linear model constructed in advance based on wind tunnel test data, and determining a volume influence factor and a volume constant based on a fitting result to obtain a volume linear model.
Specifically, the relation between different vehicle internal volume values and definition values can be obtained through wind tunnel test, and wind tunnel test data can be obtained. It can be assumed that the wind noise speech intelligibility is related to the volume of the acoustic cavity in the vehicle, a linear model AI% = k is defined 1 *X v + b 1 Input quantity X of (a) v = car volume Wherein car volume The volume value of the acoustic cavity in the vehicle is represented, and the output quantity is defined as the voice definition which is independently influenced by the volume of the acoustic cavity in the vehicle.
Further, according to the volume value of the acoustic cavity in each vehicle and the corresponding definition value, linear fitting can be performed on the linear model, and the volume influence factor k in the linear model can be obtained 1 And a constant b of volume 1 From this, a volumetric linear model can be obtained.
Furthermore, for the target vehicle type, the volume data of the sound cavity in the vehicle of the target vehicle type can be input into the volume linear model, and the corresponding first prediction definition can be calculated through the volume influence factor and the volume constant. The first prediction definition is voice definition which is independently influenced by the volume of an acoustic cavity in the vehicle.
In a specific embodiment, the building of the external modeling network model includes the following steps:
step 21, sampling the value of each external modeling characteristic parameter to obtain a plurality of characteristic data combinations;
step 22, performing transient simulation on each characteristic data combination to obtain a definition value corresponding to each characteristic data combination;
step 23, training a pre-constructed first weight network based on definition values corresponding to all feature data combinations to determine parameters of a deep learning layer in the first weight network and weight coefficients corresponding to all external modeling feature parameters in an output layer respectively, so as to obtain an external modeling network model;
The deep learning layer in the first weight network is used for predicting the voice definition corresponding to each external modeling characteristic parameter respectively, and the output layer in the first weight network is used for fusing the voice definition corresponding to all the external modeling characteristic parameters respectively according to each weight coefficient.
Specifically, various values of the external modeling characteristic parameters can be obtained by using a geometric deformation method, and then a Latin hypercube sampling method is used for obtaining characteristic data combinations formed by the values of the characteristic parameters.
Further, the speech definition affected by each feature data combination, that is, the corresponding definition value, may be obtained by a transient simulation method, so as to establish a modeling data set composed of the feature data combination and the corresponding definition value, where, for example, the feature data combination and the corresponding definition value are expressed as:
；
in the method, in the process of the invention,the value of the characteristic parameter 1 to the characteristic parameter n in the mth characteristic data combination is +.>And combining the corresponding definition value for the mth characteristic data.
Further, a first weight network of feature parameters and speech intelligibility can be created by using the kriging method, where the feature parameters are defined as(j has a value of 1-n, n represents the number of characteristic parameters), and the first weight network may be expressed as:
；
In the method, in the process of the invention,for output, i.e. speech intelligibility affected by all external modeling parameters, < >>Is characteristic parameter->Corresponding weights, ++>Is characteristic parameter->Corresponding speech intelligibility.
Specifically, the first weight network may include a deep learning layer and an output layer, where the deep learning layer is configured to output speech definitions corresponding to each external modeling feature parameter, and the output layer is configured to combine weight coefficients corresponding to each external modeling feature parameter, perform weighted calculation on speech definitions corresponding to all external modeling feature parameters, and obtain a final definition prediction value.
The method includes the steps that each characteristic data combination and a corresponding definition value are input into a first weight network, so that a deep learning layer in the first weight network outputs voice definition corresponding to each external modeling characteristic parameter respectively, the output layer fuses all voice definition output by the deep learning layer to obtain a definition predicted value, a loss function is calculated according to the definition predicted value and the definition value, parameters of the deep learning layer and weight coefficients corresponding to each external modeling characteristic parameter in the output layer are reversely adjusted according to a calculation result of the loss function, and the process is iterated until the calculation result of the loss function converges to obtain an external modeling network model.
Further, according to the target vehicle model, the external modeling characteristic data of the target vehicle model can be input into the external modeling network model, and the corresponding second prediction definition is obtained through the deep learning layer and the output layer. The second prediction definition is speech definition which is independently influenced by the external modeling characteristic parameters.
In a specific embodiment, the construction of the reverberation linear model includes the following steps:
step 31, obtaining the in-vehicle reverberation time value and the corresponding definition value corresponding to each preset vehicle type through a semi-anechoic room test;
and step 32, fitting a pre-constructed linear model based on the in-vehicle reverberation time value and the corresponding definition value, and determining a reverberation influence factor and a reverberation constant based on a fitting result to obtain a reverberation linear model.
Specifically, the in-vehicle reverberation time values (the reverberation time corresponding to 1000Hz can be selected) and the corresponding definition values of different preset vehicle types can be obtained through a semi-anechoic chamber test. Wherein the in-vehicle reverberation time value car ReverT The value of (2) may be 0.07. Ltoreq.car ReverT ≤0.13。
Assuming that wind noise speech intelligibility is related to in-vehicle reverberation time, a linear model AI% = k is defined 2 *X T + b 2 Input quantity X of (a) T = car ReverT And defines the output as speech intelligibility solely affected by the in-vehicle reverberation time.
Further, according to the reverberation time value in each vehicle and the corresponding definition value, linear fitting can be performed on the linear model, and the reverberation influence factor k in the linear model can be obtained 2 And reverberation constant b 2 Thus, a reverberant linear model can be obtained.
Furthermore, for the target vehicle model, the in-vehicle reverberation time data of the target vehicle model can be input into the reverberation linear model, and the corresponding third prediction definition can be calculated through the reverberation influence factor and the reverberation constant. The third prediction definition is speech definition which is independently influenced by the in-vehicle reverberation time.
In a specific embodiment, the construction of the acoustic network model includes the following steps:
step 41, sampling the values of the sound insulation parameters under each transmission path to obtain a plurality of sound insulation data combinations, and carrying out statistical energy simulation on each sound insulation data combination to obtain a corresponding definition value;
step 42, training a pre-constructed second weight network based on definition values corresponding to all sound-insulation data combinations to determine weight coefficients corresponding to parameters of a deep learning layer in the second weight network and sound-insulation parameters in each transmission path in an output layer respectively, so as to obtain a sound-insulation network model;
The deep learning layer in the second weight network is used for predicting the voice definition corresponding to the sound insulation parameters under each transmission path, and the output layer in the second weight network is used for fusing the voice definition corresponding to the sound insulation parameters under all the transmission paths according to each weight coefficient.
Specifically, multiple values of sound insulation parameters of each transmission path under different vehicle types and different configurations can be obtained through a semi-anechoic room test, a Latin hypercube sampling method is utilized to obtain sound insulation data combinations formed by the values of the sound insulation parameters, and a definition value corresponding to each sound insulation data combination is obtained through SEA. Wherein the sound insulation parameter may be an average value of PBNR at all frequency bands, i.e. an average sound insulation amount. Illustratively, the sound deadening data combination and corresponding sharpness values are expressed as:
；
in the method, in the process of the invention,the values of the sound insulation parameters 1 to n in the mth sound insulation data combination are +.>And combining the corresponding definition value for the mth sound insulation data.
Further, a second weight network of sound insulation parameters and voice definition can be created by using the kriging method, and the sound insulation parameters are defined as(j has a value of 1 to n, n represents the number of sound insulation parameters), and the second weight network may be represented as:
；
In the method, in the process of the invention,for output, i.e. speech intelligibility affected by all sound-insulating parameters,/>For sound-insulating parameters->Corresponding weights, ++>For sound-insulating parameters->Corresponding speech intelligibility.
Specifically, the second weight network may include a deep learning layer and an output layer, where the deep learning layer is configured to output speech definitions corresponding to the sound insulation parameters under each transmission path, and the output layer is configured to combine weight coefficients corresponding to the sound insulation parameters under each transmission path, perform weighted calculation on speech definitions corresponding to all the sound insulation parameters, and obtain a final definition prediction value.
The method includes the steps that each sound insulation data combination and a corresponding definition value are input into a second weight network, so that a deep learning layer in the second weight network outputs voice definitions corresponding to sound insulation parameters under each transmission path, all voice definitions output by the deep learning layer are fused by an output layer, a definition predicted value is obtained, a loss function is calculated according to the definition predicted value and the definition value, the weight coefficients corresponding to parameters of the deep learning layer and the sound insulation parameters under each transmission path in the output layer are reversely adjusted according to a calculation result of the loss function, and the process is iterated until the calculation result of the loss function converges, so that a sound insulation network model is obtained.
Further, for the target vehicle model, the sound insulation parameter data of the target vehicle model can be input into the sound insulation network model, and the corresponding fourth prediction definition is obtained through the deep learning layer and the output layer. The fourth prediction definition is speech definition which is influenced by the sound insulation parameters under each transmission path.
In a specific embodiment, the construction of the noise source network model includes the following steps:
step 51, obtaining definition values corresponding to the external noise source values;
step 52, training a pre-constructed third weight network based on all external noise source values and corresponding definition values to determine parameters of a deep learning layer in the third weight network and weight coefficients corresponding to external noise sources in an output layer respectively, so as to obtain a noise source network model;
the deep learning layer in the third weight network is used for predicting the voice definitions respectively corresponding to the external noise sources, and the output layer in the third weight network is used for fusing the voice definitions respectively corresponding to all the external noise sources according to the weight coefficients.
Specifically, SEA can be utilized to analyze the influence of different noise sources on the sound pressure level in the vehicle, and various external noise source values and corresponding definition values can be collected. For example, the external noise source value and the corresponding sharpness value are expressed as:
；
In the method, in the process of the invention,is the value of noise source 1-noise source n in the mth external noise source value, +.>Is the definition value corresponding to the mth external noise source value.
Further, a third weight network of noise sources and speech intelligibility can be created by the Kriging method, and the noise sources are defined as(j is 1-n, n represents the number of characteristic parameters), and the third weight network may be represented as:
；
in the method, in the process of the invention,for output, i.e. speech intelligibility affected by all noise sources,/->Is a noise source->Corresponding weights, ++>Is a noise source->Corresponding speech intelligibility.
Specifically, the third weight network may include a deep learning layer and an output layer, where the deep learning layer is configured to output speech definitions corresponding to each noise source, and the output layer is configured to combine weight coefficients corresponding to each noise source, perform weighted calculation on speech definitions corresponding to all noise sources, and obtain a final definition prediction value.
By way of example, all external noise source values and corresponding definition values may be input into the third weight network, so that the deep learning layer in the third weight network outputs the speech definitions corresponding to the noise sources respectively, so that the output layer fuses all the speech definitions output by the deep learning layer to obtain a definition prediction value, a loss function is calculated according to the definition prediction value and the definition value, the parameters of the deep learning layer and the weight coefficients corresponding to the noise sources in the output layer are reversely adjusted according to the calculation result of the loss function, and the process is iterated until the calculation result of the loss function converges to obtain the noise source network model.
Further, for the target vehicle model, external noise source data of the target vehicle model can be input into the noise source network model, and the corresponding fifth prediction definition is obtained through the deep learning layer and the output layer. Wherein the fifth predicted intelligibility is speech intelligibility solely affected by the noise source.
It should be noted that, for the external modeling network model, the sound insulation network model, and the noise source network model, the weight coefficient in the output layer reflects the rule of influence (or sensitivity, correlation) between the corresponding parameter and the definition value, and can be obtained through learning by an intelligent optimization algorithm.
For the external modeling network model, a response surface between each external modeling characteristic parameter and the definition value can be constructed based on the definition value corresponding to each characteristic data combination, sensitivity between each external modeling characteristic parameter and the definition value is obtained based on the response surface, and then the weight coefficient corresponding to each external modeling characteristic parameter is determined according to the sensitivity, so that the influence rule of each external modeling characteristic parameter on the definition value is obtained.
Aiming at the sound insulation network model, a response surface between the sound insulation parameters and the definition values under each transmission path can be constructed based on the definition values corresponding to each sound insulation data combination, sensitivity between the sound insulation parameters and the definition values under each transmission path is obtained based on the response surface, and then the weight coefficient corresponding to the sound insulation parameters under each transmission path is determined according to the sensitivity, so that the influence rule of the sound insulation parameters under each transmission path on the definition values is obtained.
For the noise source network model, a response surface between each external noise source and the definition value can be constructed based on the definition value corresponding to each external noise source value obtained through simulation, sensitivity between each external noise source and the definition value is obtained based on the response surface, and then the weight coefficient corresponding to each external noise source is determined according to the sensitivity, so that the influence rule of each external noise source on the definition value is obtained.
Besides the intelligent optimization algorithm based on the response surface, the correlation between each parameter and the definition value can be analyzed, and the weight coefficient corresponding to each parameter can be determined based on the correlation.
S130, determining corresponding definition contribution quantity according to the air tightness data and a pre-constructed air tightness grading model, and analyzing and fusing the first prediction definition, the second prediction definition, the third prediction definition, the fourth prediction definition, the fifth prediction definition and the definition contribution quantity through an intelligent optimization algorithm to obtain the wind noise voice definition of the target vehicle type.
In a specific embodiment, the construction of the airtight classification model comprises the following steps:
step 61, determining the air tightness value and the corresponding definition value of each preset vehicle type through an air tightness test and a wind tunnel test;
Step 62, determining a reference vehicle model with zero definition contribution from all preset vehicle models according to the definition value corresponding to each preset vehicle model, and obtaining a reference air tightness range of the reference vehicle model;
and 63, determining each grading air tightness range according to the air tightness value of each preset vehicle type, and determining the definition contribution quantity corresponding to each grading air tightness range based on the gap between the definition values of each preset vehicle type and the reference vehicle type to obtain an air tightness grading model.
Specifically, the air tightness value and the corresponding definition value of each preset vehicle model can be acquired through an air tightness test and a wind tunnel test, and then the reference vehicle model is determined according to the definition values corresponding to all the preset vehicle models, and the reference air tightness range is determined according to the air tightness value of the reference vehicle model. Wherein, the definition contribution of the reference vehicle model is zero.
Further, each classified airtight range can be determined according to the airtight value of each preset vehicle model, and the definition contribution quantity corresponding to each classified airtight range is obtained by combining the difference between the definition value of each preset vehicle model and the definition value of the reference vehicle model, so that an airtight classification model is constructed. The constructed air tightness grading model comprises a reference air tightness range, each grading air tightness range and corresponding definition contribution quantity of each range. Illustratively, table 1 shows a gas tightness classification model.
Table 1A gas tightness classification model
As shown in Table 1, the reference vehicle type has three levels, the reference air tightness range is 90-100, and the corresponding definition contribution is zero. Aiming at a target vehicle type, according to the air tightness data of the target vehicle type, inquiring the corresponding air tightness range in the air tightness grading model, and further inquiring the corresponding definition contribution AI% at For example, AI% at = +x1, +x2, 0, or-X4.
Specifically, for a target vehicle model, after the first prediction definition, the second prediction definition, the third prediction definition, the fourth prediction definition, the fifth prediction definition and the definition contribution quantity output by each model are obtained, the first prediction definition, the second prediction definition, the third prediction definition, the fourth prediction definition and the fourth prediction definition can be fused through an intelligent optimization algorithm, so that wind noise voice definition influenced by the cavity volume, the external modeling characteristic, the in-vehicle sound absorption environment, the sound insulation performance, the external noise source and the air tightness of the whole vehicle is obtained.
By way of example, the influence rules and causal response relations of the first prediction definition, the second prediction definition, the third prediction definition, the fourth prediction definition, the fifth prediction definition, the definition contribution amount, the sensitivity between the wind noise voice definition, the contribution amount and the like can be explored in advance through an optimization algorithm, so that each weight in the definition fusion model can be obtained, for example, a response surface between each prediction definition, the definition contribution amount and the wind noise voice definition is constructed, the sensitivity between each prediction definition, the definition contribution amount and the wind noise voice definition is obtained through the response surface, and then each weight in the definition fusion model is obtained based on the sensitivity. Further, the wind noise voice definition can be output by fusing each predicted definition and definition contribution through each weight in the definition fusion model.
For example, the sharpness fusion model may be expressed by the following formula:
；
in the method, in the process of the invention,for the first prediction definition,/>For the second prediction definition is calculated,for the third prediction definition +.>For the fourth prediction definition +.>For the fifth prediction definition +.>For clarity contribution, ∈>The wind noise voice definition output by the definition fusion model;
wherein i is a variable, the value of i is 1 to 6, and the value of i is 1For the first prediction definition, i is taken as 2 +.>For the second prediction definition, i is taken as 3 +.>For the third prediction definition, i is taken as 4 +.>For the fourth prediction definition, i is taken as 5 +.>For the fifth prediction definition, i takes a value of 6Contributing to definition; />For the weights corresponding to the first prediction definition, the second prediction definition, the third prediction definition, the fourth prediction definition, the fifth prediction definition or the definition contribution amount, each weight can reflect the sensitivity between the wind noise voice definition, and the larger the weight is, the larger the sensitivity is, namely the larger the influence on the wind noise voice definition is.
Specifically, for a target vehicle model, the first prediction definition, the second prediction definition, the third prediction definition, the fourth prediction definition, the fifth prediction definition and the definition contribution amount of the target vehicle model can be input into a definition fusion model, and weighting calculation is performed by combining the corresponding weights, so that the final wind noise voice definition is obtained.
In the embodiment of the invention, the wind noise performance of the whole vehicle is considered to be strongly related to the appearance design of the whole vehicle, and is closely related to the sound insulation performance, the in-vehicle sound absorption environment, the in-vehicle sound cavity volume, the air tightness of the whole vehicle and the like, so that the wind noise performance of different vehicle types is analyzed by combining the sound cavity volume, the external modeling characteristics, the in-vehicle sound absorption environment, the sound insulation performance, other external noise sources and the air tightness of the whole vehicle, and the analysis accuracy of the wind noise performance is ensured. The wind noise voice definition of each vehicle type can be obtained by respectively predicting different vehicle types, so that the wind noise performance of each vehicle type can be analyzed conveniently, wherein the higher the wind noise voice definition is, the stronger the wind noise performance is represented.
The invention has the following technical effects: the corresponding first prediction definition, second prediction definition, third prediction definition, fourth prediction definition, fifth prediction definition and definition contribution are determined by acquiring volume data of an acoustic cavity in a vehicle of a target vehicle type, external modeling characteristic data, reverberation time data in the vehicle, sound insulation parameter data, external noise source data and air tightness data and further respectively combining a pre-built volume linear model, an external modeling network model, a reverberation linear model, a sound insulation network model, a noise source network model and an air tightness grading model, the wind noise voice definition of the target vehicle type is obtained, the rapid prediction of the wind noise voice definition of the whole vehicle is realized, the rapid evaluation of the wind noise performance of different vehicle types can be realized, the support of wind noise performance intelligent design, target standard evaluation, weight reduction and cost reduction performance difference analysis and the like can be provided for research and development engineers in the vehicle type development process, the wind noise performance development and verification work is preposed to the concept importing stage, the wind noise performance design is rapidly developed, the wind noise performance can be rapidly evaluated without simulation analysis, wind tunnel test or anechoic room test, the development efficiency and the development precision are greatly improved, and the development cost is reduced; in addition, the method combines the volume of the sound cavity, the external modeling characteristic, the sound absorption environment in the vehicle, the sound insulation performance, other external noise sources and the air tightness of the whole vehicle, analyzes the wind noise performance of different vehicle types, and can ensure the analysis accuracy of the wind noise performance.
Fig. 2 is a schematic structural diagram of an electronic device according to an embodiment of the present invention. As shown in fig. 2, electronic device 400 includes one or more processors 401 and memory 402.
The processor 401 may be a Central Processing Unit (CPU) or other form of processing unit having data processing capabilities and/or instruction execution capabilities and may control other components in the electronic device 400 to perform desired functions.
Memory 402 may include one or more computer program products that may include various forms of computer-readable storage media, such as volatile memory and/or non-volatile memory. The volatile memory may include, for example, random Access Memory (RAM) and/or cache memory (cache), and the like. The non-volatile memory may include, for example, read Only Memory (ROM), hard disk, flash memory, and the like. One or more computer program instructions may be stored on the computer readable storage medium, and the processor 401 may run the program instructions to implement the method for predicting the wind noise and speech clarity of the whole vehicle according to any embodiment of the present invention described above and/or other desired functions. Various content such as initial arguments, thresholds, etc. may also be stored in the computer readable storage medium.
In one example, the electronic device 400 may further include: an input device 403 and an output device 404, which are interconnected by a bus system and/or other forms of connection mechanisms (not shown). The input device 403 may include, for example, a keyboard, a mouse, and the like. The output device 404 may output various information to the outside, including early warning prompt information, braking force, etc. The output device 404 may include, for example, a display, speakers, a printer, and a communication network and remote output devices connected thereto, etc.
Of course, only some of the components of the electronic device 400 that are relevant to the present invention are shown in fig. 2 for simplicity, components such as buses, input/output interfaces, etc. are omitted. In addition, electronic device 400 may include any other suitable components depending on the particular application.
In addition to the methods and apparatus described above, embodiments of the present invention may also be a computer program product comprising computer program instructions which, when executed by a processor, cause the processor to perform the steps of the overall wind noise speech intelligibility prediction method provided by any of the embodiments of the present invention.
The computer program product may write program code for performing operations of embodiments of the present invention in any combination of one or more programming languages, including an object oriented programming language such as Java, C++ or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device, partly on a remote computing device, or entirely on the remote computing device or server.
In addition, an embodiment of the present invention may also be a computer readable storage medium, on which computer program instructions are stored, which when executed by a processor, cause the processor to execute the steps of the method for predicting wind noise and speech clarity of a whole vehicle provided by any embodiment of the present invention.
The computer readable storage medium may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may include, for example, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium would include the following: an electrical connection having one or more wires, a portable disk, a hard disk, random Access Memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of the present application. As used in this specification, the terms "a," "an," "the," and/or "the" are not intended to be limiting, but rather are to be construed as covering the singular and the plural, unless the context clearly dictates otherwise. The terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method or apparatus comprising such elements.
It should also be noted that the positional or positional relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. are based on the positional or positional relationship shown in the drawings, are merely for convenience of describing the present invention and simplifying the description, and do not indicate or imply that the apparatus or element in question must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as limiting the present invention. Unless specifically stated or limited otherwise, the terms "mounted," "connected," and the like are to be construed broadly and may be, for example, fixedly connected, detachably connected, or integrally connected; can be mechanically or electrically connected; can be directly connected or indirectly connected through an intermediate medium, and can be communication between two elements. The specific meaning of the above terms in the present invention will be understood in specific cases by those of ordinary skill in the art.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention.
Claims (10)
1. The method for predicting the wind noise voice definition of the whole vehicle is characterized by comprising the following steps of:
acquiring volume data, exterior modeling characteristic data, interior reverberation time data, sound insulation parameter data, external noise source data and air tightness data of an interior acoustic cavity of a target vehicle type;
determining a corresponding first prediction definition based on the in-vehicle acoustic cavity volume data and a pre-built volume linear model, determining a corresponding second prediction definition based on the exterior modeling feature data and a pre-built exterior modeling network model, determining a corresponding third prediction definition based on the in-vehicle reverberation time data and a pre-built reverberation linear model, determining a corresponding fourth prediction definition based on the sound insulation parameter data and a pre-built sound insulation network model, and determining a corresponding fifth prediction definition based on the external noise source data and a pre-built noise source network model;
And determining corresponding definition contribution according to the air tightness data and a pre-constructed air tightness grading model, and analyzing and fusing the first prediction definition, the second prediction definition, the third prediction definition, the fourth prediction definition, the fifth prediction definition and the definition contribution through an intelligent optimization algorithm to obtain the wind noise voice definition of the target vehicle type.
2. The method of claim 1, wherein the constructing of the volumetric linear model comprises:
acquiring wind tunnel test data, wherein the wind tunnel test data comprise volume values of sound cavities in all vehicles and corresponding definition values;
and fitting a linear model constructed in advance based on the wind tunnel test data, and determining a volume influence factor and a volume constant based on a fitting result to obtain the volume linear model.
3. The method of claim 1, wherein the building of the extranet model comprises:
sampling the value of each external modeling characteristic parameter to obtain a plurality of characteristic data combinations;
performing transient simulation on each characteristic data combination to obtain a definition value corresponding to each characteristic data combination;
Training a pre-constructed first weight network based on definition values corresponding to all feature data combinations to determine parameters of a deep learning layer in the first weight network and weight coefficients corresponding to all external modeling feature parameters in an output layer respectively, so as to obtain an external modeling network model;
the deep learning layer in the first weight network is used for predicting the voice definition corresponding to each external modeling characteristic parameter respectively, and the output layer in the first weight network is used for fusing the voice definition corresponding to all the external modeling characteristic parameters respectively according to each weight coefficient.
4. The method of claim 1, wherein the constructing of the reverberation linear model comprises:
acquiring an in-vehicle reverberation time value corresponding to each preset vehicle type and a corresponding definition value through a semi-anechoic chamber test;
and fitting a linear model constructed in advance based on the in-vehicle reverberation time value and the corresponding definition value, and determining a reverberation influence factor and a reverberation constant based on a fitting result to obtain the reverberation linear model.
5. The method of claim 1, wherein the constructing of the acoustic network model comprises:
Sampling the values of the sound insulation parameters under each transmission path to obtain a plurality of sound insulation data combinations, and carrying out statistical energy simulation on each sound insulation data combination to obtain a corresponding definition value;
training a pre-constructed second weight network based on definition values corresponding to all sound-insulation data combinations to determine weight coefficients corresponding to parameters of a deep learning layer in the second weight network and sound-insulation parameters under each transmission path in an output layer respectively, so as to obtain the sound-insulation network model;
the deep learning layer in the second weight network is used for predicting the voice definition corresponding to the sound insulation parameters under each transmission path, and the output layer in the second weight network is used for fusing the voice definition corresponding to the sound insulation parameters under all the transmission paths according to each weight coefficient.
6. The method of claim 1, wherein the constructing of the noise source network model comprises:
acquiring definition values corresponding to the external noise source values;
training a third weight network constructed in advance based on all external noise source values and corresponding definition values to determine parameters of a deep learning layer in the third weight network and weight coefficients corresponding to external noise sources in an output layer respectively, so as to obtain a noise source network model;
The deep learning layer in the third weight network is used for predicting the voice definitions respectively corresponding to all external noise sources, and the output layer in the third weight network is used for fusing the voice definitions respectively corresponding to all external noise sources according to all weight coefficients.
7. The method of claim 1, wherein the constructing of the airtight classification model comprises:
determining the air tightness value and the corresponding definition value of each preset vehicle type through an air tightness test and a wind tunnel test;
according to the definition value corresponding to each preset vehicle model, determining a reference vehicle model with zero definition contribution from all preset vehicle models, and obtaining a reference air tightness range of the reference vehicle model;
and determining each grading air tightness range according to the air tightness value of each preset vehicle type, and determining the definition contribution quantity corresponding to each grading air tightness range based on the difference between the definition values of each preset vehicle type and the reference vehicle type to obtain the air tightness grading model.
8. The method according to any one of claims 1-7, further comprising:
and respectively constructing a corresponding volume linear model, an external modeling network model, a reverberation linear model, a sound insulation network model, a noise source network model and an air tightness grading model aiming at different vehicle types.
9. An electronic device, the electronic device comprising:
a processor and a memory;
the processor is configured to execute the steps of the whole vehicle wind noise voice definition prediction method according to any one of claims 1 to 8 by calling a program or instructions stored in the memory.
10. A computer-readable storage medium storing a program or instructions that cause a computer to execute the steps of the whole vehicle wind noise speech intelligibility prediction method of any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311615551.4A CN117313249B (en) | 2023-11-30 | 2023-11-30 | Whole vehicle wind noise voice definition prediction method, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311615551.4A CN117313249B (en) | 2023-11-30 | 2023-11-30 | Whole vehicle wind noise voice definition prediction method, equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117313249A true CN117313249A (en) | 2023-12-29 |
| CN117313249B CN117313249B (en) | 2024-01-30 |
Family
ID=89281463
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311615551.4A Active CN117313249B (en) | 2023-11-30 | 2023-11-30 | Whole vehicle wind noise voice definition prediction method, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117313249B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117690303A (en) * | 2024-02-04 | 2024-03-12 | 四川三元环境治理股份有限公司 | Noise early warning system, device and early warning method based on traffic data acquisition |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE19813285A1 (en) * | 1998-03-26 | 1999-09-30 | Bayerische Motoren Werke Ag | Method of automatic measurement of the clarity of speech in a motor vehicle |
| US20160138967A1 (en) * | 2013-06-21 | 2016-05-19 | Brüel & Kjær Sound & Vibration Measurement A/S | Method of determining noise sound contributions of noise sources of a motorized vehicle |
| CN107084851A (en) * | 2017-04-18 | 2017-08-22 | 常州大学 | Method based on psychoacoustic parameter in statistics Energy Flow Analysis prediction high-speed train |
| CN110398338A (en) * | 2018-04-24 | 2019-11-01 | 广州汽车集团股份有限公司 | Wind is obtained in wind tunnel test to make an uproar the method and system of speech intelligibility contribution amount |
| CN110688712A (en) * | 2019-10-11 | 2020-01-14 | 湖南文理学院 | Evaluation index for objective annoyance degree of automobile wind vibration noise sound quality and calculation method thereof |
| WO2021064193A1 (en) * | 2019-10-04 | 2021-04-08 | Katholieke Universiteit Leuven | Method and apparatus for determining a measure of speech intelligibility |
| CN115497454A (en) * | 2022-09-21 | 2022-12-20 | 安徽江淮汽车集团股份有限公司 | In-vehicle language definition optimization space recognition method |
| WO2023005741A1 (en) * | 2021-07-29 | 2023-02-02 | 中国第一汽车股份有限公司 | Vehicle wind noise insulation performance evaluation method, device, apparatus and storage medium |
-
2023
- 2023-11-30 CN CN202311615551.4A patent/CN117313249B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE19813285A1 (en) * | 1998-03-26 | 1999-09-30 | Bayerische Motoren Werke Ag | Method of automatic measurement of the clarity of speech in a motor vehicle |
| US20160138967A1 (en) * | 2013-06-21 | 2016-05-19 | Brüel & Kjær Sound & Vibration Measurement A/S | Method of determining noise sound contributions of noise sources of a motorized vehicle |
| CN107084851A (en) * | 2017-04-18 | 2017-08-22 | 常州大学 | Method based on psychoacoustic parameter in statistics Energy Flow Analysis prediction high-speed train |
| CN110398338A (en) * | 2018-04-24 | 2019-11-01 | 广州汽车集团股份有限公司 | Wind is obtained in wind tunnel test to make an uproar the method and system of speech intelligibility contribution amount |
| WO2021064193A1 (en) * | 2019-10-04 | 2021-04-08 | Katholieke Universiteit Leuven | Method and apparatus for determining a measure of speech intelligibility |
| CN110688712A (en) * | 2019-10-11 | 2020-01-14 | 湖南文理学院 | Evaluation index for objective annoyance degree of automobile wind vibration noise sound quality and calculation method thereof |
| WO2023005741A1 (en) * | 2021-07-29 | 2023-02-02 | 中国第一汽车股份有限公司 | Vehicle wind noise insulation performance evaluation method, device, apparatus and storage medium |
| CN115497454A (en) * | 2022-09-21 | 2022-12-20 | 安徽江淮汽车集团股份有限公司 | In-vehicle language definition optimization space recognition method |
Non-Patent Citations (3)
| Title |
|---|
| 张辉;刘浩;周;: "车身密封性对风噪声声品质影响试验分析", 汽车工程师, no. 08 * |
| 杨春清;苏丽俐;李洪亮;王海洋;边科胜;: "高速工况车内语言清晰度优化", 汽车工程师, no. 01 * |
| 程志伟;: "语音清晰度和平均混响时间在车内声学的应用", 应用声学, no. 06 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117690303A (en) * | 2024-02-04 | 2024-03-12 | 四川三元环境治理股份有限公司 | Noise early warning system, device and early warning method based on traffic data acquisition |
| CN117690303B (en) * | 2024-02-04 | 2024-04-26 | 四川三元环境治理股份有限公司 | Noise early warning system, device and early warning method based on traffic data acquisition |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117313249B (en) | 2024-01-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117313249B (en) | Whole vehicle wind noise voice definition prediction method, equipment and storage medium | |
| CN116227042B (en) | Vehicle windage coefficient determination method, apparatus and storage medium | |
| CN111707351A (en) | Abnormal position positioning method and system based on noise vibration source of truck chassis | |
| Lepley et al. | A computational approach to evaluate the vehicle interior noise from greenhouse wind noise sources | |
| Huang et al. | Multitarget prediction and optimization of pure electric vehicle tire/road airborne noise sound quality based on a knowledge-and data-driven method | |
| JP2019114114A (en) | Vibration analysis device for panel member of vehicle | |
| Moron et al. | A CFD/SEA approach for prediction of vehicle interior noise due to wind noise | |
| JP4156639B2 (en) | Apparatus, method, and program for supporting voice interface design | |
| CN113886974B (en) | Method for predicting sound path noise of in-vehicle structure | |
| Wang et al. | A hybrid prediction for wind buffeting noises of vehicle rear window based on LES-LAA method | |
| Siano et al. | Sensitivity analysis and correlation Experimental/Numerical FEM-BEM for Noise Reduction assessment of an engine beauty cover | |
| Morales et al. | Efficient wave-based acoustic material design optimization | |
| US20230274052A1 (en) | Method to reduce a vehicle pass-by noise | |
| CN113158331B (en) | In-vehicle noise optimization method based on genetic algorithm | |
| Ganty et al. | Numerical simulation of noise transmission from A-pillar induced turbulence into a simplified car cabin | |
| CN113127969A (en) | Automobile damping plate simulation modeling method, device, equipment and storage medium | |
| Leroy et al. | Uncertainty assessment for outdoor sound propagation | |
| Wu et al. | Multi-objective prediction and optimization of vehicle acoustic package based on ResNet neural network | |
| Pulvirenti et al. | A perceptual evaluation of numerical errors in acoustic FEM simulation for sound quality applications | |
| Zhang et al. | Virtual reality of interior noises of vehicles based on boundary element and neural networks | |
| Freeman et al. | Integrating CFD and Experiment: The Jaguar Land Rover Aeroacoustics Process | |
| CN117725765B (en) | Vehicle suspension multi-objective optimization method, device and medium based on response analysis | |
| CN113297675A (en) | Method and device for decomposing wind noise performance index of automobile, vehicle and storage medium | |
| KR20110028800A (en) | Layout for speeker in vihicle | |
| Nopiah et al. | Optimisation of acoustical comfort in vehicle cabin using goal programming |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |