CN-117694894-B - Risk degree assessment method for user abnormal sound complaints

CN117694894BCN 117694894 BCN117694894 BCN 117694894BCN-117694894-B

Abstract

The invention provides a risk degree assessment method for user abnormal sound complaints, which comprises the steps of obtaining abnormal sound audio data of the user complaints, calculating energy saliency A of the abnormal sound, constructing a modulation abnormal sound matrix, wherein the modulation abnormal sound matrix comprises a plurality of modulation abnormal sound saliency, the value range of the modulation abnormal sound saliency is A-100% -A+100%, obtaining subjective complaint degree data, associating the modulation abnormal sound matrix with the subjective complaint degree data, determining psychological saliency B, and predicting the risk degree of the abnormal sound complaints caused by the user in the market by using a neural network model, wherein the risk degree of the user complaints is represented by the psychological saliency B. The method can predict the possible degree of complaints of users on the market, is helpful for taking measures in advance, and reduces the after-sale cost and the risk of damaging the brand image.

Inventors

LIU ZHAO
LIU BIN
YU XIAOFEI
SHU YUAN
ZHAO MENG

Assignees

上汽大众汽车有限公司

Dates

Publication Date: 20260505
Application Date: 20231213

Claims (7)

1. A risk level assessment method for user abnormal complaints, comprising the steps of: s1, acquiring abnormal sound audio data complained by a user; s2, calculating the energy saliency A of abnormal sound; S3, constructing a modulation abnormal sound matrix, wherein the modulation abnormal sound matrix comprises a plurality of modulation abnormal sound saliency, and the value range of the modulation abnormal sound saliency is A-100% -A+100%; s4, obtaining subjective complaint degree data; s5, associating the modulation abnormal sound matrix with subjective complaint degree data to determine psychological saliency B; Step S6, predicting the risk degree of the abnormal sound causing the user complaint in the market by using a neural network model, wherein the risk degree of the user complaint is represented by a psychological prominence B; the step S1 specifically includes: Locating abnormal sounds complained by users, and identifying and reproducing the abnormal sounds complained by the users; analyzing and disassembling the working condition with abnormal sound; A data acquisition system is used for acquiring abnormal sound audio data of the whole vehicle under an abnormal sound working condition; The step S2 specifically includes: Performing FFT analysis on the abnormal sound audio data by using signal processing software to obtain a spectrogram of the abnormal sound audio, wherein the abscissa of the spectrogram represents frequency; Determining an abnormal sound center frequency f c , an abnormal sound complaint band upper limit frequency f c,u and an abnormal sound complaint band lower limit frequency f c,l ; Calculating a reference frequency band upper limit frequency f u,u and a reference frequency band lower limit frequency f l,l according to the abnormal sound center frequency f c , the abnormal sound complaint frequency band upper limit frequency f c,u and/or the abnormal sound complaint frequency band lower limit frequency f c,l ; Calculating an energy saliency A of abnormal sound according to the abnormal sound complaint frequency band upper limit frequency f c,u , the abnormal sound complaint frequency band lower limit frequency f c,l , the reference frequency band upper limit frequency f u,u and the reference frequency band lower limit frequency f l,l ; the calculation formula of the upper limit frequency f u,u of the reference frequency band is as follows: ; Wherein when In the time-course of which the first and second contact surfaces, , , , , When (1) In the time-course of which the first and second contact surfaces, , , , , When (1) In the time-course of which the first and second contact surfaces, , , , , ; The calculation formula of the reference band lower limit frequency f l,l is: ; Wherein when In the time-course of which the first and second contact surfaces, , , , , When (1) In the time-course of which the first and second contact surfaces, , , , , When (1) In the time-course of which the first and second contact surfaces, , , , , ; The calculation formula of the energy saliency A of the abnormal sound is as follows: 。
2. the risk assessment method for user abnormal sound complaints according to claim 1, wherein the sampling rate of the data acquisition system is more than 2.56 times of the analysis bandwidth when the abnormal sound audio data of the whole vehicle is acquired under the abnormal sound working condition.
3. The method of claim 1, wherein the signal processing software is sound vibration analysis software, including HEAD ARTEMIS, LMS testlab and/or BBM PAK.
4. The risk assessment method of user abnormal sound complaints according to claim 1, wherein the modulation abnormal sound matrix in step S3 includes 8 modulation abnormal sound saliency levels, a-10%, a-5%, a+5%, a+10%, a+15%, a+20%, a+25%, a+30%, respectively.
5. The risk level assessment method of user abnormal sound complaints according to claim 1, wherein in the step S4, the experimenter performs subjective complaint level assessment on the abnormal sound audio after the playback for a plurality of times to obtain subjective complaint level data, and specifically includes: selecting test persons according to gender, age, education level, and experience of silent study evaluation, wherein the ages are 20-29 years old, 30-39 years old, 40-49 years old, and 50 years old and above, and the education level is divided into special departments, gramineae, major, and doctor; And (3) carrying out subjective complaint degree evaluation after continuously playing back abnormal sound audio for more than 20 times by a tester, wherein the subjective complaint degree comprises 10 scores which are respectively 1, 2, 3,4,5, 6, 7, 8, 9 and 10, the lower the score is, the higher the complaint degree of the tester on the abnormal sound is, 1-5 is that the tester is not acceptable on the abnormal sound, and 6-10 is that the tester is acceptable on the abnormal sound.
6. The risk level assessment method of user abnormal complaints according to claim 5, wherein the step S5 specifically includes: modulating abnormal sound audio, so that the modulated abnormal sound audio respectively meets the modulation abnormal sound saliency; the experimenter carries out subjective complaint degree evaluation on the modulated abnormal sound audio; and selecting modulated abnormal sound audio with the most approximate subjective complaint degree of the abnormal sound audio after repeated playback, wherein the corresponding modulation abnormal sound saliency A+x% is the psychological saliency B of the abnormal sound.
7. The risk assessment method for user abnormal complaints according to claim 1, wherein the neural network model in step S6 is a BP neural network; the input variables of the neural network model include abnormal sound center frequency, frequency bandwidth, energy prominence and/or subjective complaint degree; The output variables of the neural network model include psychological saliency.

Description

Risk degree assessment method for user abnormal sound complaints Technical Field The invention relates to the field of vehicle abnormal sound evaluation, in particular to a risk degree evaluation method for user abnormal sound complaints. Background In the automotive industry, the problem of complaints about abnormal sound of the whole car has become a widely existing challenge. Once a user finds a particular abnormal sound, due to the masking effect, the abnormal sound hearing threshold is increased, and the user can more easily notice the abnormal sound again in later use, so that the perception and the dysphoria of the user are deepened. Currently, research on the problem of abnormal sound of automobiles is mainly focused on the engineering field, and analysis is performed on physical parameters obtained through sensors and test equipment, such as sound intensity, frequency spectrum, and the like. However, these parameters do not necessarily reflect the subjective perception of the abnormal sound by the user, and do not describe well the degree of boredom of the complaint abnormal sound by the user, and evaluate the degree of psychological complaints of the customer. And the abnormal sound is repeated for a plurality of times, so that the hearing threshold of the user can be changed, the aversion degree of the user is deepened, and the user cannot describe the user by using conventional physical parameters. With the penetration of psychoacoustic research, more and more attention is focused on the influence of sound on the psychological state and emotion of an individual, and a new possibility is provided for evaluating the perception and complaint degree of abnormal sound by users. The invention CN114764526A discloses an evaluation method for abnormal sound performance of a vehicle, which comprises the steps of subjectively evaluating the receiving degree of abnormal sound in the vehicle by a tester under a test condition, recording the subjectively evaluated data, detecting and processing the abnormal sound in the vehicle by an objective test module under the same test condition as the subjectively evaluated data to obtain data of sound quality parameters of the abnormal sound in the vehicle, matching the recorded subjectively evaluated data with the obtained sound quality parameter data, and establishing an evaluation matrix corresponding to the subjectively evaluated data and the sound quality parameter data. The method utilizes the evaluation matrix to evaluate the abnormal sound performance of the vehicle, is too complicated to evaluate, does not relate to the association of the abnormal sound in the development stage and the after-sales stage, cannot evaluate the abnormal sound risk in the development stage, and can cause that a large amount of resources are input for solving the abnormal sound problem in the development, so that the optimization of the utilization rate of research and development resources is not facilitated. Disclosure of Invention The invention aims to provide a risk degree assessment method for user abnormal sound complaints, so as to solve the problems, predict the possible degree of the user abnormal sound complaints in the market, help take measures in advance and reduce the after-sale cost and the risk of damaging brand images. The invention provides a risk degree assessment method for user abnormal complaints, which comprises the following steps: s1, acquiring abnormal sound audio data complained by a user; s2, calculating the energy saliency A of abnormal sound; S3, constructing a modulation abnormal sound matrix, wherein the modulation abnormal sound matrix comprises a plurality of modulation abnormal sound saliency, and the value range of the modulation abnormal sound saliency is A-100% -A+100%; s4, obtaining subjective complaint degree data; s5, associating the modulation abnormal sound matrix with subjective complaint degree data to determine psychological saliency B; Step S6, predicting the risk degree of abnormal noise causing user complaints in the market by using a neural network model, wherein the risk degree of user complaints is represented by a psychological saliency B. In one embodiment, the step S1 specifically includes: Locating abnormal sounds complained by users, and identifying and reproducing the abnormal sounds complained by the users; analyzing and disassembling the working condition with abnormal sound; and acquiring abnormal sound audio data of the whole vehicle under the abnormal sound working condition by using a data acquisition system. In one embodiment, when abnormal sound audio data of the whole vehicle under the abnormal sound working condition is collected, the sampling rate of the data collection system is more than 2.56 times of the analysis bandwidth. In one embodiment, the step S2 specifically includes: Performing FFT analysis on the abnormal sound audio data by using signal processing software to obtain a spectrogram of the abnormal sou