TY - GEN
T1 - Evaluation of Objective Sound Quality Feature Extraction with Kernel Principal Component Method in Electric Drive System
AU - Huang, Xin
AU - Qiu, Zizhen
AU - Wang, Fang
AU - Zhiguo, Kong
AU - Li, Jifang
AU - Ji, Xiang
N1 - Publisher Copyright:
© 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
PY - 2023
Y1 - 2023
N2 - This paper takes the electric drive system used in the electric vehicle as the research object, in which the objective sound quality of noise samples is extracted and evaluated based on the kernel principal component (KPCA) analysis method. Seven different power-level prototypes and their related parameters are firstly presented, while the sample library under different operational conditions has been established. Secondly, the KPCA method is employed to extract the contributions of eight objective psychological features. The results show that the KPCA method can effectively achieve multi-dimensional feature extraction. The cumulative contribution of sharpness and tonality is meeting 98.18%, which can fully represent the objective sound quality. Moreover, the sharpness and tonality are more sensitive to the speeds under different load conditions. Especially, tonality obtains a different pattern with SPL-A above 10000 r/min. This work can provide a theoretical and practical basis for predicting and optimizing the objective and subjective sound quality in electric vehicle applications.
AB - This paper takes the electric drive system used in the electric vehicle as the research object, in which the objective sound quality of noise samples is extracted and evaluated based on the kernel principal component (KPCA) analysis method. Seven different power-level prototypes and their related parameters are firstly presented, while the sample library under different operational conditions has been established. Secondly, the KPCA method is employed to extract the contributions of eight objective psychological features. The results show that the KPCA method can effectively achieve multi-dimensional feature extraction. The cumulative contribution of sharpness and tonality is meeting 98.18%, which can fully represent the objective sound quality. Moreover, the sharpness and tonality are more sensitive to the speeds under different load conditions. Especially, tonality obtains a different pattern with SPL-A above 10000 r/min. This work can provide a theoretical and practical basis for predicting and optimizing the objective and subjective sound quality in electric vehicle applications.
KW - electric drive system
KW - experimental evaluation
KW - kernel function principal component analysis
KW - objective psychological feature
KW - sound quality
UR - https://www.scopus.com/pages/publications/85161430651
U2 - 10.1007/978-981-99-1365-7_20
DO - 10.1007/978-981-99-1365-7_20
M3 - Conference contribution
AN - SCOPUS:85161430651
SN - 9789819913640
T3 - Lecture Notes in Electrical Engineering
SP - 277
EP - 287
BT - Proceedings of China SAE Congress 2022
PB - Springer Science and Business Media Deutschland GmbH
T2 - Society of Automotive Engineers - China Congress, SAE-China 2022
Y2 - 22 November 2022 through 24 November 2022
ER -