摘要
从声品质客观评价、声品质主观评价、声品质客观量化模型三方面介绍了当前高速列车车内声品质评价研究现状和结果,归纳了高速列车车内声品质主、客观评价方法,总结了高速列车声品质客观量化模型,探讨了不同模型的优缺点,展望了高速列车车内声品质评价未来的发展方向。分析结果表明:现阶段高速列车声学标准和声学设计目标大多采用A计权声压级作为车内噪声评价指标,但在大多数情况下高速列车车内噪声以中低频率为主,此时A计权声压级不能很好地表征人耳对高速列车车内噪声的主观感受,应考虑使用声品质对高速列车车内噪声进行主、客观评价;未来应重点关注声品质客观参量对高速列车车内声品质适用性的研究,如何提取关键的声品质客观参量是高速列车车内声品质评价研究的重要方向之一;现有的传统客观心理声学参量不能很好地与机器学习模型结合实现声品质的准确评价,将传统声音信号进行特征提取,并结合机器学习模型进行声品质评价分析是未来高速列车车内声品质评价的发展趋势;传统的高速列车车内声品质主观评价方法评价时间长,可重复性差;建立高精度声品质客观量化模型代替传统主观评价方法,以缩短评价时间,提高评价准确性,是未来高速列车车内声品质评价研究的重点方向;传统的多元线性回归模型不能很好地评价高速列车车内声品质,随着机器学习的迅速发展,未来选择合适的机器学习模型结合智能算法优化,开发更准确、高效的声品质评价预测模型是高速列车车内声品质研究的重要内容。
The current research status and results of the evaluation of sound quality in high-speed trains were introduced from three aspects,including objective evaluation,subjective evaluation,and objective quantitative model.The subjective and objective evaluation methods of sound quality,as well as the objective quantitative models of sound quality in high-speed trains were summarized.The advantages and disadvantages of different models were discussed,and the future development directions of evaluation of sound quality in high-speed trains were forecasted.Analysis results show that at the present stage,most of the acoustic standards and acoustic design objectives for high-speed trains use A-weighted sound pressure level as the interior noise evaluation index,but in most cases,the noise is dominated by the low and medium frequencies.At this time,the A-weighted sound pressure level fails to well characterize the human ear's subjective feeling of the noise,so the use of the sound quality should be considered to carry out the subjective and objective evaluations of the noise.In the future,it is necessary to pay more attention to the applicability study of the objective parameters of the sound quality on the sound quality in high-speed trains,and the extraction of the key objective parameters of the sound quality is one of the important directions of the research on the evaluation of sound quality.The existing traditional objective psychoacoustic parameters cannot be well combined with the machine learning model to realize the accurate evaluation of sound quality.Extracting the features of traditional sound signals and combining it with the machine learning model for sound quality evaluation and analysis are the development trends of the future evaluation of sound quality in high-speed trains.The traditional subjective evaluation method of sound quality in high-speed trains has a long evaluation time and poor repeatability.Therefore,establishing a high-precision objective quantitative model of sound quality to replace tradit
作者
钱堃
沈政华
谭璟
刘珂
段继英
杜习康
赵剑
QIAN Kun;SHEN Zheng-hua;TAN Jing;LIU Ke;DUAN Ji-ying;DU Xi-kang;ZHAO Jian(School of Mechanical Engineering,Dalian University of Technology,Dalian 116023,Liaoning,China)
出处
《交通运输工程学报》
EI
CSCD
北大核心
2024年第5期154-172,共19页
Journal of Traffic and Transportation Engineering
基金
国家重点研发计划(2019YFE0121300)
中央高校基本科研业务费专项资金项目(DUT22RC(3)002)
中国博士后科学基金(2019M650657)。
关键词
高速列车
噪声
声品质
评价模型
神经网络
智能算法
high-speed train
noise
sound quality
evaluation model
neural network
intelligent algorithm
