乘用车用柴油机噪声声品质预测技术

Prediction of Passenger Car Diesel Engine Noise Quality

以某款直列4缸乘用车用柴油机为研究对象,选取描述发动机噪声的响度、尖锐度等5个心理声学特征进行客观评价分析;以主观满意度为评价标准,应用成对比较法进行主观评价测试研究;应用多层感知器神经网络(MLPNN)方法构建噪声心理声学特征与主观满意度之间的非线性映射关系;基于MLPNN预测模型分析5个特征参量对主观满意度的贡献度大小,从而得出响度、粗糙度和尖锐度为影响乘用车柴油机噪声品质前3个重要指标的结论。

With a four- cylinder inline passenger car diesel engine as a research object, we choose five psychoacoustic parameters（loudness, sharpness, etc） for objective evaluation and analysis. Taking the Sensory Pleasantness Index（SPI） as the criterion for subjective evaluation, the paired comparison is applied in the subjective evaluation and test research. The nonlinear relation between psychoacoustic parameters and SPI is established by using Multilayer Perceptron Neural Networks（MLPNN）. Based on the MLPNN prediction model, the contribution of psychoacoustic parameters to engine noise is analyzed, and the importance ranking of psychoacoustic parameters（loudnessroughnesssharpness） is obtained.