Improvement of interior sound quality for passenger car based on optimization of sound pressure distribution in low frequency

The low-frequency acoustic response model of car is established with finite element method (FEM) and boundary element method (BEM), and the model is validated via road test. Utilizing the equal rectangle bandwidth critical band within 20-200 Hz, the 16 sound samples are designed by orthogonal experiment with the collected signal of driver at 50 km/h as example. The psychoacoustical indices of each sample are calculated through the program compilation, and the annoyance of each sample is obtained via subjective test. The prediction model for sound quality is completed under the comprehension of genetic algorithm (GA), particle swam optimization (PSO) and support vector machine (SVM). Taking the sound pressure level of each equal rectangle bandwidth band in low frequency as variables and taking the minimal subjective annoyance of samples as optimization objective, an optimization model for sound quality is established. Subsequently, the optimal sound quality is obtained. An extra subjective evaluation validates this proposed optimization method.