Evaluation and optimization of sound quality in high-speed trains

Abstract The rapid development of high-speed railways has been tempered by various technical challenges. As more high-speed lines carrying more high-speed trains have opened in China, abnormal train noise has become an intermittent problem, affecting the quality of high-speed trains and the subjective comfort of passengers and crew. To mitigate the abnormal noise in the conductor room of a high-speed train, this paper proposes a method for evaluating and optimizing the sound quality (SQ) inside a high-speed train. First, the vibrational noise in the train was tested and analyzed under passenger-carrying operating conditions. Second, the SQ was objectively evaluated by quantifying the interior noise with one traditional physical acoustic parameter and six psychoacoustic parameters. In a statistical analysis of a subjective listening questionnaire, the low-frequency noise was best described by a verbal expression (the Dichen descriptor in Chinese), which serves as an evaluation index in jury tests. A subjective jury (30 participants) scored the values of various sound samples. Although the A-weighted sound pressure level (A-W SPL) in the conductor room was relatively small, it was perceived as extremely irritating. The auditory perception of the human ear was not objectively described by the traditional A-W SPL, but was quantified by the psychoacoustic parameter evaluation. In a correlation analysis of the subjective and objective evaluations, the correlation coefficient between tonality and the Dichen sensation was 0.983. The tonality is the key parameter of perceived SQ in high-speed trains. Therefore, an objective quantitative model of the subjective Dichen sensation was established through multiple linear regression theory with tonality as a variable. To determine the source of the discomforting noise, the vibro-acoustic characteristics in the conductor room were analyzed and a finite element analysis of the end wall was performed. The abnormal noise was traced to a high-intensity 40-Hz pure tone caused by an externally sourced structural modal excitation near the end wall. Based on hybrid finite element–statistical energy analysis, a vibro-acoustic radiation simulation model of the end wall was established. After optimizing the end-wall structure, the SQ in the high-speed train was significantly improved. The optimization effect was verified in a noise test conducted in the interior of a real high-speed train, and in a vibro-acoustic radiation simulation model. Our results can guide the evaluation and optimization of SQ in high-speed trains.