Sound Quality Evaluation Research On Motorcycle Engine

With the development of economy and technological, we are requires not only the motorcycle have excellent maneuverability, but also it have a comfortable environment for us. The NVH performance of motorcycle is an important factor to affect us desire to purchase. Sound quality of engine is good or bad that have an important influence on NVH performance of motorcycle. The sound quality evaluation influenced by personal factors, and requires a lot of manpower, material and financial. Based on objective and subjective parameters, established subjective annoyance model to predict sound quality is bad or good, so we do not spend a lot of manpower and material resources.This thesis summary that the research of achievements about sound quality of domestic and aboard at first, and describes the theoretical basis sound quality: introduce the objective evaluation parameters of sound quality in detail, and the method of sound quality evaluation.According to the requirement and task of this thesis, use labview platform to design acoustic signal test and analysis software, include signal test, signal read, signal save, signal spectrum analysis, filter, psychoacoustic parameters analysis module.Playback and audition the sample data, so we are take 30 acoustic samples sound quality subjective and objective analysis, this thesis choose grouping of paired comparison method. This thesis set error checking, the weight factor of consistency to ensure the results reliability of experimenter, eliminated some large deviation data. At last, the weight factor of consistency of experimenter greater than 80.7407%.Calculated the psychoacoustic parameters of engine acoustic signal(A-weighted sound pressure level, loudness, sharpness, roughness, fluctuation strength and annoyance), associated with the subjective evaluation to correlation analysis and partial correlation analysis, so that it can determine intrinsic link between the subjective and objective parameters. The correlation coefficients of A-weighted sound pressure level, sharpness and subjective annoyance is great, indicating A-weighted sound pressure level and sharpness significantly affect the human ear to evaluation the subjective annoyance of engine.According to subjective annoyance evaluation score guide the engine acoustics research, choose some samples that have different score spectral analysis, the psychoacoustic parameters analysis, and filtered to determine the frequency bands that influence the engine sound quality, Certifying the engine sound quality subjective evaluation can guide to improved NVH performance of the engine.Established sound quality subjective annoyance model based on multiple linear regression and BP neural network. Studies shown: the maximum relative error of BP neural network model predicts is significantly lower than the predicted value of the regression model, that the maximum relative error. The performance index of BP neural network model is significantly best than the performance index of regression model. Determined BP neural network subjective annoyance model is superior to multiple linear regression model.