Nonlinear, statistical models of tire-pavement noise

Tire-pavement noise is the result of a complex system of noise generation mechanisms and is affected by several different pavement and atmospheric parameters. Accurately predicting tire-pavement noise from given a set of parameters has proven difficult for researchers. The purpose of this research was to explore a wealth of pavement, atmospheric and noise data taken at the MnROAD pavement test facility and to develop amodel to predict tire-pavement noise on asphalt pavements. Using a series of sub-models, variables significant to noise generation were identified. Finally, two variations of a model of noise generation were developed, each capable of predicting one-third octave band on-board sound intensity (OBSI) spectra. The model was developed using a hybrid statistical-experimental approach and was able to predict overall OBSI levels to within 1.5 dB for 80-90% of the pavements tested.