Parametric study on the road traffic noise prediction models

In order to investigate complex and broad impacts, many studies have been conducted using noise mapping. Noise mapping have used several noise prediction models, but error between measured value and estimated value in applying in Korea have been confirmed. This error depends on properties of noise prediction models or uncertainties of the elements by the model users. This paper was aimed to offer a basic study on intrinsic effect of parameters as comparing assuming scenarios and noise prediction models. The variation in sound power level to changes in the main factors was explored by parametric study and sensitivity analysis. Sound power level of KHTN was increased by increasing each input factor in all scenarios. The pattern of output behavior was also similar for all scenarios. Traffic speed had the most effective on sound power level of KHTN model for all scenarios except for scenario 5. The different units among the main factors could be mentioned as the reason which sensitivities of traffic flow were quite low at all scenarios. In the case of comparison of noise prediction models, RLS90 relatively overestimated the sound power level compared to the other models for all main factors and all scenarios. The variation of PWL of RLS90 model was relatively sensible. This approach will give a better insight on the noise prediction method.