A new steady-state traffic noise model for estimating L10(h) on free flow roads using Reference Energy Mean Emission Levels

Abstract

In this paper, a new steady-state model was developed for predicting L10(h) on free flow roads based on a simulated sound level function. Initially, Reference Energy Mean Emission Levels (REMELs) were used in the simulation process to determine the sound level generated by a single vehicle at a specified distance. Then, a random distribution was used for the time each vehicle reached the receptor, where a time-dependent function for estimating sound level was established. Next, a steady-state model was calibrated with data obtained from the simulated sound level function. This model was developed to estimate L10(h) based on traffic volume, average speed, distance, percentage of heavy vehicles, and angles of the road segment. Finally, the new model was validated with the measured data, with the accuracy of the model verified. To better evaluate the proposed model, the values of L10(h) were also calculated by CoRTN model. The mean absolute error for CoRTN model was 1.71 dB A, while it was 0.96 dB A for the proposed model. Since the measured data were not used for calibration, and the new steady-state model was based on a simulated sound level function, the new model had greater accuracy and fewer restrictions than other steady-state models. Using the proposed method, the need for measured data to calibrate the traffic noise models is reduced. Instead, there is a wide and unlimited range of data, which makes it possible to find more accurate relationships between L10(h) and independent variables of the model.