Modelling the Traffic Noise Emission in a Metropolis, the study case of Goiania

Abstract

The noise pollution have been the center of political concerns around the world, since its social impacts are inherently linked with health problems and economical losses. Therefore, this paper aims trough the modeling techniques, applying the Principal Components Analysis and non-linear regression with the Non-Linear Least Square Method to modelling and to assess the individual influence of cars, motorcycles and heavy vehicles on the Equivalent Sound Pressure Level (SPL) in an urban areas of a megalopolis, which was Goiania - Brazil, and simultaneously verify the adequacy of the SPL with the municipal regulatory agency. Then, a set of experimental measurements of the SPL where made on the streets and the traffic flow stratification of cars, motorcycles and trucks counting, served as the input variables to the regression modelling evaluated, which were the linear, quadratic, exponential and power models, with one and two degrees of freedom. The results showed, with 93% of certainty the model can only be expressed as function of the cars numbers then neglecting the motorcycles and heavy vehicles variables, it was showed that the power model with one degree of freedom, best descried the dynamics of the SPL as function the cars only. Besides, approximated 85% of the measurements site had 67 dB(A) which is bigger to the maximum 55 dB(A) allowed by the regulatory agency. In conclusion, the proposed methodology showed to be consistent and can be exploited to verify its accuracy over new cities so asses the quality of the proposed models.