STUDY OF HORN EFFECT IN AN INTERSECTION TRAFFIC NOISE MODEL

Abstract

Industrialization and urbanization produce environmental noise that is undesirable. These unwanted or excessive sounds adversely affect our health and environment, despite not being noticed by us. Transportation is the most important noise source we must deal with every day, as it is the fastest growing and the most difficult to avoid. The most significant contributor to transportation noise is highway traffic. Sound levels were analysed using A-weighted sound levels. The traffic sound levels are thus expressed as dB(A). There has been an increase in ambient noise level in New Delhi (Leq) exceeding the permissible limits for almost every type of land use. During the daytime, the average equivalent noise level is 81 dB(A) in the silence zone, 79 dB(A) in the residential zone, and 81 dB(A) in the commercial zone and 78 dB(A) in the industrial zone, while the CPCB recommends 50, 55, 65, 75 dB(A) for the respective areas. Noise on the highway’s accounts for almost 70% of all noise. Therefore, it is essential to predict traffic noise in advance. In some ways, the different traffic noise prediction models around the world differ from one another, but their fundamental methodology is the same. The basic noise levels are based on adjusting a series of parameters to consider geometric factors, traffic flows, observer distances and barriers. Traffic noise levels are predicted for highway intersections using a Multiple Linear Regression (MLR) model and an Artificial Neural Network (ANN). Data Collection & Site Selection: The site was selected based on a variety of factors, including: (i) an unsignalized highway intersection (ii) the site should be removed from known noise sources, such as airports and construction sites (iii) there should be no noise barriers installed at the site. Manglaur intersection is on NH 334 between Haridwar and Delhi that was studied for data collection. During the data collection period, 18 hours of data were collected, at 15-minute intervals. The parameters used to predict traffic noise are (i) volume of traffic (ii) type of vehicle (iii) speed of the vehicle. It is concluded that ANN gives considerably better results than MLR in this study. Based on the analysis of the ANN and MLR, the mean absolute error is * 0.36 dB(A) and * 2.02 dB(A), respectively.