Artificial Neural Networks for Predicting Noise from Construction Sites

Abstract

Exposure of populations to noise at their workplace and at home is a major concern for academia and industry. This concern has resulted in a number of standards for predicting the level of noise and studying the adverse noise-induced effect on the health of the work force. Among these standards is the assessment of noise for open construction sites. Current British Standard methods for noise assessment are limited in their application. The paper explains the need for noise prediction and demonstrates the limitation of the existing methods. A background introduction to Artificial Neural Networks (ANN) methods with a methodology for developing ANN applications is provided. ANN is proposed as an alternative approach to construction noise prediction methods. A model for construction noise prediction is presented and its structure is described.