Identification of Sound Sources Via Genetic Algorithms

Abstract

An adequate work environment must offer conditions that do not compromise the health of employees. According to the Ministry of Labor and Employment, high levels of noise can cause serious health, physical and psychological problems. The Laboratory of Acoustics and Vibrations of the Federal University of Uberlândia developed a program aimed at controlling industrial noise. For this, a correct identification of the Sound Power levels radiated by the equipment to be treated is necessary. This work presents a methodology for identifying the NWS (Sound Power Levels) of a compressor set (compressor, rotation amplifier and induction motor) through the SPL (Sound Pressure Levels) measured at a constant spacing between measurement points and distance from the equipment. The methodology uses an optimization routine via Genetic Algorithms, which minimizes the error between the measured SPLs and the estimated SPLs. Through the use of this procedure, coherent estimated values ​​were found for the NWS of the compressor set, thus enabling a reliable identification for the use of the computer program to treat the sources.