Prediction of noise emission from power plant by a mathematical model

Abstract

The principles of outdoor sound propagation were used to a develop power plant noise prediction model in order to illustrate the accuracy level of the mathematical model. Sound Pressure Level (SPL) measurements were conducted during dry season at the Rayong Combined Cycle Power Plant (RPP) following the ISO 1996/1 and equal angle methods. Measurement and calculation methods for determining the significant sound source of the prediction model involved the determination of the Sound Power Level (PWL) using SPL measurement in outdoor environment according to ISO 37.46. The representative noise sources were the main buildings and the cooling towers. The PWLs of both sound sources were calculated by Colenbrander’s method and the area surface method. The PWLs of the left, right, and central parts of the main building were 114.7, 112.9 and 118.2 decibels, respectively. The PWL of the cooling tower was 116.7 decibels. During measurement at any point in the outdoor environment, the transfer function data of each transmission path were collected. The author developed a computer program using the Visual Basic programming language in order to perform this model calculation. The program can use measured SPLs for calculating the PWL of an industrial sound source, or it can use the PWL from the machine’s database and transfer function data to examine the SPL at any mimission point in outdoor environment. Those predicted SPLs were compared with on-site measured SPLs. The comparison results were used to investigate the accuracy level of this model. SPLs. The comparison results were used to investigate the accuracy level of this model. Results showed that the accuracy level of the model is within 10 decibels from the measured data. It was also found that the measured SPL at immission positions under the influence of upwind conditions were lower than predicted levels and the directivity correction caused by source positions and the environment affects the predicted value.