The complete image-source method for the prediction of sound distribution in non-diffuse enclosed spaces

An image-source based computer model is presented for the prediction of sound distribution in non-diffuse fitted enclosed spaces, such as factories. The new model, CISM, was developed on the basis of the sound propagation predictions of the Ondet and Barbry RAYCUB model in laboratory and industrial spaces. CISM uses an optimum representation, based on the accuracy of the RAYCUB results, so that the prediction accuracy is maintained while run-time and data input requirements are reduced. The basis of CISM is described, and predictions for various configurations of a fitted laboratory space, an empty factory incorporating a single barrier and a theoretical space with absorptive patches are presented. For comparison, sound propagation measurements were obtained and RAYCUB derived models, RAYCUB-DIR and RAYCUB-DIR REDIR, were used to predict these sound levels. This work demonstrated that the CISM model is as accurate as the ray-tracing model RAYCUB in the described enclosed spaces. The advantages of the CISM model are a reduced run-time (for a typical factory, CISM is approximately ten times faster than RAYCUB-DIR) and reduced data input (CISM uses less than half the information required by RAYCUB-DIR). (C) 1997 Academic Press Limited.