Noise identification in reverberant sound field by using simulated annealing

Abstract

Noise control is important and essential in an enclosed machine room where the noise level is restricted by the occupational safety and health act. Before the appropriate noise abatement is performed, the identification work of location and free-field sound energy of equipment inside the reverberant sound field become crucial and an absolute prerequisite. Research on new techniques of single noise control and sound absorption system has been well addressed and developed; however, the research work on sound identification for an existing multi-noise enclosed room is rare and observably insufficient. Without the actual location and pure free-field noise level, noise control work will be improper and wasted; therefore, the numerical approach of noise recognition from the reverberant sound field becomes necessary and obligatory. In this paper, the novel technique of simulated annealing (SA) in conjunction with the method of minimized variation square is applied in the following numerical optimization. In addition, various sound monitoring systems in detecting the noise condition within the echo area is also introduced. Before noise identification can be carried out, the accuracy of the mathematical model in a single-noise enclosed system has to be checked by SoundPlan (a professional simulation package in sound field). Thereafter, the SA recognition of three kinds of multi-noise systems has to be exemplified and fully explored. The results reveal that either the locations or sound power levels (SWLs) of noises can be precisely distinguished. Consequently, this paper may provide an efficient and rapid way in distinguishing the location and free-field noise level of equipment in a complicated sound field.