Sound monitoring system of operating machines in power plant

Despite the current technology in place, the power plant operator´s job is still very demanding. Even most minor of operating errors can lead to distar. This paper introduces a sound detection system which is used to detect unusual sounds and relay the sound information of the operating machine´s status to the plant operators. The hardware system consists of microphones, the multichannel A/D converter, the sound track audio DSP24, and Pentium IV PC. The Visual C++ and matlab of program languages are used for the software development. They comprise packages of functions that are WavelnOpen, MmioOpen, WaveOutOpen, and PostThread Message which are all under the Window 2000 multimedia library. We developed the real-time multichannel sound monitoring algorithm consisting of 3 classes which save files, output the sound files to the speakers, and control the threads. The frequency characteristics of the major operating machines in the plant are reported.