NEURO BASED CLASSIFICATION OF FACILITY SOUNDS WITH BACKGROUND NOISES

作者：Shibata Akihiro^*; Konishi Masami; Abe Yoshihiro; Hasegawa Ryuusaku; Watanabe Masanori; Kamijo Hiroaki

来源：International Journal of Innovative Computing Information and Control, 2010, 6(7): 2861-2872.

• Acoustic analysis
• Diagnosis
• Neural network
• Classification of sounds

The detection of abnormality in a facility is vital for plant operation. Humans feel the change of surroundings by various sensing such as eyes, nose, and ears. The diagnosis by the sound has the advantage of being able to detect the wide-ranging abnormalities. The purpose of this study is the recognition of various facility sounds with background noises. Also realization of the preventive maintenance of pipelines is studied. There are great needs for the diagnosis of gas leakage. Sounds of 9 facilities in, the plant are recorded. Adding to this, gas leakage simulator used to generate gas leakage sounds is made for various crack sizes and pressures. The recorded sounds of facilities are preprocessed by applying Fast Fourier Transformation. The features of sounds are extracted and classified by a Neural Network. As a result of the test, sounds of 9 facilities are recognized with over a certainty of about 94[%]. Moreover, the equipment in the factory is diagnosed by applying the recognition system. Classification and discrimination of cracks are carried out using a Neural Network. Through the acoustic experiments, it is proved that the method of the acoustic diagnosis can classify a leakage sound of a pipeline.

• 出版日期2010-7