As machinery and plants install with a large scale and high efficiency in modem advanced information technology society, malfunctions and accidents of them give tremendous impacts for such as security, environment and economy system, etc. And so detection technologies of abnormality of them are regarded as important as an effective method of prevention of malfunctions. The characteristics, reliability and problems of technologies of the detections and predictions of abnormality for machinery and plants were investigated to utilize abnormality models and given considerations of criterion of judgment for detections of abnormality, wide area and simultaneous detection systems of abnormality, etc, in this paper.
Fatigue tests of forged low-alloy steels, two kinds, of KSFA110 (NK standards, σB=1, 080MPa class), with various surface conditions were carried out to clarify the fatigue properties in the ultra-high-cycle regime over 108 cycles of repeated stress. Interior fractures occurred in the ultra-high-cycle regime, which were affected by the residual stresses. Fatigue crack propagation lives of the interior fractures were calculated assuming that the crack propagation life is predominant in the fatigue life. Parameters for the fatigue crack propagation law of the interior fracture were estimated as m=10.4 and C=2.33 x 10-18. Threshold stress intensity factor range of the interior fatigue crack propagation depended on the initial crack size.
If abnormalities of the marine engine are recognized in the early stage, the engine trouble and accident can be avoided. This paper describes the method of searching the position where the abnormality occurred on the engine by using the generating sound. A small-sized diesel engine was used as a test engine, and the experiment was made by generating the sounds which were regarded as an abnormal sound near the test engine under operation. The abnormal sound which had a sine-wave characteristic was generated with a loud speaker. The frequency of the abnormal sound was detected by using Maharanobis' generalized distance, and the abnormal position was searched by means of the frequency data. As a result, it was possible to search exactly the abnormal position by using the above method.