Fig. 12 Scatter diagram of observed value and estimated value (normal and abnormal)
Conclusion
Much information is included in the vibration generated in one cycle of the engine, the vibration of the side wall of the 2-cycle diesel engine for marine use was obtained for a long time, and its characteristic was examined, and also, it was examined how it was changed. The radiant sound and the vibration in one cycle were frequency-analyzed, and the sound and the vibration attributable to the combustion and the slide of the engine were clearly confirmed in a characteristic pattern from the three-dimensional figure. In reviewing the slide of the piston of the engine, it was judged from the pattern in the three-dimensional figure that the information from the vibration directly propagated from the engine body is more than that obtained from the radiant sound. In this experiment, the information from the vibration was used from the above-mentioned reason, but it does not necessarily lead to that the diagnosis by the radiant sound is inferior. For example, in a case of the extensive monitoring, the radiant sound is more advantageous in the viewpoint of the simultaneous monitoring, and this point will also be examined. The multivariate analysis was tried to examine the relationship between the running time of the engine and the change in the vibration. As a result, the normal model for each engine running time could be prepared by achieving the multiple regression analysis of the vibration in a transition mode under the normal running condition of the engine. In the example, the vibration pattern under the normal condition is modeled, and the vibration pattern of the engine can be confirmed for the running time, and it is possible to examine whether the engine is in a normal running condition if the method is advanced.
References
1) Kimura, Nakai & Kishimoto : Abnormal Sound Detection by Neural Network in the Diesel Engine. Bulletin of the M.E.S.J.,Vol.26, No. 1 (1998)
2) Kimura, Nakai & Mizutani : Diagnosis System of Diesel Engine by Statistical Vibration Analysis. Bulletin of the M.E.S.J., Vol.24, No. 1 (1996)
