The Application of Correlation Dimensions of Vibration Signals in Analysis of the Technological State of Diesel Engines

The fault-free operation of diesel engine in a normal state is a kind of self-excited vibration, but with the change of its technological states in the operation, the vibration will become much more like a random vibration. All the variable information of these technological states is included in the vibration signals of diesel engines, which can be used to judge the different technological states. However, the existence of a lot of noise signals at the same time may affect the accuracy of the judgment. Aiming at this problem, this article puts forward that an ameliorated wavelet packet algorithm can be used to decompose and reconstruct the measured signals before using the fractal theory to make analysis and judgment on them. The experiment shows that, such a method can effectively improve the signal-to-noise ratio, and as a result, can accurately judge the technological states and fault signatures of diesel operation.