| Petrography monitoring technique is an important fault diagnosis method. It is applied widely because it have broad scope of measure and high sense efficiency. But the mainkind subjective factors have bad effect on the development of the technique. And because of the complicated morphology features, there is no a quantitative parameter that can characterize the wear particles objectively and fully even now. It is found that the ferrography wear particles are multiscale fractal in microcosmic state. With the development of the fractal theory and computer technique, it is possible that ferrography wear particles are quantitative analyzed and identified automatically.The traditional experiment device of ferrography analysis is reformed, which is a image digital processing system by assembling ferrography ferrographic microscope, CCD video camera and the computer. That is the hardware foundation for the alteration from the artificial analysis of the ferrography wear to the automatic analysis. This paper programs a software based on the step size method and the box dimension, and calculates the samples of the ball particles , the cutting particles, the chip particles, the blocky particles and the laminar particles which are collected 100 every from three ways: the experiment, the field collection, and the photo of wear particles. The scope of the profile fractal dimensions of the five kinds of wear particles are received. And the box dimensions of the debris group of the cutting particles, the ball wear particles and both are calculated through emulation in process of wear.The result of calculation provides that the five kinds of wear particles are statistically fractal, and the profile fractal dimension obeys the normal distribution, i.e., the ball wear particles: D (D, 2) ~D (1.0231, 8.727E-05), the cutting particles: D (D, 2) ~D( 1.0990, 0.001526), the chip particles: (D, 2) ~ D( 1.0945, 0.000734), the blocky particles: (D- 2) ~D ( 1.1744, 0.001562), the laminar particles:D( D, 2)~D( 1.2585, 0.001399); and the box dimensions of the debris group is altered corresponsively with the amount and the kind of thewear particles. In the first phase of wear, the box dimension is gruadully increasedwith the decrease of the amount of the wear particles. The studying results of thispaper rich the extent of the ferrography technique, and it is very important to theon-line automatic recognition of the wear particles and the monitor of wear statusin theory and practical fields. |