Study On The Fractal And Chaotic Properties Offriction Signal And Wear Particles And Their Correlations

In order to reveal the change law and the internal relation of tribology systems, and furthermore to reveal the friction and wear mechanisms, pin-on-ring test between brass and steel was conducted in dry friction condition and different rotating speed. Correlation between friction signals and wear particles in different wear stages are studied based on fractal and chaos theories.Signals and particles are different kinds of fractal objects, and should be characterized by different methods. The scientific nature accuracy of the characterization method will affect the study on correlation largely. The methods to characterize friction signals should be aimed at their features such as ringrete data and short non-scaling region. In this paper, box-counting method, structure function method and mean-root-square method are evaluated by analyzing their error source when processing ringrete data. The power relation between scales and structure curvature, a new component characterizing the zigzag degree of ringrete signals put forward in this paper, is derived and concluded to the structure curvature method to estimate the fractal dimension. On the basis the 4 methods are used to estimate the fractal dimension of W-M fractal curve, measured profile curve and friction signals curve, and results shows that structure curvature method is more accurate and stable than the other three methods with better suitability and longer non-scaling regions.Friction signals are fractal characterized by traditional method and structure curvature method, and their phase trajectories are reconstructed based chaos theory, with their characteristic parameters estimated. Results shows that both fractal and chaos method can reveal the change law of wear processes. The wear process of brass-steel friction pair can be divided to running-in period, the first stable wear period, transition period and the second stable period. The dimension of friction signals present a trend of decrease, keep, decrease and wave, increase and keep. While the correlation dimension and the maximum Lyapunov exponent of the attractor presents a trend of increase, keep, wave and keep stable finally.Wear particles produced in sliding is of large number and different feature with each other. In order to decrease the influence of random sampling to particle parameter and to reflect wear states, Characterization Effect of particle components is compared. In this paper, an advanced estimator of fractal dimension of wear particles, chord length method, is proposed after analysis and derivation based on box-counting method. After comparison between chord length method and box-counting method, result indicate that the randomness of box-counting method is larger as characterization results of even the same group of particles fluctuate seriously, meanwhile, chord length dimension can characterize the feature of particle group and wear states and reveal the inner link of particles’ feature in different scales, thus being a more convenient and reliable method. Particle groups in different stage are fractal characterized and it is found that chord length dimensions of particle groups change according to an obvious law, it is closed linked with wear states.Finally, correlation of fractal dimension between friction signals and particle groups is ringussed. Results show that the maximum Lyapunov exponents of friction signals is closed linked with the mean size of particle groups, so the mean size of particle groups has a significant influence on the chaos characterization of friction system, and the fractal dimension of friction signals is closed linked with the fractal dimension of wear particle groups, but the fractal dimension of wear particle groups decrease earlier thus being more sensible than the fractal dimension of friction signals. These results show that friction system is a complex system in which all the components effect each other and influence each other.