Nonlinear Friction Parameters Identification Research Based On Genetic Algorithm

As a common and complicated nonlinear physical process, friction is mostly generated in the relative motion between the non-ideal smooth surface. Friction may bring in obvious negative effects such as viscous hinder, limit oscillation and energy waste during the running process of most precise mechanical systems, which seriously restrict the whole system’s controlling and precision performance. Considering that, it is a continuous focus both at home and abroad that how to comprehensively and precisely cognitive the friction process, and effectively inhibit and weaken its negative effects.The research in this paper is mainly focus on the friction process during free-segmented motion of the moving mass on the cantilever. Firstly, time-varying compensation polynomial is brought in the classic stribeck model to build time-varying compensation model1, based on the theory of classic friction model and Flourier transform. After that, wavelet De-nosing and multi-self-adaptive-strategy Genetic Algorithm are used for model parameters identification to prove model1’s excellence. At last, considering model1’s limitation on model-order increasing, a new time-varying compensation strategy based on the real dynamic motion process, is applied to build the time-varying compensation model2. It is proved that model2solves model1’s limitation well by numerical calculation. The friction parameters of the mass on cantilever free-section movement are acquired by setting up time-varying compensated friction modal and nonlinear identification based on genetic algorithm, which provides critical damping data for dynamic analysis on cantilever and has an effective result during multiple projects research.