Study On Friction Modeling And Compensation On An AC Servo Feed System Based On Signal Feature

A table feed system is widely used in machine building and machining process. Several nonlinearities of the system, such as friction, backlash, hysteresis, saturation, non-model dynamics and external disturbance etc. bring big influences on system performance, among which nonlinear friction has the most evident impact on the performance of system motion control. After pitch error of ball-screw and backlash eliminated, friction will be the dominate factor influencing precision of motion control of a high-precision table servo feed system. So how to eliminate effectively influences of friction has been a key problem of study. To realize the expected operating performance by controlling influences of friction is a research task of significance and practical application value.A subject investigated in this dissertation is an AC servo driven table feed system consisting of ball-screw and linear rolling guide, on which friction measurement, diagnosis, modeling and compensation technologies are studied. From view of theory and practice, the methods of analysis, modeling and compensation of friction based on the feature of multi-signals are attempted to be presented.Firstly, pitch error of ball-screw is measured and analyzed by a laser Doppler displacement meter. Based on analysis of measuring results and theoretical analysis of compensation methods, different errors are classified and respectively compensated with the software and hardware method. Installation of an origin switch and improvement of the method of origin trapping effectively enhance repetitive positioning precision of the system. An error feed-forward compensation control means is employed in terms of periodical error and casual error of the system, and its amelioration of control precision and algorithm implementation. Compensation experimental results on the XY worktable verify effectiveness of the compensation method and correctness of theoretical analysis so that positioning precision of the computer-controlled table is improved well without addition of cost.Secondly, each section of the CNC AC servo table feed system is investigated in detail, then a dynamical model of the system is built and system parameter identification is carried out. By utilizing torque monitoring signal provided by the AC servo driver, the relationship between friction and steady velocity is determined. A simplified Stribeck friction model is established based on these experimental data.