Disturbance Compensation for Linear Drive System Using Inverse-Model Based Approach

Lim, Kai Zhi (2019) Disturbance Compensation for Linear Drive System Using Inverse-Model Based Approach. Final Year Project (Bachelor), Tunku Abdul Rahman University College.

Text LIMKAIZHI.pdf Restricted to Registered users only Download (3MB)

Abstract

Tracking performances of drive system in machine tools are always a great concern in industry associated with its great effect on the quality of products. However, the tracking performances of drive system often affected by the external disturbances that normally existed during the machining processes. Large tracking error or contour error are caused by these external disturbances. Two general types of disturbance forces are namely; friction forces, and cutting forces resulted from the interaction between cutting tool and workpiece during cutting process. This proposal mainly focuses on the compensation of cutting forces through the design of control algorithm. A cascade P/PI controller with inverse-model-based approach will be designed to compensate the cutting forces acting on the linear drive single axis positioning system. The classical cascade P/PI controller consists of two loops, namely; the velocity loop that controlled using PI controller, and position loop that controlled using P controller. Design of cascade P/PI controller is based on numerical analysis on the gain margin and phase margins, Nyquist plot and bandwidth. The inverse-model-based disturbance observer is designed as an add-on model to the Cascade P/PI controller. Inverse-model based disturbance observer will estimate and cancel out the disturbances using difference between control input signal and signal obtained from filtered system output with inverse of the plant model that obtained from previous work. The tracking performances of designed cascade P/PI controller with inverse-model based observer are numerically compared with traditional cascade P/PI controller in terms of tracking error reduction, maximum tracking error, and transient response. The results obtained showed that Cascade P/PI with add-on module have better compensate ability than Traditional Cascade P/PI. The tracking error shown that when the input disturbance frequencies is 10Hz,20Hz and 30Hz, the percentage of improvement from Cascade P/PI to Cascade P/PI with inverse-model based approach are 71%, 48% and 33.89%. For the RMS error, the result show that Cascade P/PI with inverse-model based have lower error than traditional cascade P/PI and the percentage reduction after add on inverse-model based approach are 12.62%. Lastly, for the FFT tracking error analysis, this FFT result show that after implement the inverse-model based, the FFT tracking error is lower compare with traditional cascade P/PI. The result show that the percentage reduction after the inverse-model based approach are implement is 59.82%, 56.79% and 28.69% at input disturbance frequencies 13Hz, 16Hz and 32Hz.