Research On Linear Motor Internal Model Control System Base On Kernel Ridgelet Regression

In many industry control fields, the controlled objects often must be moved in linear. However, due to the immature development of the linear motion driver, we have to transfer the rotary motion of the rotary motor to linear motion by mechanical transform facility for a long time. Fortunately, linear motor can be straightly driven directly without mechanical conversion mechanism so that the linear motor has many unique advantages. Due to the lack of intermediate mechanical conversion links, uncertain disturbance, such as load disturbance, parameter perturbation and thrust ripple, could affect the operation of linear motor without any buffer or attenuation links, which will deteriorate the servo performance, and will be difficult to meet the requirements of high precision and high speed servo drive performance. Internal model control has been widely used in industrial control field, for its popular advantage, such as simple construction, intuitive and simple designing, less online tuning parameter, clear adjusting method, and so on. While conventional internal model control could not meet the high-precision tracking requirements of the linear motor, due to its low internal model accuracy and poor matching degree between the internal model and the controller. Aim at getting good robustness and strong anti-interference ability of the linear motor servo system, this paper designs a new internal model control system based on conventional internal model control.The internal model control system designed in this paper has the following two characteristics:First, the internal model of the linear motor is built by kernel ridgelet regression method. Then by analyzing the structure of the internal model, the process of calculating the controller is converted to root a nonlinear equation. And secant method is used to implement it. In this way, a good matching degree between the internal model and the controller is made. Then the stability and steady-state error analysis is done. The simulation results show that, compared to the conventional internal model control whose controller is designed directly based on sampled data, the new internal model control method could have higher control precision and better anti-jamming, and there is no steady-state error for step input signal.Second, due to the characteristic of the direct drive, when the linear motor is tracking the periodic input signal, the displacement-dependent periodic thrust ripple and modeling error will reduce the servo performance. Aim at the above problem, and according to the conclusion in chapter four, that the tracking error is directly relate to the non-matching between the linear motor and the internal model, so, a compensator based on repetitive control theory is designed to compensate the repetitive disturbance, such as thrust ripple and other uncertainty. The simulation results indicate that, this control scheme could improve the tracking accuracy of the internal model control system effectively.