Composite Tension System Modeling And Control Strategy

Tension control is a very important part in web transport system. High precision's tension control is the key performance index of keeping products' quality. Our main concern is always how to control independently tension and linear speed in spite of their strong coupling.This paper takes laminator as our researching object, which includes four sub -systems: unwinding,jointing,printing and rewinding. Unwinding subsystem takes magnetic particle brake as its actuator, and the others all take DC motors and DC drives as their actuators. Unwinding and rewinding subsystems both take load cells as their measuring devices, but jointing-printing subsystem takes dancer roll. Laminating system includes most usual actuators and measuring devices in tension control system, so it is a good researching platform of tension control.Laminating product line is a typical three segment tension control system which is driven by three DC motors. This paper specifies a general modeling methods of tension systems in detail, concludes laminator's block diagram and state space model. From its models, we can see that it is a nonlinear,time-variant,strong cross-coupling,multi-variables control systems, all of these reasons make it more difficult to design a appropriate controller.Because web transport system has a typical characteristics which is unwinding and rewinding radius and inertia time-variant, it makes conventional control strategies difficult to achieve satisfying performance index. So this paper puts forward parameters adaptive PID control algorithm and gives simulation results which take unwinding subsystem for example. And then, an observer-based new sensor-less scheme is proposed, which aimed at lowering down hardware cost and increasing product line's velocity, based on the dynamic torque balance equations of rollers. Then, SIMULINK simulation results are compared between the conventional scheme and the proposed one. They indicate that the proposed scheme not only performs equivalently dynamic response as classical feedback control does, but also owns higher steady precision. It fully validates that the proposed algorithm is feasible. Because laminating system is a multi-variables and time-variant control system, a linear quadratic regulator control algorithm is proposed and gain scheduling control strategies is incorporated. Finally, MATLAB simulation study validates the proposed algorithm's feasibility.