Page 29

EDNE MAY 2015

Spin Cycle Control systems using the traditional PID approach need either robust system modelling or extensive empirical tuning to achieve optimum results. An alternative approach has become available, however, that can simplify control system tuning to as little as one parameter. Called active disturbance rejection control (ADRC), this approach can compensate for changing system parameters in real time. Consider, for example, an industrial winding machine. Winders are used in manufacturing paper, plastic films, metals, wire and textiles. These products are then used in other manufacturing processes, such as packaging and printing. Winders must produce tightly wound, uniform rolls of material to ensure the follow-on process’ success. Figure 1 depicts a winder. A speed controller is used to drive the motor spindle on which the material is wound. Position controllers are used to hold the dancer arms in the vertical position shown in the diagram. The dancer arms control the tension of the material as it is wound on the spindle. Winders are notoriously difficult to control due to constantly changing dynamics. The first challenge is that the material being wound may have deviations in quality. The position controllers for the dancer arms must control the tension equally well across varying product quality so as not to distort the product. For example, when producing thin plastic film, there are Simplify complex control problems using disturbance rejection By Adam Reynolds, LineStream Technologies Figure 1 Essential elements of a winder system. variations in resin quality, and these imperfections must not be amplified during the winding process. The second challenge is that the inertia of the winder’s motor shaft changes by an order of magnitude as the spool transitions from empty to full. The tension must remain constant as the material is wound. It is fairly easy to maintain the correct dancer position when the motor spindle is empty, but control of the dancer arms becomes more difficult as the spool of material grows. It is also difficult to control the dancer arm position as the winder accelerates and decelerates. These dynamic changes during the winding process are very difficult to tune and control with standard PID-based control. PID-based methods can only be statically tuned for single snapshots of an evolving system. Adequate PID tuning for winder applications requires significant time and expertise. Active Disturbance Rejection Control (ADRC), however, can actively compensate for changing conditions, such as inertia changes and material imperfections. Traditional PID control simply reacts to deviations in a system's output from a desired setpoint by applying a correction signal based on a system model or tuned to a specific system configuration. That correction is a linear combination of the error term, its rate of change, and its time integral (to correct for a DC bias). ADRC works by actively estimating what is disturbing the system from its desired behaviour, including changes within the system itself. It compensates for that disturbance using a non-linear combination of error terms. (For a detailed description of ADRC, see From PID to Active Disturbance Rejection Control by Jingqing Han.) This column continues with a comparison of the ADRC control technique with PID, using LineStream Technologies SpinTAC Motion Control Suite. Complete article, here 28 EDN Europe | MAY 2015 www.edn-europe.com


EDNE MAY 2015
To see the actual publication please follow the link above