Real-time estimation and irregularity control of the sliver draft system

Abstract

High regularity is important in manufacturing spun yarn. Among controllers used in regularity control, existing PID controller is suitable for linear time-invariant systems but not for the control of sliver draft systems with non-linear characteristic. Thus, the present study designed a knowledge-based fuzzy controller in consideration of disturbance and non-linear characteristic of sliver draft systems. RLS (Recursive Least Squares) was used as an estimation algorithm to formulate a model used in computer simulation for implementing the controller, and real time estimation was made using real data from draft systems to get a model close to actual systems. This research used a estimation model in designing a knowledge-based controller that minimizes the variation of control response to the sliver linear density of the system, and the result of linear density control was used to prove the superiority of the controller in control performance through CV%.