改良型積分器於伺服馬達之定位控制應用
Abstract
本論文提出一種改良型重置積分的方法，以改善系統輸出響應，並以實作方式驗證其可行性。以往控制器採用的積分器多為線性積分器(linear integrator, LI)，可降低系統的穩態誤差，但同時易使系統輸出產生超越量。文獻中已提出一種重置(reset)的方法，當系統誤差等於零的瞬間，立即將積分器的輸出值歸零，稱之為Clegg積分器(Clegg integrator, CI)。CI於暫態時將積分器輸出值歸零可降低超越量，但是於穩態時積分值的歸零造成控制系統無法抵抗外部直流干擾。本文提出改良型重置積分器的方法，其在暫態時可進一步降低LI控制系統輸出的超越量，並且加快系統安定時間；穩態時亦可抑制干擾，改善CI控制系統的穩態響應。
本文實驗平台共有兩個系統，其一以無刷伺服馬達安裝平衡負載或偏心負載，以分別建立線性或非線性實驗平台；另一以無刷伺服馬達結合導螺桿組成之線性平台，進行直線運動定位控制驗證。以上兩個平台均採用TI TMS320C6713 DSP與Xilinx可程式閘陣列(FPGA)結合而成之控制器硬體核心，並以C語言與硬體描述語言(VHDL)作為控制器設計之發展工具。將本文所提出之改良型積分器於此實驗平台驗證，並且由實驗結果可知本方法具有實用性。
This paper presents an improved reset integration method, in order to refine system output responses. In conventional controllers, a linear integrator (LI) is introduced to improve the steady-state responses. But the LI simultaneously causes system overshoot. J. C. Clegg firstly proposed a reset control element, called Clegg integrator (CI), which overcomes the limitations of the linear control. The CI resets the integrator’s output to zero whenever the CI’s input crosses zero. This reset motion reduces system overshoot and settling time, but leads to weak disturbance rejection. The purpose of this paper is to modify this reset control in order to combine the benefits of the LI and the CI. This paper employs two experimental systems: the first one consists of a brushless servo motor and a pair of inertial loads, so that the system can have either symmetric or eccentric payloads. Another one contains a brushless servo motor and a commercially available single-axis ball screw. In the experimental system, the control kernel is a DSP/FPGA system, and the C language and VHDL are utilized as developing tools for the servo control system. Experiments of the reset method have been conducted and proven to have better transient and steady responses than past approaches.
This paper presents an improved reset integration method, in order to refine system output responses. In conventional controllers, a linear integrator (LI) is introduced to improve the steady-state responses. But the LI simultaneously causes system overshoot. J. C. Clegg firstly proposed a reset control element, called Clegg integrator (CI), which overcomes the limitations of the linear control. The CI resets the integrator’s output to zero whenever the CI’s input crosses zero. This reset motion reduces system overshoot and settling time, but leads to weak disturbance rejection. The purpose of this paper is to modify this reset control in order to combine the benefits of the LI and the CI. This paper employs two experimental systems: the first one consists of a brushless servo motor and a pair of inertial loads, so that the system can have either symmetric or eccentric payloads. Another one contains a brushless servo motor and a commercially available single-axis ball screw. In the experimental system, the control kernel is a DSP/FPGA system, and the C language and VHDL are utilized as developing tools for the servo control system. Experiments of the reset method have been conducted and proven to have better transient and steady responses than past approaches.
Description
Keywords
滑動積分回授, 定位控制, Clegg積分器, Integral sliding feedback, Position control, Clegg integrator