Optimal reset control and disturbance compensation for advanced hard disk drives

The demand for better servo control technology for hard disk drives (HDDs) never ceases. One important element for HDDs servo control is fast read/write speed which demands the fast and swift motion of HDD servo system. The other important requirement for HDDs is better precision for future drives w...

Full description

Saved in:
Bibliographic Details
Main Author: Li, Hui
Other Authors: Du Chunling
Format: Theses and Dissertations
Language:English
Published: 2011
Subjects:
Online Access:https://hdl.handle.net/10356/46315
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:The demand for better servo control technology for hard disk drives (HDDs) never ceases. One important element for HDDs servo control is fast read/write speed which demands the fast and swift motion of HDD servo system. The other important requirement for HDDs is better precision for future drives with ultrahigh densities. In order to achieve fast and high accurate performance, a framework of optimal reset control is proposed for HDD. In such theory framework, LQR optimal reset control and H2 optimal reset control are included. LQR optimal reset control can improve transient response performance; while H2 reset control can achieve better control accuracy and disturbance attenuation performance. A unified reset control is design to combine the merits of the two control and achieve a unified performance via a soft switching mechanism. The dual stage hard disk can achieve higher bandwidth than the single one. The LQR optimal reset control is also designed for the dual stage HDD servo systems to achieve better transient performance. In order to attenuate di®erent disturbances, two add-on disturbance rejection methods are proposed. One "add-on" disturbance method which can cancel most of the components in periodic disturbance has been proposed. This method has rebuilt a signal which can approximately represent the disturbances thus cancel the disturbances by adding the opposite value of this signal. The other "add-on" disturbance compensation method is a general form of disturbance observers. This method does not need to solve the plant model inverse, and uses H1 control method to design the Q-¯lter in the disturbance observer. After using these two methods, the disturbances can be compensated well in each case.