Y. Hori

A Novel Induction Machine Flux Observer and its Application to a High Performance AC Drive System

Abstract We proposed an induction machine flux observer based on Gopinath’s minimum order observer theory. In this paper, we generalize the design methods showing that up to 10 different kinds can be constructed and that they can be categorized into 4 groups. We construct a flux feedback type vector controller incorporating the proposed flux observer, and confirm its effectiveness with both numerical simulations and experiments. Finally, we introduce a novel approach to reducing parameter sensitivity based on using all degrees of freedom inherent in the flux observer’s gain matrix.

Implementation of robust flux observer based field orientation (FOFO) controller for induction machines

The authors present a flux-observer-based field-orientation controller. The controller is centered about the flux observer, which performs the direct coordinate transformation without using any flux sensors. The advantage of this system is in the extremely low sensitivity of the generated torque to changes in machine parameters. The key is the desensitization of the flux observer, for which a variable pole allocation method for the flux observer is developed. A digital signal processor is used for actual implementation.<<ETX>>

Robust DC servosystem design based on the parameterization of two degrees of freedom control systems

The authors propose a general robust servosystem design method based on the two-degrees-of-freedom (TDOF) controllers. With this approach, it is possible to specify the command input response to an arbitrary reference model while keeping a strong robustness to system variations and a high suppression performance with respect to disturbances. These can be designed independently of the command input response. The TDOF method is generalized to unstable plants, and two types of practical parameterizations for the TDOF controller are proposed in order to design the speed and position servosystems, respectively. Their excellent control performances have been confirmed by experimental results. The application of one of the proposed servocontrollers to the motion control of robotic manipulators is described.<<ETX>>

Robust Flux Observer Based Field Orientation ( FOFO ) Controller

Abstract We succeeded in the implementation of the Flux Qbserver based Field Qrientation (FOFO) controller which we proposed at 10th World Congress in Munich. The flux observer estimates the unmeasurable rotor flux components without using any additional sensors and performs the direct coordinate transformation required for the field orientation. This system is extremely robust to changes in machine parameters such as the rotor resistance . To this aim, we developed the novel variable pole allocation method of the flux observer. DSP is used for actual implementation. In this paper, we describe its theoretical development on the complex number system based on the phase segregation/spiral vector method.

Position and mechanical impedance control method of robot actuators based on the acceleration control

A novel hybrid control method for the position and mechanical impedance of robot actuators is proposed and is based on the acceleration control previously presented by the author (1988). Because the acceleration controller suppresses the disturbance forces effectively and makes each joint fully independent of the others, simple decentralized controllers can be designed based only on the angle command for each joint. The author shows the superiority of the proposed system by comparing it with the resolved-acceleration control method.<<ETX>>

Optimal Control of Looperless Tandem Hot Strip Rolling Mills

Abstract A looperless tandem hot strip rolling mill system must be treated as a single multivariable system due to the strong interaction between stands. In this paper we develop a model of such a looperless system which consists of some 100 dynamic state equations and several hundred algebraic equations. The optimal control feedback matrix is calculated, and its relation to conventional control methods is discussed. Next, as a more realistic method, a decentralized optimal control technique is developed by using parameter optimization technique which greatly simplifies the controller hardware. Finally a novel method utilizing statistical methods is developed for obtaining an appropriate weighting matrix which forms the performance index for the optimal control.

A Novel Decentralized Control of Large Scale Power System Based on the Disturbance Observer

Abstract A novel decentralized control method of a Large scale power system is proposed. The whole system has been divided into some generator sub-systems. The interferences on the sub-system are treated as unknown disturbances. Any sub-system can be decoupled from the other ones by means of the fast feedforward compensation based on the estimated disturbances obtained by the disturbance observer. This compensation allows us to design the small-scale decentralized controller. This design method has the advantage that the controller is invariant even when the external power systems structure changes. Finally, the effectiveness of the proposed method is confirmed by numerical simulations.