Peter Cafuta

Nonlinear control of synchronous servo drive

The complete control design for a permanent magnet synchronous (PMAC) motor, derived from the input-output linearization, is presented in the paper. The motor model, written in the rotors d-q coordinates, is nonlinear with respect to the state variables and linear in the control. The input-output linearization makes it possible to write the motors model in Brunovski decoupled canonical form which makes the synthesis of linear controllers possible. The proposed control structure allows perfect tracking of smooth references in the case of nominal parameters. In relation to the bounded parameter perturbations, the robustness of the feedback system is improved by introducing an additional compensation signal which assures the stability of the perturbed system in Lyapunovs sense. The influence of parameter variations that prevents exact compensation of the control plant nonlinearities was analyzed for the PMAC motor. A description of the laboratory setup is given, and the experimental results of the proposed PMAC servo drive control are presented.

IM Torque Control Schemes Based on Stator Current Vector

This paper proposes an induction machine torque control derived from the model in the stator current vector reference frame. The required torque is produced by simultaneously manipulating the magnitude and the rotation speed of the stator current vector, thus forcing the rotor flux linkage vector to change implicitly in such a way that overall stability is preserved. Additional control features include maximal torque-per-ampere ratio in steady state and almost perfect command tracking even if the machine is magnetically saturated. The control adopts a cascaded structure and is based on a partial dynamic inversion of the reduced model that assures existence and uniqueness of the inverse mapping between the required torque, the rotor flux linkage vector, and the stator current vector. Singularity at zero rotor flux linkage represents no restriction for the control performance in the admissible machine operating range. The implementation of the proposed control requires the estimation of the torque-producing rotor flux component and cascaded stator current controllers. Experimental results confirm the key expectations and show the potential and benefits of the proposed control schemes.

Non-Holonomy in Induction Machine Torque Control

In this brief, induction machine (IM) torque control is studied as an example of a 3-D non-holonomic integrator including drift terms. By expanding Brocketts controller derived for the driftless systems, a control structure is proposed that provides simultaneous modulation of both the amplitude and the frequency of the sinusoidal stator current vector. Although not explicitly controlled or programmed, the rotor flux linkage vector is implicitly forced to track natural periodic orbits satisfying non-holonomic constraints of the IM. The proposed control assures high dynamics in the torque response, maximal torque per amp ratio during transients and in steady-state and global asymptotic stability. The overall IM control scheme includes cascaded high-gain current controllers based on measured electrical and mechanical quantities together with rotor flux linkage vector estimator. Simulation and experimental results illustrate the main characteristics of the proposed control.

Modelling of the robotic Powerball®: a nonholonomic, underactuated and variable structure-type system

The Powerball® is the commercial name for a gyroscopic device that is marketed as a wrist exerciser. The device has a rotor with two underactuated degrees of freedom, which can be actuated by the appropriate motion of human or robot wrist axes. After the initial spin, applying the appropriate motion and torques to the housing leads to a spin-up of the rotor. Finding these torques intuitively is an easy task for human operators, but a complex task for a technical consideration, for example, in robotics. This articles main contribution is a novel dynamic model that considers friction effects. The presented model includes all three working principles of the device: free rotor mode and both modes of rotor rolling in the housing. The work introduces models with one and two degrees of freedom actuation, both of which are suitable for laboratory control experiments. An estimation of the friction is discussed, and both the simulation and the experimental results are presented to evaluate the models.

Pulsating torque reduction for permanent magnet AC motors

Control methods for the pulsation of torque reduction for the surface-mounted permanent magnet motors are discussed. The pulsation torque is a consequence of the non-sinusoidal flux distribution and due to interaction of the rotors permanent magnets with the changing stator reluctance. The proposed control method is an estimator based. To ensure parameter convergence the Lyapunov direct method is used in the estimator design for the flux Fourier coefficients. A novel nonlinear torque controller based on the flux/torque estimate is introduced to reduce the influence of the flux harmonics. The influence of the cogging torque is considerably reduced at lower motor speed using the internal model principle and adaptive feedforward compensation technique. An overall control scheme and experimental results are also presented.

A contribution to the control of the non-holonomic integrator including drift

In this paper a three-dimensional non-holonomic integrator (NI) with drift terms is considered. For this type of plant, the question how to obtain desired piecewise constant output functions with minimum norm control inputs is explored. It is shown that this can be achieved by a nonlinear controller that provides simultaneous modulation of both the amplitude and the frequency of the harmonic input vector. The internal states are implicitly forced to follow natural periodic orbits satisfying the non-holonomic constraints of the plant. Global asymptotic stability and high dynamics in the output response are achieved. The problem of singularity at zero initial state is solved by a time optimal control scheme for the internal states. By combining the nonlinear controller and the time-optimal controller using an appropriate switching strategy, a powerful control concept can be established. The torque control of an induction machine is considered as an illustrative example for the application of the control scheme. Experimental results of the closed loop feedback control system are presented.

Control of nonholonomic robotic load

Powerballreg is a commercial name for gyroscopic device that is marketed as a wrist exerciser. The device has rotor with two unactuated degrees of freedom and can be actuated with suitable rotational motion of human or robot wrist axis [1]. After substantial initial rotors spin, the properly applied torque and motion about two wrist axis lead to spin-up of the rotor. Finding this torque intuitively is easy job for most peoples, but not so easy for technical consideration for example in robotics. In the paper, a modeling of different modes of the device is presented first. A dynamic model with nonholonomic rolling connection which appear in normal operational mode is discussed. With the rotor and housing connection the additional friction effect is observed. When the nutation reaction torque is to low, only dissipation of energy is observed and rotor stops. But when the reaction torque causes normal forces on a connection in a degree that the friction is high enough, the rotor shaft begins to roll. The connection with the housing takes place in a gap so two connecting pairs are possible. One is up-down for left precession and another is down-up for opposite precession rotation. Developed models are used as a basis for control structure in the second part when the experiments are performed with the robot. The control strategy is oriented towards underactuated system, active and passive robots degrees of freedom.

Discrete boundary layer current control for VSI-IM drive

This paper discusses the time-discrete switched variable structure field oriented current control of an IM (induction motor) drive. VSI (voltage source inverter) is directly controlled, exploiting boundary layers and quantized field rotation velocity for addressing the switching table. In current control design, the fact that the calculation of the stator voltage vectors from the d-q control components does not have a unique solution is exploited. This is considered as design freedom, which is further extended by introduction of preassigned-order switching according to the time scale separation of magnetization and the generation of torque currents. Boundary layers with different reaching dynamics introduce motor stator zero voltage, which is selected by finite automaton from nearest VSI state.

New IM torque control with implicit rotor field tracking and efficiency maximization

The paper proposes an IM torque control based on the stator current vector reference frame. The introduced control concept rejects the paradigm of decoupled control with constant rotor flux vector. Instead of this widespread assumption the proposed control is derived from augmented dynamic equilibrium conditions assuring the existence and uniqueness of inverse mapping between the required torque end stator current vector. The rotor flux trajectories are forced to change only inside the globally stable sector under all operating conditions by manipulating simultaneously the magnitude and the rotation speed of the stator current vector in an appropriate way. Additional features include the maximal torque per ampere steady state ratio, almost perfect command tracking at non-zero rotor flux and insensitivity to the bounded perturbation of the rotor time constant mark introduced control. Implementation of this control requires estimation of the machine torque and cascaded current controllers. Experimental results confirm the key expectations and show the potential and benefits of the proposed control.

PWM-less current controlled VSI-IM drive

In this paper the PWM-less current controller for a voltage source inverter is discussed. The controller is of boundary layer type and a switching table realized. It allows employment of positive definite and also positive semi-definite vector function determining reaching dynamics. In its design the fact that the calculation of the stator voltage vectors from the d-q control components does not have a unique solution is exploited. This is considered as design freedom, which is further extended with logic control introducing preassigned-order switching according to the time scale separation of magnetization and the generation of torque currents. Zero voltage vectors are selected by finite automaton to the closest previous active voltage in order to minimize the inverter switching losses, reduce chattering and gives O(T/sub s//sup 2/) error type of current.