Raye A. Sosseh

Effects of the torque model on the control of a VR spherical motor

This paper presents the effects of the torque model on the control of a variable reluctance spherical motor (VRSM) that offers several attractive features by combining multi-DOF motions in a single joint. A general form of the torque model for a VRSM is derived using the principle of energy conversion. The torque models for two specific design configurations developed at Georgia Tech are compared. The first has been based on an existing design characterized by a torque model in quadratic form. For feedback control of the spherical motor, the quadratic form of the torque model requires the use of nonlinear optimization schemes for computing the stator coil current inputs. The second design incorporating high coercive permanent magnets has a linear torque-current relationship and thus allows a closed form solution for both forward and inverse torque models. The effects of the torque model on a PD-controlled VRSM prototype has been studied both numerically and experimentally. Experimental results agree well with the computation derived analytically.

Finite element torque modeling for the design of a spherical motor

This paper presents the method of modeling the torque generated by a variable reluctance spherical motor (VRSM) that presents some attractive possibilities by combining pitch, roll, and yaw motion in a single joint. Unlike prior works on the torque formulation of a VRSM, which were based on a lumped-parameter approach using equivalent magnetic circuits (widely used in developing force/torque models for electromechanical devices), this paper presents a distributed-parameter approach to predict the motors magnetic field distribution for formulating the torque of a VRSM. A detailed three-dimensional finite-element (FE) analysis has been performed on a VRSM configuration that consists of both permanent magnet (PM) poles and air-cored electromagnets. The model obtained using FE methods offers more insight and an accurate representation of torque generated by a spherical motor.

Effects of fixture dynamics on back-stepping control of a VR spherical motor

This paper presents the effects of fixture dynamics on the control of a variable-reluctance spherical motor (VRSM). The cascaded VRSM dynamics has the appropriate structure for the so-called back-stepping controller, however, requires the computation of the exact motor dynamics at each sampling time. This computational burden has an effect on the performance of the control algorithm and imposes constraints on the hardware/ software architecture of the control system. By only computing the dominant parts of the motor dynamics, this computational burden can be reduced. We have developed a robust back-stepping controller to compensate for uncertainties account for imperfect modeling and intentional computational simplification. The performance of the robust back-stepping controller has been evaluated experimentally against a classical PD controller and a standard back-stepping controller, which serve as bases for comparison.

Effects of the Torque Model on the Control of a VR Spherical Motor

Abstract This paper presents the effects of the torque model on the control of a variable reluctance spherical motor (VRSM) that offers several attractive features by combining multi-DOF motions in a single joint. A general form of the torque model for a VRSM is derived using the principle of energy conversion. The torque models for two specific design configurations developed at Georgia Tech are compared. The first has been based on an existing design characterized by a torque model in quadratic form. For feedback control of the spherical motor, the quadratic form of the torque model requires the use of nonlinear optimization schemes for computing the stator coil current inputs. The second design incorporating high coercive permanent magnets has a linear torque-current relationship and thus allows a closed form solution for both forward and inverse torque models. The effects of the torque model on a PD-controlled VRSM prototype has been studied both numerically and experimentally. Experimental results agree well with the computation derived analytically.