Sajeeva Abeywardena

Scott Beinke

Ben Cazzolato

ECP Control Moment Gyroscope

Sajeeva N. Abeywardena, Scott Beinke and Ben S. Cazzolato

Background

In 2008 the Sir Ross and Sir Keith Smith Fund awarded the School of Mechanical Engineering a grant to purchase an Educational Control Products Control Moment Gyroscope rig for teaching rotational dynamics and control. This rig complements control rigs previously purchased with grants from the Smith Fund; Quanser 2DOF Heli and Quanser 3DOF Hover. This gyroscope has now been fully commissioned and is used for teaching both undergraduate and postgraduate students. The teaching tools developed in the School of Mechanical Engineering have been posted here so others may benefit.

Control Moment Gyroscope

The four axis Control Moment Gyroscope is a dynamically rich system that provides superb demonstrations of multi-DOF rigid body control. Elementary experiments are readily performed that readily show the fascinating phenomenon of gyroscopic torque and its use in precision high authority control. More advanced topics range from MIMO linear control to fully general nonlinear control with singularity avoidance. Thus the system yields demonstrations that are intriguing to the layman and post-doctorate alike! In addition, the plant may be used to emulate the control of satellite attitude. Stimulating experiments first show the open loop nutation modes and then demonstrates their effective control.

The apparatus includes low friction slip rings at axes 3 and 4 for unlimited range of motion, and precision encoders for feedback of all position and velocity states. High torque density rare earth magnet motors drive the rotor and inner gimbal. A host of safety features such as fail-safe brakes, inertial switches, and real-time watch-dog monitoring provide for safe operation of the apparatus

Simulink Non-linear and VR Model

A Simulink model of the Control Moment Gyroscope system has been built employing an embedded m function of the dynamics derived by ECP and a model provided by Dr. Richard G. Cobb at the Air Force Institute of Technology, Wright Patterson AFB. The Simulink model is based on the default model that comes with the ECP real-time Simulink Option. A Virtual Reality Model (VRML) has been developed by the School of Mechanical Engineering, The University of Adelaide to allow one to visualise the plant response in virtual reality. You will need the VRML Texture Image file for the VRML model. Save all these in your working directory.

Simulink Real-Time Windows Target

It is possible to operate the CMG using Mathworks Real-Time Windows Target. This also requires the ECP real-time Simulink Option. We have developed a real-time Simulink model which initialises the control card, then commences a number of different control examples based on the tests in the manual. The initialisation m file is also needed. Please note that this file is still being developed and needs further work.