Design, modelling and control of a reaction wheel actuated system

The stability and control of spacecrafts has been an significant topic, since the beginning of the space age. The spacecrafts (S/C) can be required to be stable in some missions depending on the requirements. The S/C can also be required to rotate with a given speed, to observe a point in space, to keep its orientation. In these cases, depending on the orbit of the S/C, the attitude of the S/C should be changed with an active control method. The attitude control of a S/C can be realized with active and/or passive methods. Spin stabilization is a passive method and used for the control of some S/C in the beginning of the space age. This method utilizes the gyroscopic rigidity of a spinning object to reject the disturbances on the rotating axis. Gravity gradient stabilization is another passive stabilization technique. In this method, the S/C should be longer in one axis comparing it to the other axis. The gradient in the length causes the gradient in the gravity. In this way, the longest axis is aligned with the gravity vector and stability can be maintained through the orbit. However the S/C can rotate freely about the gravity axis, due to no control effort on that axis. Therefore it is not suitable for precise missions. In active methods, the attitude of the S/C can be changed with actuators. Magnetic torquers are used to exert a moment on the S/C by applying a current on an inductor. They can be used to dump the excess momentum of the momentum exchange devices. Thrusters can be used in the orbit transfers due to their high torque/force output. They can be used to apply high angle maneuvers on the S/C, also the momentum dumping operations can be done using thrusters. Momentum exhange devices uses the reaction torque generated when the speed or the rotating axis of a high inertia flywheel is changed. Reaction wheel (RW) is composed of a flywheel, an electric motor to drive it and electronic components to change the voltage/current applied on the motor in order to change the speed of the flywheel. Control moment gyroscopes (CMG) can be considered as a RW attached to a gimbal, and on that gimbal axis there is a high torque capable electric motor. The rotation axis can be changed by this gimbal and high control moments comparing to the RW, can be achieved by CMGs. RWs and CMGs are used in high precise missions. CMGs are more complex than the RWs and they are much more expensive. Depending on the mission requirements an active or a passive method and proper actuators can be chosen. In this thesis, reaction wheels, their electro-mechanical structure is researched. A literature survey is conducted on RWs and the control applications with RWs. In the literature, it is observed that most of the research is done on computer simulation environment. It is important to test the systems both mechanical and electronic, on a realistic test environment before real operations. On a test system, the sensors and actuators can be tested and a possible failure can be prevented beforehand. In this research, for the purposes of modelling an control, a RW and a RW actuated test system is designed on computer aided drafting (CAD) programme and then manufactured. As a result, it become possible to observe the behavior of this dynamic system. A mathematical model is developed and validated with the experimental data. Then, the desinged controller is tested on the platform. In the first section, the current research in the literature is presented. The purpose of the thesis is explained. In the second section, the design process of RW and the test platform is explained with the CAD drawings. The testbed can be rotated in two axis, pitch and yaw. On each axis, there is a RW to control the orientation. There are encoders on each axis to measure the angles. The measured angles are processed with a controller and the control signal is applied on the RWs. The processes of measurement, calculation, data recording, controller implementation are done on a micro-controller. In the third section, the mathematical model for a RW and the testbed is obtained. RW is modelled using DC motor model. The testbed is modelled with the Newton's second law, the friction is taken as viscous friction and embedded in the equation of motion. During tests, it is observed that the simulation results and the real data does not match. An update for the friction model is required. Literature is surveyed on the topic of friction. It is decided that the Stribeck friction model best represent the nonlinear behavior of the testbed. In the fourth section, controllers are designed. The first one is an open loop control law to determine the time when to increase speed of the flywheel and when to decrease it. In order to check if the desired angle is reached a feedback controller is needed. The second controller is A PD based controller designed using parameter space approach. In the last section, the contribution of the thesis is explained briefly and future work is discussed.