Control for an active magnetic bearing machine with two hybrid electromagnet actuators

Abstract

This thesis work begins with the revision of state of the art about active magnetic bearings (AMB), the mathematical methods used to obtain geometric and physical parameters that will influence in the mechanical, electrical design and control system proposed by this prototype. The control system will activate the magnetic bearing to center its shaft, for which it is joined a variable load in order to study the best control performance under different load over the rotor proposed by requirements. When the rotor is not controlled in its own axis even though variable load, a position error will occur that will be corrected by the program of a control system that will center the shaft (rotor). For this design was evaluated generalized AMB models [2], [3], [4] to validate the best identification for this design, furthermore as a consequence to get the best performance for the control system as it was achieved by generalized models and it was evaluated the advantage of this AMB machine through “Two hybrid electromagnet actuators” and variable load fixed to its shaft. For this reason, it was necessary to test a simple AMB with only one electromagnet actuator [4], due to compare enhancement of hybrid characteristics for the electromagnet actuators, for which, also it was evaluated how many actuators could be necessary to join to an AMB system with the target to get the control. It means, in this work there are comparisons between a simple AMB, generalized AMB models and this design, owing to show the achievements of this design. In order to show experimental results in state of the art, it is known that Siemens presented Simotics Active Magnetic Bearings technology for wear free operation in large – machine applications, regulated magnetic fields hold the rotor in suspension precisely without oil or contact, to make this task, sensors capture the position of the shaft 16000 times per second and a regulator adjusts the magnetic field to keep the rotor hovering precisely in the bearing center [1]. By other side the author [4] describes the experimental results in which is proposed that at low speed the bearing parameters are mainly determined by the controller characteristics. While at high speed, the bearing parameters are not only related to the control rule but also related to the speed. This may be due to the influence of eddying effect. [4] Furthermore, by author [3], the algorithm to get fast responses in front of disturbances, the disadvantages of these algorithms are given by not enough memory space to execute them, due to computing time is short compared with rotor displacement response time, and it is defined that it could be possible to execute the control algorithm through a real-time operating system to obtain the desired response [3]. Finally, in reference [6] it is described about filtering every noise as additive white Gaussian noise, by a predictive filter, which is obtained by analyzing Least Mean Square (LMS) and feedback/feedforward algorithm.