MODELING AND CONTROL OF THE ELECTRICAL ACTUATION SYSTEM OF AN ACTIVE HYDROMAGNETIC JOURNAL BEARING (AHJB)

摘要

The architecture of the electrical actuation module driving a magnetic-hydraulic bearing system is presented. The bearing is intended to be scaled for use in applications of all sizes in industries like shipboard for support of the engine-propeller shaft or in power-plants for the shaft through which the prime mover, e.g. steam or gas turbine, is driving the electric generator. The benefits of this new bearing is first and foremost its superb performance in terms of low down to practically no friction losses since there is no direct contact between the supporting bearing surface and the rotating shaft supported. Other benefits include the potential of active, inline, real-time balancing and alignment. To implement such concept of a magnetic-hydraulic bearing, the following tasks need to be carried out. First, identification of mechanical, electrodynamical and circuit properties of the bearing's electromagnets in the system is necessary. Toward such identification, a series of experiments needed to be carried out. To be able to carry out these experiments, a specific power electronic converter is developed to drive each electromagnet. The power electronic drive is a quad MOSFET circuit based on full-bridge converter topology and outfitted with appropriate sensory instrumentation to collect and record measurements of all the physical variables of interest. Special care has been taken to compensate for magnetic hysteresis of the electromagnets, mitigate any induction heating effects and maintain operation within the material's linear region i.e. without significant saturation occurring. The use of a power transistor bridge allows rapid changes to be applied on the electromagnet's load force which could compensate disturbance or misalignment developed on the shaft supported. The data series from these experiments are useful for formulating a possibly nonlinear model of the electromagnetical and electromechanical processes involved in the bearing's operation. Such a model can then be employed to help design a digital microcontroller system which could effectively drive the power electronics and electromagnets to perform their required tasks as part of the bearing. Besides, the model could also be used for the synthesis of the nonlinear, sampled-data (discrete-time) control law which will be programmed on the microcontroller system board.