|
romac |
|
University of Virginia Rotating Machinery and Controls Laboratory P.O. Box 400746 122 Engineer’s Way Charlottesville, VA 22904 (434)924-3292 (434)982-2246(fax) |
|
Rotating Machinery and Controls Laboratory |
|
High Speed Centrifugal Compressor Test Rig
The objectives of this test rig is to determine the bearing loads when the compressor is operated at flow/pressure combinations other than the "design point" and the examination of the forces arising from aerodynamic cross couplings. At the design point, the bearing loads due to aerodynamic flows are zero but at other operating points, they can be substantial. There are some design guidelines available for estimating off-design loading but they often produce very poor estimates. This is a problem because resizing bearings after the compressor design is complete is very difficult. Consequently, the evolution of compressor designs is slow and very conservative. One of the goals of the current research is to give accurate measures of off-design bearing loading and the aerodynamic cross couplings. The existing test rig provides a test bench to determine the bearing loads and the aerodynamic cross couplings using magnetic bearings as load cells. Such data could be used to calibrate existing models or to validate CFD based predictions. Better predictive tools would permit more aggressive design and would, hopefully, contribute to substantially better performance across the technology.
The other objective of this project is to use the sensitivity of the compressor characteristic to the tip clearance of the impeller to design a controller which is capable of actively suppressing the oscillations induced by compressor surge. This would be achieved, using a magnetic thrust bearing to modulate the tip clearance of the impeller. This way, the stable region of the compressor operation will be significantly increased and the compression system will not enter surge cycles for a wide variety of disturbances that might occur downstream of the compressor pipeline. The principal advantage of the proposed approach over conventional surge control methods lies in that, in machines already equipped with AMB thrust bearings, the method can potentially be implemented by simply modifying controller software. This dispenses with the need to introduce additional hardware, permitting adaptation of existing machinery at virtually no cost.
|