Quick-Look

Dynamic Stability of Large Centrifugal Compressors, 18-9102

Principal Investigators
Anthony J. Smalley
Justin R. Hollingsworth

Inclusive Dates: 12/02/98 - 04/02/99

Background - This project advanced the technology for predicting unstable subsynchronous vibration of large centrifugal compressors. SwRI’s extensive experience in analyzing compressor stability has enabled the development and successful application of a prediction approach that incorporates empirical factors in the values for the destabilizing forces. Large propane refrigeration compressors for liquefied natural gas (LNG) plants fall well outside the parameter ranges on which the original empirical factors are based. In addition, these compressors require a change from the original structure because they typically have multiple inlet side streams. The integrity of such machines has substantial significance to the economic viability of large LNG plants, and avoiding severe vibration is essential to this integrity. After SwRI predicted the potential for subsynchronous vibrations on a critical new machine, the operating company, engineering company, and manufacturer challenged SwRI to restructure and reconfirm the validity of the method for this class of machine.

Approach - Under this project, structural changes to the empirical factor calculation were formalized to reflect the presence of side streams. In addition, an SwRI engineer measured vibrations on the machine in question in the manufacturer's test stand. After assembly of a relevant database, the project compared predictions and observations of subsynchronous vibrations for this and a number of other large propane machines. A simplified nonlinear analysis was also performed with the goal of characterizing limit cycle vibrations.

Accomplishments - Measurements confirmed the predictions for the most recently analyzed machine and for every machine in the database. The new calculation structure makes the empirical factors convenient to determine for these complicated compressors. The nonlinear analysis showed it could qualitatively and quantitatively reproduce observed data with appropriate selection of nonlinear parameters. As a result of this work (completed in April 1999), SwRI has received four projects to improve the predicted stability of machines being designed for new LNG refrigeration trains.

Fluid and Machinery Dynamics Program
1999 IR&D Home SwRI Home