 In this project, the focus is on the identification of linear and non-linear mechanical rotor-systems, and the development of damage detection tools. The methods being used for the early detection of damage in rotating shafts are predominantly based on pure signal analysis. If a certain threshold value, e.g. a maximum amplitude, is overshot, a warning signal or alarm will be triggered. In this case, only one particular parameter is being observed, the actual cause of the damage remaining, at this time, unknown. In the approach introduced in the project, a model-updating algorithm is used, in which a computational model is run in parallel with the measurement of physical data. Several categories of damage, such as cracks in the rotor, can be simulated in the calculation model.
With the help of identification methods, the location and extent of the damage can be determined by solution of the inverse problem. An alternative application of the method is the balancing of flexible, high-speed rotating shafts, running in plain bearings.In this project, the focus is on the identification of linear and non-linear mechanical rotor-systems, and the development of damage detection tools. |
