The present paper discusses the various movement patterns during rotor stator contact. Both rotor and stator are assumed to be flexible damped single degree of freedom systems. The contact is described by a flexible viscoelastic model. Dry friction between rotor and stator is taken into account. Despite strong non-linearity due to contact, rotor unbalance causes purely synchronous motions. However, in some circumstances, the synchronous motion may become unstable and the rotor motion turns into a non-synchronous state, which can be very destructive. Non-synchronous motions include backward whirl, sub- and super-harmonic vibration and chaotic motion. The influence of various system parameters on the different types of motion is investigated by numerical simulation. The transients between synchronous to non-synchronous motions are exemplarily demonstrated by run-up and run-down processes. It is shown that different motion types may co-exist. Even in speed regions where the synchronous whirl is stable, non-synchronous motions with rotor stator contact are possible.

No abstract

Normally rotor unbalance causes synchronous forward whirl of rotor‐stator systems, even if rub occurs due to rotorstatorcontact. This synchronous forward whirl has to be stable in order to avoid destructive self‐excited dry friction backward whirl, chaotic motions or sub‐ and superharmonic vibrations. However, friction between rotor and stator can cause the synchronous forward whirl to become unstable within certain rotor speed ranges. In the present paper the stability of the synchronous forward whirl caused by unbalance is investigated for rotor motions under contact with the stator. To analyse the stability of synchronous forward whirl the equations of motion are linearised around the stationary synchronous motion. The characteristic polynomial of the perturbations is calculated and the stability is checked by the Hurwitz criterion.

The natural frequencies of blades depend on the rotational speed of the rotor train as the stiffness changes with centrifugal loading. In the case of low pressure turbines with shrunk-on-disc design the coupled rotor-blade torsional natural frequencies can also show this property. For proper analysis of the speed dependency, a complete rotor-blade model which takes the elasticity of the blades into account is required. In this paper the torsional natural frequencies calculated with a complete rotor-blade model are compared with those calculated with a model in which blade elasticity is not included. The analysis clearly demonstrates that calculations without blade elasticity lead to different natural frequencies. By modeling the complete rotor and taking blade elasticity into account, it is demonstrated that the torsional natural frequencies of a complete rotor-blade model can also become speed dependent. As a consequence, a distinction between the natural frequencies at nominal speed and natural frequency at critical speeds becomes necessary. In the following, measured torsional natural frequencies at different rotating speeds of an individual low pressure rotor are presented. A comparison of the measured speed dependency of the torsional natural frequency with calculation results thereby taking the blade elasticity into account is conducted. The analysis shows that the measured speed dependency can be predicted with a high level of accuracy and can become important for modes which are dominated by the blades of the last stages. As a consequence of this analysis, a clear distinction between natural frequency at nominal and at critical speed has to be made for certain rotor and blade designs. It is shown that the use of the Campbell diagram is highly beneficial for designing rotor trains with large blades with regard to their torsional vibration behavior.

No abstract