8.1.3 Bearing performance
The first step towards understanding the performance of a bearing under dynamic conditions is to know the condition of the bearing oil film in vibration; as described by Smith , this is different from the steady state hydrodynamically-generated oil film noted earlier.
When a sufficiently large journal vibration is superimposed on the steady state running condition, changes in the oil film thickness and circumferential movement of the pressure bearing oil film take place, accompanying the movement of the journal. The way in which the oil film is redistributed depends on the character of the vibration. This is one of the sources of complication of the dynamics of turbine bearings. Figure 1.122 illustrates the steady running state film and other film regimes which may apply to the pressure bearing oil film with the journal vibrating. Figure 1.122 (a) illustrates the steady state film, with the film starting at the position of maximum clearance, (b) an oscillating film, (c) rotating trailing film, (d) rotating leading film and (e) rotating transition film.
The oscillating-film condition, (b), is obtained when a small amplitude of vibration is superimposed on the steady-running condition; this vibration is sustained by a fluctuating force superimposed on the steady load. The motion of the journal axis results in both squeeze and angular swing of the film. This oscillating-film condition is of frequent occurrence in bearings in service. It is closely related to the steady-running condition.
In severe vibration, the journal motion may become a whirl of such large amplitude that the position of minimum film thickness travels round the bearing with the whirl. The pressure-bearing film then travels round with the journal and the operating condition is no longer closely related to the steady running condition.
In a journal whirl of sustained magnitude at constant eccentricity, the idealised bearing is in a condition of steady running relative to axes rotating with the whirl. There is an effective speed of rotation, relative to these axes, of (N-2F), where N is the speed of rotation of the journal and F is the frequency of whirl measured in the same direction. The position of formation of the pressure-bearing film depends on the direction of the mean peripheral speed of bearing and journal relative to the rotating axes, that is on whether N is greater or smaller than 2F.
When the frequency of whirl is less than half the journal speed, the pressure-bearing film is built up on the trailing side of the minimum film thickness, as in Fig 1.122 (c). When the frequency of whirl is greater than half the journal speed, the pressure-bearing film is built up on the leading side of the minimum film thickness, as in Fig 1.122 (d). The diagrams show, for each case, the direction of the reaction applied by the journal to the bearing.
When a rotating load is superimposed, on a much smaller steady load, there is a variation of eccentricity during each whirl rotation, but the film regimes obtained resemble those given by whirl of constant eccentricity.
With rotating trailing film the reaction of the bearing on the journal (opposite to the reaction just mentioned) has a component normal to the attitude radius in the direction of the whirl. This component tends to increase the whirl. An explanation is thus offered of why a rotating trailing film is associated with tendency to instability. As reaction force increases and minimum film thickness decreases, the direction of the reaction swings closer to the attitude radius, so that if there is damping elsewhere in the system, a steady condition of whirl may be attained.
With rotating leading film the corresponding component tends to damp the whirl. A bearing in this regime does not excite instability.
The transition between these last two conditions is of importance, because it may exhibit a sharply resonant tendency to maintain rotor whirl when there is a natural frequency close to half running speed. Figure 1.122 (d) gives a diagrammatic representation of this film condition; the angular extent of the pressure-bearing film is not known.
Prediction of rotor dynamic behaviour depends on an adequate understanding of the bearing performance, for which full-size bearing test rigs are commonly used. Dynamic behaviour can also be influenced by the stiffness of the support structure and the effectiveness of the connection between bearing pedestals and the turbine foundations. In this latter respect, bolting to steel foundations is more predictable than bolting to soleplates grouted into mass concrete foundation blocks.