"Rising Vector" by contact seal 

A company in the chemical sector experienced start-up problems with an integrated gear compressor. The machine had recently been overhauled. The installed protection system (Bently Nevada 3300) registered high vibrations in phases 3 and 4. These resulted in an outage.

Obervation during start-up 1 

The machine was started and what immediately catches the eye is that the vibrations in phase 3 have a fluctuating pattern. 

Trend Direct values phase 3:

Trend 1xRPM values phase 3 X-probe:

Trend 1xRPM values phase 3 Y-probe:

We don’t have phase information, due to the lack of a Keyphasor. The 1xRPM was determined by using a Sim Keyphasor on the known RPM of phases 3 and 4. However, we do suspect that this is the image of a Rising Vector. By this we mean that there is a moment of contact somewhere, probably around Seal Phase 3. This issue will probably be analogous to that of phase 4, namely that wear is impeding the radial mobility of the carbon rings.

Trend Direct values phase 4 

After a slight initial variation, phase 4 shows stable vibration values. In the following plot these occur around 40µ, but eventually they rise towards 50µ and then they stabilise ? too high for this number of RPM. These initial vibrations are more pronounced than before the overhaul. The runout values are acceptable (around 5µ).

Thermal progress: 

The gap tensions on phases 3 and 4 progress over time. This is probably the result of thermal expansions. 

• V17231X: upon start-up 6.0V, stabilised to 7.0V --> Difference 1V --> 128µ 
• V17231Y: upon start-up 6.7V, stabilised to 8.4V --> Difference 1.7V --> 217µ 
• V17241X: upon start-up 6.0V, stabilised to 8.0V --> Difference 2V --> 256µ 
• V17241Y: upon start-up 6.6V, stabilised to 7.1V --> Difference 0.5V -->64µ 

If these were due to deformation of the housing following expansions, they would be extremely high values. The gap tensions become more and more negative --> larger distance between probes and axle. Possibly the housing is being pushed upwards. In principle, the cooler (at the other side of the pipe) would not be fixed any more (can slide until the restriction), but I think that its weight is still too high, causing it to remain where it is, and thus causing the housing to absorb the bulk. A quality Carbon Seal should be flexible enough to deal with this change.

Observations during start-up 2 

The machine was restarted. Due to a power failure the initial start-up data have not been recorded. On 15 January we notice a change in the vibration image. Phase 4 shows a temporary decrease, while the vibrations of phase 3 increase dramatically. 

Trend Phase 4 direct values:

Trend Phase 3 direct values:

Trend Phase 3 1xRPM values: Before the increase we see the alternating behaviour of the vibrations (= moment of contact). After the high values this disappears and the 1xRPM is stable --> no contact.

Trend Gap phase 3: 

Leading up to the increase in vibrations, we notice a change in Gap tension. In other words, the position of the axle changes. We suspect that in the end the ‘extra support’ of a fixed seal is lost (maybe radially coming loose after all), which results in the axle taking up a position where it is stable itself. The way to the stable position could involve high vibrations.

Orbit phase 3 at the time of the high vibrations 

Fairly round shape --> no preload

Orbit phase 3 at the time of the low vibrations: 

Very low values

Advice after measurement 

We advise the following course of action at the next opportunity: 

• Inspection of Seal phase 3. 
• Replace Concept seal. This way, even after hours of operation, the rings can’t wear out a groove, causing the seal to be radially stuck.
• Provide the machine with compensators, which can set off thermal expansion.
• Provide Keyphasors, allowing for more options when analysing. 
• The vibrations on phase 4 are around 50µ, which is too high for this number of RPM. We advise balancing the rotor at the next opportunity.

Case conclusion 

By using our knowledge and our measuring equipment relating to the monitoring of plain bearings, we have clearly isolated the problem. In consultation with the mechanical services of the company we have drawn up a clear action plan.