Page 260 - Compression Machinery for Oil and Gas

Reciprocating Compressors Chapter 5 245

possible to tune the VFD in the field to avoid the generation of certain alternat-
ing torques (see Ref. [18], appendix c6 for further information).
As a last resort, it may be necessary to avoid operation at some conditions
(often referred to as “lock-out” speeds). Field data can be very helpful for iden-
tifying the exact speed ranges and conditions to avoid. In some lightly damped
systems, it may only be necessary to change the operating speed slightly to
avoid the resonance causing the high shaft stress levels. This approach can
potentially allow for larger allowable speed ranges (smaller “lock-out” ranges),
as less uncertainty would be involved than if the recommended speeds were
based on a torsional analysis alone.

Testing Methods for Torsional Systems
Field testing of torsional systems can involve significant challenges. Typical
package instrumentation is focused on the monitoring of lateral vibration
probes, and most often does not include provisions for monitoring torsional
oscillation. An exception to this general statement is when a torsional mode
involves significant participation in gearbox shafting. Due to the physical nature
of the gear mesh, a mechanism is provided for torsional vibration to couple with
lateral vibration. As a result, torsional modes can sometimes be monitored in the
lateral probe signals from a gearbox. This section provides a brief summary of
typical testing methods and instrumentation used for documenting torsional
behavior.
Fig. 5.55 illustrates typical field data that indicates the excitation of a tor-
sional critical speed. The plot clearly shows resonant behavior as fifth- and
sixth-order excitation energy from a reciprocating compressor transitions
through the appropriate frequency. A waterfall plot can be very useful for esti-
mating separation margins between excitation of critical speeds by various
orders and running speeds, and also for estimating the damping of the modes
in question. In this particular case, a large third-order component is also present,
which is primarily a function of phasing within the reciprocating compressor
involved and not related to a resonance.
Historically, one reason that torsional vibration measurements were not
prevalent involved the relative lack of durability or difficult installation require-
ments of typically available instrumentation. The subsequent paragraphs pro-
vide some further information about the most common types of torsional
instrumentation and effective torsional data capture techniques.
One predominant type of instrumentation for gathering torsional data, used
widely in the industry, is a strain gage telemetry system, as shown in Fig. 5.56.
The system consists of the strain gages on the shafting, a rotating collar attached
to the coupling hub or spacer, and a stationary antenna/receiver. This system is
intended for short-term use only, and each collar has an operating speed limi-
tation due to centrifugal force. The data records strain levels, which can be
directly related to shaft stress, and evaluated for acceptability.