Page 475 - Practical Machinery Management for Process Plants Major Process Equipment Maintenance and Repair

452 Major Process Equipment Maintenance and Repair

What is meant by condition monitoring? Quite simply it is the observa-
tion and analysis of externally observable parameters to estimate the in-
ternal condition of the machine. As the word “estimate” implies, this is
not an exact science, but one that requires a great deal of skill and experi-
ence. The technology of condition monitoring is advancing rapidly as
lower cost computers make sensing, trending, and analyzing turbine data
easier and more accessible. The parameters most closely monitored are
performance and vibration.
Monitoring may be as simple as the analysis of a few key variables
which are manually recorded on log sheets. It may also be as extensive as
the computerized collection and analysis of 150 variables for a single gas
turbine train. Process plants which place a high value on outage time and
operate unspared single train units can easily justifj the monitoring of every
conceivable variable around a gas turbine on the basis of proper identifica-
tion of the cause of nips alone (Figure 9-5).
Vibration monitoring of gas turbines presents a few difficulties not
normally encountered on other types of twbomachinery. High tempera-
tures in and around the engine require that special precautions be taken to
ensure probe wires remain cool. On industrial turbines with noncontact
type probes many times this means running the probe leads in the bearing
oil drain piping, since this path is cooled by the flow of oil. Monitoring
with seismic probes is another option employed by some industrial users
and is the only practical method for aircraft derivative units. On large
industrials establishing good locations for these probes in close proximity
to the bearing is sometimes difficult. Many users have chosen to retrofit
noncontact probes with good success when unavailable through the man-
ufacturer.
Performance monitoring is carried out in basically the same fashion on
each engine, although the analysis of the data taken is quite different de-
pending on the type of turbine and control scheme employed. Basically
one tries to determine the engine’s ability to produce power for a given
amount of fuel. If carried out properly this analysis should not only hdi-
cate the deterioration of thermodynamic performance, but give an indica-
tion of the mechanical health of the engine. The simplest form of perfor-
mance monitoring involves the trending of key variables (e.g., pressure
ratio, compressor speed, turbine speed, fuel flow) associated with the en-
gine and comparing them with a baseline which has been developed. This
method is cQmmonly used on fixed geometry engines, both industrial and
derivative. kiable geometry engines, however, do not lend themselves
easily to this type of analysis making trending extremely difficult in com-
parison to the fixed geometry engines. Thermodynamic gas path analysis
is really the only good way to establish the performance of a variables
geometry turbine. This involves an iterative process of matching the per-