renewable
energy
Conditionmonitoringsystem
requirementsforwindturbines
10
uptime
m
e
gazine · 02/08
Wind turbine generators
(WTG) are expected to deliver
energy with optimal availability
and minimal Operation & Main-
tenance cost just like any oth-
er energy generating turbine.
Condition monitoring systems
have proven invalu-
able in reducing
maintenance
costs and
maximizing
uptime in con-
ventional pow-
er stations over the
years, but can the same ma-
chine condition monitoring sys-
tem be used on a WTG?
W
ind energy is the fastest grow-
ing renewable energy technol-
ogy in the world. Production is cur-
rently increasing around 30% each
year, and there is no sign of it slow-
ing down. Wind power generation
has increased five-fold between 2000
and 2007, and it is expected that
world wind energy generating capac-
ity exceeds 100 GW this year. Be-
cause WTG play such an increasingly
important role in the energy market,
there is increasingly more pressure to
be price competitive with other forms
of energy. That means wind turbine
availability has to be maximized and
Operation & Maintenance costs mini-
mized, as is done in the other types
of power stations.
Plant-wide condition monitoring sys-
tems have been successfully reduc-
ing downtime and maintenance costs
in nuclear and thermal power stations
over the last 15 years. The same sys-
tems are flexible enough to be used
also in the petrochemical and proc-
ess industries. We have been doing
this for the last 15 years. So it is only
natural to use the same monitoring
experience and system technology for
the emerging wind turbine generator
market, isn't it? The answer is no!
Condition Monitoring System
Configuration
A modern plantwide condition moni-
toring system, like COMPASS, will
use approximately 400 input/out-
put channels for monitoring a 1000
MW power plant with four steam tur-
bines. The input/output channels will
be hardwired to 90 monitoring mod-
ules arranged in 8 cabinets with a to-
tal of 16 racks. The monitoring mod-
ules perform signal conditioning on
the input signals for generating sca-
lar measurements for both safety
and condition monitoirng purposes.
A monitoring server is remotely con-
nected to the racks over a network
for storing and displaying data in
plots.
A similar plant wide condition moni-
toring monitoring system for a WTG
windfarm would be both impractical
and uneconomical. First of all, there
is no instrumentation hall in a wind
turbine park, so all monitoring hard-
ware has to be inside the WTGs.
Each wind turbine requires between
10-12 channels, which corresponds
to one - possibly two - monitoring
modules in a rack-mounted system.
The rack is almost empty! This can
be very expensive if the wind tur-
bine park has 30 or more WTGs in it.
Not only are the half-empty racks not
very cost-effective, but they use up
valuable space within what is consid-
ered to be already cramped quarters
inside the nacelles. It would be com-
pletely impractical to connect cables
from several WTGs to a single rack
to fill it up. The botton line is that a
WTG requires a dedicated hardware
system to be both cost and space ef-
fective, and fit for the purpose.
Machine Construction
The relatively simple machine con-
struction of a steam turbine train
rests on a rigid foundation and turns
at high speeds. The turbo-generator
sets have proven designs that have
been operating for many years. The
potential failure modes of these ma-
chines are well documented, and rel-
atively easy to detect and diagnose.
A WTG drive train on the other hand
rests on a non-rigid foundation and
employs a complex planetary gear
box that turns at slow speeds. The
potential failure modes here are not
so easy to detect and diagnose. They
require a more comprehensive moni-
toring strategy to identify weak sig-
nals where there is heavy dynamic
variation that overlaps with signals
from other vibration sources. Many
of the WTG components are new de-
signs with very little maintenance his-
tory, so some components potentially
fail before their life expectancy has
been reached. This requires a
t