Upload
others
View
5
Download
0
Embed Size (px)
Citation preview
Performance of two nacelle-
mounted ice detectors: A case study
Katja Hynynen Irene Romero
Svetlana Afanasyeva Jordi Armet
Olli Pyrhönen Alstom Renovables España S.L
Lappeenranta University of Technology
WinterWind 2016, Åre, 10.2.2016
Background
− Amount of wind power in cold climate and icyconditions is increasing
− Ice on the blades causes
− Safety risk
− Power losses
− Ice on the turbine blades is sensed by
− nacelle mounted detectors
+ Easy to install on existing turbine
+ Low investment costs
- Sense ice on nacelle, not on blades.
− blade mounted sensors
+ Sense ice on blades themselves
- More reliable results of the performancewould be required
Picture: http://www.tuulivoimayhdistys.fi/hankelista
Scope
− The scope of the study is to investigate
− What type of ice and how often occurs in Muukko wind farm?
− In which weather conditions does ice occur?
− Performance of two nacelle mounted ice detectors
− The study is part of a research
project ’Wind power in cold climate
and complex terrain’ carrying out by
− Lappeenranta university of
technology
− Alstom Renevables España S.L
− TuuliMuukko and
− TuuliSaimaa
Scope
WP Sub. WP Title
1.WP1. WIND FARM DATA
ACQUISITION1 Wind farm data acquisition
2.
WP2. POWER &
OPERATIONAL
PERFORMANCE ANALYSIS
AND OPTIMIZATION
2.1 Power performance in cold climates
2.2 Operational performance in cold climates
2.3Evaluation of WTG loads and dynamics due to ice
accretion
2.4 Analysis turbine control for cold climate
3.WP3. ICING SENSORS AND
ICE DETECTION
3.1 State of the art
3.2 Performance analysis of ice detection sensors
Sensors and measurements (1/2)
− Ice detector, Sensor A
− Freezing rain sensor
− Ice detection is based on change of resonantfrequency of vibrating probe
− Ice detector, Sensor B
− In-cloud icing sensor
− Ice detection is based on ultrasonic principle.
− Both sensors have heatingswitched on at the time of alarm
− Web camera mounted on nacelle and shooting blades
Sensors and measurements (2/2)
Other measurements used in the study
− Weather data of nearby Lidar
− Weather data from FMI
− SCADA data of the turbine
− Measurement period 18.10.2014 – 15.4.2015 (6 months)
Ice types found by web camera (1/2)
− Rime ice on leading edge found in 20 days
Ice types found by web camera (2/2)
− Snow-type of ice on the blades in 47 days
− In total, some type of ice (may be several types) found in 59 days
− Camera far from the blade
Not possible to see
glaze or other thin
formation of ice
− Poor visibility during dark,
fog and heavy precipitation
− Camera with motion
detector does not take
pictures when turbine is
standing
Most probably there has
been
− ice more often than in
59 days
− also glaze
Challenges in detecting ice using web camera
01.10. 11.10. 21.10. 31.10. 10.11. 20.11. 30.11. 10.12. 20.12.-10
0
10
Tem
pera
ture
, °C
01.10. 11.10. 21.10. 31.10. 10.11. 20.11. 30.11. 10.12. 20.12.
0
0.5
1
Sensor
B
01.10. 11.10. 21.10. 31.10. 10.11. 20.11. 30.11. 10.12. 20.12.
0
0.5
1
Sensor
A
Alarms of ice detectors
− 76 alarms in total
− Overall duration of
alarms 264 hours
− Each alarm over 1 hour
was analyzed
− Weather conditions
− Ice seen by camera?
− Performance of ice
detectors
− Did ice detectors
detect when there
occurred ice?
− False alarms?
− Missing alarms?
− In-cloud icing occurred
especially on November 2014
− In total, 13 in-cloud icing
cases were found
− Rime ice accreted according
to the camera
In-cloud icing causing rime ice
12:00 16:00 20:00 00:00 04:00 08:00 12:00 16:00 20:00 00:00 04:00 08:00 12:00-2
0
2
Tem
pera
ture
, °C
Temp
Dew point
12:00 16:00 20:00 00:00 04:00 08:00 12:00 16:00 20:00 00:00 04:00 08:00 12:0096
98
100
Hum
idit
y,
%
FMI
Lidar
12:00 16:00 20:00 00:00 04:00 08:00 12:00 16:00 20:00 00:00 04:00 08:00 12:000
20
40
60
80
waw
a
12:00 16:00 20:00 00:00 04:00 08:00 12:00 16:00 20:00 00:00 04:00 08:00 12:002
4
6
Win
d s
pe
ed
12:00 16:00 20:00 00:00 04:00 08:00 12:00 16:00 20:00 00:00 04:00 08:00 12:00
0
0.5
1
Ice P
rese
nt
Sensor A
Sensor B
Precipitation causing icing
− 24 cases where
precipitation near zero
degrees caused icing
− Several types of ice
accreted: rime, snow-type
− Different types of
precipitation; drizzle, rain,
snow, no clear evidence of
freezing rain
− In 4 cases, most probably
wet snow
05:00 07:00 09:00 11:00 13:00 15:00 17:00 19:00 21:00 23:00-2
-1
0
1
Tem
pera
ture
, °C
Temp
Dew point
05:00 07:00 09:00 11:00 13:00 15:00 17:00 19:00 21:00 23:0090
95
100
Hum
idit
y,
%
FMI
Lidar
05:00 07:00 09:00 11:00 13:00 15:00 17:00 19:00 21:00 23:000
20
40
60
80
waw
a
05:00 07:00 09:00 11:00 13:00 15:00 17:00 19:00 21:00 23:00
4
6
8W
ind
sp
eed
05:00 07:00 09:00 11:00 13:00 15:00 17:00 19:00 21:00 23:00
0
0.5
1
Ice P
rese
nt
Sensor A
Sensor B
Performance of ice detectors
− All the ice occurrences seen by web camera were detected by both ice sensors or at least one of them.
− Sensor A detected 95% (72/76) of icing cases
− Sensor B detected 97% (74/76) of icing cases
− In many cases, sensor A gave alarm before sensor B
− Reason probably higher threshold value of sensor A
− However, during in-cloud icing, sensor B often gave alarm before sensor A
− Reason most probably that in-cloud icing sensor B is better capableof sensing in-cloud icing
− Additionally, sensor B was faster or only detecting sensor in 4 possible wetsnow cases.
− Both sensors also gave some ”false alarms” and no ice was found withcamera.
− May have been accreted glaze or other thin layer of ice not visible bycamera
Conclusions
− Both nacelle mounted sensors A and B performed very good.
− All the proven ice occurrences were detected by the sensors.
− No power loss caused by icing was observed by the first
alarm.
− In future, in order to ensure proper threshold value for the ice
alarm, the performance of nacelle mounted sensors should be
analyzed together with blade mounted sensor and power loss
of turbine
Thanks to the project partners:
TuuliMuukko