a new design wind speed for a wind turbine generator (wtg) considering typhoon loads garciano,...
TRANSCRIPT
A New Design Wind Speed for a Wind Turbine Generator (WTG)
considering Typhoon Loads
Garciano, Lessandro Estelito O.Graduate Student*
Koike, TakeshiProfessor*
* Department of Civil Engineering Musashi Institute of Technology
Introduction
Proposed 100 MW wind farm
Proposed 30 MW wind farm
Proposed 40 MW wind farm
South East Asia’s first 25 MW wind farm
Proposed 40 MW wind farm
Strong typhoons in the Philippines
1970 – 76 m/s (JOAN)1995 – 72 m/s (ANGELA)
1995 – 72 m/s (IRMA)1985 – 67 m/s (DOT)
1990 – 67 m/s (AMY)
1991 – 57 m/s (RUTH)
WTG failures in Okinawa Japan due to super typhoon Maemi
Tower buckling failure Blade failure
Footing-tower connection failure
A proposal to mitigate WTG buckling failure due to typhoons
Reliability-based analysis will be used to assess the probability of failure of a WTG tower
)ˆ(vSf
)(vSf
The load S will be based on distributions from a non-typhoon and typhoon prone areas
)ˆ(vSf)(vSf
The resistance is derived as a function of wind speed)(vRf
)(vRf
A proposal to mitigate WTG buckling failure due to typhoons
)ˆ(vSf
)(vRf
Using the relationship between load and resistance, we have
)ˆ()(1
vSvRPPF
)()(2
vSvRPPF
13
)()( FnewF PvSvRPP
221)( )()()( vRvSnewvR vS
newvRf )(
)(vSf
)(vRf
We introduce so that newvRf )(
Proposed mitigation of WTG failure due to typhoons
0)( 504 eF VvRPP
4
0)( 505 FnewenewF PVvRPP
neweVnewvR
neweVnewvRFP
5022
)(
50)(
4
)( 250
2)(4)(50 newVenewvRnewvRnewVe
11.050 )/(4.1)( hubnewrefnewe zzVzV
Extreme Wind Load Models(a) from typhoon-prone area
The Generalized Extreme Value (GEV) distribution is used to model annual extreme wind speeds
Extrapolate simulated samples from the GEV distribution to WTG hub height using logarithmic law
/1
1exp),,;(x
vG
)/ln(
)/ln(
or
or zz
zzvv
Extreme Wind Load Models(b) from non typhoon-prone area
2,102
1hubw UP
))/)ˆ(exp(exp(),;ˆ(0
vvS
The Gumbel distribution is used to model annual extreme wind speeds
U10 is simulated using the mean wind pressure equation below
nvv ˆˆ ,...,1The annual wind speed maxima are taken from U10 which are blocked in years
Buckling resistance of WTG tower
])2([592.016
33 tDDE
f
t
DfR
yyo
ycrssyssE
ssyyn fXXtDD
EX
Xf
t
DR ,
33
,
, ])2()[592.01(6
1
hIckAvCS amppsTo5.02 )21(5.0
Strength of tubular members from ISO recommendation & Kato et al.
Introducing uncertainties in the model (Sorensen et al)
Moment effect at base of WTG (Sorensen et al)
Buckling resistance of WTG tower
strstaerodynamppTn XXXhXXIckAvCS exp5.02 )21(5.0
5.0
exp5.0)21(
2)(
strstaerodynamppT
n
XXXhXXIckAC
RvR
Introducing uncertainties in the model
Resistance in terms of wind speed
Numerical Simulation of
)(vGfGEV modeling of
)(vSf Distribution of
)(vGf
)(vSf
Numerical Simulation of
)( 10USf
One-year distribution of U10
40-years of simulated annual maxima
)ˆ(vSfDistribution of
)ˆ(vSf
Numerical Simulation of
Variable Distribution type
Expected value
c.o.v.
D (m) 3.0 to 4.0
t (mm) 50 and 75
Fy (MPa) LN (lognormal)
5.88 0.05
E (MPa) 2.1e5 0.05
Xy,ss LN 1 0.02
XE,ss LN 1 0.02
A (m2) 2123
kp 3.3
Variable Distribution type
Expected value
c.o.v.
camp 1.35
Xdyn LN 1 0.05
h(m) 60
Xdyn LN 1 0.10
Xexp LN 1 0.20
Xst LN 1 0.10
Xstr LN 1 0.03
5.0
exp5.0)21(
2)(
strstaerodynamppT
n
XXXhXXIckAC
RvR
)(vRf
Numerical Simulation of
D t D/t R(V) R(V)
3.0 50 60 111 44
75 40 138 55
3.5 50 70 129 51
75 47 160 64
4.0 50 80 147 73
75 73 183 58
Simulation results for D = 3.5 and t = 75 mm
Results of buckling resistance analysis
)(vRf
Results for buckling failure analysis
D t PF1 1 PF2 2
3.0 50 0.0317 1.8566 0.0832 1.3841
75 0.0085 2.3876 0.0357 1.8026
3.5 50 0.0123 2.2472 0.0459 1.6863
75 0.0028 2.7744 0.0174 2.1110
4.0 50 0.0048 2.5873 0.0028 2.7744
75 0.0008 3.1426 0.0003 3.4615
)ˆ()(1
vSvRPPF )()(2
vSvRPPF
New buckling resistance results
D t
Initial estimate of R(V)new
Final estimate of R
(V)new
R(V) new PF3 R(V) new PF3
3.0 50 140.98 0.0144 128.83 0.0317
75 190.57 0.0010 161.51 0.0085
3.5 50 174.43 0.0023 151.44 0.0123
75 236.55 0.0001 189.03 0.0028
221)( )()()( vRvSnewvR vS
New buckling resistance results
R(V) new R(V) new PF4 4 Vrefnew Ve50new
128.83 44 0.1530 1.024 62 87
161.51 55 0.0588 1.565 65 91
151.44 51 0.0785 1.415 64 90
189.03 64 0.0255 1.952 67 94
0)( 504 eF VvRPP
)( 250
2)(4)(50 newVenewvRnewvRnewVe
11.050 )/(4.1)( hubnewrefnewe zzVzV
Results from the other 49 wind stations
Station
ID Station Name
GEV Parameters
135 Basco, Batanes 42.45 9.93 0.18
232 Aparri, Cagayan 32.04 12.34 0.22
446 Virac Synop, Catanduanes 36.33 16.37 -0.04
531 San Jose, Occidental Mindoro
30.49 7.68 -0.48
The PF1 from these stations increased when typhoons were considered
Results from the other 49 wind stations (D = 3.0 & t = 50 mm)
Station ID R (v’) S(v’) PF2 R(V) new Vrefnew Ve50new
232 42.49 23.23 0.0465 121.09 57 79
135 50.30 17.33 0.0514 120.53 56 79
531 32.31 7.14 0.0391 91.98 37 52
446 44.68 19.26 0.0530 120.77 56 79
New design wind speed map (D = 3.0 & t = 50 mm)
Kriging of ArcGIS was used to interpolate results of Ve50new to other areas
The areas in red indicates an increase in probability of buckling failure if typhoon loads are considered
Using the proposed mitigation scheme, these areas will have Ve50new > 70 m/s
The blue and yellow areas indicate that the PF2 < PF1 even when typhoon loads were considered
Concluding Remarks
Strong typhoons occur in the Asian Region Based extreme wind data and the recent experience of a
wind farm in typhoon prone areas, survival wind speed (Ve50) may be exceeded during the economic life of a wind farm
In view of this, the authors proposed a mitigation scheme by introducing a new buckling resistance (R(v)new) in order to maintain the same probability of failure
Based on this new resistance, a 50-year design extreme wind speed (Ve50new) and a new reference wind speed (Vrefnew) can be derived
Concluding Remarks
The authors also analyzed the probability of buckling failure of a WTG tower using typhoon data from other wind stations (using D = 3.0 m & t = 50 mm)
The results showed that only the PF1 from 4 stations increased
Based on the (R(v)new) of each station, a Ve50new and Vrefnew were also derived
Using ArcGIS kriging method, a new 50-year design extreme wind speed map was developed
This map will be useful for future owners of commercial size or small scale wind farm
Thank you