Elec ical Powe and Ene gy Sys ems 153 (2023) 109376
A ailable online 20 July 2023
0142-0615/© 2023 The Au ho (s). Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by-
nc-nd/4.0/).
Con en s lis s a ailable a ScienceDi ec
In e na ional Jou nal o Elec ical Powe and Ene gy Sys ems
jou nal homepage: www.else ie .com/loca e/ijepes
In luence o yaw con ol on licke p oduced by wind u bines
Koldo Redondo a,∗, Jose Julio Gu ie eza, Izaskun Azca a e b, Mikel Le u iondo a,
Jose An onio U igüen b, So ía Ruiz de Gauna a, Pu i icación Saiz a
aCommunica ions Enginee ing Depa men , Uni e si y o he Basque Coun y (UPV/EHU), Plaza Ingenie o To es Que edo 1, Bilbao, Spain
bApplied Ma hema ics Depa men , Uni e si y o he Basque Coun y (UPV/EHU), Plaza Ingenie o To es Que edo 1, Bilbao, Spain
ARTICLE INFO
Keywo ds:
Wind u bine
Yaw con ol
Powe quali y
Vol age luc ua ions
Flicke
ABSTRACT
IEC 61400-21-1 desc ibes he p ocedu e o measu ing he licke p oduced by a wind u bine. This is a complex
p ocedu e, as i in ol es he p ocessing o ol age and cu en ime se ies eco ded in a g id connec ed wind
u bine o i s en i e ope a ing ange. The s anda d assumes ha he licke p oduced by a wind u bine is
ela ed o wind a ia ions and swi ching ope a ions. The ol age and cu en signals eco ded in a 2 MW Type
III wind u bine loca ed a a 32 MW wind powe plan in Spain we e used o he s udy. The wo k shows
ha mos o he licke p oduced by he wind u bine is due o he powe consump ion o he mo o used by
he yaw con ol sys em o o ien he nacelle. These ope a ions gene a e ol age changes whose ampli ude is
independen o he gene a ed powe . As he gene a ed powe inc eases, so does he numbe o yaw con ol
ope a ions and hus he licke emi ed. A simple analy ical me hod is p oposed o es ima e he licke p oduced
by yaw con ol ope a ions. The esul s con i m ha inco po a ion o his s a egy in o he s anda d could
conside ably simpli y he cu en licke measu emen p ocedu e.
1. In oduc ion
The ha es ing o wind ene gy by wind u bines (WTs) is con-
di ioned, among o he aspec s, by he yaw e o . This is de ined as
he misalignmen in ho izon al plane be ween he wind di ec ion and
he WT o o axis. The yaw e o causes di e en undesi ed e ec s:
he powe cap u ed dec eases as he e o inc eases [1,2]; i causes
c osswind, which gene a es ib a ions in he WT s uc u e due o
asymme y loads [3]; and i modi ies he di ec ion o he wake and,
he e o e, he cha ac e is ics o he wind ha eaches he o he WTs o
he wind powe plan (WPP) [4,5].
Yaw con ol s a egies a e used o sol e he e ec s ou lined abo e.
Some o hem maximize powe gene a ion [2,6], using di ec ly he
wind di ec ion ob ained by con en ional senso s o o ien he WT as
a unc ion o he calcula ed yaw e o [7–11]. These me hods esul
in equen yaw ope a ions and WT pe o mance wi h signi ican yaw
e o s [12]. O he me hods p edic he wind di ec ion, and conse-
quen ly he op imal nacelle o ien a ion, based on Lida senso s [13,14]
o ime se ies models [6,15]. The e a e s a egies ha ope a e wi hou
senso s, based on acking o he maximum powe poin o he op imal
o o speed [16–19]. O he s a egies aim o minimize he a igue load
expe ienced by he WT [20,21]. Finally, some s a egies ocus on he
wake e ec so ha yaw con ol is applied o he whole WPP, o imp o e
o e all powe gene a ion [22,23], as in he case wi h ex emum seeking
con ol [24], as well as o minimize he o e all a igue loads [23].
∗Co esponding au ho .
E-mail add ess: [email p o ec ed] (K. Redondo).
Yaw con ol ac ua o s use one o mo e elec ically d i en gea s o
o ien he nacelle in he se di ec ion. Each nacelle o ien a ion esul in
a sudden powe consump ion a mo o s a up. These powe changes
a e injec ed in o he g id gene a ing ol age luc ua ions, esul ing in
licke emissions om he WPP. Flicke is de ined as he imp ession o
ins abili y in isual sensa ion due o luc ua ions in he b igh ness o
ligh sou ces caused by luc ua ions in hei supply ol age [25]. IEC
61000-4-15 es ablishes he speci ica ions o he implemen a ion o he
licke me e [26]. This p ocedu e de ines a pa ame e ha e alua es
he annoyance p oduced by ligh luc ua ions in a 10-min in e al,
called licke se e i y, 𝑃𝑠𝑡 [27]. Fluc ua ions p oducing alues o 𝑃𝑠𝑡 >1
a e conside ed annoying.
The in e na ional s anda d IEC 61400-21-1 [28] es ablishes he
p ocedu es o cha ac e izing he licke o WTs connec ed o he
g id. The s anda d de ines wo scena ios: con inuous ope a ion, when
luc ua ions a e due o he in e ac ion be ween WT and changes in wind
cha ac e is ics [29–33]; and swi ching ope a ions, when apid ol age
changes a e gene a ed [34,35]. The implemen a ion o he s anda dized
p ocedu e equi es he compliance wi h an exhaus i e es p o ocol
ha gua an ees high measu emen accu acy and he collec ion o la ge
amoun o elec ical da a o e weeks o cha ac e ize he ull ope a ing
powe ange [36–38].
h ps://doi.o g/10.1016/j.ijepes.2023.109376
Recei ed 10 Ma ch 2023; Recei ed in e ised o m 26 May 2023; Accep ed 11 July 2023
In e na ional Jou nal o Elec ical Powe and Ene gy Sys ems 153 (2023) 109376
2
K. Redondo e al.
Fig. 1. Flicke coe icien measu emen diag am acco ding o he IEC 61400-21-1 [28].
The objec i e o his wo k was, on he one hand, o cha ac e ize he
licke p oduced by he mo o s a up o he yaw con ol sys em. On
he o he hand, since he IEC 61400-21-1 s anda d does no cu en ly
conside i , he wo k p oposes a simple p ocedu e o es ima e he licke
due o yaw con ol, wi hou he need o connec he WT o he g id. Fo
he wo k, he h ee phase- o-neu al ol ages and he h ee line cu en s
we e eco ded du ing 35 days a a WPP in Spain.
The manusc ip is s uc u ed as ollows: Sec ion 2desc ibes he
licke measu emen p ocedu es acco ding o IEC 61400-21-1 ha will
be used o ob ain esul s, as well as he cha ac e is ics o he WT
s udied and he da abase ob ained. Sec ion 3p esen s he esul s o
he wo k: i s , he e ec ha he powe consump ion due o he
mo o s a up has on ol age and on ins an aneous licke pe cep ion;
second, he in luence o he yaw ope a ions on he licke se e i y
bo h in con inuous and in swi ching ope a ions o he WT; hi d, a
simply ma hema ical me hod o es ima e he licke p oduced by yaw
ope a ions based on he 𝑃𝑠𝑡 = 1 cu e. Finally, Sec ion 4summa izes
he conclusions and main con ibu ions o he wo k.
2. Ma e ials and me hods
2.1. Flicke measu emen acco ding o he IEC 61400-21-1 s anda d
Fo bo h con inuous and swi ching ope a ions, he licke coe i-
cien s, 𝑐(𝜓𝑘), a e calcula ed acco ding o Fig. 1. The ime se ies o line
cu en , 𝑖𝑚(𝑡), and phase- o-neu al ol age, 𝑢𝑚(𝑡), eco ded a he WT
e minals, a e used as inpu signals.
Block A implemen s he in e ac ion be ween he WT and he g id,
aiming a calcula ing he ol age 𝑢𝑓 𝑖𝑐 (𝑡), ha con ains hose luc ua ions
p oduced exclusi ely by he WT. The s anda d speci ies ha 𝑢𝑓 𝑖𝑐 (𝑡)mus
be ob ained o di e en g id impedance, de e mined by ou phase
angles 𝜓𝑘(30◦, 50◦, 70◦and 85◦) and he sho -ci cui appa en powe
o he g id, 𝑆𝑘,𝑓 𝑖𝑐 .
Block B implemen s he IEC llicke me e [26]. The p ocedu e e-
p oduces he esponse o he lamp-eye-b ain sys em [39–41]. The
incandescen lamp was aken as he e e ence since i was he mos
sensi i e ligh ing echnology o ol age luc ua ions. The ins an aneous
licke pe cep ion, 𝑃𝑖𝑛𝑠𝑡, is ob ained as he ou pu signal o he model.
To calcula e he annoyance p oduced by ligh luc ua ions, a s a is ical
and empo al in eg a ion o he 𝑃𝑖𝑛𝑠𝑡 alues in 10-min is pe o med,
ob aining he sho - ime licke se e i y alue 𝑃𝑠𝑡,𝑓 𝑖𝑐 .
Block C ob ains he licke coe icien 𝑐(𝜓𝑘)by no malizing each
alue o 𝑃𝑠𝑡,𝑓𝑖𝑐 acco ding o [28]:
𝑐(𝜓𝑘) = 𝑃𝑠𝑡,𝑓𝑖𝑐 ⋅𝑆𝐶𝑅 =𝑃𝑠𝑡,𝑓𝑖𝑐 ⋅
𝑆𝑘,𝑓𝑖𝑐
𝑆𝑛
(1)
whe e 𝑆𝑛is he a ed appa en powe o he WT. The s anda d sugges s
alues o he sho -ci cui powe a io, 𝑆𝐶𝑅, be ween 20 and 50.
Du ing con inuous ope a ion, licke coe icien s o 10-min ime
se ies a e classi ied in o 11 powe bins (0, 1, 2, up o 10), wi h he
midpoin o hese bins being 0%, 10%, 20%, up o 100% o he nominal
powe , 𝑃𝑛, espec i ely. Measu emen s a e collec ed con inuously wi h
a minimum o 21 ime se ies o each o he powe bins, disca ding he
Table 1
Reco ded WT cha ac e is ics.
Pa ame e Desc ip ion Value
𝑈𝑛Nominal ol age 690 V
𝐼𝑛Ra ed cu en 1500 A
𝑃𝑛Ra ed powe 2 MW
𝑃 𝐹 Powe ac o ange 0.98 CAP - 0.96 IND
𝑓0Main equency 50 Hz
𝑣𝑖𝑛 Cu -in wind speed 4 m/s
𝑣𝑟𝑎𝑡𝑒𝑑 Ra ed wind speed 13 m/s
𝑣𝑜𝑢𝑡 Cu -ou wind speed 25 m/s
Table 2
Summa y o WT measu emen s da a base.
Pa ame e Desc ip ion Value
𝑁𝑇To al ime se ies 4914
𝑁𝐶𝑂 Con inuous ope a ion 4380
𝑁𝑆𝑊 Swi ching ope a ion 294
𝑁𝑁𝐺 Non-gene a ion 240
𝑃𝑎𝑣𝑔 A e age powe 0.54(0.26–0.99) MW
ime se ies which con ain swi ching ope a ions o which co espond
wi h non gene a ion mode o he WT.
Cha ac e iza ion o ol age luc ua ions in swi ching ope a ions is
pe o med o a du a ion o 𝑇𝑝seconds, long enough o con ain he
en i e swi ching ope a ion a oiding he e ec s o he esul ing con in-
uous ope a ion. This assessmen is made by calcula ing he licke s ep
ac o , 𝑘𝑓(𝜓𝑘), acco ding o [28]:
𝑘𝑓(𝜓𝑘) = 1
130
⋅𝑐(𝜓𝑘)⋅𝑇0.31
𝑝(2)
and he ol age change ac o , 𝑘𝑢(𝜓𝑘), acco ding o [28]:
𝑘𝑢(𝜓𝑘) = √3⋅
𝑈𝑓𝑖𝑐,𝑚𝑎𝑥 −𝑈𝑓𝑖𝑐,𝑚𝑖𝑛
𝑈𝑛
⋅
𝑆𝑘,𝑓𝑖𝑐
𝑆𝑛
(3)
whe e 𝑈𝑛is he nominal ol age o he WT and 𝑈𝑓𝑖𝑐,𝑚𝑎𝑥 and 𝑈𝑓𝑖𝑐,𝑚𝑖𝑛
a e he maximum and minimum ms alues o he ic i ious ol age
𝑢𝑓 𝑖𝑐 (𝑡). Calcula ion o bo h pa ame e s mus be ca ied ou o a leas
15 ope a ions, ob aining he mean alue as he inal esul .
2.2. Da a collec ion
Measu emen s om a 2 MW ype III WT loca ed a a 32 MW WPP
in Spain we e analyzed. WT cha ac e is ics a e summa ized in Table 1.
I is a pi ch egula ed, upwind WT wi h ac i e yaw con ol, h ee-blade
o o , and high-e iciency 4-pole doubly ed gene a o wi h wound o o
and slip ings. The yaw con ol sys em consis s o ou gea s elec ically
ope a ed by a 2.2 kW 6-pole asynch onous mo o . I akes decisions
based on he in o ma ion ecei ed om he ul a-sonic anemome e s
moun ed on op o he nacelle.
The h ee phases- o-neu al ol ages and he h ee line cu en s
we e eco ded (a a sampling a e o 20 kHz) du ing 35 days, ob aining
a o al o 4914 10-min ime se ies (see Table 2). The connec ion and
shu -down e en s o he WT we e iden i ied, which allowed he ime
In e na ional Jou nal o Elec ical Powe and Ene gy Sys ems 153 (2023) 109376
3
K. Redondo e al.
Fig. 2. His og am o 10-min ime se ies du ing con inuous ope a ion (blue), as a unc ion o powe in e als. The cases o swi ching ope a ion (g een) and non gene a ion ( ed)
ha e also been g ouped sepa a ely.
se ies o be classi ied in o h ee g oups: con inuous ope a ion (CO),
swi ching (SW) and non-gene a ion (NG). Almos 89% o hem we e
con inuous ope a ions (𝑁𝐶𝑂 =4380), app oxima ely 6% con ained
swi ching ope a ions (𝑁𝑆𝑊 =294) and he emaining 5% co esponded
o non-gene a ion unc ioning mode (𝑁𝑁𝐺 =240). The a e age powe
𝑃𝑎𝑣𝑔 o each ime se ies was ob ained, wi h a median alue o 0.54 MW
and in e qua ile ange (IQR) o 0.26–0.99 MW. Fig. 2 shows he
dis ibu ion o he ime se ies ela ed o unc ioning mode o WT. In
con inuous ope a ion, powe gene a ion was close o 𝑃𝑛(bins 8–10)
du ing 9% o he ime, and i was below 50% o 𝑃𝑛du ing 75% o
he ime. The numbe o eco ded ime se ies me he equi emen s o
IEC 61400-21-1. Fig. 2 shows ha he en i e WT ope a ing ange was
eco ded, wi h a minimum o 126 ime se ies pe bin. Mo eo e , 230
connec ions o WT we e iden i ied om he swi ching ope a ion ime
se ies.
3. Resul s
All he esul s desc ibed in his sec ion we e ob ained o 𝜓𝑘= 85◦,
close o he g id impedance epo ed by he WPP ope a o , and 𝑆𝐶𝑅 =
20.
3.1. Cha ac e iza ion o yaw con ol ope a ions
Each mo o s a up ha o ien s he nacelle owa ds he se di ec ion
gene a es an ene gy consump ion ha p oduces ol age changes and,
consequen ly, licke . Fig. 3 illus a es he empo al e olu ion o he
main magni udes a ec ed by a yaw ope a ion. The mo o s a up (a
0.38 s) equi ed an ins an aneous cu en consump ion o 115 A (7%),
which educed he gene a ed powe (a change o 110 kW was obse ed,
which ep esen ed a educ ion o 5.7%). The powe consump ion o he
mo o p oduced a sha p ol age d op (𝛥𝑉 =1.6 V) which co esponded
o a ela i e ol age change ampli ude o 𝑑𝑣 =𝛥𝑉 ∕𝑉= 0.4%.
This p oduced a maximum ins an aneous licke pe cep ion alue o
𝑃𝑖𝑛𝑠𝑡,𝑚𝑎𝑥 = 0.54.
Vol age changes due o men ioned mo o s a ups appea ed egu-
la ly. Fo he o al eco ded ime se ies mo e han 53,000 s a ups we e
iden i ied acco ding o IEC 61000-4-30 s anda d [42]. Fig. 4 p esen s
he analysis o all he 10-min ime se ies included in he s udy. The
a e age ampli ude o he ol age d ops, 𝑑𝑣, did no depend on ei he
he powe o he WT unc ioning mode, wi h median (IQR) alues
o 0.379 (0.374–0.384)%. Numbe o yaw ope a ions in 10-min, 𝑁,
p esen ed a high dispe sion in all h ee WT modes. Only du ing con-
inuous ope a ion a dependence o 𝑁wi h powe was obse ed. Fi s ,
𝑁 emained cons an (10 ope a ions in 10-min) up o 1.5 MW, and
hen alues inc eased om 1.5 MW upwa ds, exceeding 30 ope a ions
in 10-min a 𝑃𝑛. The magni ude o he ol age d op was mainly due o
he cha ac e is ics o he mo o used o he o ien a ions o he nacelle.
The ela ionship be ween 𝑁and powe indica es ha he ac i a ion
a e o he ac ua o s depended on he implemen ed con ol s a egy.
The consump ion de i ed om mo o s a up had a negligible in lu-
ence on he gene a ed ene gy. Indeed, by simula ing he beha io o a
10-min ime se ies wi h an a e age powe o 𝑃𝑎𝑣𝑔 = 2 MW and 𝑁= 30,
each p oducing a powe educ ion o 110 kW and a ol age d op o
𝑑𝑣 = 0.37%, we ob ained an ene gy loss o 𝐿𝑦= 92 Wh, ep esen ing
0.03% o he gene a ed ene gy. Howe e , om he licke poin o iew,
we ob ained a licke se e i y alue o 𝑃𝑠𝑡 = 0.2, which, al hough s ill
a om he h eshold (𝑃𝑠𝑡 = 1), ep esen ed a ele an pe cen age o
i .
3.2. Flicke measu emen du ing con inuous ope a ion
Panels (a), (b) and (c) o Fig. 5 show h ee cases o 10-min ime
se ies, classi ied in o di e en powe bins. The beha io explained
in Sec ion 3.1 was ep oduced in a e y simila way o each yaw
ope a ion. Ab up inc eases in 𝑃𝑖𝑛𝑠𝑡 had simila alues o he h ee
ime se ies (be ween 0.5 and 0.6), and well abo e he negligible alues
p oduced by he changing wind cha ac e is ics.
Resul s o hose cases a e summa ized in Table 3. The numbe
o ope a ions g ew as he gene a ed powe inc eased. Ene gy losses
we e e y low in all cases. The calcula ion o licke coe icien s due
o yaw ope a ions, 𝑐𝑦, we e es ima ed ma hema ically conside ing he
p opo ionali y be ween he ampli ude o he luc ua ion and he licke
se e i y [43], as ollows:
𝑐𝑦=𝑃𝑠𝑡,𝑦 ⋅𝑆𝐶𝑅 =𝑑𝑣
𝑑𝑣𝑟𝑒𝑓 (𝑁)
⋅𝑆𝐶𝑅 (4)
whe e 𝑑𝑣 was he a e age o ampli udes 𝑑𝑣 (%) and he denomina o
𝑑𝑣𝑟𝑒𝑓 (𝑁)(%) is he ampli ude o 𝑁 ol age d ops in 10 min ha
p oduce 𝑃𝑠𝑡 = 1 acco ding o [26]. Fo he h ee cases, 𝑐𝑦 alues o 2.74,
3.46 and 3.86 we e ob ained, espec i ely. Ex ac ing om he 𝑃𝑖𝑛𝑠𝑡
signal he sho ime in e als con aining he ol age d ops p oduced
by he mo o s a up, he licke coe icien s o each case we e also
calcula ed by IEC licke me e . The ob ained alues had a de ia ion o
less han 5% wi h espec o he alues calcula ed using Eq. (4). The
licke coe icien s exclusi ely due o wind a ia ions, 𝑐𝑤, we e also
calcula ed by emo ing om 𝑃𝑖𝑛𝑠𝑡 he sho in e als co esponding o
he ol age d ops p oduced by mo o s a up, wi h alues 1.28, 1.28
and 1.39 o he h ee cases.
In e na ional Jou nal o Elec ical Powe and Ene gy Sys ems 153 (2023) 109376
4
K. Redondo e al.
Fig. 3. A segmen o 2 s con aining a yaw ope a ion. F om op o bo om: ms alue o cu en (A); gene a ed powe (MW); ms alue o ol age (V); and ins an aneous licke
pe cep ion 𝑃𝑖𝑛𝑠𝑡.
Fig. 4. Cha ac e iza ion o yaw con ol ope a ions a each ime se ies du ing CO (blue), SW (g een) and NG ( ed). Top: he a e age ampli ude o ol age d op in 10-min, 𝑑𝑣(%).
Bo om: numbe o yaw con ol ope a ions, 𝑁(in 10-min).
Fig. 6 shows he esul s o he whole se o ime se ies, i con i med
ha he licke due o he yaw ope a ions, 𝑐𝑦, inc eased wi h he
gene a ed powe due o he inc ease o 𝑁(Fig. 4), while he licke
due o wind a ia ions, 𝑐𝑤, emained nea ly cons an o mos o he
powe bins. Conside ing a quad a ic law o he summa ion o di e en
sou ces o licke , desc ibed in IEC 61000-3-7 [44], 𝑐𝑦 alues we e
esponsible o mos o he licke p oduced by he WT, abo e 80%
o he o al licke a 𝑃𝑛. Howe e , 𝑐𝑤 alues we e conside ably lowe ,
ep esen ing less han 13% o he o al licke a 𝑃𝑛.
The e o e, mos o he licke du ing con inuous ope a ion was due
o yaw ope a ions, and depended di ec ly on 𝑁. Fu he mo e, i is
possible o p ope ly es ima e he licke p oduced by yaw ope a ions
In e na ional Jou nal o Elec ical Powe and Ene gy Sys ems 153 (2023) 109376
5
K. Redondo e al.
Fig. 5. Th ee cases o 10-min ime se ies o con inuous WT ope a ion: panel (a), (b) and (c) co espond wi h bin 5, 8 and 10, espec i ely. On each panel, he op plo is he
deli e ed powe , he middle plo is he ms alues o ol age 𝑢𝑓 𝑖𝑐 (𝑡)and he bo om plo is he ins an aneous licke pe cep ion 𝑃𝑖𝑛𝑠𝑡(𝑡).
Fig. 6. The median alues o he licke coe icien s ela ed o he powe bins. The
blue as e isks ep esen he licke coe icien , 𝑐. The ed ci cles ep esen he es ima ed
licke coe icien due o o ien a ions o he nacelle, 𝑐𝑦. The g een squa es ep esen
he licke coe icien due o wind a ia ions, 𝑐𝑤.
Table 3
Resul s om cases in Fig. 5.
Case (a) (b) (c)
Powe bin 5 8 10
𝑃𝑎𝑣𝑔 (MW) 0.91 1.57 1.93
𝐿𝑦(%) 0.02% 0.02% 0.03%
𝑐3.16 3.82 4.26
𝑑𝑣 (%) 0.38 0.38 0.37
𝑁(cpm) 0.9 2.0 3.0
𝑑𝑣𝑟𝑒𝑓 (𝑁)(%) 2.80 2.20 1.93
𝑐𝑦2.74 (75%) 3.46 (82%) 3.86 (82%)
𝑐𝑤1.28 (16%) 1.28 (11%) 1.39 (11%)
by means o Eq. (4), i.e., using he cu e 𝑃𝑠𝑡 = 1 and knowing 𝑁and
𝑑𝑣.
3.3. Flicke measu emen in swi ching ope a ions
Fo licke cha ac e iza ion in swi ching ope a ions, an e alua ion
pe iod o 𝑇𝑝= 5 s was used.
Fig. 7 shows ou cases o swi ching ope a ions o he WT. The
e en s associa ed o swi ching a e ma ked wi h wo e ical g een dash
lines. The i s line indica es he elease o he o o b akes (a 0.83 s,
0.57 s, 1.27 s and 0.82 s o he panels (a),(b),(c) and (d), espec i ely),
In e na ional Jou nal o Elec ical Powe and Ene gy Sys ems 153 (2023) 109376
6
K. Redondo e al.
Fig. 7. Fou swi ching ope a ions wi h 𝑇𝑝= 5 s. (a) SW wi hou mo o s a up. (b) SW wi h mo o s a up ahead. (c) SW wi h mo o s a up in middle. (d) SW ollowed by mo o
s a up. On each panel, op plo ep esen s he ac i e powe , middle plo ep esen s 𝑢𝑓𝑖𝑐 (𝑡) ms ol age and bo om plo ep esen s ins an aneous licke pe cep ion 𝑃𝑖𝑛𝑠𝑡.
and he second co esponds o he gene a o connec ion, ha is when
he o o eached he gene a o synch onism speed (a 2.41 s, 2.39 s,
2.90 s and 2.42 s o panels (a),(b),(c) and (d), espec i ely). These wo
e en s a ec ed powe , ol age and 𝑃𝑖𝑛𝑠𝑡 signals. Bu when he gene a o
connec ion occu ed, he ins an aneous ol age d op esul ed in a mo e
signi ican ise o 𝑃𝑖𝑛𝑠𝑡. F om ha momen on, powe gene a ion s a ed,
p oducing a powe su ge, highe ol age oscilla ions and, he e o e,
highe 𝑃𝑖𝑛𝑠𝑡 alues han be o e he gene a o connec ion.
The e en s co esponding o he mo o s a up a e ma ked wi h
e ical ed dash lines o he cases o panels (b), (c) and (d). I s
e ec on licke was conside ably mo e ele an han he impac o he
swi ching ope a ion. The excu sion o 𝑃𝑖𝑛𝑠𝑡 was almos 10 imes la ge
and, depending on he ins an o he onse , e en masked he e ec o
he swi ching ope a ion. In panel (b) he mo o s a up occu ed a
𝑡= 0.23 s, be o e eleasing he o o b akes. By 𝑡= 2.00 s he e ec
o yaw ope a ion on 𝑃𝑖𝑛𝑠𝑡 anished and i s alue was compa able o
he e ec o gene a o connec ion a 𝑡= 2.39 s. In panel (c) he mo o
s a up occu ed a 𝑡= 2.38 s, be ween he elease o he o o b akes
and he gene a o connec ion. This connec ion p oduced a delay in he
descen o 𝑃𝑖𝑛𝑠𝑡 a 𝑡= 2.90 s, bu i s e ec was masked by he e ec o
he yaw ope a ion. Finally, in panel (d) he mo o s a up occu ed a
𝑡= 3.06 s, a e gene a o connec ion a 𝑡= 2.42 s. The e ec o yaw
ope a ion was supe imposed on he e ec s om con inuous ope a ion,
which was main ained a e gene a o connec ion.
The alues o 𝑘𝑓and 𝑘𝑢ob ained om Fig. 7 cases and he max-
imum alues o 𝑃𝑖𝑛𝑠𝑡 in he in e al 𝑇𝑝a e lis ed in Table 4. Flicke
Table 4
Flicke esul s o he swi ching ope a ions om Fig. 7.
Case 𝑃𝑖𝑛𝑠𝑡,𝑚𝑎𝑥 𝑘𝑓𝑘𝑢
(a) SW 0.05 0.03 0.05
(b) SW +yaw 0.66 0.09 0.11
(c) SW +yaw 0.57 0.09 0.10
(d) SW +yaw 0.62 0.09 0.08
du ing he swi ching ope a ions wi hou he in luence o he yaw
ope a ion was clea ly lowe han licke a ec ed by i . Fo case (a), he
alue o 𝑃𝑖𝑛𝑠𝑡,𝑚𝑎𝑥 was mo e han 10 imes lowe han o cases (b), (c)
and (d); he alue o 𝑘𝑓was 3 imes lowe ; and he alue o 𝑘𝑢almos
2 imes lowe .
Finally, Fig. 8 shows he licke alues o he whole se o swi ching
ope a ions, classi ied acco ding o whe he o no hey we e a ec ed
by a yaw ope a ion. O he o al 230 swi ching ope a ions, 96 we e
a ec ed (42%). Median 𝑘𝑓 o bo h classi ica ions was 0.04 and 0.09
and median 𝑘𝑢was 0.05 and 0.10.
The licke emission due o a yaw ope a ion u ned ou o be much
mo e signi ican han he licke emission due o a swi ching ope a-
ion. In addi ion, almos hal o he swi ching ope a ions coincided
wi h a yaw ope a ion. This gene a ed a mischa ac e iza ion o he
e ec on licke due o he swi ching ope a ion when he p ocedu e o
IEC 61400-21-1 s anda d [28] was employed. Tha is, he high alues
In e na ional Jou nal o Elec ical Powe and Ene gy Sys ems 153 (2023) 109376
7
K. Redondo e al.
Fig. 8. Dis ibu ions o 𝑘𝑓(panel a) and 𝑘𝑢(panel b) o he whole se o he WT swi ching ope a ions wi h o wi hou coinciden yaw ope a ion.
o 𝑘𝑢and 𝑘𝑓we e caused by he p esence o a yaw ope a ion and no
by he swi ching ope a ion i sel .
4. Discussion and conclusions
Acco ding o he IEC 61400-21-1, he licke p oduced by a WT mus
be e alua ed du ing con inuous ope a ion and in swi ching ope a ions.
Howe e , i does no speci ically conside he licke p oduced by yaw
ope a ions.
The wo k analyzes he whole ope a ing ange, bo h in con inuous
ope a ion and in swi ching ope a ions, o a ype III WT a a WPP in
Spain, wi h he aim o s udying he e ec on he licke p oduced
by he consump ions de i ed om he s a up o he mo o used o
yaw ope a ions. Conside ing he yaw ope a ion as an isola ed e en ,
i s e ec on licke was h ee imes mo e signi ican han he e ec
p oduced by he swi ching ope a ion. Du ing con inuous ope a ion,
he yaw ope a ions occu wi h high egula i y, and he esul s showed
ha hey should be conside ed as sou ces o ol age luc ua ions. In
his sense, he numbe o yaw ope a ions and he measu ed licke
g ew consis en ly wi h he gene a ed powe . Disagg ega ing he licke
exclusi ely due o yaw ope a ions, i accoun ed o mo e han 80%
o he o al licke . Howe e , he con ibu ion due o changing wind
cha ac e is ics emained almos cons an a le els below 15% o he
o al licke .
Cu en ly, licke educ ion in wind powe gene a ion has been
achie ed h ough e ec i e ac i e and eac i e powe managemen [45,
46], inc eased use o complex con ol s a egies [47,48] and he
widesp ead use o a iable speed WTs. Fo hese easons, mos o he
licke co esponds o he consump ion de i ed om yaw ope a ions.
Employing yaw con ol s a egies ha minimize he numbe o yaw op-
e a ions [6] and he powe consump ion a mo o s a up appea s o be
he simples ways o educe he licke p oduce by yaw ope a ions. Re-
cen s udies p opose he applica ion o yaw con ol in he whole WPP,
aking in o accoun he wake e ec [24]. Inco po a ing a educ ion
in he numbe o yaw ope a ions in o his app oach would maximize
powe gene a ion, minimize loads, as well as u he minimize licke
emissions om he en i e WPP.
Since he ol age d ops p oduced by mo o s a ups a e o cons an
ampli ude, i is possible o use he 𝑃𝑠𝑡 = 1 cu e o make an accu a e
ma hema ical es ima e o licke se e i y. Gi en he complexi y o
IEC 61400-21-1 and conside ing ha mos o he licke is due o yaw
ope a ions, upda ing he s anda d using he p oposed ma hema ical
es ima e would ha e signi ican ad an ages. Fi s , i would be an
assessmen p io o connec ing he WT o he g id. Secondly, since i
would no be necessa y o measu e o weeks, he economic cos s would
be conside ably educed. Finally, i would simpli y he p ocedu e,
educing i o he analy ical compa ison o he yaw con ol sys em da a
wi h he 𝑃𝑠𝑡 = 1 cu e.
CRediT au ho ship con ibu ion s a emen
Koldo Redondo: In es iga ion, Visualiza ion, W i ing – o iginal
d a . Jose Julio Gu ie ez: Concep ualiza ion, Supe ision, Funding
acquisi ion. Izaskun Azca a e: W i ing – e iew & edi ing, Fo mal
analysis. Mikel Le u iondo: Da a cu a ion. Jose An onio U igüen:
W i ing – e iew & edi ing. So ía Ruiz de Gauna: W i ing – e iew
& edi ing. Pu i icación Saiz: Fo mal analysis.
Decla a ion o compe ing in e es
The au ho s decla e ha hey ha e no known compe ing inan-
cial in e es s o pe sonal ela ionships ha could ha e appea ed o
in luence he wo k epo ed in his pape .
Da a a ailabili y
Da a will be made a ailable on eques .
Acknowledgmen s
This wo k was suppo ed by he Basque Go e nmen (Basque Coun-
y, Spain) h ough he g an IT1590-22.
Re e ences
[1] Bu on T, Jenkins N, Sha pe D, Bossanyi E. Wind ene gy handbook. Wiley.
[2] Yang J, Fang L, Song D, Su M, Yang X, Huang L, Joo YH. Re iew o
con ol s a egy o la ge ho izon al-axis wind u bines yaw sys em. Wind Ene gy
2021;24(2):97–115. h p://dx.doi.o g/10.1002/we.2564.
[3] Manwell JF, McGowan J, Roge s A. Wind ene gy explained, heo y, design and
applica ion. Wiley; 2009.
[4] Ba helmie RJ, F andsen ST, Ra hmann O, Hansen KS, Poli is ES, P ospa hopou-
los J, Cabezón Ma ínez D, Rados K, Van De Pijl SP, Schepe s JG, Schlez W,
Philips J, Neube A. Flow and wakes in la ge wind a ms in complex e ain
and o sho e. In: Eu opean wind ene gy con e ence & exhibi ion (EWEC 2008).
2008.
[5] Shakoo R, Hassan MY, Raheem A, Wu Y-K. Wake e ec modeling: A e iew o
wind a m layou op imiza ion using jensens model. Renew Sus ain Ene gy Re
2016;58:1048–59. h p://dx.doi.o g/10.1016/j. se .2015.12.229.
[6] Song D, Yang J, Fan X, Liu Y, Liu A, Chen G, Joo YH. Maximum powe
ex ac ion o wind u bines h ough a no el yaw con ol solu ion using p edic ed
wind di ec ions. Ene gy Con e s Manage 2018;157:587–99. h p://dx.doi.o g/10.
1016/j.enconman.2017.12.019.
[7] Wu KC, Joseph R, Thupili N. E alua ion o classical and uzzy logic con olle s o
wind u bine yaw con ol. In: P oceedings. he i s IEEE egional con e ence on
ae ospace con ol sys ems. 1993, p. 254–8. h p://dx.doi.o g/10.1109/AEROCS.
1993.720937.
[8] Bu F, Huang W, Hu Y, Xu Y, Shi K, Wang Q. S udy and implemen a ion
o a con ol algo i hm o wind u bine yaw con ol sys em. In: 2009 wo ld
non-g id- connec ed wind powe and ene gy con e ence. 2009, p. 1–5. h p:
//dx.doi.o g/10.1109/WNWEC.2009.5335830, iSSN: 2162-1063.
[9] K agh KA, Fleming PA, Scholb ock AK. Inc eased powe cap u e by o o speed–
dependen yaw con ol o wind u bines. J Sola Ene gy Eng 2013;135(031018).
h p://dx.doi.o g/10.1115/1.4023971.
In e na ional Jou nal o Elec ical Powe and Ene gy Sys ems 153 (2023) 109376
8
K. Redondo e al.
[10] Fa ag W, El-Hosa y H, El-Me wally K, Kamel A. Design and implemen a ion
o a a iable-s uc u e adap i e uzzy-logic yaw con olle o la ge wind u -
bines. J In ell Fuzzy Sys ems 2016;30(5):2773–85. h p://dx.doi.o g/10.3233/
IFS-152030.
[11] As ol i D, Cas ellani F, Becche i M, Lomba di A, Te zi L. Wind u bine sys ema ic
yaw e o : Ope a ion da a analysis echniques o de ec ing i and assessing
i s pe o mance impac . Ene gies 2020;13(9):2351. h p://dx.doi.o g/10.3390/
en13092351.
[12] Bakhshi R, Sandbo n P. Maximizing he e u ns o LIDAR sys ems in wind
a ms o yaw e o co ec ion applica ions. Wind Ene gy 2020;23(6):1408–21.
h p://dx.doi.o g/10.1002/we.2493.
[13] K agh KA, Hansen MH, Mikkelsen T. P ecision and sho comings o yaw
e o es ima ion using spinne -based ligh de ec ion and anging. Wind Ene gy
2013;16(3):353–66. h p://dx.doi.o g/10.1002/we.1492.
[14] K agh KA, Hansen MH. Po en ial o powe gain wi h imp o ed yaw alignmen .
Wind Ene gy 2015;18(6):979–89. h p://dx.doi.o g/10.1002/we.1739.
[15] Ouyang T, Kusiak A, He Y. P edic i e model o yaw e o in a wind u bine.
Ene gy 2017;123:119–30. h p://dx.doi.o g/10.1016/j.ene gy.2017.01.150.
[16] Fa e F, P i sche L, Be na don D. Senso less ac i e yaw con ol o wind
u bines. In: IECON’01. 27 h annual con e ence o he IEEE indus ial elec onics
socie y (Ca . No.37243), Vol. 2. 2001, p. 1370–5. h p://dx.doi.o g/10.1109/
IECON.2001.975981, ol.2.
[17] Xin W, Yanping L, Wei T. Modi ied hill climbing me hod o ac i e yaw con ol
in wind u bine. In: P oceedings o he 31s chinese con ol con e ence. 2012,
p. 6677–80, iSSN: 2161-2927.
[18] Ka akasis N, Mesemanolis A, Nalmpan is T, Mademlis C. Ac i e yaw con ol in
a ho izon al axis wind sys em wi hou equi ing wind di ec ion measu emen .
IET Renew Powe Gene 2016;10(9):1441–9. h p://dx.doi.o g/10.1049/ie - pg.
2016.0005.
[19] Ye Z, Wang X, Chen Z, Wang L. Uns eady ae odynamic cha ac e is ics o a
ho izon al wind u bine unde yaw and dynamic yawing. Ac a Mech Sinica
2020;36(2):320–38. h p://dx.doi.o g/10.1007/s10409-020-00947-2.
[20] Ekelund T. Yaw con ol o educ ion o s uc u al dynamic loads in wind
u bines. J Wind Eng Ind Ae odyn 2000;85(3):241–62. h p://dx.doi.o g/10.
1016/S0167-6105(99)00128-2.
[21] Jeong M-S, Kim S-W, Lee I, Yoo S-J, Pa k KC. The impac o yaw e o on
ae oelas ic cha ac e is ics o a ho izon al axis wind u bine blade. Renew Ene gy
2013;60:256–68. h p://dx.doi.o g/10.1016/j. enene.2013.05.014.
[22] Ahmad T, Basi A, Ahsan M, Coupiac O, Gi a d N, Kazem ab izi B, Ma hews PC.
Implemen a ion and analyses o yaw based coo dina ed con ol o wind a ms.
Ene gies 2019;12(7):1266. h p://dx.doi.o g/10.3390/en12071266.
[23] Knudsen T, Bak T, S ens up M. Su ey o wind a m con ol—powe and a igue
op imiza ion. Wind Ene gy 2015;18(8):1333–51. h p://dx.doi.o g/10.1002/we.
1760.
[24] Kuma D, Ro ea MA, Aju EJ, Jin Y. Wind plan powe maximiza ion ia
ex emum seeking yaw con ol: A wind unnel expe imen . Wind Ene gy
2023;26(3):283–309. h p://dx.doi.o g/10.1002/we.2799.
[25] Lode i S, Azca a e I, Gu ié ez JJ, Le u iondo LA, Redondo K, Sáiz P, Mele o JJ,
B una J. Flicke o mode n ligh ing echnologies due o apid ol age changes.
Ene gies 2019;12(5):865. h p://dx.doi.o g/10.3390/en12050865.
[26] IEC 61000-4-15: elec omagne ic compa ibili y (EMC) - Pa 4-15: es ing and
measu emen echniques - licke me e - unc ional and design speci ica ions.
In e na ional Elec o echnical Commission; 2010.
[27] Gu ie ez JJ, Saiz P, Le u iondo LA, Azca a e I, Redondo K, Lazkano A. Flicke
measu emen in eal scena ios: Reducing he di e gence om he human
pe cep ion. Elec Powe Sys Res 2016;140:312–20. h p://dx.doi.o g/10.1016/
j.eps .2016.06.010.
[28] IEC 61400-21-1: wind ene gy gene a ion sys ems - pa 21-1: measu emen
and assessmen o elec ical cha ac e is ics - wind u bines. In e na ional
Elec o echnical Commission; 2019.
[29] La sson A. Flicke emission o wind u bines du ing con inuous ope a ion, ene gy
con e sion. IEEE T ans 2002;17(1):114–8.
[30] Lopez C, Blanes J. S a is ical analysis o licke p oduced by wind a ms wi h
ixed speed asynch onous gene a o s in con inuous ope a ion. In T ans Elec
Ene gy Sys 2013;23(8):1440–51. h p://dx.doi.o g/10.1002/e ep.1670.
[31] Fooladi M, Akba i Fo oud A. Recogni ion and assessmen o di e en ac o s
which a ec licke in wind u bines. IET Renew Powe Gene 2016;10(2):250–9.
h p://dx.doi.o g/10.1049/ie - pg.2014.0419.
[32] Liang X. Eme ging powe quali y challenges due o in eg a ion o enewable
ene gy sou ces. IEEE T ans Ind Appl 2017;53(2):855–66. h p://dx.doi.o g/10.
1109/TIA.2016.2626253.
[33] Sinsel SR, Riemke RL, Ho mann VH. Challenges and solu ion echnologies o
he in eg a ion o a iable enewable ene gy sou ces - a e iew. Renew Ene gy
2020;145:2271–85. h p://dx.doi.o g/10.1016/j. enene.2019.06.147.
[34] La sson A. Flicke emission o wind u bines caused by swi ching ope a ions,
ene gy con e sion. IEEE T ans 2002;17(1):119–23. h p://dx.doi.o g/10.1109/
60.986448.
[35] Gu ie ez JJ, Azca a e I, Saiz P, Redondo K, Le u iondo LA, Ruiz de Gauna S.
Imp o ing he de ec ion o RVCs o a be e assessmen o hei in luence on
licke . IEEE T ans Powe Deli 2021. h p://dx.doi.o g/10.1109/TPWRD.2021.
3068330.
[36] Redondo K, Lazkano A, Saiz P, Gu ie ez JJ, Azca a e I, Le u iondo LA. A s a egy
o imp o ing he accu acy o licke emission measu emen om wind u bines.
Elec Powe Sys Res 2016;133:12–9. h p://dx.doi.o g/10.1016/j.eps .2015.11.
040.
[37] Redondo K, Gu ie ez JJ, Saiz P, Le u iondo LA, Azca a e I, Lazkano A. Accu a e
di e en ia ion o imp o ing he licke measu emen in wind u bines. IEEE
T ans Powe Deli 2017;32(1):88–96. h p://dx.doi.o g/10.1109/TPWRD.2016.
2527840.
[38] Redondo K, Gu ie ez JJ, Saiz P, Azca a e I, Le u iondo LA, Lazkano A. A
p oposal o e i ica ion es s o he licke measu emen p ocedu e o g id-
connec ed wind u bines. Measu emen 2017;95:116–27. h p://dx.doi.o g/10.
1016/j.measu emen .2016.09.037.
[39] De Lange Dzn H. Expe imen s on licke and some calcula ions on an elec ical
analogue o he o eal sys ems. Physica 1952;18(11):935–50. h p://dx.doi.o g/
10.1016/S0031-8914(52)80230-7.
[40] Aille e P. Dé e mina ion des lois expé imen ales du papillo emen ( licke ) en
ue de leu applica ion aux éseaux basse ension su lesquels les cha ges a ien
p iodiquemen ou aléa oi emen (soudeuses déma ages de mo eu s). Bull Soc F
Elec 1957;7:257–62.
[41] Rashbass C. The isibili y o ansien changes o luminance. J Physiol
1970;210(1):165–86. h p://dx.doi.o g/10.1113/jphysiol.1970.sp009202.
[42] IEC 61000-4-30: Elec omagne ic compa ibili y (EMC) - Pa 4-30: es ing and
measu emen echniques - powe quali y measu emen me hods. In e na ional
Elec o echnical Commission; 2021.
[43] Gu ie ez J, Ruiz J, Ruiz de Gauna S. Linea i y o he IEC licke me e
ega ding ampli ude a ia ions o ec angula luc ua ions. IEEE T ans Powe
Deli 2007;22(1):729–31. h p://dx.doi.o g/10.1109/TPWRD.2006.886767.
[44] IEC 61000-3-7: Elec omagne ic compa ibili y (EMC) - Pa 3-7: limi s - assess-
men o emission limi s o he connec ion o luc ua ing ins alla ions o MV, HV
and EHV powe sys ems. In e na ional Elec o echnical Commission; 2008.
[45] Amma M, Joos G. Impac o dis ibu ed wind gene a o s eac i e powe beha io
on licke se e i y. IEEE T ans Ene gy Con e s 2013;28(2):425–33. h p://dx.doi.
o g/10.1109/TEC.2013.2256425.
[46] Masca ella D, Venne P, Gue e e D, Joos G. Flicke mi iga ion ia dynamic
ol /VAR con ol o powe elec onic in e aced WTGs. IEEE T ans Powe Deli
2015;30(6):2451–9. h p://dx.doi.o g/10.1109/TPWRD.2015.2394237.
[47] Zhang Y, Chen Z, Hu W, Cheng M. Flicke mi iga ion by indi idual pi ch
con ol o a iable speed wind u bines wi h DFIG. IEEE T ans Ene gy Con e s
2014;29(1):20–8. h p://dx.doi.o g/10.1109/TEC.2013.2294992.
[48] Liu Y, Liu S, Zhang L, Cao F, Wang L. Op imiza ion o he yaw con ol e o o
wind u bine. F on Ene gy Res 2021;9:5. h p://dx.doi.o g/10.3389/ en g.2021.
626681.
