Au ho s: Da id Schlip , Feng Guo, Maa en an den B oek, Simon Weich
Con en
In oduc ion ............................................................................................................................ 2
Backg ound ............................................................................................................................. 2
Ou Ca ego ies ........................................................................................................................ 2
E alua ion and awa ds............................................................................................................. 3
Ou Disciplines ....................................................................................................................... 3
The 30 s Sp in ..................................................................................................................... 3
The 18 m/s Hu dles .............................................................................................................. 7
Ou imeline ............................................................................................................................ 9
How o ge s a ed ................................................................................................................... 9
Suppo ................................................................................................................................... 9
Re e ences ............................................................................................................................ 10
Task 52
LAC Summe Games 2025
2
In oduc ion
Following he success o he Lida -Assis ed-Con ol (LAC) Summe Games in 2024, his new
edi ion builds on he lessons lea ned ha we e p esen ed a he Wind Ene gy Science Con e ence
in June 2025 [1]. F om an educa ional pe spec i e, he Summe Games 2024 p o ed o be a g ea
oppo uni y o p oblem-based lea ning and encou aging collabo a ion be ween uni e si ies and
p o essionals om he indus y. In o al, 15 con ibu ions om s uden s and p o essionals ac oss
6 coun ies ha e been submi ed o h ee di e en disciplines. A he same ime, he expe ience
showed how challenging i is o design LAC exe cises ha a e ealis ic enough o be meaning ul,
ye manageable wi hin a limi ed ime ame – especially o s uden s a uni e si ies. Fo he
Summe Games 2025, we wan o con inue his collabo a i e o ma wi h upda ed scena ios and
mo e ealis ic wind condi ions.
Backg ound
Wind cha ac e is ics a e inhe en ly dynamic, subjec o sudden changes in speed, di ec ion, and
u bulence. Consequen ly, wind u bines a e highly dynamic sys ems, exci ed by such s ochas ic
in luences and mus be designed o handle hese a ia ions and dis u bances. T adi ional
con olle s eac only a e wind changes ha e impac ed he wind u bine, esul ing in an o se
ope a ion un il he con olle handles he dis u bance. Con a y, lida -assis ed con ol concep s
use p e iew in o ma ion o he wind o an icipa e changes, hence eac ing be o e he dis u bance
app oaches he o o and he e o e imp o ing he ope a ion o he u bine.
The mission o ou IEA Wind Task 52 wo king g oup is o push o wa d he lida -assis ed con ol
echnology and simpli y i s applica ion by ecommended p ac ices, open-sou ce ools, and
joined exe cises. In line wi h ou mission, we a e e y glad o launch ou Summe Games 2025
a e he success ul Summe Games 2024 o con inue o encou age s uden s and p o essionals
in he a o designing and deploying lida da a p ocessing algo i hms and lida -assis ed
con olle s.
By pa icipa ing in he games, you can inc ease you knowledge in LAC and enjoy he design
p ocess by compe ing wi h o he s. Mo eo e , we aim o igge you c ea i i y in h ee di e en
disciplines and wi h i mo i a e he de elopmen o new concep s. Du ing he compe i ion, i is
in ended ha he pa icipan s lea n he exis ing open-sou ce ools and hei applica ion will esul
e en ually in he u he de elopmen o he ools. Addi ionally, du ing he games you will be able
o exchange knowledge wi h di e en pa ies o he wind ene gy communi y. We a e looking
o wa d o s eng hening you passion o ad ancing he ision o mode n wind u bines and
ge ing he bes LAC ou o you! The ools a e based on p e ious publica ions [2], [3] and buil up
on he IEA 15 MW e e ence wind u bine [4] and he open-sou ce con olle ROSCO [5]. We in end
o w i e a publica ion on he esul s simila o [1] and in eg a e he lea nings in ou planned IEA
Recommended P ac ices on Lida -Assis ed Con ol.
Ou Ca ego ies
Fo he Summe Games we ha e he ollowing ca ego ies:
1. S uden s (including PhD s uden s)
2. Resea che s
Pa icipa ion in g oups is possible. I all g oup membe s a e s uden s, he i s ca ego y applies.
3
E alua ion and awa ds
In e e y discipline (see below) we will ank he con ibu ions o each o he wo ca ego ies
men ioned abo e depending on he cos : o he 30 s sp in and he 18 m/s Hu dles, he lowes
cos will win (minimiza ion p oblem). The esul s will be sha ed on he IEA Task LinkedIn page and
pa icipan s will also ha e he oppo uni y o p esen hei solu ion o he IEA Wind Task 52 LAC
wo king g oup, which consis s o a selec ion o academic and indus y expe s.
Ou Disciplines
In his sec ion, you will ind he di e en disciplines whe e you can pa icipa e wi h you p oposed
lida -da a p ocessing algo i hms and lida -assis ed con olle s. Rega dless o you chosen
ca ego y, we encou age you o submi ou comes o as many disciplines as you belie e align wi h
you capabili ies. In he ollowing subsec ions, each o he disciplines a e p esen ed oge he wi h
he e alua ion me hods used o e alua ing he pe o mance o he algo i hms you send.
The 30 s Sp in
In he i s challenge, like a 100-me e sp in , we'll es how well you con olle handles an
Ex eme Cohe en gus wi h Di ec ion change (ECD) a 12.5 m/s ollowing he Design Load Case
(DLC) 1.4 o he IEC 61400-1 in e na ional s anda d on wind u bine design equi emen s, see
Figu e 1. The challenge is add essing mainly con ol expe s and he in en ion is o in oduce you
o LAC and o igge new con ol concep s. In his discipline, you con olle esponse will be
es ed unde he mos challenging condi ions o any con olle e alua ion. The objec i e is o
cancel ou he e ec o he gus and o ha e a smoo h o o speed and owe - op mo ion.
Fo his pu pose, he wind p e iew and he e o e he o o -e ec i e wind speed a e gene a ed
h ough lida -simula ions, ep esen ing ealis ic measu emen condi ions. Two di e en lida
con igu a ions a e conside ed as baseline cases: a ou -beam pulsed sys em (4BeamPulsed) and
a con inuous-wa e ci cula scanning sys em (Ci cula CW). Bo h se ups p o ide di e en spa ial
a e aging cha ac e is ics, allowing a compa ison o hei impac on he con ol pe o mance. The
scan dis ance has been op imized o he Summe Games 2024 o p o ide he bes p e iew
quali y (smalles de ec able eddy size) o ozen u bulence simula ions.
The sc ip RunExample.m in he Sp in olde (wi hin he “Summe Games2025” b anch) p o ides
an example using a baseline collec i e pi ch eed o wa d con ol [1], which is designed o
compensa e he impac o he gus on he o o speed and wo ks well o high wind speeds bu
p o ides plen y o oom o imp o emen s o wind speeds close o a ed wind, see Figu e 3. The
co esponding ECD wind ield cha ac e is ics – including i s magni ude, la e al, and e ical
componen – as de ined in IEC 61400-1, a e also illus a ed in Figu e 1. To simpli y he ask, we
educed he deg ees o eedom o he wind u bine o he o o mo ion (GenDOF), he i s o e-
a owe bending-mode (TwFADOF1), and he i s side-side owe bending-mode (TwSSDOF1).
The e alua ion c i e ion o his discipline is he minimiza ion o he combined ela i e educ ion
o he o o and owe mo ion de ined as:
𝐽 = max (|Ω−Ω0|)
Ω0+ max (|𝑀𝑦T−𝑀𝑦T0|)
𝑀𝑦T0 ,
whe e Ω is he o o speed and Ω0 i s s eady-s a e alue. Fu he , 𝑀𝑦T is he o e-a owe base
bending momen and 𝑀𝑦T0 i s s eady-s a e alue.
4
The esul s o he wo baseline lida se ups a e:
• 4BeamPulsed: Cos = 0.731996
• Ci cula CW: Cos = 1.217946
The ollowing ules apply:
1. You can eplace he eedback con olle (ROSCO) and use you own combined eedback
and eed o wa d con olle , model p edic i e con ol e c. adjus ing collec i e/indi idual
pi ch angle and gene a o o que.
2. Changes in he OpenFAST inpu iles o he han adjus ing he BLADED INTERFACE sec ion
in he Se oDyn inpu ile a e no allowed.
3. The lida da a p ocessing and lida simula ion iles can be changed, bu should
co espond o comme cially a ailable lida sys ems.
In addi ion o using he in o ma ion o he lida o con inuously imp o e he con ol (e.g.
eed o wa d con ol o model p edic i e con ol), we encou age pa icipan s also o enable
ex eme e en de ec ion and mi iga ion s a egies.
Besides he con ol challenge, ano he pu pose o his discipline is o showcase he di icul ies in
simula ing DLC 1.4. The ollowing assump ions ha e been made:
1. The p opaga ion wind speed (in he lida inpu iles Lida File_4BeamPulsed.da and
Lida File_Ci cula CW.da ) is se o 20 m/s, he mean o he wind speed magni ude o e
he simula ion.
2. This p opaga ion wind speed is also used o de e mine he bu e imes in
FFP_ 1_4BeamPulsed.IN and FFP_ 1_Ci cula CW.IN which assumes ha he
p opaga ion wind speed is known o he con olle .
3. The wind ield componen s a e o a ed indi idually and p opaga ed wi h he p opaga ion
wind speed, see Figu e 2. No u bulence is added so a .
4. While he DLC 1.4 should be simula ed a a ed wind speed and a ed wind speed ±2 m/s,
we only selec ed he a ed wind speed + 2 m/s case.
We also encou age pa icipan s o commen on he ollowing ques ions:
1. Is he DLC 1.4 as simula ed he e a aluable scena io o LAC?
2. I so, how should we simula e he DLC 1.4 mo e ealis ically?
3. I no , wha should we simula e ins ead?
5
5.
Figu e 1: Wind- ield componen s (magni ude, la e al, and e ical) o he ECD simula ion o lida -assis ed con ol o e
di e en heigh s o e g ound. The ime co esponds o he owe bo om cen e . The wind ield and he igu e a e
gene a ed by he sc ip W i eECD2BladedWind.m in he Sp in Wind olde (wi hin he “Summe Games2025” b anch).
Figu e 2: Possible simula ion op ions o he ECD. The le one (used in he Summe Games 2025) o a es he wind
ec o s in a wind ield wi h ho izon al in low angle 𝛼H and p opaga es hem wi h he mean wind speed 𝑢. Ano he op ion
( igh ) would be o o a e he wind ield a ound a ce ain o a ion cen e (he e he wind u bine).
6
Figu e 3: Simula ion esul s o he baseline cases “4BeamPulsed” ( op) and “Ci cula CW” (bo om), showing he
esponse o he con olle du ing he ECD. The igu e illus a es he s a ing poin o “The 30 s sp in ” challenge, whe e
he goal is o igge new con ol concep s in o de o educe he impac o he ECD on o o and owe mo ion.
Gene a ed by he sc ip RunExample.m in he Sp in olde (wi hin he “Summe Games2025” b anch).
7
The 18 m/s Hu dles
In he second challenge, we'll es how well you lida -da a-p ocessing algo i hm is able o de ec
obs acles a high wind speeds, he e 18 m/s. The challenge is mos ly add essing lida expe s and
he in en ion is o in oduce you o lida -da a-p ocessing o con ol and o igge new lida -da a
p ocessing algo i hms and lida scan con igu a ions. The objec i e is o educe he measu emen
e o o he p edic ed o o -e ec i e wind speed (REWS) wi h wo seconds o p edic ion ime,
ypically needed o eed o wa d con ol o his u bine.
The sc ip s RunExample.m/.py in he Hu dles olde p o ide examples using a baseline lida -da a
p ocessing [1] o a ypical con inuous wa e lida (ci cula scan) and a pulsed lida (4 beams).
Building up on he Summe Games 2024, we added blade blockage and wind e olu ion using he
comme cial lida simula o solis. This simula o can ei he be un s and-alone wi h p e-
de e mined inpu s o as an ex ension o a ious wind u bine simula ion so wa es such as
OpenFAST o HAWC2. Fo he Summe Games 2025, he line-o -sigh wind speed measu emen s
include he e ec s o p obe- olume a e aging. Blade blockage e ec s a e modelled based on a
cons an o o speed and blade dimensions om he IEA 15 MW e e ence wind u bine (he e, he
isValid lag is se o alse). The wind ields ha ha e been sampled a e based on Kaimal spec a
u bulence gene a ed using Tu bSim. The e ec s o spa ial u bulence e olu ion a e included wi h
an exponen ial cohe ence model as desc ibed in [6].
The e alua ion c i e ion o his discipline is he minimiza ion o he mean absolu e e o (MAE)
be ween he REWS om he wind ield shi ed wo seconds in o he u u e (REWSWF) and he
p ocessed REWS es ima e om he lida (REWSLP). Addi ionally, we emo e he mean alues o
he e o o be independen o o se s:
e o = REWSWF − REWSLP
𝐽 = mean(|e o − mean(e o )|) .
In he examples, he REWS is il e ed by a low pass il e and hen bu e ed. The MEA is a e aged
o e he las 600 seconds o six simula ions wi h di e en u bulen wind ields gene a ed by
Tu bSim. The used baseline lida da a p ocessing LDP_ 3.m is based on he Summe Games
2024, igno ing measu emen s wi h a alse isValid lag, causing he “blade ghos e ec ” [7].
The esul s o he wo baseline lida se ups a e:
• 4BeamPulsed: Cos = 0.515998 m/s
• Ci cula CW: Cos = 0.463842 m/s
The ollowing ules apply:
1. The lida con igu a ions a e cu en ly limi ed o he wo men ioned abo e o simplici y.
2. New lida da a p ocessing ideas should be implemen ed wi h he same in e ace as he
unc ion LDP_ 3.m.
8
Figu e 4: S a ing poin o “The 18 m/s hu dles” challenge o he baseline cases “4BeamPulsed” ( op) and “Ci cula CW”
(bo om) o he i s seed: Goal is o educe he measu emen e o . He e, he o o -e ec i e wind speed (REWS) om
he wind ield is shi ed by wo seconds in o he u u e and compa ed o he p ocessed lida es ima e ( op and cen e ).
We would like o igge new lida da a p ocessing algo i hms and scan con igu a ions. Gene a ed by he sc ip
RunExample.m in he Hu dles olde (wi hin he “Summe Games2025” b anch).
9
Ou imeline
The Summe Games 2025 ha e al eady s a ed and will un un il Feb ua y 28, 2026 (end o win e
semes e o mos uni e si ies).
Wha should be p o ided:
1. Fo he discipline “The 30 s Sp in ”: please upload he zipped ou b iles o he Teams olde
Uploads (WG 2 – LAC>Summe Games2025>Uploads) in he Sp in olde using he
ollowing ilename o you zipped ile: <Name>_<YYYYMMDD>.
2. Fo he discipline “The 18 m/s Hu dles”: please upload MAT iles wi h you REWS es ima e
om you lida da a p ocessing algo i hm.
3. I possible: In addi ion o he iles abo e, DLLs and o he iles necessa y o un simula ions
o lida da a p ocessing algo i hms ( ia a Gi o k). Please check wi h someone
independen ly o ensu e ha he sc ip s un and s a e clea ly which sc ip p oduces he
submi ed iles.
We will hen e alua e he esul s oge he in ou wo king g oup mee ings and announce he
ankings in he wo disciplines and wo ca ego ies.
Any changes o he ules o submission da es can be discussed in he discussion o um:
h ps://gi hub.com/IEAWindTask52/Lida Assis edCon ol/discussions.
How o ge s a ed
To ge s a ed wi h ou Lida -Assis ed Con ol open-sou ce ools, you need o clone he Gi Hub
eposi o y o you local compu e using ei he Gi o Gi Hub Desk op. Please make su e you
change o he “Summe Games2025” b anch. In he eposi o y, we p o ide a ious example
sc ip s o LAC in bo h Ma lab and Py hon.
To ge s a ed wi h he Ma lab sc ip s you need o na iga e o he Sp in o Hu dles olde and hen
you can s a wo king based on he example sc ip s (see disciplines abo e). The sc ip s use
unc ions which a e p o ided in a olde We iMa labFunc ions and added o he MATLAB pa h.
To ge s a ed wi h Py hon sc ip s unlike Ma lab use s need o ins all a ious modules like numpy,
pandas, o scipy. To make i simple we ha e p o ided a ‘se up.py’ sc ip ha ins alls all he
necessa y modules in you en i onmen .
Suppo
I you ha e ques ions ega ding he Summe Games in gene al o ega ding he code which migh
be in e es ing o o he s as well, please use ou Fo um on Gi Hub:
h ps://gi hub.com/IEAWindTask52/Lida Assis edCon ol/discussions.
I you equi e u he suppo , please don’ hesi a e o con ac ou suppo eam ia email:
• Simon Weich simon.weich@hs- lensbu g.de
• Da id Schlip da id.schlip @hs- lensbu g.de
