Centrifuge testing of offshore wind turbine foundations

Cen i uge es ing o o sho e wind u bine ounda ions Madhu i Mu ali1, F ancisco G ajales1, Gio anna Biscon in2, Cha les Aubeny1 1 Texas A&M Uni e si y, 2 Uni e si y o Camb idge, I is widely ecognised ha a mix o enewables will be equi ed o mee mains eam ene gy needs and o b ing abou he bene i s o a di e si ied ene gy po olio. O sho e wind a ms a e becoming inc easingly popula in he ques o enewable sou ces o ene gy. Many o he a eas wi h he g ea es po en ial o o sho e wind ene gy de elopmen a e in coas al wa e s whe e shallow dep hs ex end ela i ely a o sho e (DOE, 2008). T adi ional o sho e s uc u es ha e e y la ge e ical g a i y loads and smalle ho izon al loads. Wind owe s, on he o he hand, will ha e mode a e e ical g a i y loads wi h la ge ho izon al loads and momen s in p opo ion o e ical loads. Design pa adigms which e ol ed in he o sho e oil and gas indus y a e no eadily applicable o he design o wind owe ounda ions o sho e. Design codes o o sho e wind owe s a e s ill e y much in hei in ancy and do no p o ide su icien guidance o designe s. The cos o he suppo s uc u es accoun s o abou 24% o he o al cos o building and ope a ing an o sho e wind plan . Dec easing he cos o he ounda ion sys em can be an impo an pa o making wind a iable sou ce o ene gy. Howe e , i will equi e be e unde s anding o he soil- ounda ion sys em esponse and be e ools o model ealis ically, accu a ely, and e ec i ely he in luence o he soil and ounda ion on he whole sys em (By ne e al., 2003). Cu en ly, plans o o sho e wind a ms ac oss Eu ope and No h Ame ica a e in a ious s ages o de elopmen , wi h mos o he p ojec s being limi ed o less han 30-m deep wa e s in he No h and Bal ic Seas, p ima ily in he 5 o 12-m wa e dep h ange (Musial e al., 2006). Va ious s uc u al con igu a ions ha e been de eloped wi h monopiles o g a i y based s uc u es being he mos common solu ions o wa e dep hs shallowe han 30 m (By ne e al., 2003). Al hough piled ounda ions unde o sho e loading condi ions ha e been s udied in de ail o e he yea s, he esponse o sho aspec a io piles (<L/D = 8) is no well unde s ood. A p incipal aspec o pile esponse ha equi es assessmen is he ul ima e pile capaci y o ensu e ha he ounda ion has su icien s eng h in an ex eme loading condi ion (Ande sen, 2004). A se ies o cen i uge model es s we e ca ied ou o in es iga e he esponse o sho aspec a io monopiles (L/D = 2) wi h bo h ixed and o a ing head subjec ed o la e al loads in so , no mally consolida ed clay. The cen i uge expe imen al es ing consis ed o he ollowing componen s: 1. Design and cons uc ion o he es ing ools and model ounda ions. 2. Cen i uge es ing o model caissons in so clay (kaolin) a 70 g o e a ange o eccen ici ies, loading sequences and displacemen ampli udes. The cen i uge es ing was ca ied ou a he Ne wo k o Ea hquake Enginee ing Simula ions (NEES) acili y a Rensselae Poly echnic Ins i u e. Loads we e applied on he model ounda ions using he 4- deg ee o eedom in- ligh obo wi h a cus omized ool, consis ing o an adap o designed and ab ica ed o la ch on o he in- ligh obo . This adap o is used wi h 2 di e en ypes o pile caps o achie e bo h pinned and igid connec o s (Figu e 1). The connec o s o ans e he load on o he sho monopile we e designed and ab ica ed o allow bo h o a ion and ansla ion. 1. A ball and socke join was de eloped wi h a sphe ical ball on he op o he pile cap and a cylind ical socke on he adap o allowing he pile o o a e eely o a pinned join (Figu e 1a). The pinned connec ion was designed o apply momen wi h ou di e en eccen ici ies; 1.2D, 1.5D, 2.5D and 3.5D 2. A igid locking connec ion was designed wi h a la pla e o i in o a g oo e in he adap e p o iding a igid connec ion (Figu e 1b). Figu e 1: Adap e and pile cap connec o s o o a ion and ansla ion. Single dimensional Memsic 10 g accele ome e s based on Mic o-Elec o-Mechanical Sys ems (MEMS) we e used o measu e he il by moun ing hem o he model ounda ions, Linea Va iable Displacemen T ansduce ’s used o measu e displacemen . The model piles we e also s ain gauged o measu e he applied la e al and e ical o ce. The shea s eng h o he clay bed was cha ac e ized in ligh by using a T-ba pene ome e de eloped a he Uni e si y o Wes e n Aus alia a ailable a RPI (Randolph, 2006). The es s we e ca ied ou using a T-ba ha was 5 mm in diame e and 20 mm in leng h, a a pene a ion a e o 2 mm/s. Wa e con en p o iles we e also measu ed along he leng h o he es bed o compu e shea s eng h. All piles we e es ed a a cen i ugal accele a ion o 70 g by mo ing hem la e ally by a la ge displacemen o ob ain he ul ima e la e al capaci y unde displacemen con ol a a a e o 2mm/s. The e ec s o momen loading applied a di e en eccen ici ies and ho izon al loading on piles was examined. T-ba pene ome e es s we e ca ied ou be o e and a e each pile es o ob ain he s eng h p o ile o he clay nex o each pile. The sel -weigh o he pile and connec o s simula ed he cons an e ical load o a ligh s uc u e, while he la e al load was applied o ep oduce he en i onmen al loads. The a io o e ical load o ho izon al load (V/H) o hese pile es s was compu ed o ange be ween 2.1- 3.2. Exis ing alues o his a io o o sho e wind u bines in he li e a u e a e ound o be anging om 2.1 o 5.8. The load-de lec ion cu e o he pile es ed in pu e ansla ion is p esen ed in Figu e 2a along wi h he calcula ed ul ima e la e al capaci y using me hods p oposed by Mu and Hamil on (1993), he API me hod (API, 2000) and ini e elemen analysis o compa ison. The la e al head load, H, is no malized by he p oduc o he p ojec ed e ical a ea, LD, and a shea s eng h p o ile (su a g) based on he T-ba and wa e con en s eng h p o iles o e he dep h o pile embedmen . The mono onic esponse o he monopile subjec o o a ion was examined o ou di e en eccen ici ies; 1.2D, 1.5D, 2.5D and 3.5D and is p esen ed in Figu e 2b. Simila ly o he pile es ed in ansla ion, he la e al head load, H, a he op o he pile cap was compu ed and no malized by he p oduc o he p ojec ed e ical a ea, LD, and an a e age shea s eng h p o ile o e he dep h o pile embedmen , (su a g). The la e al displacemen , y, was compu ed a he mudline using he il and displacemen measu emen s and no malized by he pile diame e , D. All he piles we e pushed la e ally a he op o he ball and socke connec o o a displacemen ampli ude equal o 30% o he pile diame e . Thus he pile displacemen ampli ude a he mudline a ied depending on he eccen ici y. As expec ed he ul ima e la e al capaci y o he piles dec eased wi h inc easing eccen ici y. Also obse ed is ha he piles es ed in o a ion mobilize inc easing s eng h wi h inc easing displacemen s due o he o a ional ailu e mechanism o sho aspec a io piles. Expe imen al da a on model o sho e ounda ions is pa icula ly necessa y o s udy combined loading and he in luence o e ical loads on he la e al bea ing capaci y o hese ounda ions. Figu e 2: Fo ce-displacemen cu es o pile es ed in; a) ansla ion; b) o a ion. Be e ools a e needed o model ealis ically, accu a ely and e ec i ely he in luence o he soil and ounda ion on he o sho e wind u bine sys em. Good expe imen al da a gi es us a be e unde s anding o he soil- ounda ion sys em esponse enabling he de elopmen o mo e e icien nume ical models leading o mo e con idence in ou designs. Acknowledgemen s The au ho s also acknowledge he Na ional Science Founda ion, NEES, and he p ojec Capaci y and Pe o mance o Founda ions o O sho e Wind Towe s, Awa d Numbe : 1041604. Re e ences API, 2000. Recommended p ac ice o planning, designing, and cons uc ing ixed o sho e pla o ms wo king s ess design. Ande sen, K., 2004. Cyclic Beha iou o Soils and Lique ac ion Phenomena. Taylo & F ancis G oup, London, Cyclic clay da a o ounda ion design o s uc u es subjec ed o wa e loading. By ne, B., and Houlsby, G., 2003. “Founda ions o o sho e wind u bines”. Phil. T ans. R. Soc. Lond. A, 361, pp. 2909–2930. DOE, 2008. 20% wind ene gy by 2030: Inc easing wind ene gy con ibu ion o u.s. elec ici y supply. Tech. Rep. DOE/GO-102008-2567, Depa men o Ene gy. Mu , J., and Hamil on, J., 1993. “P-ul ima e o und ained analysis o la e ally loaded piles.” ASCE Jou nal o Geo echnical Enginee ing, 119 (1), pp. 91–107.