International Parallel Tests on Bender Elements at the University of Porto, Portugal

REPORT In e na ional Pa allel Tes s on Bende Elemen s a he Uni e si y o Po o, Po ugal C is iana Fe ei a PhD s uden , Resea ch Assis an [ [email p o ec ed] ] An ónio Viana da Fonseca Associa e P o esso , Di ec o o LabGEO o FEUP [ [email p o ec ed] ] FEUP, Janua y 2005 LABORATÓRIO DE GEOTECNIA FACULDADE DE ENGENHARIA DA UNIVERSIDADE DO PORTO Rua Robe o F ias, s/n | 4200-465 Po o PORTUGAL Tel.: 225081988/225081728; Fax: 225081835 Email: labgeo@ e.up.p ; Websi e: h p://www. e.up.p /sgwww/labgeo 1 In oduc ion The Geo echnical Labo a o y (LabGEO) o he Facul y o Enginee ing o he Uni e si y o Po o (FEUP) is one o he pa icipan s o he INTERNATIONAL PARALLEL TEST ON THE MEASUREMENT OF Gmax USING BENDER ELEMENTS, o ganized and suppo ed by he Technical Commi ee TC-29 o ISSMGE. This ini ia i e p omo es he pa icipa ion o ins i u ions and labo a o ies wo ldwide cu en ly using his s a e-o - he-a echnique, in a se ies o bende elemen es s using he well-known Toyou a sand ( om Japan), acco ding o p e-es ablished speci ica ions. The aim o he In e na ional pa allel es s using bende elemen s is o dissemina e he expe ience o di e en p ac ices (in e ms o da a acquisi ion, analysis and in e p e a ion o esul s) used wo ldwide and o de elop he ecommenda ion o p ocedu es e e ed o bende elemen s es . The p esen epo desc ibes and de ails he esul s ob ained o a sho se ies o es s ca ied ou a he Geo echnical Labo a o y o FEUP, unde he speci ied es ing p ocedu es. 1 2 Backg ound o bende elemen es ing a LabGEO o FEUP The LabGEO o FEUP has been using bende elemen s, since 2000, when a ecen ly acqui ed iaxial sys em equipped wi h bende elemen s and comp ession ansduce s, om ISMES- Enel.Hyd o, was se up and pu in o ope a ion. Subsequen ly, ano he con en ional iaxial cell was locally modi ied and new bende and ex ende elemen s we e implemen ed. These piezoelec ic ansduce s o igina ed om he Uni e si y o B is ol, whe e signi ican e o s on he de elopmen o hese de ices we e being made, unde he supe ision o D . Da id Nash and colleagues (Viana da Fonseca & Fe ei a, 2002). These equipmen s ha e been con inuously in use, mainly o esea ch pu poses. The i s and mos impo an applica ion o his es ing echnique has been he compa a i e assessmen o sampling quali y on young esidual soils om g ani e. This echnique p o ed o show subs an ial di e ences on shea wa e eloci y Vs, hence shea modulus Gmax, o samples e ie ed by di e en sample s, wi h dis inc induced dis u bances (Fe ei a e al., 2004). The o he esea ch opic in ol ing bende elemen s has been he issue o in e p e a ion o he measu emen s. I is widely accep ed ha he e is some unce ain y in he in e p e a ion o he bende elemen aces and ha howe e simple he ansmi e wa e is, a a mo e complex wa e will be ecei ed (Moncas e , 1997). Fo his pu pose, an ex ensi e se o es s mainly on na u al esidual soils has been pe o med, using di e en in e p e a ion me hods, in o de o u he unde s and he di e ences in he esul s, owa ds de e mining which me hod mos closely es ima es he ue a el ime. In his amewo k o esea ch, i is he e o e wi h en husiasm and g ea expec a ions ha he LabGEO o FEUP has emb aced his ini ia i e. 3 Ou line o he es appa a us The iaxial es appa a us was used o ca ying ou hese es s. Cylind ical samples o Toyou a sand we e es ed, wi h 70mm diame e and 140mm heigh . The iaxial es ing de ice, equipped wi h bende elemen s and comp ession ansduce s, was o iginally de eloped a he I alian Resea ch Cen e ISMES, Spa and la e comme cialized by he company Enel.Hyd o. Besides he iaxial chambe i sel , his sys em consis s o a e y simple p essu e con ol panel, a unc ion gene a o , an inpu -ou pu ampli ie wi h da a acquisi ion and an oscilloscope. The con ining p essu e is applied by means o an ai -wa e 2 in e ace, inside he chambe , whe e he comp essed ai is egula ed in he con ol panel. Figu e 1a shows his appa a us in use a he LabGEO o FEUP. 2mm 2mm 136mm 70mm 2mm 2mm 136mm 70mm a) b) Figu e 1 – T iaxial appa a us a LabGEO o FEUP: a) iaxial chambe ; b) scheme o he ip- o- ip dis ance and p o usion o he bende elemen s 4 De ails o he bende elemen s In e ms o he piezoelec ic ansduce s, only he bende elemen s will be he eby de ailed. The bende elemen s ha e been p o ided al eady moun ed in he pla ens o he iaxial cell. Each bende elemen consis s o wo pla es glued oge he and pola ised so ha , when s essed by an elec ical signal, one elonga es and he o he con ac s. Acco ding o he manu ac u e , he dimensions o he ansduce s a e abou 20mm x 10mm wi h a hickness o 3mm. I is possible o di ec ly measu e he p o usion o he bende elemen s, which is app oxima ely 2mm on each pla en. Hence, he ip- o- ip dis ance is calcula ed by sub ac ing 4mm o he sample heigh , as shown in Figu e 1b. As o he elec ical connec ions, he manu ac u e in o ms ha he ansduce s can be used as ansmi e s o as ecei e s since hey ha e been buil in he same way. The elec onic equipmen s used o unning he bende elemen measu emen s consis o : ß A p og ammable unc ion gene a o (TTi, Thu lby Thanda Ins umen s-TG1010) wi h se e al wa e shapes (sinusoidal, squa e, amp o o he ) and con igu a ions (con inuous, epea ing pulses, sweep, e c.); ß An in eg a ed uni wi h inpu and ou pu ampli ie s (speci ically de eloped by ISMES-Enel.Hyd o), which au oma ically ampli ies 40x he inpu ol age and 3 enables he selec ion o ou le els o ou pu ampli ica ion (1x, 2x, 5x and 10x); and, ß An oscilloscope (Tek onix TDS 220), di ec ly connec ed o he ampli ie uni , which eco ds he signals and enables he immedia e iden i ica ion o he a el ime; his de ice is also connec ed o a PC and, by means o speci ic so wa e, i is possible o ans e and sa e he ob ained esul s unde a ious o ma s, o u he analysis and in e p e a ion; ß Al e na i ely o he oscilloscope, a da a acquisi ion wi spec al analysis de ice has been used (PICOScope ADC 216), also connec ed o a PC, whe e speci ically designed so wa e was used o acquisi ion and con ol. In Figu e 2, a pho og aph o he ansduce s and a simpli ied scheme o he layou o he elemen s o he bo om pla en a e p esen ed. 1. Comp ession ansduce 2. Insula ing suppo 3. Epoxy esin 4. Bende elemen 5. Silicon ubbe 6. Elec ical wi es 7. Po ous s one 8. D ainage line 9. S ainless s eel bo om pla en (no o scale) Figu e 2 – Comp ession ansduce s and bende elemen s in he iaxial cell 5 Tes desc ip ion and p ocedu e Fo his In e na ional Pa allel Tes se ies on bende elemen measu emen s, h ee ex ensi e es s ha e been conside ed alid, named RR1, RR2 and RR3. The es p ocedu e speci ica ions p o ided we e closely ollowed. De ails o each es can be ound in he espec i e es o m a ached. Toyou a sand samples we e ai plu ia ed, esul ing on oid a ios o abou 0.70. All hese samples we e d y h oughou es ing. Fou iso opic s ess s ages (50, 100, 200 and 400 kPa) we e applied, acco ding o he ins uc ions p o ided. The nomencla u e used o designa e each es s age was “S ageNumbe -Tes Name”; o example, he hi d s age (200kPa) o he second es is “03-RR2”. 4 6 In e p e a ion o bende elemen measu emen s Time and equency domain measu emen s we e aken, o he applica ion o he di e en in e p e a ion me hods, namely: ß Fi s di ec a i al, a a ious equencies ß Disc e e me hod o phase equencies [π poin s] ß Con inuous sweep inpu equency [ABETS] In wha ollows, each o hese echniques will be b ie ly desc ibed. 6.1 Time domain Fi s di ec a i al me hod, a a ious equencies [ ime domain] The i s di ec a i al me hod is he mos simple, common and usual p ocedu e o in e p e ing bende elemen measu emen s. I consis s on he iden i ica ion o he i s ins an o a i al o he wa e in he ou pu signal, simila ly o he echniques used in geophysical es s (namely C oss-Hole and Down-Hole). While i is some imes easy o de e mine i s a i al, i is o en he cause o much unce ain y. Fo ins ance, A oyo (2001) has es ima ed unce ain ies o up o 100% in es ima ion o he small s ain shea s i ness (Gmax). Many au ho s (Sanchez-Saline o e al., 1986; Viggiani & A kinson, 1995; Jo ičić e al., 1996; A ulna han e al., 1998) ha e epo ed se e al sou ces and ac o s o e o and inhe en nea ield e ec s, which mask and comp omise he iden i ica ion o he a i al poin . Figu e 3 exempli ies a ypical bende elemen ace, whe e he a i al ins an can be de e mined wi hin a ma gin o e o o abou 20%. s 10 V 100 mV s 10 V 100 mV Figu e 3 – De e mina ion o he a el ime wi h he i s di ec a i al me hod 5 Se e al solu ions ha e been p oposed in o de o minimize he subjec i i y o his me hod, mainly by educing he ini ial dis o ion o he signal. This can be achie ed by changing he equency and shape o he inpu wa e, whils ensu ing he accomplishmen o a numbe o echnical equi emen s and bounda y condi ions. Acco ding o Jo icic (2004), hese echnical equi emen s comp ise:  good elec onics equipmen , namely a unc ion gene a o and an oscilloscope wi h sc een;  well shielded and g ounded cables;  p ope ly connec ed and encased bende elemen s;  leak ee connec ions;  noise ee en i onmen ; Special a en ion should be pu on design de ails, leaks and wea -and- ea issues. On he o he hand, he bounda y condi ions e e o he spa ial condi ions (such as he alignmen o he bende elemen s, he e lec ions o he wa e on he edges o he sample, he nea ield e ec o he ela i e dis ance be ween he ansmi e and he ecei e ), he con ac be ween he bende elemen and he soil (which migh induce poo coupling) and he o e shoo ing (a high equencies, he bende elemen changes i s mode shape and he esponse becomes e y complex). In he p esen es s, he i s a i al me hod was used and se e al inpu equencies a ele an le els we e applied, as well as simul aneous P- and S- wa e eco dings, in o de o ake in o conside a ion he nea ield e ec . Figu e 4 – Fi s di ec a i al me hod a a ious inpu equencies, conside ing nea ield e ec 6 This p ocedu e has e idenced he p esence o nea ield e ec , enabling he iden i ica ion o he comp essional componen o he shea wa e. I should be no ed ha , o all he es s ca ied ou a FEUP, he pola i y o he ecei ed wa e is nega i e, i.e., in e ed in ela ion o he inpu wa e. Fo his eason, he i s a i al poin co esponds o he beginning o he i s signi ican ough. 6.2 F equency domain The use o con inuous signals which equi e he shea wa e eloci y o be decoded om measu emen s o ela i e phase o ansmi ed and ecei ed signals is gaining in popula i y (e.g. Blewe e al., 1999). While he echnique is used e y widely ac oss a wide ange o ields, Viggiani and A kinson (1995) we e he i s o apply phase-delay me hod o bende elemen es ing. Acco ding o G eening e al. (2003), hese me hods ha e a numbe o ad an ages o e adi ional pulse-based measu emen s. Chie amongs hese is ha i is con enien o c ea e an algo i hm o de e mine a el ime by es ablishing he g adien o a g aph o phase di e ence agains equency. Phase delay me hods can be pe o med eliably using " adi ional" equipmen i.e. a signal gene a o and oscilloscope (Kaa sbe g, 1975). A con inuous ha monic sinusoid is used as he inpu signal. The equency o he signal is changed and he equencies a which he ansmi ed signal and ecei ed signal a e exac ly in and ou o phase wi h one ano he (so called π-poin s) a e no ed. G eening and Nash (2003) showed ha same in o ma ion could be es ablished less one ously using b oadband (sweep) inpu signal and a spec um analyse . Whe eas he con inuous sweep inpu me hod enables he acquisi ion o a con inuous phase angle e sus equency ela ionship, o he disc e e me hod only speci ic poin s can be obse ed, he π-poin s; hence, his is a simpli ied e sion o he mo e comple e con inuous sweep inpu equency me hod. In he p esen wo k, o he equency domain, wo di e en me hods we e conside ed: he disc e e me hod o phase equencies and he con inuous sweep inpu equency. Disc e e me hod o phase equencies [π poin s] The disc e e me hod o phase equencies me hod consis s on he iden i ica ion o he π- poin s, i.e., he equencies o which he inpu and ou pu signals a e ei he in o ou o phase. The phase angle co esponds o a measu e o his ela ion be ween he inpu and he ou pu signals, aking a alue mul iple o π, when an in o ou o phase ela ion occu s. 7 In he absence o spec al analysis equipmen , he iden i ica ion o he phase equencies is pe o med manually, by de ining disc e e poin s o which he phase angle is known o be mul iple o π. The sequence p ocedu e o he applica ion o his simpli ied app oach o he equency domain can be di ided in he ollowing s ages:  Inpu o a con inuous sine wa e, a a e y low equency, om he unc ion gene a o ;  Selec ion o a XY display in he oscilloscope;  Slow and g adual inc ease o he inpu signal equency, un il a ela ion be ween he wo inpu and ou pu wa es is clea ly isible (simila o any in Figu e 5);  Regis a ion o he equencies alues o he π-poin s, de ined by he obse a ion in he oscilloscope o he Lissajous igu es (Figu e 5), when he ela ion o he wo wa es becomes linea ; his p ocess is epea ed o as long as i can be clea ly obse ed (gene ally no u he han 20kHz): In- and ou -o -phase equencies: π φ k = )(  Plo o hese equency alues e sus he espec i e numbe o wa eleng hs N, which inc ease sequen ially o 0.5 in each phase equency; since he slope is only dependen on he ela i e di e ence be ween he o dina es and, as long as he sequen ial inc ease o 0.5 is main ained, he absolu e alue o he i s N can be a bi a y; Numbe o wa eleng hs: 22 )( k V LL N== ⋅ == π φ λ  De e mina ion o he a e age slope o he ob ained plo , de ined by linea eg ession o all, o a selec ion o , he poin s o he plo ; he slope alue coincides wi h he a el ime. Wa e eloci y: )( 2 L V φ πλ ⋅=⋅= T a el ime: N V L == (≡ slope) x y y xx yy y xx in phase ou o phase a) b) c) d) e) Figu e 5 – Di e en oscilloscope XY displays o he ansmi ed and ecei ed signals (Rio, 2002) 8 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 equency (Hz) a el ime (ms) 0.5 kHz 1 kHz 2 kHz 4 kHz 6 kHz i s a i al i s a i al Figu e 14 – Time cha s o es s age 04-RR1 Besides enabling an o e iew o he a el ime in a con inuous manne , he ime cha also ep esen s a new, use ul and p ac ical way o selec ing he mos adequa e a el ime, wi h less i e a ions and e o . The e o e, i is clea ly an imp o emen o he sweep me hod. Ha ing been concei ed in an a emp o unde s and he i s and mos complex es esul s, i is a his poin conside ed a o una e ui o such inciden . 8 Discussion o he in e p e a ion me hods The i s impo an conclusion o be de i ed om his exe cise is ha he esul s om he equency domain a y signi ican ly o sligh changes in he equency ange, meaning ha he use o hese me hods may equi e an educa ed judgemen and a c i ical assessmen o he ob ained alues. As a inal compa ison, he a el ime esul s o he ime domain and he equency domain ha e been ga he ed in a g aph agains he applied iso opic e ec i e s ess, shown in Figu e 15. The con e gence o he esul s in he highe s esses is e iden . Despi e ollowing simila ends, he compa ison be ween he di e en in e p e a ion me hods esul s in epe i i e es s is no conclusi e. In ac , he di e ences in he esul s o he i s s ess s ages a e qui e a iable among he h ee es s. While he RR1 es , he di e ence be ween he wo domains is almos double, o RR2 i is less han 20%. Bea ing in mind ha hese es s we e ca ied ou in iden ical samples, he e is ce ainly oom o u he in es iga ions. 15 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0 50 100 150 200 250 300 350 400 450 iso opic e ec i e s ess (kPa) a el ime (ms) RR1- i s a i al RR2- i s a i al RR3- i s a i al RR1-sweep RR2-sweep RR3-sweep Figu e 15 – T a el ime esul s o all es s, in he ime and equency domains Such disc epancies in he esul s clea ly in o m ha he e is s ill much o unde s and ega ding how he inpu wa e a els wi hin he sample, whe he and how much o i is e lec ed in i s bounda ies, and how i is picked up by he ecei e on he o he end. In ou opinion, hese se ies o es s ha e shown much unce ain y and many o he agili ies in he cu en ly a ailable in e p e a ion me hods. Ne e heless, hese ha e also led o a con inuing and eno a ed in e es in a deepe esea ch on his inno a i e and p omising labo a o y es . I should be poin ed ou ha , alike all o he es ing equipmen s and echnologies, he e is a scope o ma e ials and condi ions o which he esponse o he sys em is mo e eliable and adequa e. Ou side ha scope o applicabili y, he e is a high p obabili y o inconsis en esul s. We do no know whe he his is he case o he bende elemen es ing me hod. In any case, o he geoma e ials ha e been es ed a labGEO o FEUP as well as in o he labo a o ies, which show less dispe si e and complex esul s. F om his exe cise de i es he need o unde s and in g ea e dep h how he inpu wa e is ac ually a eling wi hin he sample and he u gency o know he igh ack, whe he i is he ime o he equency domain, o any o he . 16 9 Re e ences A oyo, M. (2001). Pulse es s in soil samples. PhD Thesis Uni e si y o B is ol, UK. A ulna han, R.; Boulange , R.W.; Rieme , M.F. (1998). Analysis o bende elemen es s. Geo echnical Tes ing Jou nal, Vol. 21, No. 2, pp. 120-131. Blewe J, Blewe I. J. & Woodwa d P.K (1999). Measu emen o shea wa e eloci y using phase sensi i e de ec ion echniques. Canadian Geo echnical Jou nal, No. 36, pp. 934-939. Fe ei a, C.; Mendonça, A.A. & Viana da Fonseca, A. (2004). Assessmen o sampling quali y in expe imen al si es on esidual soil om g ani e in Po o (in Po uguese). P oceedings o he 9 h Po uguese Geo echnical Cong ess, A ei o, Po ugal, SPG, Lisboa, Vol. 1, pp. 27-38. G eening, P. D.; Nash, D. F. T.; Benahmed, N.; Fe ei a, C. & Viana da Fonseca, A. (2003). Compa ison o shea wa e eloci y measu emen s in di e en ma e ials using ime and equency domain echniques. P oceedings o he 3 d In e na ional Symposium on De o ma ion Cha ac e is ics o Geoma e ials, IS-Lyon '03, F ance. Sep embe 2003. Di Benede o e al. (Eds). Swe s & Zei linge , Lisse. Vol. 1, pp. 381-386. G eening, P.D & Nash, D.F.T. (2003). F equency domain de e mina ion o G0 using bende elemen s. ASTM Jou nal o Geo echnical Tes ing. Jo ičić, V. (2004). Rigo ous bende elemen es ing. Wo kshop on Bende Elemen Tes ing o Soils, UCL, London. Jo icic,V.; Coop, M.R.; Simic, M. (1996). Objec i e c i e ia o de e mining Gmax om bende elemen es s. Géo echnique, Vol. 46, No. 2, pp. 357-362. Kaa sbe g, E.A. (1975). Elas ic-Wa e Veloci y Measu emen s in Rocks and O he Ma e ials by Phase- Delay Me hods. Geophysics No.40, pp. 955-901. Moncas e , A.M. (1997). The shea modulus o sand a e y small s ains. MSc hesis. Depa men o Ci il Enginee ing, Uni e si y o B is ol, UK. Rio, J. (2002). Ad ances in labo a o y geophysics using bende elemen s. In e nal Repo , Uni e si y College London. Sánchez-Saline o I., Roesse J.M. & S okoe II K.H. (1986). Analy ical s udies o wa e p opaga ion and a enua ion Geo echnical epo Viana da Fonseca, A. & Fe ei a, C. (2002) Bende elemen s: excelen labo a o y echniques o he e alua ion o e e encial geo echnical pa ame e s (in Po uguese). P oceedings o he 8 h Po uguese Geo echnical Cong ess, LNEC, Lisboa Viggiani, G. & A kinson, J.H. (1995). In e p e a ion o bende elemen es s. Géo echnique, Vol. 45, No.1, pp.149-154. FEUP, Janua y 2005 17