Experimental study of the transitional behaviour of the silty soils from the Venice Lagoon

(1)

Experimental study of the transitional behaviour of the silty soils

from the Venice Lagoon

By Agnese Marcosanti

Thesis submitted to the Università di Bologna in partial fulfilment of the requirements for the degree of Doctor of Philosophy in the Faculty of Engineering

Supervisors: Guido Gottardi

Co-supervisors:

Matthew Richard Coop Daniele Costanzo

(2)

(3)

INTRODUCTION

In this w ork of thesis the m echanical behaviou r of granu lar soils consisting of mixtures of different percentages of sand and fines is shown. The soils come from the Venice lagoon, p articu larly from Trep orti, a site in the north east of the lagoon. A su m m ary of som e im p ortant contribu tions to research on Venice soils is p resented in Chapter 1.

The m ain aim of the research w as to investigate the resp onse of reconstitu ted sam p les in norm al com p ression and shearing in ord er to id entify a p ossible transitional behaviou r (Martins et al., 2002; N ocilla et al., 2006). Moreover, it w as stu d ied w hether a link betw een the p rop erties at the m acro-scale and m icro-stru ctu ral featu res occu rred . A Scanning Electron Microscop y analysis w as performed and its main objective was to identify, if possible, some patterns or mode of orientation that cou ld exp lain at a m icro-scale level the reason for transitional behaviou r. In other term s, an investigation of m icrostru ctu re, p articu larly microfabric and particle orientations, was undertaken.

(6)

CHAPTER 1 The soils from the Venice lagoon

The soils from the Venice lagoon have long been object of stu d ies, given the historical and cu ltu ral im p ortance of the city and the need to p rotect its d elicate environment.

These soils consist of a highly chaotic interbed d ing of d ifferent sed im ents (Cola and Sim onini, 2002), w hose m ain characteristic is the p resence of a silt fraction, alw ays com bined w ith sand and / or clay. Althou gh the soils p rofiles vary from site to site, the u niqu e geological origin and com m on d ep ositional environm ent lead to qu ite u niform m ineralogical p rop erties. Several d ifferent factors, both natu ral and m an-origined , influ enced the soils d istribu tion w ithin the lagoon. Am ong the others, there are m arine transgression and regression, the d iversion of river m ou ths, the tid al m ovem ents, the w aves action, the w ind d irection, all the d ifferent kind of constructions built in the harbour areas.

Different researches p ointed ou t the high heterogeneity of the Venetian soils, as their most peculiar feature.

1.1. Morphology of the lagoon

The Venice lagoon su rrou nd s the city of Venice in the north-w est of Italy. It is located on an area of arou nd 550 km 2, betw een the Brenta river on the sou th sid e and the Sile river, also know n as the foce d i Piave vecchia (literally Piave river old m ou th) , on the north sid e. On the w est sid e its bord ers m ainly consist in ind u strial and reclaimed areas.

(7)

Fig. 1. 1: general view of the Venice lagoon at present (modified from Ricceri et al., 2002).

On the east side, the area is limited by the coast line defined by, going from south to north, Sottom arina, Pellestrina, Lid o d el Cavallino and Lid o d i Jesolo. The area is arou nd 13 km w id e and 55 km long. The lagoon has three inlets to the op en sea, Lido, Malamocco and Chioggia.

The p resent configu ration and m orp hology of the Venice lagoon, rep resented in the figu re above, are the resu lt of natu ral m od ifications bu t also of several hu m an actions that have altered the sp ontaneou s equ ilibriu m of the lacu strine environment.

(8)

Fig. 1. 2: Evolution of the Lido inlet morphology (Ricceri, 2007)

Those intervents had m ainly the aim of p rotecting the historic city of Venice and of maintaining its typical insularity (Ricceri, 2007).

The most important actions involved:

the d eviation of the Brenta, Sile and Piave rivers into canals on the ou tsid e of the lagoon, whereas their mouths were previously located inside the lagoon; the m od ifications of the inlets m orp hology aim ing at red u cing the effects of the solid transport of sediments on the navigation;

the reinforcing of the coastline.

In figu re 1.2 the evolu tion of the m orp hology of the Lid o inlet area throu gh the centu ries is rep orted . It can be noted that the first m od ifications d ate back to the early 1300s w hen the river d iversions started in ord er to p revent excessive am ou nt of fluvial sediments deposit in the lagoon.

1682

1552

1725 1811

1885 1910

(9)

1.2. Brief geological background

In ord er to stu d y the characteristics of the Venetian su bsoil an overview of the geological history of the area has been attem p ted . Accord ing to the geological and p aleontological history, the soils w ithin the Venice lagoon are characterised by a common mineralogical origin and depositional environment.

The Qu aternary basin of the Venice lagoon is arou nd 900 m d eep . Du ring this p eriod , there w as an alternation of ingression and regression of the sea in the lagoon area, therefore both flu vial and m arine sed im ents can be fou nd . The geological history of the area exp lains the chaotic d istribu tion of the soil sed im ents in layers w ith d ifferent thickness consid ering even sites close to each other. The actu al origin of the Venice lagoon is traced arou nd 6000 years ago, d u ring a m arine transgression called fland rian in the H olocene ep och, w ith the sea w ater d iffu sing into the lacustrine basin (Simonini et al., 2006).

Accord ing to Ricceri (2007), the sed im ents d ep osited d u ring the Qu aternary can be divided into four groups with respect to the depth below mean sea level:

u ntil 5-10 m below m ean sea level: the d ep ositional environm ent of these shallowest deposits is lacustrine and it can be referred to the Holocene epoch; from 5-10 m u ntil 50-60 m below m ean sea level: the d ep ositional environm ent is the result of the last Wurmian glaciations that characterised the superior Pleistocene epoch;

from 50-60 m until 300 m below sea m ean level: the d ep ositional environm ent is the resu lt of alternating lacu strine, continental and m arine sed im ents from su p erior Pleistocene epoch;

from 300 m u ntil 900-950 m below m ean sea level: the d ep ositional environm ent is mainly marine from inferior Pleistocene and Pliocene

(10)

Fig. 1. 3: reconstruction of the emerged land during Pliocene epoch in the moment of the largest marine ingression (modified from Ricceri, 2007).

Fig. 1. 4: reconstruction of the emerged land in the moment of the largest marine transgression, during the last Wurmian glaciation (modified from Ricceri, 2007).

Sed im entological stu d ies p ointed ou t that, becau se of the intrinsic d iscontinu ou s natu re of the lacu strine environm ent, it w as p ossible to d istingu ish betw een the

Rome

Venice

Coast line at present

S E A Emerged land

Emerged land Wurmian glaciers I C S E A

(11)

H olocene and the su p erior Pleistocene d ep osits only u sing a com bination of sed im entological, p aleontological, geotechnical and m ineralogical investigations among the others (Belloni et al. 2007).

Particu larly interesting for the research carried ou t in this thesis is the analysis of the u p p er hu nd red m etres below sea level, that consist in a chaotic alternation of sand , silt and clay layers d ep osited ap p arently w ithou t any p attern d u ring the su p erior Pleistocene, esp ecially d u ring Wu rm ian glaciations (Favero et al, 1973). The shallowest 10-15 metres were deposited during Holocene epoch.

The u p p er Wu rm ian d ep osit is a layer 5-12 m d eep know n as caranto. This consists of a cru st of highly overconsolid ated very silty clay, originated by essiccation p henom ena d u ring the last Pleistocenic glaciation (Sim onini et al., 2002). This is w here the historical venetian bu ild ings w ere fou nd ed becau se of its good m echanical p rop erties. The caranto is consid ered as a benchm ark betw een the Holocenic lacustrine deposits and the continental Pleistocenic sands, silts and clay. In figu re 1.5 a reconstru ction of the m ost recent p assages in the evolu tion of the Venice lagoon geology is shown through sections of the modelled subsoil.

In p articu lar, the continental d ep osits from the Pleistocene ep och and the lacu strine and coastal d ep osits from the H olocene ep och can be d istingu ished . The Pleistocenic continental sed im ents are characterised by an intrinsic heterogeneous natu re w hich is the m ost p ecu liar featu re of the Venetian soils, as it w ill be seen in the following paragraphs.

The d ep ositional p atterns of the Venetian sed im ents are therefore qu ite d ifficu lt to describe because different soil types correspond to the same depth in different areas of the lagoon.

(12)

Fig. 1. 5: Overview of the most recent geological phases with particular respect to the superior Pleistocene and the Holocene (modified from Gatto and Previatello, 1974).

In figu re 1.6 below , typ ical soil p rofiles of the Venice lagoon are rep orted . These d ata are rep resentative of a collection of the m ost im p ortant com p rehensive geotechnical and geological studies carried out on Venice soils.

lacustrine basin

the lagoon at the beginning of the Holocene

the lagoon at present

During the Wurmian glaciations (18000 years ago) the area consisted of a fluvial and marshy environment.

After the glaciations the area was subjected to a huge flooding.

The area becomes again a marshy and fluvial environment for the following 10000 years.

Marshes formation during a marine regression around 6000 years ago.

vegetation and peat deposits

coastal sands and lacustrine silts

marshy deposits

unconsolidated soil layers from the same epoch as Caranto

Overconsolidated sediments known as caranto

other continental sediments HOLOCENIC LACUSTRINE AND COASTAL

DEPOSITS

(13)

Fig. 1. 6: Typical soil profiles of the Venice lagoon (from Simonini et al., 2006).

1.3. Literature review of the Venice lagoon soils

The soils of the Venice lagoon have long been the object of stu d y, d u e to several d ifferent issu es that have d evelop ed in the area in the last sixty years and contribu ted to its p recariou s equ ilibriu m (Colom bo, 1970). The first researches aim ing at characterising the geology and the geotechnical p rop erties of the lagoon soils have been u nd ertaken betw een 1965 and 1975, p articu larly Bonatti in 1968 carried ou t m ineralogical and p aleontological stu d ies on the sed im ents at a site called Motte di Volpego.

The Venice lagoon has long been su bjected to su bsid ence, both natu ral and m an-ind u ced , the latter increased by the an-ind iscrim inate u se of the w ater sou rces in the su bsoil. Particu larly d u ring the 70s, great attention has been given to the issu es of

(14)

the Venice lagoon, esp ecially becau se of the su bsid ence effects on the frequ ency of flooding of the historic city, phenomenon known as acqua alta,.

The Italian Governm ent throu gh the Consiglio N azionale d elle Ricerche (CN R), u nd er the ad vice of UN ESCO, com m issioned a 1000 m d eep borehole, located at Tronchetto, w ith the aim of getting an insight of the geotechnical p rop erties and of the aqu ifer and aqu itard system , in ord er to p revent the increase of this issu e ( Favero et al. 1973; Row e 1973; Ricceri & Bu tterfield 1974). Stu d ies necessary for the d esign of the fou nd ations of new large ind u strial settlem ents on the m ainland lead to a thorou gh know led ge of the shallow est grou nd of the lagoon, betw een 70s and 80s, esp ecially w ith resp ect to the geotechnical p rop erties obtained throu gh standard tests.

In the following years several commissioned were set up by the Italian Government with the aim of protecting the historic city of Venice and in 1989 a project involving the constru ction of m ovable gates at the three lagoon inlets w as ou tlined (Gentilom o 1997; H arlem an et al., 2000; Sanzeni, 2006). These gates control the tidal flow and when particularly high tides occur, they have to temporarily separate the lagoon from the sea (Harleman et al. 2000).

Therefore, p relim inary geotechnical investigations w ere necessarily carried ou t to d raw relevant soil p rofiles and cross-sections at the inlets, thu s to achieve a su itable d esign of the m ovable gates fou nd ations. The resu lts of these analyses show ed that, althou gh the stratigrap hy is very heterogeneou s, there is a p red om inance of a silt fraction, com bined w ith clay and / or sand d iffu sed in the su bsoil of the lagoon. Moreover the sediments present similar mineralogical features.

Given the relevant heterogeneity of soil layering, a few test sites w ere selected as representative of the Venetian lagoon soils.

The tests carried ou t at the Malam occo Test Site aim ed essentially at investigating the soil p rop erties by m eans of in-situ investigations, w hich inclu d ed boreholes, p iezocone and d ilatom eter tests (Cola & Sim onini 1999, 2002). Som e researchers tried to u se the d ata com ing from these tests to evalu ate the reliability of the m ost w id ely u sed charts or correlative equ ations for the p iezocone and d ilatom eter interp retation on the basis of the com p arison w ith the resu lts of the laboratory tests (Sim onini & Cola 2000; Ricceri et al. 2002). The heterogeneou s layering how ever, p revented from obtaining certain correlations. A com p rehensive laboratory test p rogram w as com p leted on sam p les com ing from at Malam occo Test Site (Cola & Sim onini 2002; Sim onini et al., 2006; Biscontin et al. 2001, 2006). The results highlighted the sensitiveness of sam p les to stress relief and sam p ling d istu rbance. Moreover the heterogeneity of the Venice sed im ents resu lted in the need of a qu ite large number of tests in order to define accurately the basic soil properties.

(15)

Another site w as chosen, the Trep orti Test Site located close to the Lid o inlet. This site w as d ed icated to the stu d y of in-situ the stress-strain-tim e p rop erties of the heterogeneou s Venetian soils (Sim onini, 2004, 2006). With this aim , a circu lar em bankm ent w as constru cted , ap p lying a load of 100 kPa, and the t m easu rem ents of the ground displacements and of pore pressure evolution were collected.

Boreholes w ith u nd istu rbed sam p ling, trad itional CPTU (Gottard i & Tonni 2004) and DMT (Marchetti et al. 2004), seism ic SCPTU and SDMT (Mayne & McGillivray 2004) w ere em p loyed to characterize soil p rofile and estim ate the soil p rop erties for comparison with those directly measured in situ.

Fig. 1. 7: View of the Venice lagoon with the locations of some test sites (from Simonini et al., 2006).

Other stu d ies carried ou t on the Venice lagoon soils com p rise the w ork by Cola (1994), w ho collected d ata regard ing the m echanical p rop erties of clayey and silty soils at Fusina.

Moreover, Biscontin et al. (2006) assessed the com p ressibility of Venetian soils, as natu ral silty clayey soils, in ord er to p resent a nu m ber of constitu tive law s, relating the influ ence on the norm al com p ression behaviou r of a p article d istribu tions w ith varying coarse-grained and fines fraction.

(16)

1.4.

Althou gh the soil layers d istribu tion varies greatly from site to site of the lagoon and it is d ifficu lt to ind icate a soil p rofile , the sam e cannot be said for the basic soil features.

The soil p rofi

2002). and the Malam occo test site can be consid ered typ ical and rep resentative of the entire lagoon area. Several geotechnical investigations have been u nd ertaken at Malamocco, within t

inlets of the lagoon. Accord ing to the resu lts of these analyses d ata are available regard ing the grain size d istribu tions, shap e of the p articles and basic soil properties.

Basic material properties

Althou gh the soil layers d istribu tion varies greatly from site to site of the lagoon and it is d ifficu lt to ind icate a soil p rofile , the sam e cannot be said for the basic soil features.

The soil p rofi

2002). and the Malam occo test site can be consid ered typ ical and rep resentative of the entire lagoon area. Several geotechnical investigations have been u nd ertaken at Malamocco, within t

inlets of the lagoon. Accord ing to the resu lts of these analyses d ata are available regard ing the grain size d istribu tions, shap e of the p articles and basic soil properties.

Basic material properties

Althou gh the soil layers d istribu tion varies greatly from site to site of the lagoon and it is d ifficu lt to ind icate a soil p rofile , the sam e cannot be said for the basic soil

The soil p rofiles at the Malam occo inlet have been w id ely stu d ied (Sim onini et al., 2002). and the Malam occo test site can be consid ered typ ical and rep resentative of the entire lagoon area. Several geotechnical investigations have been u nd ertaken at Malamocco, within the p roject aim ed at the constru ction of the m ovable gates at the inlets of the lagoon. Accord ing to the resu lts of these analyses d ata are available regard ing the grain size d istribu tions, shap e of the p articles and basic soil

Basic material properties

les at the Malam occo inlet have been w id ely stu d ied (Sim onini et al., 2002). and the Malam occo test site can be consid ered typ ical and rep resentative of the entire lagoon area. Several geotechnical investigations have been u nd ertaken at he p roject aim ed at the constru ction of the m ovable gates at the inlets of the lagoon. Accord ing to the resu lts of these analyses d ata are available regard ing the grain size d istribu tions, shap e of the p articles and basic soil

Basic material properties

les at the Malam occo inlet have been w id ely stu d ied (Sim onini et al., 2002). and the Malam occo test site can be consid ered typ ical and rep resentative of the entire lagoon area. Several geotechnical investigations have been u nd ertaken at he p roject aim ed at the constru ction of the m ovable gates at the inlets of the lagoon. Accord ing to the resu lts of these analyses d ata are available regard ing the grain size d istribu tions, shap e of the p articles and basic soil Althou gh the soil layers d istribu tion varies greatly from site to site of the lagoon and it is d ifficu lt to ind icate a soil p rofile , the sam e cannot be said for the basic soil

les at the Malam occo inlet have been w id ely stu d ied (Sim onini et al., 2002). and the Malam occo test site can be consid ered typ ical and rep resentative of the entire lagoon area. Several geotechnical investigations have been u nd ertaken at he p roject aim ed at the constru ction of the m ovable gates at the inlets of the lagoon. Accord ing to the resu lts of these analyses d ata are available regard ing the grain size d istribu tions, shap e of the p articles and basic soil

(17)

Fig. 1. 8: soil profile at Malamocco test site. It can be noted that the ground level is at 10.5 m below mean sea level (modified from Simonini et al., 2006).

Three m ain soil typ es have been id entified in the irregu lar alternation of sed im ents: sand , silt and very silty clay w ith a few thin layers of com p acted p eat (Ricceri et al., 2002; Sim onini et al., 2006). Accord ing to the Unified Soil Classification System these groups were described as:

medium to fine sand (SP-SM); silt (ML);

very silty clay (CL).

In figu re 1.8, it can be noted that the sand fraction is qu ite u niform , bu t m oving tow ard s finer m aterials the grain size cu rves d isp lay a larger range of p article diameters and the soil becomes more graded (Cola and Simonini, 2002).

Fig. 1. 9: Typical grain-size distributions of the groups SM-SP, ML and CL at Malamocco test site (from Simonini et al., 2002).

From p article size analysis, the three classes of soils occu r, at Malam occo, in p rop ortions of 35% sand (SM-SP), 20% silts (ML), 40% very silty clays (CL) and 5% m ed iu m p lasticity clays and p eat (CH , OH and Pt). Percentages of silts higher than 50% are present in the 65% of the analysed samples.

With resp ect to the w hole soil p rofile at Malam occo (see figu re 1.6) the d iam eter corresp ond ing to the 60% of p assing p articles, D 60 _{varies from arou nd 0.005 m m to}

GRAVEL SAND SILT CLAY

Particle size, mm Percentage

(18)

0.1 m m ; w hereas the D10_{varies w ithin the range 0.0001 m m and 0.1 m m . Therefore} it can be noticed that both p aram eters have an oscillation of arou nd tw o ord ers of magnitude.

In ord er to take into consid eration the fact that w ith d ecreasing d iam eter, the soil sam p les are m ore grad ed , thu s the non- u niform ity coefficient U d ecreases, Cola and Sim onini (2002) introd u ced a new grain size ind ex IGS_{, cou p ling the op p osite}

variation of D50_andU_{with depth. This parameter is defined as:}

U D D IGS 0 50 /

where D 0 _{is a reference diameter equal to 1 mm.}

Fig. 1. 10: percentages of the three classes of soils at Malamocco test site (from Simonini et al., 2002).

The average sp ecific gravityGS_{is arou nd 2.77±0.03. The liqu id lim it of the cohesive}

fraction is on average 36.9% and the plasticity index is on average 14.7%.

In figu re 1.10 typ ical grain size com p ositions of the three fractions are show n. For the sand fraction (SP-SM) tw o com p ositions are rep resented , each referring to one of the tw o d ifferent m ineralogical com p osition, the carbonatic (a) and the siliceou s (b), as it will be seen in details in the following paragraph.

1.5. Mineralogical analysis

One of the m ost im p ortant p rocess that originated the soils from the Venice lagoon is the flu vial d ep osition of sed im ents. Within these d ep osits, tw o d istinct typ es of

(19)

p etrograp hic sou rces w ere fou nd : one consisting in the granitic p rovince and one consisting in the lim estone-d olom ite p rovince. The Po and Ad ige rivers contribu te to the first p etrograp hic sou rce, som etim es called Pad ana p rovince, that has a m ainly siliceou s-clastic com p osition, w hereas the Brenta, Piave, Livenza and Tagliam ento rivers contribu te to the second one, often called Veneta p rovince, w hich consists m ainly of carbonate sed im ents, w ith a p red om inance of d olom ite over calcite (Favero, 1973).

Fig. 1. 11: mineralogical compositions with respect to the grain size distributions of the three fractions. It can be noted that for the sand fraction two mineralogical composition are displayed. (modified from Cola and Simonini, 2002).

One of the m ost recent stu d y carried ou t to d eterm ine the m ineralogical com p osition of the soils from the Venice lagoon w as p erform ed on sam p les com ing from Malam occo test site throu gh X-ray d iffractom etric techniqu e by Cu rzi in 1995. The analysis w as u nd ertaken on three d ifferent kind s of sam p les com ing from a borehole 60 m -d eep . The resu lts are su m m arised in the grap h below . The m ineralogical com p osition of the bu lk sam p les, of the sand and of the clay fraction are reported.

(20)

Quartz and Feldspar Smectite CloriteKaolinite Quartz Feldspar Rock fragments Carbonates Carbonates Clay minerales Illite 5 10 15 20 25 30 35 40 45 50 55 60 D e p th b e lo w m .s .l .,

m 0 20 40 60 80 100

Bulk Sample Composition, %

0 20 40 60 80 100

Clay Fraction Composition, %

0 20 40 60 80 100

Sand Fraction Composition, %

Fig. 1. 12: mineralogical composition obtained from X Ray-diffractometric analysis referring to bulk samples, sand and clay fraction from Malamocco test site (from Simonini et al., 2006)

Consid ering the overall com p osition of the sed im ents (bu lk sam p les), carbonates, consisting m ainly in a m ixtu re of calcite and d olom ite, are the m ost significant component, but quartz, feldspar, muscovite and chlorite are quite abundant as well. It can be noted that the p ercentage of qu artz and feld sp ar increases w ith increasing d ep ths. The clay m inerals fraction is on average low er or equ al to 20%, m ainly com p osed of illite (50-60%), w ith chlorite, kaolinite and sm ectites as second ary minerals.

In the sand fraction, carbonates and qu artz are the m ain m ineralogical com p onents, w ith feld sp ar and rock fragm ents. Within the sand fraction the m inerals distribution at different depths is less uniform than the clay fraction

It can be show n that w ith d ecreasing p article d iam eter from sand to clay, the p ercentage of clay m inerals (illite, chlorite, kaolinite and sm ectite) increases whereas the carbonate and quartz-feldspar content decreases.

In the table below , the m ineralogical com p onents of bu lk sam p les rep resentative of d ifferent d ep ths at Malam occo test site are show n. It can be seen that even for sam p les com ing from d ep th valu es very close to each other the m ineralogy can be very d ifferent, as it is the case of the sam p le com ing from 34.88 m , w hich has carbonates as the d om inant fraction and the sam p le com ing from 37.82 m , w hich has quartz as the major component.

(21)

Sample depth (m) Calcite and Dolomite (%) Quart z (%) Feldspar (%) Muscov ite (%) Chlorit e and Kaolini te (%) Chlori te/ Smecti te (%) Illite/ Smecti te (%) 8.45 44.7 40.5 3.4 5.7 3.7 2.0 - 34.88 22.4 33.6 24.6 14.0 5.4 - - 59.51 19.5 66.3 9.9 1.1 2.2 - 1.0 37.82 72.2 16.5 8.7 1.3 1.3

Table 1. 1. Major components of some samples from MTS (selection of data from Simonini et al. 2006)

Fig. 1. 13: mineralogical composition of bulk samples at different depths from Malamocco test site (data from Curzi, 1995 and Simonini et al., 2006)

A com p arison betw een the d ifferent m ineralogical com p ositions fou nd in the soil p rofiles at fou r d ifferent sites (Motte d i Volp ego, Tronchetto, Fu sina, Malam occo and Trep orti ) w as carried ou t by Sim onini et al. (2006). Accord ing to this stu d y, the

0 10 20 30 40 50 60 70 80 90 100 8.45 m 34.88 m 37.82 m 59.51 m m in e ra l c o m p o si ti o n , %

Calcite and Dolomite (%) Quartz (%)

Feldspar (%) Muscovite (%)

Chlorite and kaolinite (%) Chlorite and Smectite (%) Illite/Smectite (%)

(22)

carbonate m inerals, Calcite and Dolom ite, are the m ajor com p onent w ith p ercentages higher than 50-70%. Only a thick layer of arou nd 3-5 m w ith a low carbonate content w as id entified at d ep th varying betw een 20 and 27 m below m ean sea level. A few other d ep osits w ith low carbonate content are fou nd at higher and low er d ep ths in the lagoon area, esp ecially at Malam occo test site. It is generally acknow led ged that soil layers p oor in carbonates, bu t characterised by p eat and organic m atter, can be attribu ted to lacu strine sed im entation ep isod es (Simonini et al., 2006).

Fig. 1. 14: carbonate content values at four different sites (see also figure 1.6), Motte di Volpego, Tronchetto, MTS and TTS (modified from Simonini et al., 2006).

0 5 10 15 20 25 30 35 40 45 50 55 60 0 20 40 60 80 100 D e p th b e lo w m e a n s e a le ve l, m Carbonate content, (%) Motte di Volpego Tronchetto Malamocco Treporti

(23)

The shap es of the p articles

scanning electron m icroscop y. In figu re 1.12 and 1.13 below , an overview of their shap es can be ap p reciated . The m ean rou nd ing ind ex of the grains of both the sand and the silt fractions ranges betw een 0.

to affect this p aram eter. Am ong the d ifferent m inerals, qu artz and feld sp ar grains are generally angular.

Fig. 1.

samples at Malamocco test site (modifie

Fig. 1. 15: scanning electron micrographs of the sand and clay fraction obtain samples at Malamocco test site (modifie

: scanning electron micrographs of the sand and clay fraction obtain samples at Malamocco test site (modifie

a. SAND FRACTION

100

The shap es of the p articles w ithin the tw o fractions w ere analysed by m eans of scanning electron m icroscop y. In figu re 1.12 and 1.13 below , an overview of their shap es can be ap p reciated . The m ean rou nd ing ind ex of the grains of both the sand and the silt fractions ranges betw een 0.

to affect this p aram eter. Am ong the d ifferent m inerals, qu artz and feld sp ar grains

: scanning electron micrographs of the sand and clay fraction obtain samples at Malamocco test site (modified from Cola and Simonini, 2002).

SAND FRACTION

100 m

w ithin the tw o fractions w ere analysed by m eans of scanning electron m icroscop y. In figu re 1.12 and 1.13 below , an overview of their shap es can be ap p reciated . The m ean rou nd ing ind ex of the grains of both the sand and the silt fractions ranges betw een 0.23 and 0.34. The p article size d oes not seem to affect this p aram eter. Am ong the d ifferent m inerals, qu artz and feld sp ar grains

: scanning electron micrographs of the sand and clay fraction obtain d from Cola and Simonini, 2002).

w ithin the tw o fractions w ere analysed by m eans of scanning electron m icroscop y. In figu re 1.12 and 1.13 below , an overview of their shap es can be ap p reciated . The m ean rou nd ing ind ex of the grains of both the sand 23 and 0.34. The p article size d oes not seem to affect this p aram eter. Am ong the d ifferent m inerals, qu artz and feld sp ar grains

b.SILT FRACTION

100

: scanning electron micrographs of the sand and clay fraction obtained from the bulk

SILT FRACTION

100 m

ed from the bulk w ithin the tw o fractions w ere analysed by m eans of scanning electron m icroscop y. In figu re 1.12 and 1.13 below , an overview of their shap es can be ap p reciated . The m ean rou nd ing ind ex of the grains of both the sand 23 and 0.34. The p article size d oes not seem to affect this p aram eter. Am ong the d ifferent m inerals, qu artz and feld sp ar grains

(24)

CHAPTER 2 The mechanical behaviour of soils

Introduction

In this chap ter the m ain theorical fram ew ork of reference for m echanical behavior (d rained and u nd rained ) of soils is p resented , p aying p articu lar attention to granular soils.

The critical state fram ew ork theory, established thanks to the w ork of Rend u lic, Skempton and Casagrande among the others, is reviewed for clays and sands. Several factors influ ence the m echanical behaviou r of soils: the influ ence of the initial d ensity, for granu lar soils, and of the stress history, for cohesive soils, on their shear strength and strain are described.

2.1. The Critical state framework

The Critical State fram ew ork w as d evelop ed thanks to the w orks of Roscoe, Schofield , Wroth and other researchers from Cam brid ge University. This theory allow s to d efine a com p rehensive view of the m echanics of soils, becau se it takes into accou nt the stress history, the stress p aths and the initial characteristics of the material. This theory is p ow erfu l also w ith resp ect to its ap p licability to both cohesive and granu lar soils, for d rained and u nd rained cond itions. Within this fram ew ork a p articu lar state, called Critical State is d escribed and to this cond ition soils subjected to shear stresses aim, irrespectively of their initial conditions.

The stu d ies that lead to the d evelop m ent of this theory w ere first carried ou t on reconstituted clays. According to Burland, (1990) a reconstituted clay can be defined as one that has been thorou ghly m ixed at a w ater content equ al or greater than the liquid limit wL_.

Roscoe et al., (1958) id entified a u niqu e N orm al Com p ression line (N CL) for reconstitu ted clays. that sep arates norm ally consolid ated and over-consolidated states. Moreover, consid ering shearing behaviou r, a Critical State Line (CSL) w as id entified at high strains. and it rep resents the locu s of the stress-strain states reached eventu ally in d rained and u nd rained shearing. It can be consid ered the locus of points that separate contractive and dilative behaviour.

(25)

Follow ing this research carried ou t on clays, the m echanics of sand s, w as investigated. The Critical State framework (Stationary State theory) was found to be ap p licable even to granu lar m aterials. Loose and d ense sam p les w ithin a w id e range of d ifferent initial void ratios converge to a u niqu e norm al com p ression line, that ru ns p arallel to the critical state line d eterm ined from shearing behaviou r at high stress levels.

It has to be p ointed ou t that althou gh the Critical State fram ew ork is ap p licable to both sand s and clays, the com p ression m echanism s involved are d ifferent. In clays the com p ression m od e of behaviou r relies on su p erficial, electro-static forces, whereas in sands it is due to particle breakage (Coop and Lee, 1993).

A first hint of the exp erim ental find ings is rep orted below . In figu re 2.1, the behaviou r of a sand at d ifferent initial d ensities is rep orted , p articu larly the shear strength , the volu m etric strain v _{and the void ratio as fu nctions of the axial}

strain, for d ense and loose sand s u nd er shear stresses. It can be seen that the m echanical behaviou r of loose and d ense sand s show s a great d ep end ence on the initial void ratio: a d ense sand show s first a p eak strength, then w ith increasing stresses a critical state is reached at high strains (strain softening behaviou r). For loose sand s a m onotonic increase of shear strength is observed u ntil the critical state condition is reached without any peak resistance (strain hardening behaviour). To the p eak cond ition corresp ond s a p eak shear strength angle 'p

, w hereas at the critical state this p aram eter is equ al to 'c_{. Loose sand s only have the critical shear}

(26)

Figure 2. 1 overall view of the shearing response of loose and dense sands (from Head, 2004). Both sand s reaches the critical state cond ition C, characterised by strains at constant volume and constant shear stresses.

Thus the Critical state and a critical void ratio e c _{is reached by both loose and dense}

sand s, thu s it d oes not d ep end on the initial d ensity. Moreover, nevertheless the different resp onse in shearing (d ilating for d ense sam p les,ec e0_{, and contractive} for loose sam p les,ec e0_{), as soon as they both finally reach the critical state, they} show a constant volume.

It can be conclu d ed that the critical state is an intrinsic p rop erty of the m aterial and d oes not d ep end on the initial d ensity, bu t only on the p rincip al effective stress, w hereas the p eak shear strength d ep end s on the d ilatancy, thu s on the initial void ratio. Clays tend to a critical state as w ell, bu t for these soils the mechanical resp onse in shearing is linked to the stress history and not to the initial d ensity, as seen for granular soils.

As a m atter of fact, the stu d ies by the Cam brid ge grou p focu sed on the m echanics of reconstitu ted clays and these stu d ies lead to d eep interest on the critical state of cohesive and granular soils (see paragraph 2.2).

(27)

As it w ill be d escribed in the follow ing p aragrap hs, the stress-strain behaviou r of norm ally consolid ated clays can be com p ared to the one of loose sand s and the over-consolidated clays one can be compared to the dense sands behaviour.

Figure 2. 2: Stress-strain behaviour in shearing of normally consolidated clays and over-consolidated clays in drained triaxial tests (from Head, 2004).

In ord er to characterize the m echanics of cohesive and granu lar soils, triaxial tests are p erform ed in d rained and u nd rained cond itions. If the shearing p hase is d rained , the effective stresses coincid es w ith the total stresses. For granu lar soils, this is usually the case: given their high p erm eability the ind u ced excess p ore pressures are quickly dissipated, whereas if the test is carried out on a cohesive soil, the load shou ld be ap p lied slow ly enou gh for the generated excess p ore p ressu res to dissipate.

If the shearing p hase is u nd rained , the d rainage is closed and the shearing p hase is p erform ed w ithou t any volu m e change. Thu s it is necessary to m easu re the excess pore pressures in order to determine the effective stress variation.

(28)

2.2. Mechanical behaviour of clays

In this p aragrap h, a review of the d ifferent researches that contributed to the development of the Critical State framework is presented. Particularly, the results of d rained and u nd rained triaxial tests on reconstitu ted , and su bsequ ently natu ral, clay lead to the identification of a unique state surface.

2.2. 1 Reconstituted clays

The need of a theoretical fram ew ork able to d escribe the m echanical behaviou r of clays and to be a reference for both sam p les in their reconstitu ted / rem ou ld ed and in their natural states, was one of the first challenge for soil mechanics researchers. Mod ern soil m echanics is based on the stu d ies by Rend u lic and H vorslev carried out in the 1930s on the properties of reconstituted natural soils or artificial materials su ch as illite or kaolinite. These w orks w ere the starting p oint for the d evelop m ent of the critical state framework by the Cambridge soil mechanics school.

In 1936 Rend u lic p erform ed the first laboratory stu d ies on the com p ressibility an d shear strength p rop erties of reconstitu ted norm ally consolid ated clays, Wiener Tegel clay am ong the others. H e focu sed on the relationship betw een the void ratio and the stress states d u ring d rained and u nd rained shearing p aths, both in compression and in extension.

From the results of tests in undrained and drained shearing conditions, he observed a u niqu e failu re line, w hich in the follow ing years w ou ld have been d escribed as Critical State line.

Follow ing Rend u lic s ap proach, H enkel (1956, 1960) carried ou t a com p rehensive laboratory p rogram consisting of triaxial tests on reconstitu ted sam p les of Weald clay. H is find ings essentially confirm ed Rend u lic s resu lts. H e hyp othesised that, if the stress p aths from d rained and u nd rained tests crosses each other on the q-p plane, they belong to the surface referring to a constant void ratio (specific volume). H e obtained contou rs of constant w ater content for both d rained and u nd rained tests. Du ring d rained shearing, the d rainage is allow ed and thu s the void ind ex (or the w ater content) varies w ith the volu m etric strain. For the several d rained stress p aths the w ater content contou rs (cu rves corresp ond ing to equ al w ater contents) were drawn.

(29)

A u niqu e

w ater content contou rs w ere fou nd to coincid e w ith the u nd rained tests stress paths.

Figure 2.

observed by Henkel in 1960 (from Atkinson and Bransby, 1978) With resp ect to the com p ression behaviou r, o

the

p resented by Skem p to

sam p les of estu arine clay from Gosp ort. higher than

These resu lts w ere also confirm ed in 1959 by Leonard s and Ram iah, w ho

ou t com p ression tests on the sam e m aterial. The norm al com p ression lines w ere found to converge for stress values of around 100 kPa.

A u niqu e su rface of constant void

w ater content contou rs w ere fou nd to coincid e w ith the u nd rained tests stress paths.

Figure 2. 3: contours of constant water content for drained and undrained tests on Weald c observed by Henkel in 1960 (from Atkinson and Bransby, 1978)

With resp ect to the com p ression behaviou r, o the convergence of cu rves referring to

sam p les of estu arine clay from Gosp ort. higher than 100 kPa

su rface of constant void

w ater content contou rs w ere fou nd to coincid e w ith the u nd rained tests stress

: contours of constant water content for drained and undrained tests on Weald c observed by Henkel in 1960 (from Atkinson and Bransby, 1978)

With resp ect to the com p ression behaviou r, o convergence of cu rves referring to

sam p les of estu arine clay from Gosp ort.

100 kPa, as it can be seen in figure 2.3.

su rface of constant void

w ater content contou rs w ere fou nd to coincid e w ith the u nd rained tests stress

With resp ect to the com p ression behaviou r, o convergence of cu rves referring to

p resented by Skem p ton in 1943, stu d ying norm ally consolid ated reconstitu ted sam p les of estu arine clay from Gosp ort.

, as it can be seen in figure 2.3.

su rface of constant void ratio cou ld be d eterm ined , given the fact that the w ater content contou rs w ere fou nd to coincid e w ith the u nd rained tests stress

With resp ect to the com p ression behaviou r, o

convergence of cu rves referring to sam p les w ith d ifferent initial void ratios is n in 1943, stu d ying norm ally consolid ated reconstitu ted sam p les of estu arine clay from Gosp ort.

ratio cou ld be d eterm ined , given the fact that the w ater content contou rs w ere fou nd to coincid e w ith the u nd rained tests stress

: contours of constant water content for drained and undrained tests on Weald c observed by Henkel in 1960 (from Atkinson and Bransby, 1978).

With resp ect to the com p ression behaviou r, one of the first laboratory evid ences of sam p les w ith d ifferent initial void ratios is n in 1943, stu d ying norm ally consolid ated reconstitu ted sam p les of estu arine clay from Gosp ort. A convergence w as fou nd for stresses

: contours of constant water content for drained and undrained tests on Weald c

ne of the first laboratory evid ences of sam p les w ith d ifferent initial void ratios is n in 1943, stu d ying norm ally consolid ated reconstitu ted A convergence w as fou nd for stresses

: contours of constant water content for drained and undrained tests on Weald c

ne of the first laboratory evid ences of sam p les w ith d ifferent initial void ratios is n in 1943, stu d ying norm ally consolid ated reconstitu ted A convergence w as fou nd for stresses

ou t com p ression tests on the sam e m aterial. The norm al com p ression lines w ere ratio cou ld be d eterm ined , given the fact that the w ater content contou rs w ere fou nd to coincid e w ith the u nd rained tests stress

: contours of constant water content for drained and undrained tests on Weald clay,

ne of the first laboratory evid ences of samples with different initial void ratios is n in 1943, studying normally consolidated reconstituted A convergence was found for stresses

These results were also confirmed in 1959 by Leonards and Ramiah, who carried ou t com p ression tests on the sam e m aterial. The norm al com p ression lines w ere ratio could be determined, given the fact that the water content contours were found to coincide with the undrained tests stress

ne of the first laboratory evidences of samples with different initial void ratios is n in 1943, studying normally consolidated reconstituted A convergence was found for stresses

carried ou t com p ression tests on the sam e m aterial. The norm al com p ression lines w ere

(30)

Figure 2. 4: compression tests on reconstituted Gosport clay samples (Skempton, 1943).

Roscoe et al. in 1958 p resented the first com p rehensive analysis of com p ressibility and strength of reconstituted clay samples.

With resp ect to the norm al com p ression behaviou r, they rep orted the resu lts on the '

log

e _(orv log ' _{) chart and d rew the norm al com p ression line (N CL). The} normal compression line represents the boundary of the possible states in the e (w)- logp chart. The p oints on this line rep resents stress-strain states of norm ally consolid ated clays, w hereas the p oints below rep resents over-consolid ated clays. Soils w hose behaviou r follow the norm al com p ression line have m ainly p lastic volu m etric strain, w hereas on the u nload ing cu rve overconsolid ated soils are subject to elastic deformation.

Given these experimental evidences from the stress paths in drained and undrained com p ression and shearing, they hyp othesised and d em onstrated the existence of a state bou nd ary su rface on the v p' q, w hich coincid es w ith the Roscoe/ Rend u lic surface (or the H vorslev su rface in the case of over-consolid ated clays, as w ill be discussed further).

The stress p aths reach this state bou nd ary su rface w hen the soil continu es to shear w ithou t any change in volu m e u nd er constant effective stress. This su rface d efines the so-called Critical State of the soil.

(31)

Figure 2. 5: relationship between water content and consolidation pressure for a clay (modified from Roscoe et al., 1958).

The intercep t of this su rface w ith the q 0 p lane id entifies the norm al com p ression line.

Figure 2. 6: Limit state surface in the v-p -q plane (modified from Lancellotta, 2004).

The Critical state curve represents the failure points at high strains. In the p' qchart this cu rve becom es the Critical State Line. The Critical State can also be rep resented in the ' e t-s -e chart, as in the figure below.

Logp A Water content (%) 0 w B w C D 2 w 1 w 2 p 1 p 3 p p0 NCL CSL

(32)

Figure 2. 7: critical state line and normal compression lines in different planes (modified from Atkinson and Bransby, 1978).

When consid ering d rained triaxial shearing of a norm ally consolid ated clay, the stress p aths coincid e w ith cu rves that link the N CL w ith the CSL and lie on the p lane p arallel to the v (or e) axis and its intercep t w ith the p' qp lane has a steep ness of 3:1. If u nd rained shearing is analysed , the stress p aths link the N CL and the CSL, bu t in this case they belong to a p lane p erp end icu lar to the v axis, being the specific volume constant in drained conditions (see figure 2.8).

The m echanical behaviou r of over-consolid ated clays is qu ite d ifferent from the one above d escribed for norm ally consolid ated clays. First of all, their state on the v p' chart need s not only the stress valu e, bu t also the v valu e to be u nivocally described, given that it lies on the left of the NCL.

When consid ering d rained and u nd rained com p ression and shearing of over-consolidated clays, a state bou nd ary su rface on the v p' q , is reached at failu re,

' ' e e CSL CSL ' c ' CSL ' log NCL f ' f e ef kPa 1 '

(33)

d efined as H vorslev su rface. resi

Figure 2.

Atkinson and Bransby, 1978) with detail of the Hvorslev surface for undrained paths and of the Roscoe surface for drained paths

In figu r

increasing overconsolidatio ratio

Figure 2.

d efined as H vorslev su rface.

resistance at low strains, then it decreases to a constant value.

Figure 2. 8: the three dimensional state boundary or yield surface in the

In figu re 2.9, u nd rained bou nd ary su rfaces and stress p aths are rep resented w ith increasing overconsolidatio ratio

Figure 2. 9: possible undrained stress paths and critical state line for overconsolidated c

IMPOSSIBLE STATES

POSSIBLE

d efined as H vorslev su rface.

t low strains, then it decreases to a constant value.

: the three dimensional state boundary or yield surface in the

e 2.9, u nd rained bou nd ary su rfaces and stress p aths are rep resented w ith increasing overconsolidatio ratio

: possible undrained stress paths and critical state line for overconsolidated c Increasing OCR

IMPOSSIBLE STATES

POSSIBLE STATES

d efined as H vorslev su rface. In fact, over

Atkinson and Bransby, 1978) with detail of the Hvorslev surface for undrained paths and of the Roscoe surface for drained paths.

e 2.9, u nd rained bou nd ary su rfaces and stress p aths are rep resented w ith increasing overconsolidatio ratio

: possible undrained stress paths and critical state line for overconsolidated c Increasing OCR

IMPOSSIBLE STATES

STATES

In fact, over

Atkinson and Bransby, 1978) with detail of the Hvorslev surface for undrained paths and of the

e 2.9, u nd rained bou nd ary su rfaces and stress p aths are rep resented w ith OCR

R_p

.

: possible undrained stress paths and critical state line for overconsolidated c

OCR=1

In fact, over-consolid ated clays first show a p eak t low strains, then it decreases to a constant value.

e 2.9, u nd rained bou nd ary su rfaces and stress p aths are rep resented w ith .

OCR=1

consolid ated clays first show a p eak t low strains, then it decreases to a constant value.

: the three dimensional state boundary or yield surface in the v

e 2.9, u nd rained bou nd ary su rfaces and stress p aths are rep resented w ith

consolid ated clays first show a p eak

q

p' _{space (from} Atkinson and Bransby, 1978) with detail of the Hvorslev surface for undrained paths and of the

consolid ated clays first show a p eak

space (from Atkinson and Bransby, 1978) with detail of the Hvorslev surface for undrained paths and of the

: possible undrained stress paths and critical state line for overconsolidated clays. consolid ated clays first show a p eak

Experimental study of the transitional behaviour of the silty soils from the Venice Lagoon