Embed Size (px)
Citation preview
Tunnelling and Underground Space Technology 40 (2014) 160–173
Contents lists available at ScienceDirect
Tunnelling and Underground Space Technology
journal homepage: www.elsevier .com/ locate / tust
Performance-based re-use of tunnel muck as granular material forsubgrade and sub-base formation in road construction
0886-7798/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.tust.2013.10.002
⇑ Corresponding author. Tel.: +39 011 0905612; fax: +39 011 0905614.E-mail addresses: [email protected] (P.P. Riviera), rossana.bellopede@
polito.it (R. Bellopede), [email protected] (P. Marini), [email protected](M. Bassani).
1 Tel.: +39 011 0907738; fax: +39 011 0907699.2 Tel.: +39 011 0907625; fax: +39 011 0907699.3 Tel.: +39 011 0905635; fax: +39 011 0905614.
Pier Paolo Riviera ⇑, Rossana Bellopede 1, Paola Marini 2, Marco Bassani 3
Department of Environment, Land and Infrastructures Engineering, Politecnico di Torino, Corso Duca degli Abruzzi, 24, 10129 Torino, Italy
a r t i c l e i n f o
Article history:Received 21 January 2013Received in revised form 20 September 2013Accepted 8 October 2013
Keywords:Tunnel spoilMuckTunnel boring machineVolumetric characteristicsMechanical propertiesEmbankmentsSubgradeSub-basePerformance-based specificationPrescriptive specificationRoad constructions
a b s t r a c t
Large volumes of muck are produced in the Alpine Region and bordering areas as a result of new road andrailway construction. For example, in Austria every year approximately 32 � 106 Mg of muck are pro-duced from tunnelling activities. In the near future, many other initiatives along the European corridorswill lead to further construction activity, with an inevitable increase in the environmental problemsrelated to the use or disposal of the muck generated. Therefore, there is a clear opportunity for the exten-sive re-use of muck due to the high demand for granular materials (about 3 billion tonnes in Europe, only5% of which comes from recycling), the depletion of existing quarries (approximately 24,000 in Europe),and the environmental constraints preventing or delaying the opening of new quarries.
In this scenario, a new approach to the re-use of muck is both necessary and timely. Although manytypical defects deriving from its geological nature and/or from the extraction techniques employedmay lead to its rejection as an aggregate, these same defects are of less importance in embankment, sub-grade and sub-base construction in transportation infrastructures and, indeed, in most cases they can bemitigated by granular or chemical stabilization.
The investigation described here embraces this philosophy. Starting from the chemical physical char-acterization of seven different mucks derived from tunnelling activities on the Italian side of the Alps, thepaper aims to explore the potential benefits deriving from their re use as a construction material. The testmethods used all adhere to prescriptive and performance-based construction specifications. Notwith-standing the unfavourable geological origin of some of the considered materials, they all exhibitedmechanical properties that would encourage their almost complete re-use in infrastructure constructionprojects.
� 2013 Elsevier Ltd. All rights reserved.
1. Introduction
In Europe, the transportation infrastructure system is consid-ered fundamental for the smooth operation of the internal market,the mobility of people and goods, and economic and social cohe-sion between European countries. In support of this system, theTrans-European Transportation Network (TEN-T) provides road-ways, railways, and airports as defined by the European Union inthe 1980s. In the 27 EU member countries 5,000,000 km of pavedroads, of which 65,100 km are motorways, and 212,800 km of rail-way lines are included (European Commission, 2005). In NorthernItaly, 200 km of new railway tunnels and 200 km of road tunnels ofmore than 2000 m in length are planned (World’s Longest Tunnel
Page, 2012), together with new underground lines in the largesturban areas. From 2013 to 2015, the plan for the Italian railwaysystem envisages approximately 2500 km of new infrastructures(Rete Ferroviaria Italiana, 2012).
These new constructions will lead to the excavation of greatquantities of granular materials, so the re-use of tunnel mucks,presently considered a waste according to new construction spec-ifications, could make an important contribution to the sustain-able, economic and technological development of Europeansociety.
A well-performing transportation network requires large vol-umes of natural resources such as soils and aggregates. The Euro-pean Aggregates Association indicates that, in 2010, theproduction of aggregates was of the order of 3680 � 106 Mg, ofwhich recycled aggregates accounted for 186 � 106 Mg (5%) andcrushed rock accounted for 1929 � 106 Mg (53%). In Italy thereare no figures available for the percentage of recycled aggregates(European Aggregate Association, 2012). Despite the high level ofactivity associated with the provision of infrastructures and theconsiderable need for resources, the document on the impact
P.P. Riviera et al. / Tunnelling and Underground Space Technology 40 (2014) 160–173 161
assessment on the European Transport Area (European Commis-sion, 2011a) attached great significance to the employment ofenvironmental resources but paid very little attention to theemployment of mineral raw materials (soils and rocks) and theirrecycling and re-use.
This paper promotes the consideration of tunnel muck as a sta-ble and alternative source of surrogate and soils. For this purpose,seven tunnel mucks with diverse geological origins and producedby different excavation methods were considered. Although cer-tain defects may lead to the rejection of some muck as aggregatematerial for bound pavement layers (i.e., large size, elongatedshape, flakiness), these same defects render the muck more suit-able for use in embankments, subgrades and sub-bases of transpor-tation infrastructures where significant large volumes of granularmaterials are required (see Fig. 1 the typical section of a road). Itis worth noting that the granular materials employed in the con-struction of sub-bases and subgrades contribute to the mitigationof the detrimental effects of climate and dynamic stresses inducedby traffic loads, so the unbound granular material employed intheir formation should meet with stricter specifications than thoseused for the selection of embankment materials.
2. Background and literature review
2.1. Re-use of tunnel mucks
The idea of an extensive re-use of tunnel excavated materialsoriginated in the 1990s when the growing environmental and sus-tainability problems associated with the supply of natural aggre-gates became one of the most important issues in civilconstruction (Kwan and Jardine, 1999; Gertsch et al., 2000). In re-cent years the problem has been exacerbated due to the construc-tion of a number of very long tunnels which have generatedsignificant quantities of muck to be disposed of, with an ensuingconsumption of land, and economic and environmental resources(Ritter et al., 2013). This depletion of resources is certainly not sus-tainable in the long term. Nevertheless, in spite of the large scaleimpact of the problem, only a limited number of experimentalinvestigations relating to the possibility of using muck as aggregateor soil surrogate have been disseminated in literature.
A number of these studies have focused on the effects of theexcavation technique used on the properties of spoils. Grunneret al. (2003) underlined that the usages of excavated materialsshould be evaluated on the basis of the excavation driving methodas this influences the cleanliness and shape of mucks. They notedthat the use of the Tunnel Boring Machine (TBM) led to a particlesize which was suitable for aggregate, while with classical excava-tion methods the characteristics of muck depended on the physicalstate of the original rock mass and on the blasting technology used.
Some attempts have been made to assess the possibility of re-using muck as concrete aggregate especially when the excavationprocess is carried out by means of the TBM. Using six different
PAVEMENT SURFACE
PAVEMENT (BOUND LAYERS)
SUBBASE
SUBGRADE
EMBANKMENT or IN-SITU SOIL
UN
BOU
ND
LA
YER
S
FOU
ND
ATIO
N
PAVE
MEN
T SY
STEM
Fig. 1. Principal layers of a road.
TBM mucks, Olbrecht and Studer (1998) obtained a highly-work-able concrete characterized by a greater shrinkage and a lowerelasticity modulus, approximately equal to 50% of that of conven-tional concretes. Thalmann-Suter (1999) also pointed out that therecycling of excavated debris begins with the choice of diggingmethod and requires careful and continuous control of the muckproduced to ascertain its quality with practice-friendly testmethods.
The possible re-use of excavation materials has also been eval-uated in Austria where 32 � 106 Mg of muck are produced everyyear. The research by Resch et al. (2009), supported by the AustrianResearch Promotion Agency, highlighted that the re-use of muckdepends mostly on the lithological properties of the excavatedrock, the demand for mineral raw materials within a defined dis-tance from the tunnel construction site, and the treatments whichthe tunnel mucks are subjected to after excavation.
In more recent years, some experience with its re-use has beengained with the generation of large volumes of muck during theconstruction of new tunnels in the Alpine Region. An investigationcarried out at the Gotthard Base Tunnel (Lieb, 2009) analysed spoilrecycling for the production of high quality concretes and shot-cretes. In this case, a specific testing plan was developed to assessthe quality of both the raw material and the concrete mixes pro-duced by evaluating workability time, mechanical properties anddurability.
In the Danube Lobau tunnel experience (Schröfelbauer et al.,2009) it was observed that gravel and sands obtained from spoilscan be used as aggregate for concrete production or as soil forembankments and subgrades. Silt and clay obtained from excava-tions can be used instead for embankment filling and backfilling(after suitable drying) depending on their plasticity.
Finally, Burdin and Monin (2009), who worked on material ex-tracted from the shafts of the Lyon–Turin high-speed railway, alsoremarked on all the different usages for excavated material.Depending on spoil characteristics and on the basis of technicalspecifications for rock classification, they identified three distinctquality classes. In particular, the classification of excavated materi-als is useful when selecting material for the production of concreteaggregates (class 1), for the production of soil surrogates forembankments (class 2), and finally for disposal into deposit areas(class 3); similar classifications have been considered by previousauthors (Lieb, 2009; Resch et al, 2009; Ritter et al., 2013).
Burdin and Monin (2009) also noted that the extensive recy-cling of excavated debris could lead to significant benefits includ-ing a reduction in the area required for deposit, a reduction inthe cost of aggregates and embankment materials and, above all,lower CO2 emissions.
As a result of the literature review, it can be noted that few pastexperiences focus on the evaluation of muck as a resource in theconstruction of infrastructures. Only recent papers focus on therecycling of the most valuable part of mucks, which is normallyused solely in the production of cement mixtures, while in only afew cases attention is devoted to the total volume of excavatedmaterials.
Finally, various classification systems are considered in litera-ture. They all subdivide the excavated materials on the base of geo-logical exploration results and on laboratory testing by geologists’.It is worth noting that such classification systems do not considerthe standards used to assess the performance of concrete, asphaltand granular materials for pavement and embankment applica-tions (with the sole exception of the Los Angeles test).
2.2. Recycling
The common purpose in the management of large quantities oftunnel muck is their recycling in order to provide surrogate gravel
162 P.P. Riviera et al. / Tunnelling and Underground Space Technology 40 (2014) 160–173
for constructions. To obtain satisfactory mechanical properties inline with those exhibited normally by granular materials, tunnelmuck should be first selected and then treated to improve size dis-tribution and shape, breaking down flat and elongated particlesinto more polyhedral ones (Thalmann-Suter, 1997). This operationis necessary to obtain a more suitable material, which can increasein value when its use passes from embankments, to subgrades,sub-bases, or better still to bituminous and/or cementitious mix-tures used for pavements and structures.
The European Directive 2008/98/EC (European Parliament andThe Council of the European Union, 2008) considers tunnel muckto be waste material only, even when employed in other construc-tion sites. Conversely, according to the recent European Communi-cations on Prevention and Recycling of Waste (EuropeanCommission, 2011b), one of the main expected achievements ofthe European waste strategy is a reduction in the level of wastegenerated and, indeed, its use as a resource. With regard to mate-rial availability, the construction and industrial sectors are nowfacing a general depletion in the levels of traditional raw materials.The problems involve both the quantity of raw material producedas well as their quality and are caused by the exhaustion of goodquality raw material quarries and the opening of quarries produc-ing low to medium quality raw materials (Commission of the Euro-pean Communities, 2005, 2008). The need for treatments andhigher transportation costs are playing a major role in the con-struction and industrial economy. In this context, the use of tunnelmuck represents an important step towards the much heraldedgoal of sustainable development.
Two general markets exist for tunnel muck: the first one is verysmall and internal to the construction site of origin in which it isviewed as construction material; the second one is the global‘‘aggregate and soil market’’ which is external and larger but wherethe muck is deemed to be waste. Only small quantities of muck canbe employed in the same site from which it has been excavated, sothe second destination is prevalent. In this case, every national reg-ulation attributes a specific sub-classification to excavated rocksand soils. In Italy, for example, new norms are set to be introducedin which non-hazardous excavated waste materials will be classi-fied as by-products or secondary raw materials, facilitating theirdirect employment.
In tunnel excavation generally, only small quantities of exca-vated materials are of good quality, while most are considered tobe low quality and consequently employed in non-structural appli-cations or, more frequently, disposed of in landfill or dumpingsites. As a consequence, good tunnel mucks have a negligible value.The idea of an extensive recycling of excavated materials resonateswith the consistently high demand for granular materials. More-over, in many regions most quarries are close to depletion, whilenew quarries cannot be opened as a result of environmental con-straints. Currently, in Northern Italy, up to 50% of granular materi-als employed in the formation of embankments and unboundgranular layers of pavements derive from the recycling of construc-tion and demolition waste.
In light of the aforementioned considerations, two main ques-tions arise: is there the possibility to broaden the use of alternativegranular materials, such as tunnel mucks, in the construction of ci-vil infrastructures? And secondly: does the attainment of thisobjective necessitate the adoption of different constructionspecifications?
2.3. Construction specifications
Most of the difficulties encountered in the use of alternativematerials centre on the type of specification stipulated in contracts(AASHTO Highway Subcommittee on Construction, 2003).Typically, method or prescriptive specifications are used in the
selection of materials for road and railway applications. In this case,material typology and acceptance limits are rigidly imposed in or-der to guarantee the use of specific materials, the selection of whichdepends exclusively on the results obtained from engineering testson several representative samples. This approach is based on theidea that the quality of each single material can ensure the desig-nated performance of the entire structure throughout its servicelife. In the case of prescriptive specifications, only key parametersthat demonstrate an empirical correlation with fundamental engi-neering properties are considered. With their use, expected perfor-mances can be easily achieved by traditional materials only.
On the other hand, performance-based (PB) specifications arerarely used in contracts (AASHTO Highway Subcommittee on Con-struction, 2003). They establish desired levels of fundamental engi-neering properties that must be reached to ensure the expecteddesign life. Material characteristics like resilient modulus and per-manent deformation properties are taken directly into account andused in mathematical models for the calculation of fundamentalperformance variables such as stress, strain, or distress levels un-der the prevailing traffic, environmental and structural conditions.As a result, the expected performance level can be achieved byusing any traditional, innovative or recycled materials. Therefore,the use of PB specifications does not preclude the use of any gran-ular waste or by-products such as tunnel muck.
3. Objectives and methodology
In 2007, with the aim of exploring new possibilities in the realmof muck recycling, the Regione Piemonte financed the Remuck Pro-ject, which was developed by the Politecnico di Torino (2012) incooperation with a number of private companies and publicassociations.
In the Remuck Project, seven tunnel mucks coming from theexcavation of new tunnels in the Alps and from the constructionof the new underground line in the city of Turin were considered(Table 1). In light of the different petrographical properties, exca-vation methodologies and treatment processes, the investigationsought to assess the effect of such factors on the properties ofthe derived material as an alternative source to surrogate tradi-tional aggregate and soil.
After a first step which focused on the basic characterization ofmucks, the main stages of the investigation program included vol-umetric and mechanical tests carried out both in the laboratoryand in full-scale tests. In particular, three different compactionmethodologies were considered:
� the modified Proctor method (AASHTO T180, 2010), whichentails a hammer impacting on squat cylindrical moulds;� the gyratory method (AASHTO T312, 2009), which provides a
simultaneous compressive and shear effort feed into thin cylin-drical moulds;� the rolling compaction method for the generation of full-scale
layers.
The modified Proctor procedure is currently considered in pre-scriptive specifications for the derivation of parameters such as theoptimal water content and the maximum dry density of soils. Onthe other hand, the gyratory compaction procedure was selectedhere in order to better replicate the field compaction force, andhence to meet the requirements for the rational characterizationof materials as per PB specifications.
The field operation was possible thanks to the availability inconsiderable quantities of just four mucks, which made it possibleto evaluate the in-field density parameters for layers of 25 cm inheight. In all cases, a heavy articulated vibratory roller was used.
Table 1Muck samples.
Code Infrastructure Sampling site Excavation method Treatments
S1 Turin underground, Marconi station Turin, Italy EPB EM – CPS2 Turin underground, Dante station Turin, Italy Cut & Cover EM – CPS3 Turin underground, Lingotto station Turin, Italy EPB EMS4 High Speed Rail Turin-Lyon Clarea valley, Turin, Italy Coring CPS5 Regional Road 229, Bocciol tunnel Omegna, Verbania, Italy Explosive EM – CPS6 Railway Verona-Innsbruck, Brennero tunnel Aica, Bolzano, Italy TBM EMS7 Hydroelectric plant tunnel, Torrent La Thuile, Aosta, Italy TBM EM – CP
Remarks: EM: Excavated Material, CP: Crushed in mobile Plant.
P.P. Riviera et al. / Tunnelling and Underground Space Technology 40 (2014) 160–173 163
During the field operations, loose granular materials were takenfrom the deposits and used to reproduce laboratory samples. Thesamples obtained via the two compaction techniques mentionedabove underwent mechanical tests.
Proctor samples were subjected to a California Bearing Ratio(CBR) test (AASHTO T-193, 2010), which is coherently included inprescriptive technical specifications, while gyratory samples wereused in the evaluation of the resilient modulus through the dy-namic triaxial test, which is used in PB technical specificationsand is assumed as the basis for the rational structural design ofpavements. Similarly, Light Weight Drop (LWD) tests (TP BF-StB:Part B 8.3, 2003) were performed on-site with the aim of assessingthe bearing capacity of the granular full-scale layers through theestimation of the dynamic elastic modulus.
Following a comprehensive physical, volumetric and mechani-cal characterization of the muck samples, an analysis of the test re-sults in light of the acceptance limits pertaining to prescriptive andPB specifications is proposed in this paper. Furthermore, a compar-ison with reference limits derived from traditional materials ledthe authors to final conclusions regarding the recycling possibili-ties of the investigated mucks.
Fig. 2. Mobile crushing plant used for the treatment of excavated materials.
4. Materials
Table 1 contains the essential information on the seven mucksanalysed in this experimental investigation.
The first three materials, alluvial in nature, were collected fromthe new Turin underground line. The samples from Largo Marconistation (code S1) and Lingotto Station (code S3) were excavated bymeans of the Earth Pressure Balance (EPB) tunnel boring machine.This machine permits the excavation of tunnels in soft ground con-ditions where clay, silt, and sand are present. The front shield ofthe machine is filled with debris extracted by a screw conveyor.This screw compensates for the pressure difference between thebulkhead chamber and the atmospheric pressure. Foam injectionrenders the material more homogeneous, thus facilitating its exca-vation. The second alluvial sample (code S2) was taken at the CorsoDante station and excavated with the Cut and Cover method, inwhich a trench is mechanically excavated and roofed over withan overhead support system strong enough to bear the load ofwhatever is to be built above the tunnel.
The S4 material was derived from the crushing of micascistcores collected during the exploration phase in the Clarea Valleyfor the new High Speed Railway line from Turin to Lyon, whichforms part of the TEN-6 axis. Only part of the cores taken fromthe depth of the future tunnel were taken and used to form theS4 sample. The S5 and S6 samples were both grey granite: the firstwas excavated by means of the Explosive method along the Boccioltunnel belonging to the new section of the Regional Road 229 inPiedmont, while the second was extracted from the pilot drift inAica (Alto Adige) of the Brennero base tunnel which is part of thenew High Speed Railway line from Verona to Innsbruck along theTEN-1 axis. In this latter case, a Tunnel Boring Machine (TBM)
was employed; with this technique, disc cutters on the front shieldcreate compressive stress fractures in the rock, causing it to chipaway. Finally, the S7 is a calcareous schist excavated by means ofa TBM from the Torrent-La Thuille hydroelectric plant tunnel.
The seven mucks were processed in a mobile plant. As a result,the mucks were divided into freshly excavated material (EM), andcrushed muck in the mobile plant (CP) as indicated in Table 1.
The mobile plant (Fig. 2) has a production rate of 280 Mg/h anda maximum input dimension of 600 mm for the material to betreated. It is composed of a vibrating screen placed above a jawcrusher and a magnetic separator, which is positioned on a con-veyer belt along which the output material is transported. Thematerial exiting from the crusher can be regulated to a minimumsize of 30 mm. As a consequence, the plant offers one end productonly.
4.1. Petrographic and geotechnical classification
The petrographic description and the geotechnical classificationof the mucks are reported in Table 2, while the particle size distri-bution is illustrated in Fig. 3.
The particle size distribution was performed using the wet siev-ing method for granular fractions larger than 75 lm and the air jetsieving method for fractions finer than 75 lm in accordance withEN 933-1 (1999) and EN 933-10 (2009) respectively. As indicatedin EN ISO 14688-2 (2004), two separate parameters have beenused to define the shape of the grading curve: the uniformity coef-ficient CU:
Cu ¼ d60=d10 ð1Þ
and the coefficient of curvature CC:
Cc ¼ ðd30Þ2=ðd10 � d60Þ ð2Þ
where d10, d30 and d60 denote the particle sizes corresponding to theordinates 10%, 30% and 60% by mass of the percentage of materialpassing through the sieve.
Table 2Petrographical description and geotechnical classification of mucks.
Code Petrographical description AASHTOclassification
CENclassification
CU CC Shape ofgradingcurve
Standard EN 932-3 M 145-91 EN ISO 14688-2
S1 Alluvial rock composed of quartz (30%), calceschist (20%), green stones (30%), granites,limestone, sandstones (20%), fines (10–15%)
A1-a saGr 314 2.4 Multi-graded
S2 Alluvial rocks composed of quartz (25%), calceschist (25%), green stones (18%), cemented rocks(20%), micaschist (10%) on the grains size 20–30 mm, fines (<10%)
A1-b saGr 93 0.8 Gap-graded
S3 Alluvial rock composed of quartz and green stones A1-a saGr 147 0.8 Gap-gradedS4 Mica schist N/A N/A N/
AN/A
N/A
S5 Granite composed of potassium feldspar (35%), quartz (40%), plagioclase (10%), biotite passingto chlorite (10–15%), and other materials including zircon with pleochroic halo, pyrite andwhite mica.
A1-a saGr 44 0.7 Gap-graded
S6 Granite A1-a saGr 146 1.3 Multi-graded
S7 Calcareous schist composed of carbonates (65%), quartz (25%), white mica (5%), and opaque (5%) A1-b sasiGr 454 0.9 Gap-graded
0
20
40
60
80
100
Pass
ing
[%]
Diameter [mm]
S1 - EMS2 - EMS3 - EMS5 - EMS6 - EMS7 - EM
Clay CobbleGravelSand redluoBtliS
0,002 mm
0.0001 0.001 0.01 0.1 1 10 100
0,063 mm 2 mm 200 mm63 mm
Fig. 3. Gradation curves of the excavated mucks.
164 P.P. Riviera et al. / Tunnelling and Underground Space Technology 40 (2014) 160–173
The three alluvial samples (S1, S2, and S3) presented the typicalpetrographical composition of Turin deposits, albeit with someminor variations between sites,, a composition which is quite wellappreciated in the aggregate market for concrete production. Thethree samples contained a high percentage of rounded fragmentsof hard rock. Of the three alluvial mucks, S1 contained the highestpercentage of fine grains.
Sample S4 is composed of mica schist from the Ambin Unity inthe Alps and was obtained from the crushing of core probes: hencethe reason why grading curve and geotechnical classifications arenot present in Table 2 and Fig. 3. The first sample of the two greygranites (S5) shows a certain degree of weathering mainly due tothe elevated presence of saussurrite in the feldspars; the secondone (S6) is more compact and characterized by mechanicalstrength. Saussurrite is a common, greenish mineral aggregate,produced in part by the alteration of feldspar, consisting chieflyof epidote and zoisite. Finally, sample S7 is a calcschist with alow percentage of mica and, consequently, low schistosity.
4.2. Physical characterization
Table 3 reports all the results obtained from physical tests onparticles (density, shape, flakiness, fragmentation and wear resis-tance) in accordance with current European standards includedin the European Committee for Standardization list. Table 3
includes data derived from laboratory tests performed on freshlyexcavated material (EM) and on crushed muck (CP). Furthermore,it states the category to which each material belongs as per EN13242 (2008), which is used to classify aggregates for unboundand hydraulically bound materials for use in civil engineeringworks and road construction.
The particle density (EN 1097-6, 2000) of the investigatedmucks assumed values in the typical range for granular materialscommonly used in road construction (around 2.70 Mg/m3). Thegranite samples (S5 and S6) were characterized by lower values,while the alluvial and schist mucks were characterized by higherdensity values.
Shape and flakiness indexes were evaluated in accordance withEN 933-3 (2003) and EN 933-4 (2008) respectively. The shape in-dex (SI) represents the ratio between the mass of non-cubical par-ticles and the total mass of particles tested, while the flakinessindex (FI) is the ratio of the total dry mass of elongated particlespassing through specific bar sieves to the weight of the full sampleexpressed in percentage terms (the test consists of two standard-ized sieving operations; firstly, particles are separated into varioussize fractions; secondly, each fraction is then sieved using barsieves).
The two tests provide useful indications with respect to theparameters relating to the compaction attitude of granular materi-als. In order to attain significant strength and stiffness levels, high
Table 3Physical and mechanical properties of freshly excavated (EM) and crushed (CP) materials and classification according to EN 13242.
Test Particle density Shape index Flakiness index Los Angeles Micro-DevalStandard EN 1097-6 EN 933-4 EN 933-3 EN 1097-2 EN 1097-1
Code Treatment (Mg/m3) (%) Cat. (%) Cat. (%) Cat. (%) Cat.
S1 EM 2.75 14 SI20 13 FI20 22 LA25 11 MDE20CP 4 SI20 10 FI20 24 LA25 12 MDE20
S2 EM 2.75 21 SI40 36 FI50 28 LA30 19 MDE20CP 19 SI20 16 FI20 28 LA30 18 MDE20
S3 EM 2.71 N/A – N/A – N/A – N/A –
S4 CP 2.79 30 SI40 29 FI35 N/A – N/A –
S5 EM 2.65 14 SI20 15 FI20 N/A – N/A –CP 7 SI20 10 FI20 38 LA40 23 MDE25
S6 EM 2.69 36 SI40 22 FI35 24 LA25 N/A –
S7 EM 2.74 53 SI55 44 FI50 N/A – 9 MDE20CP 29 SI40 40 FI50 27 LA30 N/A –
P.P. Riviera et al. / Tunnelling and Underground Space Technology 40 (2014) 160–173 165
percentages of flat and elongated particles are undesirable as theyinfluence the shear resistance of granular materials during thecompaction process leading to weaker granular layers under trafficloads (Mallick and El-Korchi, 2009). Normally, they must be dis-carded or limited to a specific percentage. Fig. 4 shows that muckindex values can vary and depend on their mineralogy, the excava-tion method used and the milling process applied. However, itmust be emphasised that the crushing processes led to a significantimprovement in these characteristics as clearly indicated by the ar-rows that link the EM data to the corresponding CP ones. In fact, allthe arrows indicate a decrease in the SI and FI indexes which is par-ticularly evident in the case of mucks S2 and S7 characterized byhigh values for both indexes.
The soundness of coarse granular materials was tested throughthe determination of the fragmentation resistance by means of theLos Angeles test (EN 1097-2, 2008) while the wear resistance wasdetermined by means of the Micro-Deval test (EN 1097-1, 2004).
In the Los Angeles test a sample of granular material, togetherwith some steel balls, is introduced into a rotating drum. The Mi-cro-Deval test is performed with a smaller cylinder and spheres,and the sample is put into the cylinder with water. As a conse-quence it tends to smoothen the particles, while the Los Angelestest tends to break them up. In both cases, after rolling is complete,the quantity of material retained on a 1.6 mm sieve is determined.The two indexes represent the percentage ratio of the fine materialproduced to the total initial mass of the sample.
0
5
10
15
20
25
30
35
40
45
50
0 10 20 30 40 50 60
Flak
ines
s In
dex
(%)
Shape Index (%)
EPB
Explosive
Cut & Cover
TBM
Coring
S7
S2S6
S1S5
S4
Fig. 4. Shape and flakiness indexes of mucks (the arrows indicate the variation ofindexes following the crushing process).
These fundamental tests permit the evaluation of the mechani-cal degradation of granular materials during handling, constructionand in-service time. The use of materials which are not adequatelyresistant to abrasion and polishing may lead to premature struc-tural failure.
Please note that some data could not be included in Table 3 forvarious reasons. For S4 muck, only the density test could be per-formed on particles derived from its crushing in the mobilecrusher. In other cases, the Los Angeles and Micro-Deval tests werenot performed due to difficulties encountered in the formation ofthe test samples caused by limited quantities of granular classesas required by the new EN norms. Since there are no alternativetests capable of providing the same information, it was not possi-ble to evaluate the soundness of these materials.
In general terms, the data reported in Table 3 and Fig. 4 confirmthat the investigated mucks may be employed in the formation ofunbound granular layers of the road structure, and that the millingprocess has a significant, positive effect on the shape and elonga-tion parameters (SI and FI indexes decrease after milling) withoutleading to any alterations in the mechanical properties of particles(LA and MDE do not experience any significant changes).
Regarding samples S1 and S2, which are very similar, alluvialmaterials excavated from along the new Turin underground linein two locations located 1200 m apart, further investigationsregarding the fines content were performed in order to betterdetermine the influence of the excavation method (EPB for S1and Cut and Cover for S2). In particular, in the case of sample S1derived from excavation with EPB, a foaming biodegradable agentwas used in order to reduce friction, stress and strain on tools, andto reduce blocking due to kneading of the material.
The sand equivalent (SE) test (EN 933-8, 2000) and the methy-lene blue (MB) test (EN 933-9, 2009) were conducted on the gran-ular fractions finer than 2 mm in order to assess the presence ofdangerous organic clay in the two materials. A SE test value lowerthan 30 indicates a significant quantity of fines (clay and silt),while a MB test value lower than 10 highlights the presence of anegligible amount of noxious clay.
Sample S1 exhibited a SE value equal to 36, and a MB valueequal to 1.9; while the S2 sample exhibited a SE value equal to96, and a MB value equal to 0.5. Part of the difference was certainlycaused by the different excavation methods that resulted in a high-er amount of fine grains in the S1 muck compared to S2, a findingwhich can be mainly attributed to the presence of silt. Both mucksshowed a very low clay content which, however, does not precludetheir more qualified use insubgrades and sub-bases, as well as inembankments.
Fig. 5. Proctor, gyratory and in-field roller compaction methodologies and typical results.
166 P.P. Riviera et al. / Tunnelling and Underground Space Technology 40 (2014) 160–173
5. Testing methods
5.1. Sample preparation and volumetric characterization
Referring to Fig. 5, the laboratory samples were compacted byfollowing two procedures. In the case of the Proctor method, onlyfiltering with the 19 mm sieve was considered using moulds with adiameter of 152.4 mm and a height of 116.4 mm as per AASHTOT180 (2010).
The optimal moisture content (wopt) and the maximum dry den-sity values (cd,max) were evaluated at 2.68 MJ/m3 of compaction en-ergy corresponding to 56 blows of the compaction hammer oneach of the five layers with a weight equal to 4540 g falling froma height of 0.457 m. The maximum dry density was used as a targetvalue for the production of samples at the gyratory shear compac-tor (GSC). Even though such a compaction technique is normallyused for the production of bituminous mixture samples in accor-dance with AASHTO T312 (2009), the authors included it in theexperimental program thanks to its ability to transfer shear stressto laboratory samples in the same manner that rollers operate onfull scale layers.
The samples compacted at the GSC were produced by applyingand maintaining a vertical pressure of 600 kPa on the top of themould, which gyrates at a rate of 30 gyration/min with a tilting an-gle of 1.25�. Three moisture contents corresponding to the optimalone (wopt) and two variations of 2% around the optimum (wopt -� 2%, wopt + 2%) were considered for the production of the speci-mens. The total quantities of dry granular material and waterwere calculated in advance so as to obtain the target Proctor drydensity and moisture content for samples of 200 mm in heightand 100 mm in diameter. The samples were produced fixing theheight as a mode of operation which is alternative for GSC to the
number of gyrations mode; hence, the number of gyrations atthe target height was always variable. To facilitate equal distribu-tion of the compaction energy in the sample, the loose materialwas divided into four parts, with each part then being compactedseparately in the mould adding one part over the one before.
The degree of compaction (Cg) was evaluated at a generic num-ber of gyrations for each layer using the following formula:
Cg ¼ 100 � cd � hf
cg � hgð3Þ
where cg is the particle density of the grains (EN 1097-6, 2008), andhg and hf represent the height of the sample measured at the genericnumber of gyrations (ng) and at the end of the compaction process(ngf) respectively. It is worth noting that the degree of compactionindicated in Eq. (3) is the complement to one hundred of the voidcontent expressed in percentage terms of the dry granular material.
Four compaction curves associated with each sample were ob-tained considering the dependency of the degree of compaction(Cg) to the number of gyrations (ng). In all cases the following typ-ical equation, also used to model the compaction process of bitu-minous mixtures (Bassani et al., 2009), was found to be the bestregression function:
Cg ¼ C1 þ kg � logðngÞ ð4Þ
where the regression parameters kg and C1 represent the workabil-ity and the initial degree of compaction (also called the self-com-paction) at the first gyration (ng = 1) respectively.
Field operations on full-scale layers of the investigated muckswere performed in order to assess the workability of such materi-als in the field. Compaction parameters like workability (kp) andthe initial compaction degree (C1) cannot be derived with sufficient
Δh(t)
σd(t)
σ3
h0
σX
σZ
σ3
σZσX
σ (t)
σd(t) σd,max
t
εz
εzrεzp
Δh(t)
h(t)
t
P1(t)P5 mm
P2.5 mm
h
v1 = cost.
P1(t)
2.5 mm 5 mm
v1 = cost.
h(t)
1
half-space, Ehs
upper layer, Ed
stress bulb
σd(t) σd,max
tneiliseRRBC Modulus Dynamic Modulus dleiFyrotarobaL
Fig. 6. CBR, resilient modulus and dynamic deflectometer tests and results.
P.P. Riviera et al. / Tunnelling and Underground Space Technology 40 (2014) 160–173 167
accuracy after each roller pass (np). As a consequence, the compac-tion assessment was made by comparing the field dry density tothe laboratory maximum density from the Proctor test (Fig. 5).In-field dry density was evaluated by performing the sand conetest subsequent to the completion of the compaction process con-sisting of a specified number of roller passes (npf).
5.2. Mechanical characterization
As per the schema hitherto described, the mechanical tests,adhering to prescriptive and PB specifications, were conducted toassess the bearing capacity of mucks. In the case of the prescriptiveapproach, CBR tests were performed in adherence with AASHTO T-193 (2010) on specimens compacted with the Proctor procedure.For the PB approach, tests performed included resilient modulustests (AASHTO T-307, 2007) on laboratory specimens, and dynamicLight Weight Drop tests (TP BF-StB section B 8.3, 2003) for the der-ivation of the dynamic modulus on in-field layers.
The CBR is an index of bearing capacity that is traditionally usedfor the evaluation of natural soils and granular materials employedin the formation of embankments, subgrades and sub-base layers.The index represents the highest percentage ratio between theforce (P1) necessary to penetrate to two specific depths (h, equalto 2.5 and 5 mm) in a confined specimen of compacted granular
material, and the force necessary to repeat the same procedurewith the reference Californian limestone crushed rock, character-ized by a CBR equal to 100% (Fig. 6). During the test, the stressand strain state is unknown and the resistance to penetrationcan only be assessed in relative terms.
The resilient modulus test is a dynamic triaxial test (Fig. 6)where an impulsive deviatoric pressure (rd) is applied to the uppersurface of a cylindrical laboratory specimen. The resilient modulusrepresents the ratio between the maximum deviatoric stress (rd,-
max) recorded at each load application, and the maximum recov-ered vertical strain (ez,max = Dhmax/h0). Two testing protocols areavailable in AASHTO T-307 (2007) for subgrade and sub-basematerials respectively. In this investigation the first was adoptedto test the EM samples, whereas the second was used to test theCP samples. In both cases, only particles passing through the20 mm sieve were used for the formation of test samples.
The Light Weight Drop (LWD) test is a plate loading test that isused to estimate the dynamic modulus (Ed) of subgrades and sub-bases. It consists of a falling weight that impacts on a rigid plate,0.3 m in diameter, and an accelerometer that records the maxi-mum deflection of the layer on impact. The estimate of the dy-namic modulus (Edf) is made by referring to the equivalent halfspace system through the application of the following formula(Huang, 2004):
0
20
40
60
80
100
0.01 0.1 1 10 100
Pass
ing
[%]
Diameter [mm]
S1 - CP S2 - CP S5 - CP S7 - CP Subbase acceptance limits (Dmax = 30 mm)
0
20
40
60
80
100
0.01 0.1 1 10 100
Pass
ing
[%]
Diameter [mm]
S1 - CP S2 - CP S4 - CP Subbase acceptance limits (Dmax = 70 mm)
168 P.P. Riviera et al. / Tunnelling and Underground Space Technology 40 (2014) 160–173
Edf ¼ p � rmax � r � ð1� m2Þ2 � Dhmax
ð5Þ
where rmax is the maximum pressure applied by the falling weighton the rigid plate, r is the radius of the plate, m is the assumed Pois-son Ratio, and Dhmax is the maximum deflection of the plate asmeasured by the accelerometer.
When used in the case of a two-layer system like the one inFig. 5, the modulus of the upper layer (Ed) can be calculated by con-sidering the Biroulia–Ivanov equation:
Edf ¼Ehs
1� 2p 1� 1
n3:5
� �arctan p�h1
4a � n� � ð6Þ
in which Ehs is the dynamic modulus of the lower half-space, h1 isthe thickness of the upper layer, and finally n = (Ehs/Ed)2.5.
During the LWD test performed in situ, a peak value of the test-ing force equal to 7.1 kN was applied, which corresponds to a peakstress of 100 kPa (rmax). Each layer was tested at three differentpoints; the dynamic modulus at each point was calculated, follow-ing three pre-conditioning loading applications, through therecording of deflection (Dhmax) of three further load applications.The average of the three testing point values was considered tobe representative of the entire layer.
It should be highlighted that the stress–strain conditions underthe test plate are not representative of those occurring under realtraffic conditions, where the pressure applied by trucks is 6–8times greater in the contact area between tires and pavements.As a result, the stress bulb generated under the test plate is rela-tively shallow with its depth only marginally exceeding itsdiameter.
Fig. 7. Gradation curves of the mucks following the treatment process.
6. Results analysis6.1. Classification, grading and particle shape
On examination of the tables and figures presented in Section 4,the test results on granular materials obtained from mucks showthem all to be potentially suitable materials for the formation ofembankments, subgrades and sub-bases for the reasons detailedbelow. In fact, the data reported in Table 2 show that the excavatedmaterials are classified as sandy gravel and belong to the A1 classof the AASHTO classification systems (AASHTO M145 2008). Theirgrading levels vary from multi-graded to gap-graded curves andthey exhibit a wide range of values for the uniformity coefficient(Cu) variable, even though the coefficient of curvature (Cc), whichrepresents the second moment of the grain size distribution curve,reveals well graded materials as evidenced by the fact that all val-ues fall within the two reference limits, equal to 1 and 3.
Fig. 7 reports the grading curves of the materials crushed in theportable milling machine and the two limits for sub-bases. Theselimits are reported in the technical specifications of the Ministerodelle Infrastrutture e dei Trasporti (2001) which consider twotypes of UGM, the difference between them being the maximumdiameter (Dmax). Table 1 indicates that five CP materials are largelyin compliance with the specifications, with the exception of twomaterials:
� S7 presents an excessive quantity of fine grains (d < 0.075 mm)with respect to Dmax = 30 mm;� S4 shows a lower content of sand when compared to the
Dmax = 70 mm lower limit.
The excessive quantity (of fine grains) in S7 is partly due to thehigh initial fine content in the (EM) material generated by the TBM,while in the case of S4 it should be remembered that the CP
materials derive from the crushing of cylindrical cores. In this lat-ter case, higher quantities of fine grains are to be expected follow-ing traditional excavation.
6.2. Volumetric analysis of compacted materials
Table 4 synthesized the volumetric results obtained on com-pacted materials referring to methods and test procedures shownin Fig. 5.
Despite the origins and types of selected mucks, all the materi-als require a restricted water content to ensure sufficient workabil-ity (wopt.Proctor varies between 4.05% and 6.50%), with a smallvariation when crushed materials are considered in place of theEM ones. Dry density is in line with the typical values presentedin literature, with variations that depend on grading and particledensity (Table 3). In Table 4, in addition to the parameters pre-sented and discussed in Section 5.1, in the case of Proctor compac-tion the ratio between the uniformity coefficient of the granularmaterial derived through the sieve analysis before (Cu,in) and aftercompaction (Cu,fin) has been included. Such a parameter is relatedto the sensitivities to compaction forces that lead to a grading var-iation especially in the case of tender and weak mucks, as in thecase of the spoiled gray granite (S5) which demonstrates the wid-est range in values from 1.56 (CP) to 1.91 (EM). This behaviour isalso confirmed by the high Los Angeles and Micro-Deval values re-ported in Table 3.
In the columns referring to the in-field compaction, the ratio be-tween the field and laboratory dry density (cd/cd,max) has been in-cluded to attest to the soundness of field compactions by rollersand, at the same time, the attitude of the granular materials tobe rolled in full scale layers. During compaction, the water content
R2 = 0.9578
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
9.0
9.5
10.0
64 65 66 67 68 69 70
Wor
kabi
lity,
kg
Self compaction, C1 (%)
S2-CP
S5-CP
S5-EM
S1-CP
S7-CP
S7-EM
S6-EM
Fig. 8. Relationship between self-compaction (C1) and workability (kg) parametersderived from gyratory compaction and reported in eq.4.
Table 4Compaction, workability and optimum water content values resulting from laboratory and in-field compaction studies.
Code Treatment Proctor (impulsive) Compaction In-field roller compaction Gyratory compaction
cd,max wopt.Proctor CU,fin/CU,in DCU cd cd/cd,max w C1 kg/cm3 % � % g/cm3 % % % –
S1 EM 2.233 4.85 0.75 �25.0 – – – – –CP 2.236 5.15 1.23 +23.3 – – – 68.3 8.06
S2 EM 2.146 6.50 1.18 +17.8 – – – – –CP 2.180 6.00 0.91 �9.2 – – – 67.0 8.50
S3 EM 2.231 4.05 – +75.3 2.125 95.2 4.20 72.5 6.88
S4 CP – – – – 2.073 – 4.46 – –
S5 EM 2.048 6.38 1.91 +90.9 2.014 98.3 2.14 64.7 7.16CP 2.108 6.40 1.56 +55.6 – – – 65.9 8.03
S6 EM 2.204 4.66 1.01 +0.6 2.169 98.4 2.53 69.7 6.34
S7 EM – – – – – – – 69.4 6.51CP 2.245 4.80 0.88 �11.9 – – – 68.5 7.43
P.P. Riviera et al. / Tunnelling and Underground Space Technology 40 (2014) 160–173 169
was less than the optimal value measured in the Proctor study,thus confirming that mucks may be used effectively and workedeven when the water content is not well controlled.
Although the data does not reveal a clear tendency when simplyassociated with the physical parameters included in Tables 2 and 3,self-compaction and workability are correlated as clearly indicatedin Fig. 8. Data evolve following a parabola: low workability isexhibited by the excavated samples of S5, S6, and S7 mucks, whilehigher values are shown by crushed samples. In the case of sam-ples S5 and S7, the crushing process increases the workabilityalthough different degrees of self-compaction occurred. The maxi-mum value for workability is evident in those mucks derived fromthe excavation of alluvial sandy gravel, so it cannot be excludedthat the rounded surface of most of the constituent grains contrib-uted to such a result.
On the other hand, self-compaction (C1) is mostly influenced bythe particle size distribution and by the shape and surface textureof particles. In the case of S7 EM and CP samples this is due to thehigh content of very fine particles that completely fill the space be-tween the coarse grains (the mass percentage of particles finerthan 75 lm is equal to 20.6% for EM, 22.0% for CP), while in thecase of the S6-EM muck this is a consequence of its regular contin-uous grading curve that favours the initial packing of grains.
6.3. Mechanical analysis
In Table 5 the results of the CBR test, adhering to prescriptivespecifications, are reported. It should be stressed that the high val-ues measured confirm that the CBR test is sensitive to the localconditions which characterize each particular sample, so toughparticles derived from the crushing of rock lead to very high CBRvalues. At the optimal water content, all results present valuesgreater than 80%, which is considered to be the lower limit forcrushed rock.
Considering the data for S1 and S2 mucks, the crushing per-formed in the mobile plant produced great benefits in the samplesof alluvial origin, while in the case of muck S5 a reduction in theCBR index was observed. The optimal water content for this test(wopt.CBR) is, generally speaking, approximate to the correspondingone derived from the Proctor test (wopt.Proctor in Table 4).
In Table 5 the average ratio DCBR/Dw is also reported whichillustrates the sensitivity of the investigated materials to any diver-gence in the water content levels from the optimal value. Thesevalues were calculated by considering the CBR data derived fromtests in which the water content varied by ±2% from the optimalvalue (wopt.CBR), so the results were specific for a variation in water
content above (w > wopt.CBR) or below (w < wopt.CBR) the optimal va-lue. A minimum of three CBR tests were performed on each mucksample.
The data reported in Table 5 fail to indicate any benefits result-ing from the milling process. In fact, neither the CBR values ob-tained at the optimal water content (wopt.CBR), nor those obtainedat lower or higher content values exhibit any discernible trend inrelation to the crushing treatment used.
The results highlight the very high sensitivity to water contentvariation, subverting the inference derived from the Proctor com-paction study. In fact, with a variation of only 1% in water content,the CBR of materials like S1-CP, S2-CP and S3-EM became too low,reaching values that fall outside the acceptance limit for subgradesand sub-bases.
In the case of the resilient modulus test, the investigated mate-rials fall within the typical domains for reference materials. In con-trast to the CBR test, repeated triaxial load tests involve the entirevolume of the sample, and therefore the toughness of particles haslimited influence while the surface interaction occurring at thepoints of contact between grains plays a major role.
As previously mentioned, in this experimental investigation thesubgrade protocol of AASHTO T-307 (2007) was considered for thecharacterization of EM samples, while the sub-base one was usedfor the CP samples. The synthesis of results derived from experi-mental data is given in Table 6, where the two parameters k1 andk2 were obtained via regression analysis through the Hick andMonismith (1971) equation:
Table 5Mechanical properties derived from laboratory CBR tests.
Code Treatment CBRProctor CBR wopt.CBR CBR DCBR/Dw CBR DCBR/Dw
for w < wopt.CBR for w > wopt.CBR
% % % % – % –
S1 EM 154.6 154.6 5.34 34.1 60.3 30.9 61.9CP 184.4 226.8 4.37 152.5 66.1 5.8 82.0
S2 EM 105.6 105.6 6.83 45.9 29.9 36.4 34.6CP 176.4 176.4 5.97 119.5 27.7 12.6 91.0
S3 EM 149.9 203.5 3.82 183.3 18.1 3.2 59.6
S5 EM 212.1 225.3 6.97 177.8 19.0 175.9 46.4CP 166.3 180.8 6.41 106.6 23.6 – –
S6 EM 210.3 317.3 4.78 204.7 77.1 122.6 134.8
S7 CP 201.0 215.0 4.10 126.4 178.7 38.6 62.0
Remarks: CBRProctor represents the California Bearing Ratio evaluated at optimum moisture conditions derived from Proctor study, while CBR is the maximum value recordedat the water content wopt.CBR.
Table 6Hicks–Monismith regression parameters.
Humidity Dry (wopt.c � 2%) Damp (wopt.c) Wet (wopt.c + 2%)
Code Treatment k1 k2 R2 k1 k2 R2 k1 k2 R2
S1 CP 358 0.64 0.991 323 0.69 0.994 269 0.57 0.988S2 CP 335 0.60 0.989 362 0.64 0.991 326 0.66 0.993S3 EM 359 0.28 0.983 575 0.65 0.996 227 0.28 0.983S5 EM 339 0.61 0.995 – – – 294 0.72 0.998
CP 291 0.68 0.994 291 0.75 0.996 290 0.70 0.994S6 EM 239 0.27 0.982 – – – 323 0.52 0.993S7 EM 292 0.50 0.992 178 0.30 0.983 371 0.65 0.996
CP 475 0.56 0.987 456 0.65 0.992 399 0.66 0.993
Table 7Hicks–Monismith regression parameter variations between dry, damp and wet conditions.
Humidity Variation from damp (wopt.c) to dry (wopt.c � 2%) Variation from damp (wopt.c) to wet (wopt.c + 2%)
Dk1 Dk2 Dk1 Dk2
Code Treatment % % % %
S1 CP �10.8 7.2 16.7 17.4S2 CP 7.5 6.3 9.9 �3.1S3 EM 37.6 56.9 60.5 56.9S5 CP 0.0 9.3 0.3 6.7S7 EM �64.0 �66.7 �108.4 �116.7
CP �4.2 13.8 12.5 �1.5
170 P.P. Riviera et al. / Tunnelling and Underground Space Technology 40 (2014) 160–173
MR ¼ pa � k1 �hpa
� �k2
ð7Þ
where h is the bulk stress equal to r1 + 2r3 for the triaxial condi-tions, and pa is the unit reference pressure of 1 kPa used to makethe parameters k1 and k2 non-dimensional. Eq. (7) was found tobe very accurate since the coefficient of determination was alwaysgreater than 0.98.
As originally indicated by Hicks and Monismith, the resilient re-sponse is influenced by bulk density, gradation and fines content,particle roughness and angularity, and degree of saturation (whichin turn depends on the residual voids content after compaction andon water content). In particular, when granular materials are com-pared, high quality materials have larger k1 values and smaller k2
values (Rada and Witczak, 1981).In Table 6, crushed samples of alluvial mucks S1 and S2 pre-
sented similar values and trends when water content varied fromdry to wet conditions. In contrast, the third alluvial material (S3),available only in the freshly excavated form, was more sensitiveto water content variation, with higher values of both parameters
in correlation with damp conditions. Granular materials derivedfrom the mucks show values similar to the alluvial ones; S5crushed muck exhibits a stable behaviour independently of thewater content, while S6 shows lower moduli. Finally, the crushedsample of S7 muck shows a higher resilient behaviour than the ori-ginal excavated material, in particular for dry and dampconditions.
Table 7 summarises the percentage variation in the resilientmodulus parameters k1 and k2 when the mucks pass from dampto dry and wet conditions respectively. In contrast to the CBR testresults, the resilient modulus test results confirm once again thebenefits associated with the milling process since the absolutemaximum variation of k1 is 16.7% (S1), while the maximum varia-tion for k2 is 17.4% (S1); in addition, larger variations were re-corded in the case of freshly excavated mucks.
The six graphs of Fig. 8 report the comparisons between theregression curves and the typical limits of granular sub-base mate-rials for three moisture conditions (dry, damp and wet), whichwere associated with the wopt.c � 2%, wopt and wopt + 2% for bothexcavated (EM) and crushed (CP) materials respectively. Without
)PC(slairetamdehsurC)ME(slairetamdetavacxED
ry
Dam
p W
et
0
100
200
300
400
500
600
700
800
900
1000M
R [M
Pa]
MR [M
Pa]
MR [M
Pa]
MR [M
Pa]
MR [M
Pa]
MR [M
Pa]
θθ [MPa] θ [MPa]
S3 - EMS5 - EMS5 - EM trial sectionS6 - EM trial sectionS7 - EMLimits of typical values
0100200300400500600700800900
1000
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
θ [MPa] θ [MPa] 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
θ [MPa] θ [MPa] 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
S1 - CPS2 - CPS5 - CPS7 - CPLimits of typical values
0
100
200
300
400
500
600
700S3 - EMS3 - EM trial sectionS5 - EMS7 - EMLimits of typical values
0
100
200
300
400
500
600
700S1 - CPS2 - CPS4 - CP trial sectionS5 - CPS7 - CPLimits of typical values
0
100
200
300
400
500S3 - EMS5 - EMS7 - EMLimits of typical values
0
100
200
300
400
500S1 - CPS2 - CPS5 - CPS7 - CPLimits of typical values
Fig. 9. Resilient modulus and dynamic modulus comparisons: experimental data and typical ranges for granular sub-base materials.
P.P. Riviera et al. / Tunnelling and Underground Space Technology 40 (2014) 160–173 171
referring to specific cases, it can be observed that the resilientbehaviour of the materials considered is in line with literature data(Huang, 2004). In the case of low water content (wopt.c � 2%) andEM samples, the materials derived from the crushing of rocks havea lower performance than those derived from alluvial deposits (S3).Materials with optimal or high water contents exhibited resilientmoduli values within the ranges reported in literature (Huang,2004).
Field data derived from the LWD test have been plotted in thesame graph in Fig. 9. Considering the water content data reportedin Table 4, the tests on S3 and S4 were performed on damp layerswhile the tests on S5 and S6 were performed on dry layers. In thegraphs, the values have been associated with a bulk stress equal to83.6 kPa, which is the average value in the stress bulb limited to adepth of 0.4 m (Fig. 4). With the exception of the layers composedof S5-EM and S6-EM mucks which were affected by a low watercontent (dry condition in Fig. 9), in the other two cases character-ized by damp conditions (S3-EM and S4-CP) the dynamic modulusis coherent with the stiffness values derived from resilient modu-lus tests.
7. Discussion
The excavation methodology certainly affects the grading oftunnel muck. Table 2 confirms that the materials excavated bymeans of a mechanized shield like EPB and TBM are characterizedby high values of the uniformity coefficient (CU). Looking at Table 3,the effect of mechanized excavation is notable for hard, compactrocks such as the granite S6 and the calcareous schist S7 whichhave high shape and flakiness index values. In contrast, in the caseof alluvial mucks S1 and S3, the effect of the excavation methodused is, as expected, negligible (Bellopede et al., 2011; Carduet al., 2009).
Furthermore, the use of foaming biodegradable agent, for thereduction of friction and the conservation of the excavation tools,does not preclude the mechanical properties of excavated material,as has been demonstrated in the case of the alluvial sample S1.
Regarding the treatment process, on the basis of the test resultsit is possible to affirm that the use of a mobile jaw crusher resultsin a beneficial decrease in the shape and flakiness indexes, partic-ularly in the case of schistose rocks such as S7 (Fig. 4). In fact,
172 P.P. Riviera et al. / Tunnelling and Underground Space Technology 40 (2014) 160–173
elongated and flat particles tend to hinder the internal organiza-tion of grains during the compaction process (Fig. 8).
At the same time, the mechanical resistance of grains, whichwas measured by means of the Los Angeles and Micro-Deval tests,remained unchanged (Table 3), thus assuring sufficient hardnessand toughness to resist crushing and degradation during thein situ placing and compaction phases. Furthermore, Fig. 8 high-lights how the change in grain shape occurring after the grindingprocess led to a significant improvement in terms of workabilitydespite a small variation in self-compaction. The workability ofsamples S5-CP and S7-CP is very close to that of samples S1-CPand S2-CP, which include some particles with smooth surfaces asa consequence of their alluvial origin. In the case of pavementapplications, it is important to use polyhedral particles with arough surface texture to counteract the reciprocal sliding of grainsunder the effect of shear stress induced by traffic load, thus obtain-ing layers which are more deformation–resistant throughout thepavement service life.
Although certain defects revealed by qualification tests maylead to the rejection of some mucks for the production of aggregatefor high performance composite materials (i.e., concrete or bitumi-nous mixtures), all the mucks appear to be suitable for employ-ment in subgrades and/or pavement sub-bases. In fact, with theproviso that they are first subjected to a milling process, these typ-ical defects in rock spoils can be largely alleviated when used in theabove applications.
The CBR test results would support the rejection of some mucks(S1-CP, S2-CP, and S3-EM) and the careful use of the remainingones, both treated and untreated as a consequence of their highsensitivity to variations in moisture levels, but the resilient modu-lus test (performance-based) results challenge this inference. Infact, this test demonstrates that crushed mucks are, in comparisonto freshly excavated ones, less sensitive to variations in moisture.Moreover, all the data obtained from dry, damp and wet samplesfall within the typical ranges proposed in literature for currentlyused unbound granular sub-base materials.
8. Conclusions
The excavation of tunnels is an important issue for the AlpineRegion and neighbouring areas, and one that is expected to havean even greater environmental impact in the near future due tonew initiatives with very long railway tunnels and other newtransportation infrastructures. A major aspect of the manage-ment of the significant volumes of mucks generated will bethe endeavour to find possibilities for their effective use as avaluable resource rather than as simple backfilling or wastematerial.
The paper set out to make a contribution, in the form of practi-cal solutions, to this issue. In the course of an extensive researchprogram, the paper assessed the effects of the excavation method-ology (EPB, TBM, Cut & Cover, and Explosive) and the treatmentprocess (EM/CP) on the volumetric and mechanical properties ofseveral mucks that were collected as representative samples fromsome of the main infrastructures under construction on the Italianside of the Alps. The experimental program was organized by refer-ring to prescriptive and performance-based specifications and thetest results compared to traditional unbound granular materialsand soils that are currently in use.
The work focused on the laboratory characterization of sevenmucks. Furthermore, thanks to their availability in large quantities,four mucks were fragmented using a full scale portable plant andemployed in the formation of full scale layers.
The results demonstrate that the use of performance-basedspecifications and related testing methods is fundamental for the
rational evaluation of non-traditional unbound granular materialsuch as tunnel muck. The widespread application of prescriptivespecifications for the classification and selection of excavatedmaterials should, therefore, be abandoned in favour of advancedtesting methods that allow the use of non-traditional materialsand practices.
Recently, some occasional national efforts have addressed thepromotion of performance based specifications. Unfortunately,these are not always completely understood, and frequently con-sidered interchangeable with the prescriptive ones.
When performance-based specifications are adopted, excava-tion material previously regarded as waste may be usefully em-ployed in road constructions even in the formation of structuralunbound layers such as subgrade and sub-base. This means thatthe value of tunnel muck can be substantially increased from thetypical low price quoted for poor quality backfilling or embank-ment material, attaining its optimum added value.
Acknowledgements
The investigations described in this paper were carried out inthe laboratories of the Department of Environment, Land and Infra-structures Engineering (DIATI) of the Politecnico di Torino.
The research presented in the paper refers to the activities car-ried out by the WP4 and WP6 of the Remuck Project (Title: Innova-tive methods for the eco-compatible and sustainable recycling ofmuck from tunnel excavation, also considering the potential con-tent of noxious minerals) funded by Regione Piemonte (CIPE 2006).
This research has been made possible thanks to the cooperationof: BBT, SCR Piemonte, AK Ingegneria Geotecnica S.r.l and GTT,which provided the material used to conduct the tests and RADISSpa and CO.GE.FA. S.p.A. (partners in the Remuck Project) whichwere involved in the transportation and treatment of tunnel mucksamples.
References
AASHTO Highway Subcommittee on Construction, 2003. Major Types ofTransportation Construction Specifications. A Guideline to UnderstandingTheir Evolution and Application. Report of the Quality Construction TaskForce. American Association of State Highway and Transportation Officials,Washington D.C., August 2003.
AASHTO M145, 2008. Standard Specification for Classification of Soils and Soil-Aggregate Mixtures for Highway Construction Purposes. American Associationof State and Highway Transportation Officials.
AASHTO T180, 2010. Standard Method of Test for Moisture-Density Relations ofSoils Using a 4.54-kg (10-lb) Rammer and a 457-mm (18-in.) Drop. AmericanAssociation of State and Highway Transportation Officials.
AASHTO T-193, 2010. Standard Method of Test for the California Bearing Ratio.American Association of State and Highway Transportation Officials.
AASHTO T-307, 2007. Standard Method of Test for Determining the ResilientModulus of Soils and Aggregate Materials. American Association of State andHighway Transportation Officials.
AASHTO T312, 2009. Preparing and Determining Density of Hot Mix Asphalt (HMA)Specimens by Means of the Superpave Gyratory Compactor. AmericanAssociation of State and Highway Transportation Officials.
Bassani, M., Santagata, E., Baglieri, O., Ferraris, M., Salvo, M., Ventrella, A., 2009. Useof vitrified bottom ashes of municipal solid waste incinerators in bituminousmixtures in substitution of natural sands. Advances in Applied Ceramics 108(1), 33–43.
Bellopede, R., Brusco, F., Oreste, P., Pepino, M., 2011. Main aspects of tunnel muckrecycling. American Journal of Environmental Sciences 7 (4), 338–347.
Burdin, J., Monin, N., 2009. The management of excavated materials from the Lyon-Turin rail link project. Geomechanik und Tunnelbau 2 (5), 652–662.
Cardu, M., Oreste, P., Cicala, T., 2009. Analysis of the tunnel boring machineadvancement on the Bologna-Florence railway link. American Journal ofEngineering and Applied Sciences 2 (2), 416–420.
Commission of the European Communities, 2005. Towards a Thematic Strategy onthe Sustainable Use of Natural Resources. Communication from theCommission to the Council, the European Parliament, the European Economicand Social Committee and the Committee of the Regions.
Commission of the European Communities, 2008. The Raw Materials Initiative —Meeting Our Critical Needs for Growth and Jobs in Europe. Communicationfrom the Commission to the European Parliament and the Council.
P.P. Riviera et al. / Tunnelling and Underground Space Technology 40 (2014) 160–173 173
EN 932-3, 2004. Tests for general properties of aggregates. Procedure andterminology for simplified petrographic description. European Committee forStandardisation.
EN 933-1, 1999. Tests for Geometrical Properties of Aggregates. Determination ofParticle Size Distribution – Sieving Method. European Committee forStandardisation.
EN 933-3, 2003. Tests for Geometrical Properties of Aggregates. Determination ofParticle Shape. Flakiness Index. European Committee for Standardisation.
EN 933-4, 2008. Tests for Geometrical Properties of Aggregates. Determination ofParticle Shape. Shape Index. European Committee for Standardisation.
EN 933-8, 2000. Tests for Geometrical Properties of Aggregates. Assessment ofFines. Sand Equivalent Test. European Committee for Standardisation.
EN 933-9, 2009. Tests for Geometrical Properties of Aggregates. Assessment ofFines. Methylene Blue Test. European Committee for Standardisation.
EN 933-10, 2009. Tests for Geometrical Properties of Aggregates. Assessment ofFines. Grading of Fillers (Air-jet Sieving). European Committee forStandardisation.
EN 1097-1, 2004. Tests for Mechanical and Physical Properties of Aggregates –Determination of the Resistance to Wear (Micro-Deval). European Committeefor Standardisation.
EN 1097-2, 2008. Tests for Mechanical and Physical Properties of Aggregates.Methods for the Determination of Resistance to Fragmentation. EuropeanCommittee for Standardisation.
EN 1097-6, 2000. Tests for Mechanical and Physical Properties of Aggregates.Determination of Particle Density and Water Absorption. European Committeefor Standardisation.
EN 13242, 2008. Aggregates for Unbound and Hydraulically Bound Materials for Usein Civil Engineering Work and Road Construction. European Committee forStandardisation.
EN ISO 14688-2, 2004. Geotechnical Investigation and Testing. Identification andClassification of Soil. Principles for a Classification. European Committee forStandardisation.
European Commission, 2005. Energy and Transport DG. Trans-European TransportNetwork (TEN-T Priority Axes and Projects 2005). Office for Official Publicationsof the European Communities. ISBN 92-894-9837-4.
European Aggregate Association, 2010. Estimates of Aggregates Production Data2010. <http://www.uepg.eu/statistics/estimates-of-production-data/data-2010> (accessed 24.12.12).
European Commission, 2011a. Roadmap to a Single European Transport Area –Towards a Competitive and Re Source Efficient Transport System. CommissionStaff Working Paper, Brussels.
European Commission, 2011b. Report from the Commission to the EuropeanParliament, the Council, the European Economic and Social Committee and theCommittee of the Regions on the Thematic Strategy on the Prevention andRecycling of Waste.
European Parliament and The Council of the European Union, 2008. Directive 2008/98/EC on Waste and Repealing Certain Directives. Official Journal of theEuropean Union, L. 312/3, November 22nd, 2008.
Gertsch, L., Fjeld, A., Nilsen, B., Gertsch, R., 2000. Use of TBM muck as constructionmaterial. Tunnelling and Underground Space Technology 15 (4), 379–402.
Grunner, K., Durmekova, T., Wagner, P., 2003. Exploitation possibilities of thetunnel’s muck as a construction material. Slovak Journal of Civil Engineering 1,11–25.
Hicks, R.G., Monismith, C.L., 1971. Factors influencing the resilient properties ofgranular materials. Transportation Research Record, Transportation ResearchBoard, National Research Council 345, 15–31.
Huang, Y.H., 2004. Pavement Analysis and Design, second ed. Prentice Hall.Kwan, J.C.T., Jardine, F.M., 1999. Ground engineering spoil: practices of disposal and
reuse. Engineering Geology 53 (2), 161–166.Lieb, R., 2009. Material management at the Gotthard base tunnel – experience from
15 years of construction. Geomechanik und Tunnelbau 2 (5), 619–626.Mallick, R.B., El-Korchi, T., 2009. Pavement Engineering: Principles and Practice. CRC
Press, pp. 156–157.Ministero delle Infrastrutture e dei Trasporti, 2001. Studio a Carattere Pre-
normativo delle Norme Tecniche di Tipo Prestazionale per Capitolati Specialid’Appalto. Roma (in Italian).
Olbrecht, H., Studer, W., 1998. Use of TBM chips as concrete aggregate. Materialsand Structures/Materiaux et Constructions 31, 184–187.
Politecnico di Torino, 2012. Innovative Methods for the Eco-compatible andSustainable Recycling of Muck from Tunnel Excavation, also Considering thePotential Content of Noxious Minerals. Remuck Project. <http://areeweb.polito.it/ricerca/remuck/index.php/workshop-finale.html> (accessed24.12.12).
Rada, G., Witczak, M.W., 1981. Comprehensive evaluation of laboratory resilientmoduli results for granular material. Transportation Research Record,Transportation Research Board, National Research Council 810, 23–33.
Resch, D., Lassnig, K., Galler, R., Ebner, F., 2009. Tunnel excavation material – highvalue raw material. Geomechanik und Tunnelbau 2 (5), 612–618.
Rete Ferroviaria Italiana, 2012. Migration Strategy in Italy. National Plan to Developand Implement ERMTS for the Railway Infrastructures. <http://ec.europa.eu/transport/modes/rail/interoperability/ertms/doc/edp/national_deployment_plans/italian_migration_strategy_en.pdf> (accessed24.12.12).
Ritter, A., Einstein, H.H., Galler, R., 2013. Planning the handling of tunnel excavationmaterial – a process of decision making under uncertainty. Tunnelling andUnderground Space Technology 33, 193–201.
Schröfelbauer, T., Schreitl, B., Kitzler, C., 2009. S1 Danube-Lobau tunnel – recyclingof tunnel spoil material. Geomechanik und Tunnelbau 2 (5), 633–642.
TP BF-StB: Part B 8.3, 2003. Dynamic Plate-Load Testing with the Aid of the LightDrop-Weight Tester. Technical Test Code for Soil and Rock Mechanics in RoadConstruction, Road and Transportation Research Association, Working Groupfor Foundation and Soils Engineering.
Thalmann-Suter, C., 1997. Tunnel muck: troublesome disposal material or potentialconcrete aggregate? Tunnel, 23–24.
Thalmann-Suter, C., 1999. Concrete Aggregate Production with TBM MuckExplained on the Alptransit Tunnel Projects. Utilizing Ready-Mixed Concreteand Mortar. Thomas Thelford, pp. 131–144.
World Longest Tunnel Page. <http://www.lotsberg.net/> (accessed 24.12.12).
