-1-

FIRST TIME APPLICATION OF EXPANDED POLYSTYRENE

IN HIGHWAY PROJECTS IN GREECE

Georgios Papacharalampous, Civil Engineer MSc, Sotiropoulos & Associates SA, Greece, mail:geomsot@gmail.com

Elias Sotiropoulos,

Civil Engineer MSCE, Sotiropoulos & Associates SA, Greece ABSTRACT The first application of expanded polystyrene (EPS) in civil engineering projects in Greece was completed recently for the construction of an embankment in a section of the Athens-Thessaloniki highway, very close to the historical site of Thermopylae. At a length of about 1km the highway, including an interchange, was planned on 8.5m high embankments at a flat, swampy area, with an elevation close to the sea level, covered by recent marine deposits. When the construction of the initially planned earth fill, resting on ground improved by stone columns, reached 4m height, a deep seated failure of the underlying soft clay and total collapse of the constructed embankment at its whole 200m length took place. A new design had to face short term stability, but also long term consolidation settlements and resistance to earthquake. Furthermore, due to its major importance and serious traffic problems at this area, the project had to be finished in the shortest possible time. A lightweight fill solution with EPS fulfilled all above requirements. By a substantial application of EPS100 type blocks (γ=20kg/m3, σc10 = 100kPa) at a total length of ≈1000m and a volume of ≈ 65,000m3, short and long term stability was ensured, settlements were minimized and construction time was drastically reduced to a total period of 2.5 months. The current paper presents the ground conditions in the area, design criteria and requirements, EPS material specifications, the proposed quality control and data and conclusions from the construction and monitoring. PROJECT DESCRIPTION – HISTORICAL AND GEOLOGICAL BAC KGROUND A lightweight fill by expanded polysterene (EPS) blocks has been designed and constructed for a 1km long section of Athens-Thessaloniki E75 major highway during 2008. This was the first and still is the only application of EPS in highway projects in the country. The project is located in central Greece, south of Maliakos Gulf and the city of Lamia, very close to the area of the historical battle of Thermopylae (480 B.C.), where Leonidas, king of Sparta with a few heroic warriors hindered the Persian invasion to Europe. The battle took place at a narrow passage between the sea and the mountains, a clear evidence that the extended ma-rine and deltaic deposits, which now cover the area did not exist at that time (Figure 1). The highway crosses a plain, swampy area, with very poor ground characteristics and mechani-cal properties. Ground elevation is ≈+1.0m above sea level and the ground water table is almost at surface. The area is also suffering from strong earthquakes.

-2-

Figure 2. View of failure

The initial design anticipated a highway embankment up to 7.5m high to allow underpass of lo-cal road and culverts and an even higher 8.5m approach embankment for Thermopylae interchange branch overpass. It was foreseen replacement of the top soil to the depth of 1m by granular material and soil improvement by stone columns of 80cm diameter, 11-15m length, at 3m spacing in a triangular grid. When construction of the highway embankment reached a height of 4÷4,5m in a relatively short period a global failure at its whole 200m length took place. A continuous, more than 2m wide rupture, appeared at the crown, parallel to the alignment and at a distance 10÷14m from the edge. The outer part subsided about 1m (Figures 2 and 3). GROUND CONDITIONS AND FAILURE ANALYSIS Investigation of the observed failure mechanism and the ground conditions included: (a) survey mapping and deformation monitoring, (b) geotechnical investigation of the subsoil and the stone columns by trial pits and boreholes and (c) back analysis of stability.

Ground properties were found much more adverse than the rest of the deltaic area, where both the highway and the railway have been constructed. The soil profile is shown in Figure 3. The embankment is founded on a very soft, clay, of high plastic-ity, to a depth of 10÷12m, underlain by compressible, normally consolidated, clayey deposits to a depth ≈ 25m.

Collapse is attributed to deep seated, un-drained failure of the very soft clay (Α) to depth 5-10m from GL. Failure surface is not typically circular, but more complex, due to

Figure 1. Location of the project and layout of the site

-3-

(a) (b)

Figure 4. a) Back analysis of failure, b) Recorded settlements of railway embankments

Figure 3. Typical section at failure - Soil profile

the stone columns effect. Soil resistance was evidently overes-timated and soil improvement method has not functioned as planned. The stone columns had a negligible contribution to total resistance, since the soft clay ex-hibits limited side support. They were found to retain their integri-ty out of the failed area, but with reduced diameter and inferior quality with depth. Below the failed embankment they were highly mixed with clay, with a distinct rupture at depth ≈ 3÷5m.

The undrained strength of clay was estimated by back analysis of observed failure (Figure 4a). Stone columns were introduced with a reduced diameter (0.5÷0.6m) and strength (friction angle φ=32°). A scenario of lower soil strength of the upper layer, where important organic content was identified, was also examined. A safe design had to ensure both short term and long term stability and also to minimize consol-idation settlements, expected to take place in clay layers to 25m depth. Magnitude and rate of settlements were estimated by taking into account the geotechnical investigation and testing da-ta, pre-failure settlement recordings and also monitoring data of higher embankments of the neighbouring railway project. The railway embankments suffered significant long term settle-ments (Fig.4b), although the ground conditions were quite better and plastic drains were implemented. A total settlement in the order of 1.5m for a 7.5m high earth embankment was es-timated. Analysis showed only 20÷25% of this settlement to take place during construction and the rest to be quite slow, with an expected degree of consolidation U = 50% in the first year and a total consolidation period greater than 10 years.

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

0 200 400 600 800 1000 1200

time (days)

S/H

(se

ttle

men

t/ em

bank

men

t hei

ght)

-4-

SELECTION OF A SOLUTION SATISFYING THE PROJECT’S RE QUIREMENTS The project had to fulfil the following criteria and requirements: - Stability and safe foundation on 10m thick very soft clay layer (safety factor ≥ 1.20 short term) - Post construction settlement to be ≤10cm according to Greek guidelines for highways - Necessity of rapid completion, due to serious traffic problems, with political and social impact At first a change of the longitudinal alignment was achieved, by removing the Underpass out of the critical area. Consequently the highway embankment was lowered to ≤4.5m height, though the interchange branch had to remain at 8.5m. Alternative geotechnical improvement solutions have been examined, such as: i) stage construc-tion and preloading and use of vertical drains and stabilizing berms, ii) soil improvement by a denser vibroreplacement – stone columns grid, iii) improvement by deep soil mixing combined with vertical drains and iv) lightweight EPS fill. The EPS solution has been found highly advan-tageous, in terms of bothsafety, time and maintenance cost, with a considerably lower construction cost from alternatives (ii) and (iii). Solution (i) has been considered very slow and uncertain, mainly in time programming, being almost impractical for the higher I/C embank-ment. Some basic conclusions of the comparison of the various alternatives are briefly given in Table 1 concerning only the highway section. For the much higher branch embankment the EPS solution proved the only realistic one. Table 1: Comparison of alternative solutions (only for the highway section) Solution Constr. time Cost (€) Advantages Limitations Preloading, PVD, berms, stage constr.

9 months (incl.preload)

3.6 M Cost, expe-rience

Time, uncertainties, ex-propriation, installation

Stone columns l=20m, s=1,5m

> 1 year (incl.preload)

10.6 M - Cost, time, reliability of SC, maintenance

Deep mixing l=10m, PVD l=25m

> 9 months (incl.preload)

6.2 M - Cost, time, success de-pended on ground

EPS fill 2 months 5.8 M Time, maintenance, unrelated to subsoil

Cost, lack of ex-perience

Hence authorities and involved engineers were convinced for the EPS “unconventional” solu-tion, despite their initial concern about strength, durability and sustainability to hazards like fire, water absorption, oil spill or attack by rats, being also informed about satisfactory behaviour in other projects worldwide. It should be noted that an EPS application proposal, some years earlier in 2003, for high embankments at a neighbouring area was rejected, for the reasons of doubt about above virtues of EPS. EPS fill was applied at ≈820m length of the Highway and ≈210m of the I/C branch. Solution (i) was chosen at the rest 500m, where the lower (H≤2.5m) embankments rest on slightly better ground, since construction time was found reasonable and berms were not necessary. LIGHTWEIGHT FILL DESIGN EPS fill design aimed to eliminate restrictions imposed by a heavy earth embankment, taking also into account the steep one way cross slope of the road surface (≈5%). Proposed scheme as shown schematically in typical section of Fig. 5 below comprised the following:

-5-

Figure 5: Typical section of proposed EPS fill

- Complete removal of the existing fill and the top soil to a depth 0.75m from GL - Placement of granular material at the base, of total thickness ≥50cm, comprising from bot-tom to top by: a) 25cm selected material serving as working platform, b) separation geotextile and tensile reinforcement geogrid, c) drainage-foundation and levelling granular layer of thick-ness varying from 25cm to 75cm, with a slightly inclined roof oriented parallel to road surface - Placement of EPS 100 type blocks, with minimum dimensions 0.5x1.0x2.0m, on top of the inclined levelling surface. Consecutive layers are placed at right angles to each other in order to avoid continuous vertical joints, with the upper row aligned transversal to the axis direction. A stepped side slope similar to embankment’s final slope 2:3 (v:h) is formed. Steel fasteners pre-vent blocks from moving out of position during the construction process. - Side slopes are protected by 1mm polyethylene membrane and a geotextile cloth covered with ≈0.5m lightly compacted fine granular soil and 0.3m top soil - By casting a 15cm lightly reinforced concrete slab C20/25 on top of EPS fill, protection of fill against solvents and pavement’s proper foundation and drainage is achieved. - Pavement retained its characteristics as determined by the original design: 22cm asphalt lay-ers, 40cm granular base layer, 50cm foundation and drainage layer on top of inclined slab. At the junction of lightweight and earth embankment a stepped surface is formed every 0.5m of the earth fill and a drainage pipe is placed at the base. The stepped interface is covered by granu-lar material, geotextile and sealing membrane and EPS blocks are placed on top.

Material Requirements Initially the material properties were specified on the basis of ASTM standards [2] and NRRL recommendations [3]. Gradually the recent European Standard ΕΝ 14933 [1] was adopted. Hence material type EPS 100 RF of density 20kg/m3, nominal compressive strength σc10 = 100kPa at 10% strain, allowable stress σc2 = 50kPa at 2% elastic deformation and long term wa-ter absorption Wdv ≤5% has been specified. This type satisfies the design criteria and is also the most commonly used in relevant projects worldwide. Flame-retardant material of Euroclass B (EN13501-1) was chosen, in order to exclude fire hazards. Properties are summarized in Table 2.

-6-

Table 2: Properties and Quality control for EPS Description Test method Symbol -

unit Requirement Tolerance Control fre-

quency Dimensions EN 822-3 LxWxT

(m) ≥2.0x 1.0x0.5 ± 0,5% every 50

blocks Squareness EN 824 (x,y,z) 3mm/ 500mm length Evenness EN 825 (±d/L) 10mm/ 3m or

5mm / consecutive blocks Cross section every 10m

Unit Density ΕΝ 1602 ρ (kg/m3) 20 every 1000m3

Compressive strength

ΕΝ 826 σc10 (kPa) ≥ 100 ± 10% σc2 (kPa) ≥ 50

Bending strength EN12089 σb (kPa) ≥ 150 DIMENSIONING AND ANALYSIS Dimensioning of EPS fill was made on the basis of the following analyses and checks based on European Standard [1] and USA [4] and Norwegian [3] recommendations: External Stability and Soil Settlements The global stability of the lightweight embankment interacting with the foundation soil, was checked on typical cross-sections, for circular sliding surfaces through the very soft clay layer A. Both short term, undrained conditions, long term effective stress analysis and earthquake loading according to Greek Seismic Code have been examined by Bishop’s slices method and the use of software SLIDE. An equivalent shear strength Su=100kPa and a bulk density γ = 0.5 kN/m3 was considered for the EPS fill, where soil parameters were taken according to geotechnical investi-gation results and back analysis. Analysis yielded very high safety factors for all cases. A factor of safety F.S. =4.3 was calculated for the 8.5m highest embankment for the case of undrained, deep seated failure compared to 0.73 for an earth embankment (critical case). Total and long term consolidation settlements of the compressible subsoil to 25m depth was cal-culated by 2D plane strain analysis with software FOSSA. Preloading by the 1÷1.5m previously constructed earth fill was taken into account for the estimation of post construction settlements. Total settlement was estimated ≈20cm, almost exclusively in upper layer A, since there is negli-gible influence to deeper layers. Post construction settlement after preloading was found less than 5cm compared to respective settlement >1.5m for an earth embankment. Buoyancy - uplift Base drainage layer and side ditches protect lightweight fill from the incident of flooding and uplift failure due to buoyancy forces. Further to the protection measures, the case was checked by the following calculations, based on NRRL [3] and TRB [4] recommendations: Global stability: F.S. = {Wp + Wsl + Weps }/ A = 1.96 > 1.3 (1a) Local slope uplift : F.S. = (γ Hsl )/ [γw x (D+HWL)] > 1.2 (1b) Wp = pavement and slab weight, Wsl = slope earth cover weight, γ = soil unit weight = 20kN/m3

-7-

Weps = EPS fill weigth, considering unit weight γeps =0.2kN/m3, A = buoyancy force = γw x (D+HWL) x Bw (Bw = bottom embankment width) Hsl = minimum slope cover = 0.8m, D = maximum depth of EPS below ground level = 0.25m, HWL = maximum probable water level above GL = 1.0m, γw = water unit weigth = 10kN/m3 Since construction took place in summer period and protection ditches and pumping was fore-seen neither uplift during construction nor transition due to an unbalanced water pressure, were considered as potential risks and have not been analytically examined. Uplift check due to wind forces was also omitted, due to lack of consistent methodology in the literature and reporting of relevant problems worldwide (ΤRB, [4]) Internal stability Sliding between two consecutive EPS layers was checked. The critical surface lies between the upper layers, where both minimum overburden weight and friction resistance and utmost seismic action occur. On the basis of TRB recommendations [4] three loading cases were examined: static (due to inclined layers): F.S.s = Wp x tan(δ)/Wt ≥ 2,0 (2a) wind force: F.S.w = Wp x b x tan(δ)/(Pxh+Wtxb) ≥ 1,2 (2b) seismic: F.S.e = [Wp–Ehsin(α)-Εvcos(α)] b x tan(δ) / (Wt+Εhcos(α) +Εvcos(α)) ≥ 1.0 (2c) Wp,t = overburden pressure (weight of pavement and concrete slab) acting in direction perpen-dicular and parallel to sliding surface respectively, δ = friction angle between EPS layers = 300, α = inclination (cross fall) of EPS layers = 3.15º P = wind force at fill crown, perpendicular to slope, b, h = width and height of EPS layer, Ev,h = vertical and horizontal earthquake force according to Greek Seismic Code Stresses and deformations in EPS Allowable stresses and deformations in EPS fill were checked for the three cases below: a) Control of permanent loads to allowable limits in order to avoid considerable creep defor-mations. Pavement and concrete slab loads (p) acting on top of EPS do not exceed 20-25kPa. For stresses of that order, which correspond to p < 30% σc10 = 30kPa (Frydenlund [5]) and to <1% immediate strain (TRB [4]), tolerable creep deformations are expected. b) Estimation of elastic, immediate deformation, taking place mainly during construction. Calculation formula: sEPS = pxHEPS/ EEPS ≤ 20x8/5000, assuming a deformation modulus EEPS = 5 MPa, (TRB [4]), provided settlements in the order of 1÷3cm, which is negligible compared to the settlement of the subsoil c) Deformations under repetitive traffic loads to be within acceptable elastic limits. Traffic loading (q) on top of EPS was roughly estimated by considering a load distribution coefficient of Ic = 1 for the pavement of thickness Tp,min =0,77m and Ic = 2 for the concrete slab of thickness Tc = 0,15cm. Exctracted traffic load q≤15.5kPa << 0.35 σc10 and total load q+p ≈ 35 kPa < σe =50 kPa, where σe = minimum elastic limit stress (TRB [4]), were found to be within acceptable limits with a considerable safety margin.

-8-

Figure 7: View of EPS construction

Figure 8: Positioning of blocks

Analytical calculations for Pavement design and foundation were not performed. Since this was the first application of EPS fill in Greece, it was decided to keep the original pavement scheme for the earth embankment, with a requirement of CBR=3, disregarding the presence of the con-crete slab below. That was a conservative consideration and no further check has been required. CONSTRUCTION DATA Contractor was ATTIKAT S.A. and owner of the project the Greek State, Ministry of Public Works. Design of EPS fill was made by geotechnical office Sotiropoulos & Associates SA. Technical experts K.Flaate, Τ.E. Frydenlund and A. Sagbakken from Norwegian Public Roads Administration, as consultants both during design and construction in situ, offered valuable technical support and confidence for the project’s success. Construction of lightweight fill was completed in 2.5 months (06-08/2008) and the road was opened to traffic, inaugurated by the Min-ister of Public Works at 18/09/08. A total volume of ≈ 65,000 m3 of EPS was applied at a total length of 1,030m (Figure 7), with average production rate ≈ 1000 m3/day, including protection measures and quality control. De-spite the big controversy and initial hesitations about the first use of an EPS fill, regarding its strength, suitability and durability, rapid and successful application impressed the relevant au-thorities and the press which was interested in this “peculiar” construction. Block dimensions were adapted to production availability, taking into account the large volume required and the need of construction acceleration. Hence blocks of various dimensions were used, of length 3÷5m, width 1.0÷1.35m and thickness 0.6÷1.0m, with the thicker blocks placed at the upper layer. Application design noted block types at every layer in the whole area. Adjust-ing block dimensions on site was generally avoided and wherever necessary, i.e. at junctions or adjacent to structures, it was achieved by an electric wire cutting system proved to be effective, fast and providing a fairly accurate cut with straight and even cutting surfaces. Gaps be-tween blocks were eliminated and at points of curvature triangular gaps up to 5cm were filled with sand. With the exception of edge sections a minimum of 2 layers was used. At initial construction stage the following re-finement actions took place to overcome some identified problems: - observed curvature on some blocks originated from the way blocks were stored at the factory, has been overcome by placing two longitudinal wooden beams beneath the bottom block

-9-

Figure 9: Summary of compression test results

- rearrangement of some blocks to avoid continuous transverse joints by placing the blocks shifted about half a block length, by making adjustments to block lengths and by placing the shorter blocks to the inner side of the fill (Figure 8) - random gaps between EPS fill and adjacent concrete structures was filled by polyurethane foam and sealing membrane was fixed on EPS QUALITY CONTROL AND MONITORING Quality control and frequent testing as shown on table 2 was based on relevant recommendations of EN14933 [1], whether frequency and tolerances were specified according to TRB [4] and NRRL [3] recommendations. Dimensions, squareness, density and bending and compressive strength were tested and controlled at the production plant and flatness was checked on site, fol-lowing the specifications of Table 2. The quality control performed at the production plant was both automatic and manual and well organized both regarding the actual production process and the quality control checks on the finished EPS product. In addition the owner performed individ-ual tests assisted by an independent third party laboratory. Quality certificates were submitted for blocks delivered to the construction site. Initially strength of the delivered EPS-blocks was determined at 10 % strain following EN 13163:2001 (for build-ings). A need of compliance to recent ΕΝ14933 and control of stress corre-sponding to elastic behaviour led to modification of control at three different strain levels: 2 %, 5 % and 10 %, by ex-amining the whole load – deformation curve and an additional requirement of σc 2% ≥ 50 kPa. Tests were performed on 50mm cubic samples. 3 measurements were assessed for each test result [1], on specimens taken both from inner and the outer areas of the block according to recommendations [3]. The extensive QC program proved full compliance of the design requirements. Designated strength σc10 varied from 112 to 121kPa, unit density 20.3÷21.2 kg/m3 and bending strength 200÷250 >>150kPa. Material exhibited an almost linear behavior up to 4% strain, with σc2>50kPa and σc5 ≈ 0,9σc10 ≈ 100kPa. Figure 9 shows load-deformation curves in compression by average results from various plants and independent laboratory compared to design criteria. Geotechnical instrumentation and settlement’s monitoring was provided by 11 benchmarks in-stalled on top of EPS fill, at regular spacing every 100÷150m along the road. Measurements had to be taken at weekly basis. Recorded settlements were almost uniform and very low, in the or-der of 1cm after the project’s completion, much lower than design’s prediction (Figure 10a). This is attributed to low applied loading, stiffer behaviour of preloaded soil at small strains and uniform distribution achieved through EPS and rigid base layer. Figure 10b compares recorded settlements of LW fill in the first 3 months after construction, with relevant recordings at two neighboring sections of the highway, where earth fill 4m and 7m high respectively, was founded on soil improved by stone columns. It should be noted that

0

50

0

10096

106.2

52.3

118.7

0

20

40

60

80

100

120

140

0 2 4 6 8 10

Rel. Deformation (%)

Com

pres

sive

str

ess

(kP

a)

Specification

Production Plant 1

Production Plant 2

IndependentLaboratory

-10-

ground conditions in these two areas are much better than the EPS area, emphasizing EPS solu-tion’s superiority in terms of time and maintenance. The section exhibits today, 3 years after its completion very satisfactory performance, best among the various highway sections in the valley area and no settlements or other problems have been observed. (a) (b)

Figure 10: Settlement measurements and comparison with neighbouring earth embankments CONCLUSIONS The first application of lightweight fill by expanded polystyrene in Greece was realized in an ex-tended area with very poor ground conditions, in order to face important foundation problems, to overcome observed failures and to comply with project’s time schedule. Construction of em-bankments up to 8.5m by the use of a substantial EPS amount proved very successful in terms of safety, construction time and overall post construction performance after 3 years. The effective-ness of this first application, despite initial controversy and hesitations about the solution and the use of EPS in highway projects in general, honored a positive reputation to the project and the material. EPS became quite attractive to designers, contractors and many authorities for future application in road embankments and even more in bridge abutments. REFERENCES 1. ΕΝ 14933, “ Thermal insulation and Light Weight fill products for Civil Engineering ap-

plications - Factory made products of Expanded Polystyrene (EPS)”, Specification, 2007

2. ASTM D578-95, “Standard Specification for Rigid, Cellular Polystyrene Thermal Insula-tion”, Vol. 04.06, USA, 1999

3. Norwegian Road Research Laboratory, “Expanded Polystyrene used in road embank-ments”, Code of Practice, Form 482E, Oslo, 1992

4. Transportation Research Board NA, “Geofoam Applications in the Design and Construction of Highway Embankments” , NCHRP65, Illinois, 2004

5. Frydenlund, Τ.Ε. and Aaboe, R., “Long term performance and durability of EPS as a

lightweight filling material”, EPS Geofoam 3rd Intern. Conf., Salt Lake City, Utah, 2001

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

19/09/08 09/10/08 29/10/08 18/11/08 08/12/08 28/12/08

date

settl

emen

t (c

m)

620 625

630 635

638 647

652 657

662 667

0

2

4

6

8

10

12

14

16

18

0 20 40 60 80 100 120 140 160 180

time after end of construction (days)

settl

emen

t (cm

)

EPS, avg.10 rec.SC, h=3.5-4m, avg. 2 rec.SC, h=6.5-7m, avg. 6 rec.