CONCEPTION DES CHAUSSÉES: Approache mécanique, caractérisation des matériaux et leur évaluation dans les plates-formes routières

Transcrição

1 CONCEPTION DES CHAUSSÉES: Approache mécanique, caractérisation des matériaux et leur évaluation dans les plates-formes routières António Gomes Correia Université de Minho, Portugal Président TC3 - SIMSG

2 PRESENTATION INTRODUCTION PAVEMENT MODELLING AND DESIGN NON LINEAR BEHAVIOUR OF SOILS AND UGM QUALITY ASSESSMENT - LABORATORY Simple tests Functional tests Mechanical behaviour - Cyclic triaxial tests QUALITY ASSESSMENT - FIELD Routine analysis Advanced analysis FINAL REMARKS

3 EMPIRICAL PAVEMENT DESIGN MECHANISTIC PAVEMENT DESIGN MOVE FROM EMPIRICAL SPECIFICATIONS INDEX PROPERTIES TO PERFORMANCE-BASED SPECIFICATIONS MECHANICAL PROPERTIES

4 PAVEMENT MODELLING & DESIGN

5 PAVEMENT MODELLING 2D / 3D

6 3. FEM calculations (2D / 3D) - Visco-elastic and elastoplastic models - Implemented in FEM codes and still in development to predict permanent deformations. Material properties: General stress-strain. MECHANISTIC PAVEMENT DESIGN 1. Multi linear elastic systems calculations - linear elastic models - most widely used for pavement design and back analysis calculations. Material properties: Modulus (E) and Poisson ratio (ν) FEM calculations (2D / 3D) Elastic non-linear models - Available in some codes after Science-EC project (1993). Material properties: Elastic and elastic non-linear stress-strain relationships (K(p,q); G(p,q)) variants of Boyce model. Visco-elastic and elastic non-linear stress-strain models - Implemented in FEM codes during COURAGE (1999). Material properties: Elastic non-linear stress-strain; temperature and strain rate dependent.

7 PAVEMENT DESIGN PROCESS Modèles des matériaux Dimensionnement Routine Niveaux dimensionnement Mod. réponse (mod. structurel) Mod. calage Critère de dim. Avancé Recherche Besoin d une coopération entre le responsable de la modélisation et le responsable des investigations in-situ et en laboratoire

8 QUALITY ASSESSMENT LABORATORY

9 QUALITY ASSESSMENT Laboratory tests SIMPLE TESTS: to general material assessment (petrography, flakiness, plasticity) and fragmentation assessment (LA, MBg, DSC, MDE, Gyratory) - Uses a limited portion of material s parent grading curve - Not able to predict the overall behaviour of the material. FUNCTIONAL TESTS: Construction level - to assess the bearing capacity of UGM to carry loading applied (shear strength test) and spreading the loading (RLT- stifness). Long term behavior - to assess stifness - capacity of spreading the loading (RTL stifness) and resistance to permanent deformation (RLT - permanent deformation )

10 SELECTION OF MOST APROPRIATED MATERIAL based in simple tests (SOIL AND UGM) for most part experienced based Similar to AC volumetric design (Marshall method)

11 A RATIONAL APPROACH TO LABORATORY STUDY OF MECHANICAL BEHAVIOUR Strength & Stiffness

12 NON LINEAR BEHAVIOUR OF SOILS AND UGM

13 NON LINEARITY INCREASES WITH INCREASING RELATIVE DENSITY AND GRAIN CRUSHABILITY NON LINEAR ENVELOPE OF MAXIMUM SHEARING STRENGTH - UGM Shear stress, τ Peak strength Strength at critical state φ p φ cv Normal effective stress, σ

14 NON LINEAR BEHAVIOUR DEFINITIONS OF MODULUS σ E 0 E sec E tg E E 0 E sec E tg ε lg ε

15 CYCLIC TRIAXIAL TEST PERFORMANCE TEST

16 EUROPEAN STUDY - COURAGE Grading curves Material A, granite, IST Portugal Material B, gneiss, LCPC France Material C, limestone, ZAG Slovenia 70 Passing [%] Sieve size [mm]

17 CYCLIC STRESS PATHS CYCLIC TRIAXIAL TESTS OF UGM (CEN prenv ) 600 q (kpa) E4 D4 500 E3 D3 400 E2 C4 300 D2 C3 F2 200 E1 F1 D1 C2 B C1 B2 B1 A1 A2 A3 A4 p (kpa) B3 VARIABLE CONFINING PRESSURE CEN - method A

18 CYCLIC STRESS PATHS CYCLIC TRIAXIAL TESTS OF UGM 300 (CEN prenv ) q (kpa) CONSTANT CONFINING PRESSURE CEN - method B p (kpa)

19 CYCLIC TRIAXIAL TEST (ECP) Local strain measurements

20 CYCLIC TRIAXIAL TEST (ECP) Results of strain measurements

21 CYCLIC TRIAXIAL TEST (LPC)

22 CYCLIC TRIAXIAL TEST (UMinho)

23 SOIL / SOIL+ROCK / ROCK

24 TYPICAL MOISTURE & DENSITY CONDITIONS FOR UGM - BASE WATER CONTENT w DRY w OPM - 4 w OPM - 2 w OPM - 1 DENSITY 1 Specimen 100 % ρ d OPM 1 Specimen 2 Specimens 1 Specimen 97 % ρ d OPM 1 Specimen 95 % ρ d OPM

25 TYPICAL RESULTS OF CYCLIC TRIAXIAL TEST Plastic strains STABILISATION AFTER A NUMBER OF CYCLES? QUASI-ELASTIC BEHAVIOUR? 200 [ε 1 p - ε 1 p (100)] (10-4 ) 0 [ε 3 p - ε 3 p (100)] (10-4 ) MFC : 5 % F γ d = 97 %, w = 3.7 % γ d = 97.2 %, w = 3.9 % γ d = 97 %, w = 4.1 % γ d = 96.9 %, w = 3.2 % γ d = 96.8 %, w = 2.1 % Nº (cycles) Nº (cycles)

26 TYPICAL RESULTS OF CYCLIC TRIAXIAL TEST Plastic strains Permanent strains* (x10-4) % Fines Number of cycles 95.0%; 3.82% 95.0%; 3.54% 97.0%; 3.65% 97.2%; 3.87% 97.0%; 4.1% 100.0; 3.70% 100.0%; 3.74%

27 Courage test results PERMANENT STRAIN PARAMETER A1 x 10-4 (m/m) PERMANENT STRAIN PARAMETER, A 1 RESULTS Gneiss, Control Grading Granite, Control Grading Limestone, Control Grading UK RCC&A Supply Grading A - greywacke B - carboniferous limestone C - argillaceous limestone D - ryolite E - crushed limestone F - crushed granodiorite G - basalt/dolerite RELATIVE MOISTURE CONTENT (% of OMC)

28 Wheel track University of Oulu Fil. AC 30 mm UNBOUND GRANULAR MATERIAL 300 mm (test material) mm 330 mm 2 FILTER SAND 270 mm Vertical deflection potentiometers

29 Courage test results PERMANENT STRAIN RATE PARAMETER dε p/dn x 10-5 (%/cycle) PERMANENT S TRAIN RATE RES ULTS Gneiss, Control Grading Granite, Control Grading Limestone, Control Grading UK RCC&A Supply Grading Gneiss, WT test Granite, WT test Limestone, WT test LOW STRAIN SUSCEPTIBILITY RELATIVE MOISTURE CONTENT (% of OMC) FAILED HIGH STRAIN SUSCEPTIBILITY measured at 3000 cycles

30 TYPICAL RESULTS OF CYCLIC TRIAXIAL TEST- VCP Un.-Rel. Strains (Anisotropic Boyce model; Hornych et al., 2000) ε v = p p n a * n 1 γ 3 + K 2 a + n 18 1 G a ( γ + 2 ) q p * * 2 + γ 3 G 1 a q p * * ε q = 2 3 p p *n n 1 a γ K a n 1 18 Ga ( γ 1) q * p * γ + 1 q * 6 Ga p * 50 Volumetric strains NON-LINEAR BEHAVIOUR 20 Shear strains Anisotropic Boyce model Measurements v 10 q 5 0 Anisotropic Boyce model Measurements P (kpa) P (kpa)

31 TYPICAL RESULTS OF CYCLIC TRIAXIAL TEST- CCP Un.-Rel. Strains ε = 0,001% CFMSG

32 Variation UGM STIFFNESS du module E sec RANKING en fonction TOTAL STRESSES de σeffective v et σ v STRESSES (ECP, Coronado, 2005) Module sécant E sec ( MPa) Normalisé pour e=0.33 y = 61.34x 0.41 y = 97.12x 0.31 y = x Contrainte verticale σ v (kpa) Module sécant E sec ( MPa) Normalisé pour e=0.33 y = 36.05x 0.50 y = 46.77x 0.44 y = x Contrainte verticale effective σ' v (kpa) Soacha 30% fins d 10 =20µm Soacha 30% fins d 10 =20µm E plus grand en contrainte totales Servita 20% fins, d 10 =45µm Vista H. 10% fins, d 10 =80µm E plus petit en contrainte effectives Servita 20% fins, d 10 =45µm Vista H. 10% fins, d 10 =80µm

33 TYPICAL RESULTS OF COMPRESSION TRIAXIAL TEST- CCP PECULIAR BEHAVIOUR OF COMPACTED UGM

34 EMPIRICAL versus MECHANICAL PARAMETERS QUALITY ASSESSMENT DIFFERENT RANKING! Processed materials behave in a different way to natural materials - Los Angelos and Micro-Deval. They generally give better mechanical performance in the field than would be expect from the results of such tests. Therefore performance-related tests (CLTx)

35 QUALITY ASSESSMENT FIELD

36 UGM QUALITY ASSESSMENT BY IN-SITU TESTS SIMPLE TESTS: to density and moisture content material assessment FUNCTIONAL TESTS: Construction level - to assess the bearing capacity of UGM to carry loading applied and the capacity of spreading the loading (LPT, FWD, SASW, ). Long term behavior - to assess stifness - capacity of spreading the loading (LPT, FWD, -representative state conditions?) and resistance to permanent deformation (LPT - permanent deformation).

37 NEED TO MOVE FROM EMPIRICAL TO MECHANISTIC-BASED SPECIFICATIONS MECHANISTIC-BASED SPECIFICATIONS (focus on the mechanical properties) FACILITATE QUANTITATIVE EVALUATIONS (necessary to alternative construction practices and materials reclaimed materials) Performance-based specification is required to control the long-term functional and structural performance

38 MECHANICAL-BASED IN SITU TESTS, AIPCR, C12, 1995 PLATE BEARING TEST CALIFORNIA BEARING RATIO TEST - CBR DYNAPLAQUE FALLING WEIGHT DEFLECTOMETER FWD DEFLECTOGRAPH CLEGG IMPACT HAMMER MEXECONE PENETOMETER DYNAMIC CONE PENETROMETER COMPACTION METERS DYNAMIC PLATE BEARING TESTS VARIABLE IMPACT TESTER

39 STIFFNESS - TARGET VALUES (Chaddock & Brown, 1995); Zaghloul, 1997; Flemming, 2000; Nunn et al., 1997; GTR, 1997) STIFFNESS 50 MPa Formation (GTR, 1997) 50 to 65 MPa Formation (Flemming et al., 1998) 100 MPa Foundation (Flemming et al., 1998) 80 MPa Formation - FWD stress=200 kpa; D=450 mm (Chaddock & Brown, 1995) 45 MPa Ev2 DIN - Foundation 120 MPa Ev2 DIN Sub-base (light traffic) 150 MPa Ev2 DIN Sub-base (heavy traffic)

40 CONSTRUCTION TECHNOLOGIES Rockfill (experience from dam dam construction) Soil and rock mixtures

41 ROCKFILL COMPACTION Punctual control: density+water content

42 SOIL COMPACTION Punctual control: density+water content

43 REQUIREMENTS FOR HOMOGENEITY Brandl (1977) Parameters Dry density Moduli (Ev1, Ev2) Subgrade 5 30 Coefficient of variation v (%) Sub-base base 4 20 Base 3 15

44 ROUTINE ANALYSIS

45 MODULUS FROM FIELD TESTS (Jamiolkowski et al., 1988, 2003) G 0, E 0 (ε < ε l ) G sec, E sec (ε sec > ε l ) INDEPENDANT FROM: Strain level Stress history DEPENDANT ON: Relative density Ambient stress Compressibility Aging & Fabric 1 n n G = S pa F( e) p 0 DEPENDANT ON: Strain level Stress history Relative density Ambient stress Compressibility Aging & Fabric Strain rate CORRELATIONS WITH G 0, E 0 are more reliable than with G sec, E sec

46 Punctual control: LPT (ASTM, DIN, LCPC) Pression appliquée (kpa) Module de plaque (kpa) Module de plaque Valeur du cycle de rechargement (kpa) Déplacement moyen (mm) E = 1, 5Q D 2 1 appl δ ( υ 2 )

47 PLATE LOAD TESTS ON THE UGM BASE Response of Soil and Granular Base resilient vertical strain (mdef) pressure (kpa) 3,2 2,8 2,4 2,0 1,6 1,2 0,8 0,4 0,0 resilient surface deflection (mm) soil - gauge V7-12/02/97 soil - gauge V4-13/02/97 soil - gauge V7-14/02/97 granular base - gauge V8-14/02/97 granular base - gauge V9-14/02/97 granular base - deflectometers

48 Spectral-Analysis-of-Surface-Waves (SASW) (Nazarian et al., 1987; Rix & Stokoe, 1990; Stokoe, 2004; Allen et al., 2000) Principle: Measuring the propagation velocity of surface waves of Rayleigh type, generating an experimental dispersion curve (wave velocity versus frequency), and evaluating shear wave velocity profile (G0) ) by matching theoretical dispersion curve with experimental curve. Advantages: It is non destrutive. Any layer thikness can be evaluated. Soft layers after compaction may be detected after construction. Stiffness evaluations represent moduli values which are independent of strain level (e<0,001%). Stiffness is a mechanical property showing how pavement will deflect under traffic loads. Reduction factor can be applied to convert the small strain shear modulus in a serviceability modulus. DENSITY TESTS AND SASW TESTS COMPLET EACH OTHER AND CAN BE USED TOGETHER TO GAIN A COMPREHENSIVE ASSESSMENT OF THE QUALITY OF THE FINAL COMPACTED MATERIAL

49 PAVEMENT EVALUATION Punctual control: Stiffness Seismic pavement analyser (SPA)

50 (FWD) Punctual control: Stiffness

51 SOIL COMPACTION Punctual control: Stiffness (LDW)(

52 SASW versus LPT (Allen et al., 2000) STIFFNESS FROM SASW MAY NEED TO BE CORRECTED TO THE SAME STRAIN LEVEL OF LPT

53 FWD versus LDW Ev2 (FWD, LDW) - CBRin situ (DCP) (Livneh & Goldberg, 2001) FWD LDW (3) = 300 E v ln (1/1,41) 300 6,019 CBR (4) = 300 E v ln (1/1,41) 300 4,0354 CBR German code: Ev2 LPT (5) Es CBR = k m DISCREPANCY IN THESE RESULTS SHOW THE NEED TO USE MORE MECHANISTIC APPROACH

54 100 EFFECT OF LOADING SPEED ON BC & UGM STRAINS 50 strains (µdef) 0-50 V2 L11 T12 ε zzugm ε xxbc ε yybc speed (km/h)

55 STRAINS MEASURED BY FWD AND WHEEL LOAD TESTS (40 km/h) 40 strains due to FWD load (µdef) L11 T12 V strains due to wheel load (µdef)

56 CONTINOUS COMPACTION CONTROL (LPC; Quibel, 1998)

57 ROLLER-INTEGRATED CONTINOUS COMPACTION CONTROL (CCC) OPTIMISATION OF COMPACTION

58 CCC A PROMOTION TO NON-CONVENTIONAL DESIGN PHILOSOPHY Reduce differential settlements ( for H until 120m total relative settlement is less of 1/1000) In many cases embankments exhibit significant advantages over bridges: facilitate balance of cut and fill volume during construction; environment-friendly negligible long term maintenance Reduce differential settlements between bridge deck and adjacent embankment fill.

59 ADVANCED ANALYSIS

60 PMT HSM - PLAXIS Pressure (kpa) Mod. 1 : E 50 ref =33 MPa, E oedo ref =35 MPa, E ur =100 MPa, ν=0,25, φ=38º, ψ=0º, c=70 kpa, k 0 =1,2 Mod. 2 : E 50 ref =32 MPa, E oedo ref =32 MPa, E ur =96 MPa, ν=0,25, φ=36,5º, ψ=0º, c=65 kpa, k 0 =1,2 HSM Mod. 1 HSM Mod. 2 Depth=0,46 m dv/v

61 HSM Mod3: γ =18 kn/m3, E50ref =27,5 MPa, Eoedoref =27,5 MPa, Eur=82,5 MPa, m=0,5, c=70 kpa, φ =38º, ψ =0º, n=0,25, k0=1,2 PLT HSM - PLAXIS Applied pressure (kpa) Medium displacement (mm) PLT 1 PLT 2 HSM Mod1 ( PMT1 values) HSM Mod2 (PMT2 values) HSM Mod3 (Adjusted values)

62 PMT PLT - TXSimulation ROUTINE & ADVANCED ANALYSIS (HSM( PLAXIS) (Gomes Correia et al., 2004) Modulus (kpa) Eur (PLT) = 3 X Eur (PMT) 0 Etan P/D=2 1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 log(ε v ; δ/d) E Menard (PMT1) E Menard (PMT2) E ur (PMT1) E ur (PMT2) E ur (PLT) E PLT (settlement/d) - ECP 1 E PLT (settlement/d) - ECP 2 E PLT (settlement/d) - simulation E sec (PLT) - simulation Esec P/D=0,1 Etan P/D=0,1 Esec P/D=0,5 Etan P/D=0,5 Esec P/D=1 Etan P/D=1 Esec P/D=2 Esec "Classical" triaxial Etan "Classical" triaxial

63 FINAL REMARKS

64 FINAL REMARKS (1/4) NEED TO MOVE FROM EMPIRICAL APPROACH TO MECHANISTIC APPROACH MOVE FROM SIMPLE TESTS TO FUNCTIONAL TESTS USE OF NEW MATERIALS IN ROAD CONSTRUCTION SELECTION OF MATERIALS MUST BE BASED IN MECHANICAL TEST RESULTS USING INTEGRAL MATERIAL AT THE STATE CONDITIONS EXPECTED IN THE FIELD

65 FINAL REMARKS (2/4) NON-LINEAR BEHAVIOUR OF SOILS AND UGM COMPLICATES IN-SITU TEST INTERPRETATION MAY CONFLICT WITH SIMPLIFIED ASSUMPTIONS IN THE PAST DIRECT USE OF MODULI IN PRACTICE ONLY IF: OBTAINED FOR THE SERVICEABILITY STRESS-STRAIN STRAIN CONDITIONS ANY CORRELATION OF IN-SITU TEST RESULTS SHOULD SPECIFY THE TYPE OF EQUIPMENT AND TEST PROCEDURE, THE TYPE OF MODULUS OR ANGLE OF SHEARING RESISTANCE

66 FINAL REMARKS (3/4) CURRENT PRACTICE HAS TO BE SUPPLEMENTED BY TEST METHODS THAT WILL PROVIDE CONTINUITY BETWEEN DESIGN (laboratory) AND CONSTRUCTION (field) G0 IS THE BEST PARAMETER TO BE USED Incorporation of Seismic in the different tests will be of valuable use CORRELATIONS MUST BE ESTABLISHED WITH G0 AND NOT WITH CBR

67 FINAL REMARKS (4/4) COMPARISON BETWEEN TEST RESULTS MODULI- IS ONLY POSSIBLE BY NORMALISING STRESS & STRAIN LEVELS OF TESTS EXPERIMENTAL AND ANALYTICAL STUDY IN A GRANITE SHOW THAT: E UR (PLT) ~3 X E MUR (PMT) E M (PMT) ~1%; E(PLT) ~ 0,1% (ASTM STANDARDS) COMPAIRISON OF TEST RESULTS BACK ANALYSIS OF STRESS-STRAIN STRAIN CURVES OR LOAD-DISPLACEMENT DISPLACEMENT RESULTS