AP-R462-14

Research Report AP-R462-14

Cemented Materials Characterisation Final Report

Cemented Materials Characterisation: Final Report

Prepared By Allan Alderson and Geoff Jameson

Publisher Austroads Ltd. Level 9, 287 Elizabeth Street Sydney NSW 2000 Australia Phone: +61 2 9264 7088 [email protected] www.austroads.com.au

Project Manager Andrew Papacostas

Abstract

This report brings together five years of research into the modulus, strength fatigue characteristics of cement treated granular materials as used in road pavements. Laboratory procedures used to prepare, cure and test materials for flexural modulus, flexural strength and flexural fatigue behaviour of cemented material beams are reported together with the test results.

Based on the laboratory results, a framework for the revision of the Guide to Pavement Technology Part 2: Pavement Structural Design has been suggested and the proposed revised text of the Guide has been prepared.

About Austroads

Austroads purpose is to: promote improved Australian and New Zealand

transport outcomes provide expert technical input to national policy

development on road and road transport issues

promote improved practice and capability by road agencies.

promote consistency in road and road agency operations.

Austroads membership comprises: Roads and Maritime Services New South

Wales Roads Corporation Victoria Department of Transport and Main Roads

Queensland Main Roads Western Australia Department of Planning, Transport and

Infrastructure South Australia Department of Infrastructure, Energy and

Resources Tasmania Department of Transport Northern Territory Department of Territory and Municipal Services

Australian Capital Territory Commonwealth Department of Infrastructure

and Regional Development Australian Local Government Association New Zealand Transport Agency.

The success of Austroads is derived from the collaboration of member organisations and others in the road industry. It aims to be the Australasian leader in providing high quality information, advice and fostering research in the road transport sector.

Keywords

cemented materials, cement treated crushed rock, stabilised granular, damage exponent, flexural fatigue, flexural modulus, flexural strength, breaking strain, laboratory test methods, fatigue relationship

ISBN 978-1-925037-72-2

Austroads Project No. TT1664

Austroads Publication No. AP-R462-14

Published June 2014

Pages 127

Austroads Ltd 2014

This work is copyright. Apart from any use as permitted under the Copyright Act 1968, no part may be reproduced by any process without the prior written permission of Austroads.

This report has been prepared for Austroads as part of its work to promote improved Australian and New Zealand transport outcomes by providing expert technical input on road and road transport issues.

Individual road agencies will determine their response to this report following consideration of their legislative or administrative arrangements, available funding, as well as local circumstances and priorities.

Austroads believes this publication to be correct at the time of printing and does not accept responsibility for any consequences arising from the use of information herein. Readers should rely on their own skill and judgement to apply information to particular issues.


Summary

More than 90% of the Australian and New Zealand sealed road network consists of a sprayed seal overlying granular pavements. Increased traffic loadings are placing increasing pressure on these pavements, with some non-standard materials no longer being fit-for-purpose. In many rural areas, the use of high quality crushed rock is not a cost-effective treatment to improve the structure of these pavements. Consequently there is increasing use of treatments that enhance the existing non-standard materials by adding cementitious and bituminous binders to allow recycling of scarce resources.

The objective of the research in Austroads research project TT1664 (Cemented Materials Characterisation) was to develop improved methods to design flexible pavements with cemented materials building on the research previously undertaken in Austroads research project TT1359 (Cost-effective Structural Treatments for Rural Highways: Cemented Materials).

The outcomes and findings relating to the laboratory testing were:

Test methods were developed for flexural modulus, flexural strength and fatigue of cement treated crushed rocks and natural gravels.

Strain-based fatigue laboratory relationships were a better fit to the data than stress-based relationships and it is proposed to continue use of logN-log fatigue relationship.

Strain damage exponents from 9 to 24 were calculated from the data. The variation in flexural modulus and strength in relation to density was quantified and procedures

proposed for use in design.

Based on the findings, and in light of the recommended framework for cemented materials characterisation (Austroads 2014), the report proposes revised text for Austroads Guide to Pavement Technology Part 2: Pavement Structural Design.

Austroads 2014| page i


Contents

1. Introduction ............................................................................................................................................. 1

2. Materials Tested ..................................................................................................................................... 2

3. Equipment and Test Method ................................................................................................................. 3 3.1 Specimen Preparation Method ................................................................................................................. 3

3.1.1 Material Splitting Procedure ....................................................................................................... 3 3.1.2 Laboratory Characterisation Process ......................................................................................... 3 3.1.3 Mixing Procedure ........................................................................................................................ 3 3.1.4 Slab Compaction and Cutting Procedures ................................................................................. 4 3.1.5 Curing ......................................................................................................................................... 6

3.2 Flexural Beam Test Methods ................................................................................................................... 6 3.2.1 Introduction ................................................................................................................................. 6 3.2.2 Apparatus ................................................................................................................................... 6 3.2.3 Flexural Modulus ........................................................................................................................ 8 3.2.4 Flexural Strength and Breaking Strain ....................................................................................... 9 3.2.5 Flexural Fatigue ........................................................................................................................ 10

4. Flexural Modulus Results .................................................................................................................... 12 4.1 Introduction ............................................................................................................................................. 12 4.2 Adjustment to Standard Strain ............................................................................................................... 12 4.3 Results .................................................................................................................................................... 14 4.4 Variation in Modulus with Density .......................................................................................................... 15

5. Flexural Strength Results .................................................................................................................... 17 5.1 Introduction ............................................................................................................................................. 17 5.2 Results .................................................................................................................................................... 17 5.3 Strength Variation with Density .............................................................................................................. 19 5.4 Curing Duration ...................................................................................................................................... 20

6. Breaking Strain Results ....................................................................................................................... 22 6.1 Introduction ............................................................................................................................................. 22 6.2 Variation in Breaking Strain with Density ............................................................................................... 22 6.3 Effect of Cure Duration ........................................................................................................................... 23

7. Fatigue Results ..................................................................................................................................... 24 7.1 Introduction ............................................................................................................................................. 24 7.2 Results .................................................................................................................................................... 24 7.3 Analysis .................................................................................................................................................. 25

7.3.1 Introduction ............................................................................................................................... 25 7.3.2 Strain-based Fatigue Relationships ......................................................................................... 25 7.3.3 Stress-based Fatigue Relationships ......................................................................................... 27 7.3.4 Summary .................................................................................................................................. 29

8. Estimation Of Flexural Strength ......................................................................................................... 30 8.1 Introduction ............................................................................................................................................. 30 8.2 Predicting Strength from Properties of Constituent Materials ................................................................ 30 8.3 Flexural Strength from UCS ................................................................................................................... 33

8.3.1 Review of Data in the Literature ............................................................................................... 33 8.3.2 Measurements .......................................................................................................................... 34

9. Proposed Guide Revision .................................................................................................................... 36

10. Summary ............................................................................................................................................... 37

Austroads 2014| page ii


References ...................................................................................................................................................... 38 Appendix A Properties of Test Materials ................................................................................................... 40 Appendix B Flexural Beam Test Methods .................................................................................................. 54 Appendix C Modulus, Strength and Fatigue Results ................................................................................ 63 Appendix D Fatigue Plots ............................................................................................................................ 97 Appendix E Proposed Revision of the Guide to Pavement Technology Part 2 Section 6.4 ............... 110 Appendix F Example of use of Proposed Mechanistic Procedure for Flexible Pavements

with Cemented Materials ...................................................................................................... 122

Tables Table 2.1: Summary of material properties ..................................................................................................... 2 Table 4.1: Modulus dependency on strain .................................................................................................... 13 Table 4.2: Flexural moduli after 28 days moist curing .................................................................................. 14 Table 4.3: Flexural moduli after five months moist curing ............................................................................. 14 Table 4.4: Flexural moduli after nine months moist curing ........................................................................... 15 Table 4.5: Results of regression analysis on the variation of modulus with density ratio ............................. 16 Table 5.1: Flexural strength and breaking strain of laboratory-manufactured test beams after

28 days moist curing .................................................................................................................... 17 Table 5.2: Flexural strength and breaking strain of laboratory-manufactured test beams after

five months moist curing .............................................................................................................. 18 Table 5.3: Flexural strength and breaking strain of laboratory-manufactured test beams after

nine months moist curing ............................................................................................................. 18 Table 5.4: Change in flexural strength with curing ........................................................................................ 21 Table 7.1: Properties of fatigue beams after five and nine months moist curing .......................................... 24 Table 7.2: Strain-based fatigue relationships for each material .................................................................... 26 Table 7.3: Stress-based fatigue relationships for each material ................................................................... 28 Table 8.1: Data used in development of Equation 12 ................................................................................... 31 Table 8.2: Flexural strength and UCS test results ........................................................................................ 34

Figures Figure 3.1: Motorised rotary splitter and splitting process ................................................................................ 3 Figure 3.2: Planetary concrete mixer used to mix cemented material ............................................................. 4 Figure 3.3: BP slab compactor with rectangular mould .................................................................................... 5 Figure 3.4: Wet sawing a flexural beam specimen in the laboratory ............................................................... 5 Figure 3.5: Long-term storage of beams .......................................................................................................... 6 Figure 3.6: Cross-sectional view of flexural beam testing apparatus ............................................................... 7 Figure 3.7: Flexural beam test .......................................................................................................................... 7 Figure 3.8: Example flexural modulus load pulse of 1.0 kN ............................................................................. 8 Figure 3.9: Example flexural strength test graph of load-displacement (weathered granite 3%

specimen A4-2) .............................................................................................................................. 9 Figure 3.10: Example of breaking strain and strain at 95% of the breaking load ............................................. 10 Figure 3.11: Example flexural fatigue load pulse of 2.7 kN .............................................................................. 11 Figure 3.12: Typical modulus variation during fatigue tests (specimen from quartzite 4%) ............................. 11 Figure 4.1: Example of flexural modulus dependency on strain over a wider strain range for

individual crushed granite beams cured for five months .............................................................. 12 Figure 4.2: Example of modulus variation with density ratio .......................................................................... 15 Figure 5.1: Density ratio differences of flexural strength and fatigue beams ................................................. 19 Figure 5.2: Flexural strength variation with density ........................................................................................ 20 Figure 5.3: Change in flexural strength with curing ........................................................................................ 20 Figure 6.1: Variation in breaking strain with density ....................................................................................... 22 Figure 6.2: Change in breaking strain with curing .......................................................................................... 23 Figure 8.1: Comparison of Equation 12 predicted strengths with measured values ...................................... 31 Figure 8.2: Comparison of Equation 13 predicted strengths with measured values ...................................... 33 Figure 8.3: Flexural strength variation with UCS ............................................................................................ 34 Figure 8.4: Adjusted flexural strength with UCS ............................................................................................ 35

Austroads 2014| page iii


1. Introduction

The previous Austroads Project TT1359 Cost-effective Structural Treatments for Rural Highways investigated the fatigue performance of a range of cemented materials (Austroads 2010). An important finding was that the current fatigue relationship in the Guide to Pavement Technology Part 2: Pavement Structural Design (Austroads 2012a) needed to be revised. Currently, flexural modulus is used to distinguish the variation in fatigue performance of different qualities of cemented materials, varying from lightly stabilised subbase sand to lean mix concrete. The outcome of TT1359 was that flexural modulus may not be the best parameter to relate to the fatigue behaviour of cement stabilised materials.

Consequently research was then undertaken in Austroads project TT1664 Cemented materials characterisation with the objective of proposing revised design procedures for the Guide.

In 201213 the aims of the project were to improve procedures for designing flexible pavements containing cemented materials by addressing the following tasks:

re-analyse the published information previously reported under project TT1359 Cost-effective Structural Treatments for Rural Highways: Cemented Material (Austroads 2010) and include any new data

develop improved methods to determine design moduli for cemented materials by examination of flexural beam testing, presumptive values and correlation to other laboratory tests

revise text for Guide to Pavement Technology Part 2: Pavement Structural Design, Section 6.4 (where appropriate), including a new presumptive fatigue relationship for cemented materials.

Section 2 summarises the wide range of cemented materials investigated in projects TT1359 and TT1664. Section 3 provides details of the sample preparation and test methods and equipment used in the flexural modulus, flexural strength, breaking strain and fatigue testing. Section 4 to Section 7 summarise the test results. In the event that measured flexural strength data is not available, methods of estimating flexural strength from material properties are evaluated in Section 8. A summary of the proposed key changes to cemented materials characterisation is provided in Section 9. Appendix E is the proposed revised text for the Guide.

Austroads 2014| page 1


2. Materials Tested

A wide variety of cement stabilised crushed rocks, natural gravels and a recycled crushed concrete were tested in Austroads projects TT1359 and TT1664. Table 2.1 summarises the materials properties, which are described in detail in Appendix A.

A general purpose (GP) Portland cement was used in this project for all samples. The cement was provided in small quantities during the execution of the project by a local company in Melbourne (Victoria). Freshly manufactured cement was obtained on a regular basis and it was estimated that no cement was older than one month at the time of use.

It should be noted that previously the material referred in Austroads (2010) as siltstone sourced from Para Hills in South Australia which had been stabilised with 4% cement, has been renamed to quartzite throughout this report consistent with recent road agency advice.

Table 2.1: Summary of material properties

Material Material identifier

Cement content

(%)

Curing duration (months)

Coarse aggregate

content (% > 6.7 mm)

Fine aggregate

content (% <

4.75 mm)

Plasticity index

Basalt (Mt Gambier) BAM3 3 1, 9 42 20 0

Basalt (Purga) BAP3 3 5 40 13 6

Calcrete limestone CL3 3 1, 9 57 21 4

Calcrete limestone (repeat) CL3 3 1 57 21 4

Calcrete limestone CL5 5 1, 9 57 21 4

Granite GR3 3 5 46 20 6

Hornfels HO3 3 5 52 15 6

Laterite LAT3 3 1 34 56 21

Metagreywacke MTG3 3 1 40 52 0

Modified prior stream gravel MPSG 3 1, 9 47 32 14

Quartzite (repeat) QZ4_1 4 5, 9 52 28 8

Quartzite QZ4_2 4 5 52 28 8

Prior stream gravel PSG3 3 5 1 57 14

Prior stream gravel PSG5 5 1, 9 1 57 14

Recycled concrete RCC3 3 5 76 7 6

Weathered granite WG3 3 1, 9 35 25 2

Weathered granite WG5 5 1, 9 35 25 2



3. Equipment and Test Method

3.1 Specimen Preparation Method

Material Splitting Procedure 3.1.1The bulk materials were split into representative samples using a motorised rotary splitter shown in Figure 3.1. The splitter has a large funnel into which the bulk material can be loaded. A conveyor belt then delivers this material at a constant speed to 12 removable canisters which rotate underneath the belt, also at constant speed. The process involved splitting two or more batches of 12 numbered (10 L) buckets each. Corresponding numbers from each of the batches would then be combined for a second (final) round of splitting to improve uniformity between samples.

Figure 3.1: Motorised rotary splitter and splitting process

Laboratory Characterisation Process 3.1.2Once materials had been split uniformly a laboratory characterisation process was undertaken to collect the following properties:

particle distribution using the sieve analysis testing procedure (according to AS 1289.3.6.1) maximum dry density and optimum moisture contents using a modified compaction test (according to

AS 1289 2.1.1 and 5.2.1)

liquid limit, plastic limit and plastic index using a plasticity index test (according to AS 1289 2.1.1, 3.1.1, 3.1.2, 3.3.1 and 3.4.1).

Mixing Procedure 3.1.3Batches of aggregate material (usually of 7085 kg) were weighed and preconditioned (addition of moisture) 24 hours prior to mixing. The preconditioned mix was then combined with the remaining water and cement in a motor-driven planetary concrete mixer with a tank size of 800 mm diameter and 350 mm high (Figure 3.2).



The mixing process was as follows:

The bulk preconditioned (moisture prepared) host material was placed into the mixing tank. The mixer was run for 15 seconds to spread the material evenly in the tank. The required amount of GP cement was added. The final amount of water was added to reach the target moisture content of Modified compaction

optimum moisture content.

The mixer was run for 120 seconds then let stand for 120 seconds. The mixer was run for a further 120 seconds.

Figure 3.2: Planetary concrete mixer used to mix cemented material

Following mixing, the material was placed in containers and covered with a plastic sheet and let stand for a period of time prior to compaction. The intention was to make allowance for the commencement of the cement binder reaction and also to replicate the field placement of cemented materials which can involve quarry mixing and then delivery time prior to placement. The standing time was at most 30 minutes as beyond this time samples were too difficult to compact due to the initial set of the cement binder.

Slab Compaction and Cutting Procedures 3.1.4A BP slab compactor and a rectangular mould with internal dimensions of 400 mm long x 320 mm wide x 145 mm high were used for slab compaction as shown in Figure 3.3.

800 mm

350 mm



Figure 3.3: BP slab compactor with rectangular mould

A pre-determined mass of wet (cement-treated) material was placed in the slab mould, spread evenly and tamped manually in three separate layers to commence the compaction process. The mass was selected to target a density ratio of 95% Modified maximum dry density. The material was then compacted in a single layer to the specified height of 100 mm using the slab compactor (Figure 3.3). The slab compactor was set at an initial vertical pressure of 100 kPa while the curved steel compaction head was rocked over the material. The vertical pressure was increased at 100 kPa intervals for every 10 rocking passes until the pressure reached 600 kPa (near machine capacity). The compaction process was continued until the total number of passes reached 100 (the reduction in height beyond 100 passes had previously been found to be insignificant) or until the entire length (along the 320 mm width) of the slab had been compacted to 100 mm depth.

Immediately following compaction the slab was retained in the closed mould and covered with a wet cloth and lid to minimise moisture loss. For a minimum of two days the slab was stored at a controlled temperature of 23 C prior to de-moulding. Each slab was placed into a fog room of 23 C and humidity greater than 95%. Slabs were subsequently cut into two beams after a minimum cure period of 24 days to ensure no disintegration occurred during the wet cutting process (using a diamond tipped saw shown in Figure 3.4). Beams were replaced back into the fog room for the remainder of their curing duration.

Figure 3.4: Wet sawing a flexural beam specimen in the laboratory

320 mm

400 mm

145 mm



Curing 3.1.5Prior to testing, the beams were preconditioned in a humidity- and temperature-controlled fog room for at least 48 hours to ensure a consistent moisture condition for all tests. The beams were exposed to the ambient laboratory conditions for approximately 15 minutes while their wet densities were checked by measuring the beam dimensions (to determine the specimen volume) and the wet total mass of the beam. After the testing the moist beams were then sealed in thin plastic cling wrap (see Figure 3.5) to minimise moisture loss during extended curing.

The wrapped beams and then placed in a temperature controlled environment of 23 C and cured for either 28 days, 5 months or 9 months.

Figure 3.5: Long-term storage of beams

3.2 Flexural Beam Test Methods

Introduction 3.2.1The preferred methods for testing cemented material beams are outlined below but all conform to four-point bending of flexure (beam) specimens (with a span/depth ratio greater than three). The flexural beam test methods used in this study were devised for materials with a maximum particle size of 20 mm and tested beam samples which were 100 mm high by 100 mm wide by 400 mm long.

Apparatus 3.2.2The test apparatus uses a simply supported beam loaded at its third points (Figure 3.6 and Figure 3.7). The beam supports were 300 mm apart so as to achieve a span to depth ratio of three. The beam displacement was measured at the mid-point. A 14 kN pneumatic loading frame was used to undertake all the flexural beam tests, with typical loading in the range 1 kN to 6 kN.

The mid sample displacement was measured by two linear variable differential transducers (LVDTs) placed side by side across the width of the beam. The LVDTs were mounted in a cradle that was supported on the sample. The cradle supports were located above the beam supports and a weak restraining force was applied between the cradle and the beam support to hold the cradle in place.



Figure 3.6: Cross-sectional view of flexural beam testing apparatus

Source: AS 1012.11 (2000).

The loading and measurements were all controlled and monitored by a personal computer.

Figure 3.7: Flexural beam test

LVDT to measure beam deflection

Supporting roller

Lower platen of test machine

Loading roller

100 mm 100 mm 100 mm

100

mm



Flexural Modulus 3.2.3The flexural modulus test (Appendix B.1) involved the application of cyclic haversine load pulses of 250 ms duration. The beam deflection associated with each load pulse was recorded, together with the seating load (nominally 50 N) and the peak load. The load pulse was one second duration including a 750 ms rest period between load pulses (Figure 3.8). At least 100 load pulses were applied to the specimen.

As flexural modulus varies slightly with the applied stress/strain consideration needs to be given to load applied. The moduli in this report have been standardised to values at 50 microstrain. In the event that test beams are not subsequently tested for fatigue, the modulus testing should be undertaken at a load level that generates 50 microstrain.

Figure 3.8: Example flexural modulus load pulse of 1.0 kN

However, most of the test beams in the project were tested tor fatigue after modulus testing. In order to minimise fatigue damage to specimens during the modulus testing the magnitude of the applied load in the modulus testing was targeted to produce approximately 2030 microstrains.

After completion of the flexural modulus test, the mean peak tensile strain at the bottom of the beam and mean flexural modulus were calculated from the measured deflection.

Flexural moduli were calculated using Equation 1:

E = 23 PL3

108WL2103 1

where

E = flexural modulus (MPa)

P = maximum applied force (kN) A = distance between the load and reaction clamps (mm) L = length between supporting rollers (mm)

W = average width of test specimen (mm) H = average height of test specimen (mm) = deflection at the centre of the beam (mm)



Flexural Strength and Breaking Strain 3.2.4For the flexural strength (also known as the modulus of rupture) test the MATTA was programmed to load the specimen with a seating force of 50 N for the first six seconds, after which the load was increased at a rate of 3.3 kN per minute until the specimen failed (ruptured) as described in AS 1012.11-2000.

The vertical displacement at the beam mid-point was measured using two linear variable differential transformers (LVDT) mounted in a reference frame (Figure 3.7) placed on the specimen to enable the strain at break to be calculated using the beam dimensions. The beam deflection data was sampled at a frequency of 100 Hz, together with the applied load. After the specimen failed, the peak load and approximate location of the break point were recorded.

An example graph of the measured load-displacement is shown as Figure 3.9, with displacement in blue and load in red.

Figure 3.9: Example flexural strength test graph of load-displacement (weathered granite 3% specimen A4-2)

Equation 2 was used to calculate the flexural strength:

fcf = PL(1000)

WH2 2

where

fcf = flexural strength (MPa)

P = maximum applied force (kN)

L = length between supporting rollers (mm)

W = average width of test specimen (mm)

H = height of test specimen (mm)



Breaking strains are reported at 95% of the breaking load as it was found to be more repeatable than the strain at which the specimen breaks (Austroads 2010). Figure 3.10 shows an example of force/strain history data from a flexural strength test. In the test the specimen failed at 2.58 kN load with a breaking strain of approximately 165 microstrain. Ninety-five percent of the breaking load is 95% x 2.58 kN = 2.46 kN. Hence, the strain at 95% of the breaking load is 130 microstrain for this example.

Figure 3.10: Example of breaking strain and strain at 95% of the breaking load

Equation 3 was used to calculate the breaking strain using the displacement at the breaking load. The equation used to calculate the strain at break was:

= 10808H(1)23L2

3

where

= breaking strain (microstrain)

= mid-span vertical displacement at 95% of breaking load (mm)

H = height of test specimen (mm)

L = length between supporting rollers (mm)

Flexural Fatigue 3.2.5The flexural fatigue test involved the application of cyclic haversine load pulses similar to those explained for flexural modulus (Section 3.2.3) using equipment shown in Figure 3.7.

The beam deflection associated with each load pulse was recorded, together with the seating load (nominally 50 N) and the peak load. The pulse period of 2 Hz (twice the rate of flexural modulus testing) was adopted due to the time-consuming nature of the fatigue test procedure. This period included a 250 ms rest period between 250 ms load pulses (Figure 3.11).

Microstrain ()

0 20 40 60 80 100 120 140 160 180

Forc

e (k

N)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Pmax = 2.58 kN

95% Pmax = 2.46 kN

Breaking strainb = 130

Sample Fails



The magnitude of the load pulses was generally selected to fall in the 5090% range of the breaking load as determined from the strength tests for the material type. This load pulse range was selected to produce a variation of calculated strain values of 50100 microstrain corresponding to a range of 1000 to 1 000 000+ cycles up to failure. Note if a specimen did not fail after 1 000 000 cycles the test was programmed to automatically stop loading.

Figure 3.11: Example flexural fatigue load pulse of 2.7 kN

Historical data from project TT1359 (Austroads 2010) suggests that for flexural fatigue testing the modulus decreased rapidly from the start of the test (one load cycle, see Figure 3.12). After this initial bedding-in phase, the modulus decreased at a slow, constant rate. For the specimens that failed within the testing range a turning point was observed when the modulus attained approximately 80% of the initial modulus. After this point an accelerated rate in modulus reduction was observed leading to fracture. The point of fracture was found to repeatedly occur just following the attainment of 50% of the initial modulus. This fatigue nature of cemented samples under the four point bending test can be seen in Figure 3.12.

Figure 3.12: Typical modulus variation during fatigue tests (specimen from quartzite 4%)

The flexural fatigue test involved the application of continuous load pulses until the beam fatigued. In this project fatigue life was the number of loadings at which flexural modulus of the beam was half that of the initial flexural modulus determined after 50 loading cycles:

0

2,000

4,000

6,000

8,000

10,000

12,000

14,000

0 50,000 100,000 150,000 200,000 250,000 300,000 350,000 400,000 450,000

Elas

tic M

odul

us (M

Pa)

Load Cycles

Initial ModulusEini ~ 11500 MPa

'Turning Point'E ~ 9000 MPaE ~ 80%Eini

~ 65,000 Load Cycles

Line adjusted to constant rate of modulus reduction

'Bedding-in' phase

Accelerated Modulus Reduction

Failure

For samples that did not fail this line was extrapolated from cycles to E = 80%Eini

Constant Rate of Modulus Reduction

Elas

tic m

odul

us (M

Pa)



4. Flexural Modulus Results

4.1 Introduction The flexural modulus test method described in Appendix B.1 was used on all beams before strength or fatigue testing was undertaken. Appendix C lists the measured moduli for all materials.

4.2 Adjustment to Standard Strain In order to minimise fatigue damage to test beams during the modulus test, it is common practice to reduce the applied load such that the tensile strain is 2030 microstrain. However, such strain levels are well below those applied in-service. Hence it was decided to adjust all measured moduli to a standard strain of 50 microstrain.

Moffatt (Austroads 2011) examined the load/strain dependency of flexural modulus and reported that the modulus decreased by 35 to 80 MPa of every microstrain increase in applied strain. Consequently it was considered necessary to select a standard strain at which to report the flexural modulus results.

To extend the previous research (Austroads 2011) modulus strain dependency to a wide range of material, the flexural moduli measured at 2030 microstrain were compared to initial flexural moduli at commencement of the fatigue tests. Typically, fatigue testing induced much greater strain in the beams during testing so as to induce fatigue in the materials within a realistic timeframe. For example, Figure 4.1 illustrates the variation in modulus of individual test beams of crushed granite stabilised with 3% cement cured for five months.

Figure 4.1: Example of flexural modulus dependency on strain over a wider strain range for individual crushed granite beams cured for five months

It can be seen that the lines are reasonably parallel indicating that all beams for this material had a similar sensitivity to strain level. The slopes for the seven beams ranged from 58 MPa/ to 85 MPa/ with an average of 66 MPa/.



Table 4.1 lists the average modulus dependency on strain calculated for each material. The variation in these dependencies between materials did not appear related to their modulus values.

Table 4.1: Modulus dependency on strain

Material Material ID Average for individual beams

(MPa/)

Basalt (Mt Gambier) BAM3 28

Basalt (Purga) BAP3 32

Calcrete limestone CL3 24

Calcrete limestone CL5 26

Granite GR3 66

Hornfels HO3 54

Prior stream gravel PSG5 29

Modified prior stream gravel MPSG 70

Quartzite QZ4_1 30

Quartzite (repeat) QZ4_2 33

Recycled crushed concrete RCC 15

Weathered gravel WG3 25

Weathered gravel WG5 34

Average 37

It was concluded that for an increase in strain of 1 the modulus decreases about 40 MPa. Thus if the measured modulus is 14 000 MPa at an applied strain of 25 microstrain, the modulus adjusted to 50 microstrain is 13 000 MPa.

Based on this data Equation 4 was used to standardise the measured flexural moduli:

E50 = EM 40 x (50 M) 4

where

E50 = flexural modulus standardised to a strain of 50 (MPa)

EM = measured flexural modulus at strain M (MPa)

M = tensile strain during flexural modulus testing (microstrain)

Note that this standardisation of modulus is not required if the test beams are not subsequently fatigue tested. In such cases the modulus can be measured at 50 microstrain.



4.3 Results The measured modulus on each beam is given in Appendix C and the mean moduli are listed in Table 4.2.

The mean density ratios listed are the measured densities divided by the modified compaction maximum dry density (MDD), expressed as a percentage. The target density ratio was 95%.

Equation 4 was applied to all the individual beam results for each material tested and the mean adjusted flexural moduli were calculated. The adjusted flexural moduli values are used in the analyses in the remainder of this report.

Table 4.2: Flexural moduli after 28 days moist curing

Material(2) Number of beams tested

Mean density ratio(1)

(%)

Mean tensile strain

(microstrain)

Mean measured

flexural moduli (MPa)

Mean adjusted flexural moduli

to 50 microstrain

(MPa)

Weathered granite (WG3) 32 94.6 24 8 040 6 980

Weathered granite (repeat) (WG3_2) 6

94.6 27 8 600 7 670


Calcrete limestone (CL3) 32 96.7 22 11 240 10 130


Basalt (Mt Gambier) (BAM3) 32 97.7 25 14 670 13 600

Prior stream gravel (PSG5) 27 95.1 25 11 400 10 400

Modified prior stream gravel (MPSG) 32 93.5 26 12 970 12 000

Lateritic gravel (LT3) 6 96.9 23 10 260 9 190

Metagreywacke (MTG3) 6 94.5 27 13 010 12 110

1 Density ratio is based on modified maximum dry density. 2 General purpose (GP) cement was used for all materials.

Table 4.3: Flexural moduli after five months moist curing



(%)

Mean tensile strain

(microstrain)

Mean measured



to 50 microstrain

(MPa)



Recycled concrete (RCC) 16 96.7 25 9 700 8 670

Hornfels (HO3) 20 94.0 22 20 340 19 200

Basalt (Purga) (BAP3) 16 95.6 25 13 000 11 980

Quartzite (QZ4_2) 12 94.5 21 14 300 13 130

Granite (GR3) 16 96.5 25 16 530 15 530




Table 4.4: Flexural moduli after nine months moist curing



(%)

Mean tensile strain

(microstrain)

Mean measured



to 50 microstrain

(MPa)





Basalt (Mt. Gambier) (BAM3) 26 97.7 33 14 780 14 090


Modified prior stream gravel (MPSG) 28 94.3 26 17 240 16 260

Quartzite (QZ4_1) 35 97.1 26 15 340 14 400


4.4 Variation in Modulus with Density In the research project, the compaction for the beams was targeted at 95% of the Modified compaction maximum dry density. However this was not always able to be achieved as seen from the density ratios in Table 4.2, Table 4.3 and Table 4.4.

As this also may occur in future application of the proposed Austroads design method (Appendix E), a procedure was developed to adjust the measured moduli for density.

For each material tested after five months and nine months moist curing, the variation in modulus with density ratio was plotted. For example Figure 4.2 shows the results for calcrete limestone stabilised with 5% cement (CL5) after nine months curing.

Figure 4.2: Example of modulus variation with density ratio



For each material linear regression analysis was undertaken to quantify the variation in modulus with density ratio. The results are given in Table 4.5. Note that for three materials (BAP3, CL3 and PSG5) the results were too scattered to develop a statistically significant relationship.

Using the regression equations, the percentage changes in modulus for a 1% change in density ratio were determined as shown in Table 4.5.

Table 4.5: Results of regression analysis on the variation of modulus with density ratio

Curing period Material Slope (MPa/%)

Intercept (MPa)

Correlation coefficient (R2)

Model statistical significance

Percentage change in modulus for 1% density change

5 months HO3 1083 82 601 0.40 < 0.01 5.6

PSG3 377 27 876 0.34 < 0.01 4.3

PSG5 436 31 090 0.56 < 0.01 4.0

QZ4_2 802 62 674 0.82 < 0.01 6.3

RCC 796 68 324 0.48 < 0.01 9.2

GR3 1080 88 665 0.47 < 0.01 6.9

BAP3 436 29 753 0.43 < 0.01 3.6

9 months BAM3 774 61 511 0.80 < 0.01 5.7

CL5 575 44 615 0.62 < 0.01 4.9

MPSG 1093 86 840 0.55 < 0.01 6.7

QZ4_1 851 68 185 0.35 < 0.01 5.9

WG3 322 21 949 0.18 0.03 3.8

WG5 626 45 607 0.36 < 0.01 4.4

Average 5.5

Assuming a 5% increase in modulus for a 1% increase in density, the following Equation 5 was derived to adjust measured modulus (Etest) from the value at the test density ratio (DRtest) to a value at the in-service density ratio (DRin-service).

Ein-service = Etest (1+ 0.05 x (DRin-service DRtest)) 5



5. Flexural Strength Results

5.1 Introduction Flexural strength is a key material property to explain the fatigue performance of concrete pavements (Austroads 2012a). In addition, it has been adopted in several other overseas cemented materials fatigue relationships as previously summarised (Austroads 2010). Consequently, the flexural strengths were measured to assess whether they could explain the differences in fatigue performance between cemented materials.

During flexural strength testing a monotonic load was applied to the simply supported beam using third point loading as described in Section 3. This induced (theoretically) a constant moment in the mid third of the beam and thus the beam was most likely to break at the weakest point in this zone. During loading the applied load and the resulting deflection were constantly monitored allowing the tensile stress and tensile strain at the bottom of beam to be calculated. The test method described in Appendix B.2 was used.

5.2 Results The test results for each beam are provided in Appendix C and are summarised below in Table 5.1, Table 5.2 and Table 5.3. It should be noted that there were substantially less beams tested to determine the flexural strength compared to flexural modulus. Note high variation of 28 day breaking strain for prior stream gravel may have been due to varying amounts of micro-cracking possible due to handling of these low strength materials.

Table 5.1: Flexural strength and breaking strain of laboratory-manufactured test beams after 28 days moist curing

Material(2) Cement content

(%)

Moist curing period

Number of

samples tested

Density ratio(1) Flexural strength Breaking strain

Mean (%)

Coefficient of

variation (%)

Mean (MPa)

Coefficient of

variation (%)

Mean ()

Coefficient of

variation (%)

Weathered granite 3 28 days 6 94.0 1.1 0.58 12 196 46 Weathered granite (repeat) 3 28 days 6 94.6 0.8 0.65 18 138 29

Weathered granite 5 28 days 4 95.2 0.9 1.14 5 167 21 Calcrete limestone 3 28 days 6 96.5 1.0 0.65 9 171 20 Calcrete limestone 5 28 days 6 96.7 1.1 1.03 14 208 19 Basalt (Mt Gambier) 3 28 days 6 97.2 1.5 1.24 5 160 3 Prior stream gravel 5 28 days 6 95.2 0.5 0.91 10 383 245 Modified prior stream gravel 3 28 days 6 93.2 1.0 0.73 15 100 13

Lateritic gravel 3 28 days 6 96.9 1.4 0.67 14 115 10 Metagreywacke 3 28 days 6 94.5 1.0 0.80 9 103 16

1 Density ratio is based on the modified maximum dry density. 2 General purpose (GP) cement was used for all materials.



Table 5.2: Flexural strength and breaking strain of laboratory-manufactured test beams after five months moist curing


(%)

Moist curing period

Number of

samples tested


Mean (%)

Coefficient of

variation (%)

Mean (MPa)

Coefficient of

variation (%)

Mean ()

Coefficient of

variation (%)

Prior stream gravel 3 5 months 10 97.2 1.1 0.78 9 128 21 Prior stream gravel 5 5 months 9 96.5 1.2 1.03 12 128 11 Recycled concrete 3 5 months 5 95.7 0.9 0.69 9 152 14 Hornfels 3 5 months 7 94.1 0.9 1.57 8 125 16 Basalt (Purga) 3 5 months 6 94.5 1.8 0.98 10 150 8 Granite 3 5 months 6 96.1 1.6 1.13 9 139 14 Quartzite 4 5 months 4 94.6 2.3 1.41 20 179 7


Table 5.3: Flexural strength and breaking strain of laboratory-manufactured test beams after nine months moist curing


(%)

Moist curing period

Number of

samples tested


Mean (%)

Coefficient of

variation (%)

Mean (MPa)

Coefficient of

variation (%)

Mean ()

Coefficient of

variation (%)

Weathered granite 3 9 months 4 94.3 1.0 0.96 9 184 8 Weathered granite 5 9 months 3 94.2 1.1 1.51 11 169 5 Calcrete limestone 3 9 months 4 96.6 0.2 0.97 12 212 24 Calcrete limestone 5 9 months 4 97.3 1.0 1.50 10 208 10 Basalt (Mt. Gambier) 3 9 months 4 98.0 2.7 1.97 6 216 5 Prior stream gravel 5 9 months 4 94.2 1.5 1.19 11 130 16 Modified prior stream gravel 3 9 months 4 94.5 1.2 1.27 13 126 16

Quartzite 4 9 months 4 96.5 0.7 1.55 6 155 12




5.3 Strength Variation with Density In Austroads (2014) the usefulness of flexural strength in explaining differences in fatigue performance between materials was investigated. To undertake this investigation the flexural strength results needed to be adjusted to the same density ratio as the fatigue test beams. As can be seen in Figure 5.1 for some materials significant difference in density occurred. Hence a method was required to adjust flexural strength results for density ratio.

Figure 5.1: Density ratio differences of flexural strength and fatigue beams

Analysis was undertaken to determine the degree of correlation between the flexural strength and density ratio for each material. Due to the scatter of the data and the low number of test beams most of these regressions were not statistically significant at the 5% level. Consequently it was decided to pool the data and analyse the data as follows:

To enable the results of all materials to be pooled, it was decided to express the strength results of each material in terms of a ratio to the materials strength at a density ratio of 95%.

For each material the five months and nine months flexural strength data was reviewed to assess those materials for which the strength at 95% density ratio could be estimated without excessive extrapolation. Materials were deleted from the analysis if they did not at least have one test beam within 0.5% of a density ratio of 95%.

For each of these materials, regression analysis was used to predict the flexural strength at a density ratio (DR) of 95% (FS95).

For each material, the relative flexural strength of each beam was calculated by dividing the flexural strength by the flexural strength at a density ratio of 95% (that is, FSDR/ FS95). Similarly for each beam, its percentage density ratio was divided by 95% (that is, DR/95).

Pooling all beams of all selected materials, the variation in flexural strength values with density ratio were plotted and a relationship determined by regression analysis as shown in Figure 5.2. The slope of the regression line indicates the flexural strength increases 5.1% for a 1% increase in density ratio.

It was concluded that flexural strengths increases 5% for a 1% increase in the density ratio.



As detailed in Appendix E, this density adjustment procedure is proposed to be included in the revision of the Guide.

Figure 5.2: Flexural strength variation with density

5.4 Curing Duration Several materials were tested after different curing durations and these are shown in Figure 5.3. All eight materials exhibited the expected response with the flexural strength increasing the longer they were cured. The error bars on each column represent one standard deviation.

Figure 5.3: Change in flexural strength with curing

The changes in flexural strength were calculated and are shown in Table 5.4. The average increase in flexural strength from 28 days to 9 months was 51%. These findings relate to GP cement, different increases are anticipated for slow-setting cement binders. There was insufficient data to estimate the percentage increase between 28 days and 5 months curing.

y = 0.051xR = 0.40

-0.25

-0.20

-0.15

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

-3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0

FSDR - FS95--------------------

FS95

Density ratio - 95



Table 5.4: Change in flexural strength with curing

Material(1) Cement content (%)

Flexural strength (MPa)

Increase in flexural strength from 28 day to

9 months (%)

28 days moist curing

Five months moist curing

Nine months moist curing

Weathered granite 3 0.58 0.96 66

Weathered granite 5 1.14 1.51 32

Calcrete limestone 3 0.65 0.97 48

Calcrete limestone 5 1.03 1.50 45

Basalt (Mt. Gambier) 3 1.24 1.97 59

Prior stream gravel 5 0.91 1.03 1.19 31

Modified prior stream gravel 3 0.73 1.27 74

Quartzite 4 1.41 1.55

Average 0.95 1.22 1.36 51 1 General purpose (GP) cement was used for all materials



6. Breaking Strain Results

6.1 Introduction Breaking strain was calculated from the deflection during flexural strength testing and has been used in fatigue models overseas as discussed by Gonzalez et al. (Austroads 2010). Results for all beams tested are given in Appendix C and summarised in Table 5.1, Table 5.2 and Table 5.3.

6.2 Variation in Breaking Strain with Density A regression analysis was conducted to examine the relationship between density ratio and the breaking strain during flexural strength testing. For the 25 sets of data available, only two were statistically significant at the 5% level. Consequently similarly to the approach used for flexural strength (Section 5.3) it was decided to pool the data and analyse the data as follows:

For each material the five months and nine months breaking strain data was reviewed and any material with breaking strain strength data with at least one point not more than 0.5% different from a density ratio of 95% was selected.

For each of these materials, regression analysis was used to predict the flexural strength at a density ratio (DR) of 95% (BS95).

For each material, the relative breaking strain of each beam was calculated by dividing the measured breaking strain by the predicted breaking strain at a density ratio of 95% (that is, BSDR/ BS95).

Pooling all beams of all selected materials, the variation in these breaking strain values with density ratio plotted (Figure 6.1).

It was concluded that the breaking strain data was too scattered to quantify the variation in breaking strain with density ratio.

Figure 6.1: Variation in breaking strain with density

-0.35

-0.25

-0.15

-0.05

0.05

0.15

0.25

-3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0BSDR - BS95--------------------

BS95

Density ratio - 95



6.3 Effect of Cure Duration Several materials were tested after different curing durations and these are shown in Figure 6.2. All of the eight materials exhibited an increase in breaking strain the longer they were cured. The error bars on each column represent one standard deviation.

For eight materials tested, the average increase in breaking strain between 28 days and nine months curing was an increase of 20%, with a range of 0% to 35%. These findings relate to GP cement, different increases are anticipated for slow-setting cement binders.

Figure 6.2: Change in breaking strain with curing



7. Fatigue Results

7.1 Introduction Flexural fatigue performance of cement treated beam samples was measured using the 4-point bending apparatus (Section 3.2.4) and fatigue procedure outlined in Appendix B.3. Flexural fatigue testing was carried out at a minimum of five months cure age.

The end of fatigue life was defined as the number of load cycles required to achieve half of the initial modulus. Normally, beam samples failed (broke) before attaining half modulus or only a few load cycles were applied between half modulus condition and sample breaking.

It was intended that sufficient load would be applied to the beams to induce fatigue failure in the beam within one million loading cycles. However, about 30 beams had not reached terminal condition before the test was terminated at one million cycles. In these cases an estimate of the fatigue life was made by extrapolating the data. However, the estimated fatigue life was capped at a maximum of ten million cycles. This affected less than 5% of the fatigue lives used in the analysis.

The term initial modulus was applied to the mean flexural modulus determined between pulses 10 to 50 cycles. These initial moduli in the fatigue test varied from the flexural moduli at a strain of 50 microstrain reported in Section 4.

7.2 Results Results for all test beams are provided in Appendix C, plotted in Appendix D and summarised in Table 7.1. The mean initial modulus and the mean estimated fatigue lives have been rounded to the nearest hundred cycles.

Table 7.1: Properties of fatigue beams after five and nine months moist curing

Moist curing period

(months)



(%)

Mean initial strain

(microstrain)

Mean initial stress (kPa)

Mean initial modulus

(MPa)

Mean cycles to half initial modulus

5 Recycled concrete (RCC) 14 96.8 60 488 8 800 138 500

Hornfels (HO3) 13 93.6 58 1 075 19 500 2 663 800

Basalt (BAP3) 10 96.3 50 599 12 100 2 234 000

9 Weathered granite (WG3) 19 94.4 78 640 8 700 151 800

Weathered granite (WG5) 21 95.7 75 1 017 14 300 223 900

Calcrete limestone (CL3) 21 96.6 70 552 8 500 2 128 000

Calcrete limestone (CL5) 22 97.9 89 956 11 800 246 600

Basalt (BAM3) 16 97.2 91 1 152 13 800 387 300

Prior stream gravel (PSG5) 17 95.1 68 809 12 300 216 700

Modified PSG (MPSG) 21 94.3 49 764 16 300 199 900

Quartzite (QZ4_1) 21 97.4 82 1 149 14 700 628 900

1 Relative to modified maximum dry density. 2 General purpose (GP) cement was used for all materials.



7.3 Analysis

Introduction 7.3.1As described previously (Austroads 2010), cemented materials fatigue relationships use either applied strain or applied stress to predict fatigue performance. In addition, in some relationships the logarithm of fatigue life is related to the logarithm of strain or stress (so-called log-log models), whilst other relationships are semi-logarithmic with the logarithm of fatigue life related to strain or stress.

In analysing the fatigue results these model forms were investigated to compare their ability to explain the variation in fatigue life.

Strain-based Fatigue Relationships 7.3.2Linear regression analysis was undertaken on each material using both log-log and semi-log relationships of the forms as presented in Equation 6 and Equation 7. The resulting relationships are detailed in Table 7.2 for each material tested. The term a in Equation 6 is the strain damage exponent of the fatigue relationship. In addition, the significance of the strain parameter and the correlation coefficient (R2) values are included in Table 7.2.

b)log(a)Nlog( += 6

b)(a)Nlog( += 7

where

N = number of load repetitions to half initial modulus (fatigue life) = initial elastic strain (microstrain)

a = regression coefficient b = regression coefficient



Table 7.2: Strain-based fatigue relationships for each material


(%)

Moist curing period

(months)

Number of test beams

Mean density

ratio (%)

Fatigue equation Significance (P-Value)

Adjusted R2

Recycled concrete 3 5 14 96.8 log(N) = 34.48 17.12 log (strain) < 0.01 0.55 log(N) = 11.53 0.1253 (strain) < 0.01 0.50

Hornfels 3 5 13 94.0 log(N) = 31.21 14.56 log (strain) < 0.01 0.67 log(N) = 12.21 0.1137 (strain) < 0.01 0.71

Basalt (Purga) 3 5 10 96.3 log(N) = 46.64 24.51 log (strain) < 0.01 0.87 log(N) = 16.00 0.2192 (strain) < 0.01 0.86

Weathered granite 3 9 18 94.4 log(N) = 33.91 15.45 log(strain) < 0.01 0.76 log(N) = 10.60 0.0755(strain) < 0.01 0.73

Weathered granite 5 9 21 95.7 log(N) = 23.21 9.87 log(strain) < 0.01 0.68 log(N) = 8.77 0.0534(strain) < 0.01 0.69

Calcrete limestone 3 9 21 96.6 log(N) = 33.51 15.44 log(strain) < 0.01 0.47 log(N) = 11.26 0.0882(strain) < 0.01 0.45

Calcrete limestone 5 9 22 97.9 log(N) = 32.46 14.19 log(strain) < 0.01 0.42 log(N) = 11.23 0.0720(strain) < 0.01 0.44

Basalt (Mt. Gambier) 3 9 16 97.2

log(N) = 35.79 15.57 log(strain) < 0.01 0.78 log(N) = 12.18 0.0755(strain) < 0.01 0.80

Prior stream gravel 5 9 17 95.1 log(N) = 27.10 12.17 log(strain) < 0.01 0.77 log(N) = 9.788 0.0729(strain) < 0.01 0.77

Modified prior stream 3 9 21 94.3

log(N) = 20.69 9.32 log(strain) < 0.01 0.57 log(N) = 8.982 0.0822(strain) < 0.01 0.58

Quartzite 4 9 21 97.4 log(N) = 23.63 9.83 log(strain) < 0.01 0.56 log(N) = 8.863 0.0489(strain) < 0.01 0.55

1 General purpose (GP) cement was used for all materials.

It was concluded that there was no significant difference in the fit to the laboratory data between the logN-log and logN- models. The Guide (Austroads 2012a) currently includes a logN-log model. These results provide no support to change to a semi-logarithmic model.

It was noted that there was at least a 95% probability that the logarithm of the initial strain was related to the logarithm of fatigue life in all cases. The average strain damage exponent was calculated to be 14.4.



A linear regression analysis was also conducted on each material similar to that in Equation 6 but also including a logarithm of modulus term as shown in Equation 8:

c)Elog(b)log(a)Nlog( ++= 8

where

N = number of load cycles to failure (fatigue life) = initial elastic strain (microstrain)

E = flexural modulus at 50 microstrain (MPa)

a = strain damage exponent

b , c = regression coefficients

The value of modulus adopted for the analysis was the flexural modulus measured during a separate test using the method outlined in Appendix B.2 and adjusted to a value equivalent to that which would have been obtained if it had been tested at 50 .

In all cases, modulus was not a statistically significant factor at the 95% level of significance.

Stress-based Fatigue Relationships 7.3.3A similar analysis to that done in Section 7.3.2 was repeated substituting initial stress for initial strain. Results of each material were analysed separately using both log-log and semi-log relationships of the forms as presented in Equation 9 and Equation 10. The results of the analysis are provided in Table 7.3.

b)log(a)Nlog( += 9

b)(a)Nlog( += 10

where

N = number of load cycles to half the initial modulus (fatigue life)

= initial elastic stress (kPa)

a = regression coefficient

b = regression coefficient

It was concluded that there was no significant difference in the fit to the laboratory data between the logN-log and logN- models.

The term a in Equation 9 is the stress damage exponent of the fatigue relationship. Using the results of statistically significant relationships, an average stress damage exponent of 12.4 was calculated.



Table 7.3: Stress-based fatigue relationships for each material


(%)

Moist curing period

(months)

Number of test beams

Mean density

ratio (%)

Fatigue equation Significance (P-Value)

Adjusted R2

Hornfels 3 5 13 93.6 Log(N) = 46.80 13.60 Log(Stress) < 0.01 0.61 Log(N) = 11.83 0.00575(Stress) < 0.01 0.66

Basalt (Purga) 3 5 10 96.3 Log(N) = 55.26 17.97 Log(Stress) 0.01 0.58 Log(N) = 13.52 0.0136(Stress) 0.01 0.58

Weathered granite 3 9 19 94.4 Log(N) = 54.98 17.95 Log(Stress) < 0.01 0.62 Log(N) = 12.14 0.0117(Stress) < 0.01 0.64

Weathered granite 5 9 21 95.7 Log(N) = 38.53 11.24 Log(Stress) < 0.01 0.76 Log(N) = 9.16 0.00429(Stress) < 0.01 0.74

Prior stream gravel 5 9 17 95.1 Log(N) = 39.97 12.12 Log(Stress) < 0.01 0.66 Log(N) = 9.788 0.00619(Stress) < 0.01 0.67

Modified prior stream 3 9 21 94.3

Log(N) = 24.1 6.64 Log(Stress) < 0.01 0.43 Log(N) = 7.934 0.00387(Stress) < 0.01 0.45

Quartzite 4 9 21 97.4 Log(N) = 26.81 7.18 Log(Stress) < 0.01 0.41 Log(N) = 7.979 0.00271(Stress) < 0.01 0.39

Relationships that were not significant at 95% confidence level

Calcrete limestone 3 9 21 96.6 Log(N) = 32.8 10.11 Log(Stress) 0.11 0.09 Log(N) = 9.535 0.00802(Stress) 0.10 0.09

Calcrete limestone 5 9 22 97.9 Log(N) = 22.32 5.87 Log(Stress) 0.16 0.05 Log(N) = 7.479 0.00276(Stress) 0.14 0.06

Basalt (Mt. Gambier) 3 9 16 97.2

Log(N) = 8.24 0.95 Log(Stress) 0.80 0.07 Log(N) = 5.747 0.00036(Stress) 0.80 0.07

Recycled concrete 3 5 14 96.8 Log(N) = 35.19 11.46 Log(Stress) 0.07 0.19 Log(N) = 9.611 0.0106(Stress) 0.06 0.20

1 General purpose (GP) cement was used for all materials.

A linear regression analysis was also conducted on each material similar to Equation 9 but also including a logarithm of modulus term as shown in Equation 11:

c)Elog(b)log(a)Nlog( ++= 11

where

N = number of load cycles to half initial modulus (fatigue life)

= initial elastic stress (kPa)

a = stress damage exponent

E = flexural modulus adjusted to 50 microstrain (MPa)

b , c = regression coefficients



The value of modulus adopted for the analysis was the flexural modulus measured during a separate test using the method outlined in Appendix B.2 and adjusted to a value equivalent to that which would have been obtained if it had been tested at 50 .

As observed for the strain-based relationship, in all cases, modulus was not a statistically significant factor at the 95% level of significance. Hence these regressions are not listed in this report.

Summary 7.3.4It was concluded that strain-based equations provided a better fit to the laboratory data than the stress-based equations. It should be noted that this finding may have been influenced by the fact that in conducting the fatigue experiments the load applied to each test beam of each material was selected with the objective of testing each material over a range of applied strains. In other words the loads used favoured the generation of strain-based fatigue relationships rather than a stress-based fatigue relationship.

In terms of the strain-based equation, the logN-log and logN- models had similar ability to explain the variation in the laboratory fatigue data.



8. Estimation of Flexural Strength

8.1 Introduction As summarised in Section 9, it is proposed that the Guide (Austroads 2012a) design procedures include a method of fatigue life prediction based on the design modulus and design flexural strength.

Whilst the current Guide includes a means of estimating modulus from unconfined compressive strength (UCS), as flexural strength is currently not required to predict fatigue performance a means of estimating flexural strength is not currently provided.

Given that the equipment to measure flexural strength is not commonly available in road agency and industry laboratories at present, as part of the project, methods of estimating flexural strength from commonly specified properties were investigated.

8.2 Predicting Strength from Properties of Constituent Materials Alderson (Austroads 2013) developed a method of estimating flexural strength from these properties such as the properties of the constituent materials, cement content and moist-curing period.

Using data in Table 5.1, Table 5.2 and Table 5.3 from a wide range of cemented materials stabilised with GP cement, including lean-mix concrete previously report (Austroads 2010), the following relationship was obtained (Equation 12):

FS = 0.377 + 0.284*CC 0.060*MC 0.031*F4 + 0.001*Age + 0.037*DR 0.027*C6 12

where

FS = flexural strength (MPa)

CC = GP cement content (percentage by mass)

MC = moisture content before addition of GP cement (percentage by mass)

F4 = fine aggregate content (percentage by mass passing 4.75 mm)

Age = period of moist curing prior to testing (days)

DR = density ratio of compacted sample (%)

C6 = coarse aggregate content (percentage by mass retained on the 6.7 mm sieve)

The ranges of data used in the development of Equation 12 are given in Table 8.1



Table 8.1: Data used in development of Equation 12

Variable Minimum Mean Maximum

Flexural strength (MPa) 0.2 1.2 3.8

Cement content (%) 3.0 4.1 10.0

Moisture content (%) 2.8 10.0 15.9

Percentage by mass passing 4.75 mm (%) 24 60 99

Age (days) 27 130 377

Density ratio (relative to modified Proctor maximum dry density) (%) 91.0 95.8 100.4

Percentage by mass retained on the 6.7 mm sieve (%) 0 33 100

Plasticity index 0 8 20

It is generally believed that the plasticity index does influence the strength of the cemented materials, but this was not evident in the data.

The variables included in Equation 12 are listed in order of statistical significance with cement content being the most significant. In reviewing Equation 12, most of the correlation factors confirm expectations as follows:

Increasing the cement content increased the flexural strength. Increasing the moisture content decreased the flexural strength. Increasing the fine aggregate content decreased the flexural strength. Increasing the density ratio of the compacted sample increased the flexural strength. Increasing the curing time increased the flexural strength.

Figure 8.1 compares the measured flexural strengths of individual test beams with the values predicted using Equation 12. The standard error in predicting the strength of individual test beams was 0.26 MPa. Note that Equation 12 is not applicable to slow-setting cement binders.

Figure 8.1: Comparison of Equation 12 predicted strengths with measured values

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

Predictedflexural

strength (MPa)

Measured flexural strength (MPa)



As the uncertainty in predicting flexural strength is high, the data was reanalysed as follows:

The lean-mix concrete data was removed from the data set as the proposed Guide revision does not require flexural strength of lean-mix concrete.

As the proposed fatigue relationship based on flexural strength is not applicable to recycled crushed concrete (Austroads 2014), the recycled crushed concrete were deleted from the analysis.

Given that the design flexural strength is based on the value after 90 days moist curing strengths and as there is a significant increase in strength between 28 days and 90 days, it was decided to delete the 28 days results from the analysis.

As a method was required to predict the mean flexural strength of a material, for each material the mean measured flexural strength and the mean density ratio were used in the analysis rather than individual beam data.

Using regression analysis, the following simplified strength prediction equation was derived (Equation 13):

FS = 83.9/F6 + 0.0015*Age 0.355 13

where


F6 = fine aggregate content (percentage by mass passing 6.7 mm)

Age = period of moist curing prior to testing (days)

The standard error in predicting the mean strength was 0.18 MPa. Note that Equation 13 is not applicable to slow-setting cement binders.

Figure 8.2 compares the flexural strengths predicted using Equation 13 with the mean measured flexural strengths of each material. It is apparent that the predicted strength can vary from the measured value by more than 10% for about half of the materials. For example, the crushed granite with 3% cement (GR3) has a predicted strength of 1.19 MPa compared to its mean measured value of 1.04 MPa, that is the strength is over-estimated by 14%. If this over-estimated strength was used in the proposed fatigue relationship, the fatigue life would be over-estimated by about a factor of seven. In addition it would be expected that a predictive relationship would have included an increase in flexural strength with cement content and density ratio. In regression analysis both these factors were not statistically significant, possibly due to the limited amount of data used to derive Equation 13. Furthermore the equation related to materials stabilised with GP cement and is not applicable to slow-setting binders which limits its usefulness in design. Accordingly it is proposed not to include this method of predicting flexural strength in the revised Austroads design procedures.



Figure 8.2: Comparison of Equation 13 predicted strengths with measured values

8.3 Flexural Strength from UCS

Review of Data in the Literature 8.3.1To further explore the possibility of estimating the flexural strength (FS) from other material properties, a brief review of the correlation between the UCS test and the flexural strength test was undertaken. This study was considered as a pilot study and not a comprehensive review of the topic.

Data from several reports (Doshi & Guirguis 1983, Katsakou & Kolias 2007, Kolias, Kasselouri & Karahalios 2005, Mitchell, Dzwilewski & Monismith 1974, Thompson 1986) was reviewed and while some of these reports did not specifically investigate the relationship they did measure both properties.

Following a review of the literature, Thompson (1986) recommended the following relationship for use in design (Equation 14):

FS = 0.2 UCS 14

where


UCS = unconfined compressive strength (MPa)

More recently, Equation 14 has been adopted in the AASHTO mechanistic-empirical pavement design guide (AASHTO 2008).

0.6

0.7

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

1.9

2.0

0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0

Predicted flexuralstrength

(MPa)

Measured flexural strength (MPa)

FS = 83.9/F6 + 0.0015Age - 0.355

Line of equality

CL3 QZ4

WG3GR3

Bam3

HO3

PSG5

PSG3

Bap3

CL5

QZ4



Measurements 8.3.2A limited investigation was undertaken to assess the applicability of Equation 14 for use in the revised Austroads Guide. Five of the cemented materials previously tested for flexural strength were selected and UCS samples prepared. These were tested after 28 days moist curing using the Australian standard test method for UCS and the results compared to the 28-day flexural strength values. After moist curing for 28 days, the UCS specimens soaked in water for four hours at 2025 C and then drained before UCS testing.

The results are shown in Figure 8.3 and are summarised in Table 8.2. The UCS results for the crushed basalt with 3% cement (BAM3) were unexpectedly low and appear to be erroneous given the high quality of this crushed rock.

Table 8.2: Flexural strength and UCS test results

Material Cement content (%)

Flexural strength Unconfined compressive strength

Density ratio (%)

Flexural strength (MPa)

Density ratio (%)

UCS(1) (MPa)

Basalt (Mt Gambier) (BAM3) 3 97.2 1.24 101.7 3.75

Calcrete limestone (CL3) 3 96.5 0.65 100.4 5.01

Calcrete limestone (CL5) 5 96.7 1.03 100.6 5.97

Granite (GR3) 3 96.0 1.13 99.9 6.57

Metagreywacke (MTG3) 3 94.5 0.80 101.6 6.66

1 Samples were cured in a fog room for 28 days then soaked for four hours at 2025 C, then allowed to drain before testing.

Figure 8.3: Flexural strength variation with UCS

0.5

0.6

0.7

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5

3 3.5 4 4.5 5 5.5 6 6.5 7

Flexuralstrength

(MPa)

Unconfined compressive strength (MPa)

BAM3

MTG3

GR3

CL5

CL3



The density ratios of the UCS and flexural strength specimens differed. The flexural strength beams had an mean density ratio of 96% while the UCS cylinders had a mean density ratio of almost 101%. Hence, to assess the correlation between the two tests, for each material the measured flexural strength results were adjusted to an estimated value at the mean density ratio of the UCS samples assuming 5% increase in flexural strength for each 1% increase in density ratio (Section 5.3). The resulting data is plotted in Figure 8.4 and compared to Equation 14.

It is considered that currently there is insufficient data to conclude that Equation 14 estimates flexural strength to the required precision for use in the Guide.

Figure 8.4: Adjusted flexural strength with UCS

0.6

0.7

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5

1.6

3.5 4 4.5 5 5.5 6 6.5 7

Flexuralstrength

at 28 days(MPa)

Unconfined compressive strength at 28 days (MPa)

AASHTO (2008)FS = 0.2 UCS

BAM3

MTG3

GR3

CL5

CL3



9. Proposed Guide Revision

The objective of the project was to review the design procedures for cemented materials in the Guide to Pavement Technology Part 2: Pavement Structural Design (Austroads 2012a). Based on the research reported here a number of changes have been recommended for the Guide (Austroads 2014) in terms of the characterisation of cement treated crushed rocks and natural gravels. (Note no changes are proposed to the design procedures for lean-mix concrete).

In terms of modulus characterisation the proposed changes are:

change the definition of cemented materials design modulus to be the 90-day flexural modulus in situ a test method to manufacture laboratory test beams and measure flexural modulus the inclusion of a procedure to determine the design modulus from the measured flexural modulus the inclusion of a procedure to adjust the measured flexural modulus for differences in density between

the modulus test beams and the density in situ

amendments to presumptive moduli values.

In terms of fatigue characterisation the proposed changes are:

a test method to manufacture laboratory test beams and measure fatigue characteristics and hence determine a laboratory fatigue relationship

a test method to manufacture laboratory test beams and measure flexural modulus a procedure to estimate the laboratory fatigue characteristics from the measured flexural modulus and

flexural strength

a procedure to determine in-service fatigue relationships from the laboratory fatigue characteristics a procedure to determine in-service fatigue relationships from design flexural modulus and design flexural

strength

presumptive in-service fatigue relationships based on presumptive moduli and strengths for three types of cemented materials.

Appendix E and Appendix F contains the proposed revised text for Section 6.4 of the Guide to Pavement Technology Part 2: Pavement Structural Design.



10. Summary

The objective of the project was to review the design procedures for cemented materials in the Guide to Pavement Technology Part 2: Pavement Structural Design (Austroads 2012a). The report provides the data which was used to develop a framework for the revision of the Guide as reported elsewhere (Austroads 2014).

In developing this framework cognisance was taken of the following key findings:

It was observed that measured flexural modulus varies with the applied load/strain. Hence a procedure was developed to standardise measured modulus to a value at 50 microstrain.

In the proposed revision of the Guide a method will be provided to estimate fatigue characteristics which utilises flexural strength. The preferred method of determining flexural strength is laboratory strength measurement of test beams. Alternative methods of estimating flexural strength from material properties such as cement content, particle size distribution, plasticity, density and UCS were investigated. It was concluded that none of these methods is currently suitable for inclusion in the Guide given the high dependence of fatigue life on flexural strength.

Flexural moduli and flexural strength values vary with density ratio. Hence procedures were developed to enable measured moduli and strength to be adjusted for differences between the density of test beams and in situ densities.

In analysing the laboratory fatigue data both strain-based and stress-based fatigue relationships were fitted to the data. It was concluded that strain-based fatigue models were a better fit to the measured data than stress-based fatigue models. However, this conclusion may have been influenced by the fact that the load levels selected for the test beams of each materials were chosen to give a range of initial strains rather than a range of initial stresses.

In terms of the alternative strain-based fatigue models, it was concluded that there was no significant difference in the data fit between logN-log models and semi-logarithmic models (logN-). Based on this finding it is proposed the Guide continues to use a logN-log for both calculating laboratory fatigue relationships from measured data and for in-service fatigue relationships.

Using the laboratory results in this report a framework for revision of the Guide has been prepared (Austroads 2014). Based on this framework the proposed revised text was prepared (Appendix E and Appendix F). Note no changes are proposed to the design procedures for lean-mix concrete.



References

AASHTO 2008, Mechanistic-empirical pavement design guide: manual of practice: interim edition, MEPDG-1, AASHTO, Washington, DC, USA.

Austroads 2007, Austroads LTPP and LTPPM study: summary report for 2005-06, AP-T81-07, Austroads, Sydney, NSW.

Austroads 2010, Cost-effective structural treatments for rural highways: cemented materials, AP-T168

AP-R462-14

Documents

austroads austroads

road transport issues

road transport sector

austroads project

road pavements

road industry

austroads membership

success of austroads