Improved Specimen Preparation for Soil-Cement Design and Construction Monitoring

Transportation Research Board Webinar AFS80 Standing Committee on Stabilization

of Geomaterials and Recycled Materials

September 12, 2016

W. Griffin Sullivan, PE Engineer, Mississippi DOT

Doctoral Student, Mississippi State University 601-359-1755, wsullivan@mdot.ms.gov

Isaac L. Howard, PhD, PE Materials and Construction Industries Chair Civil and Environmental Engineering Dept.

Mississippi State University 662-325-7193, ilhoward@cee.msstate.edu

Acknowledgements • Several individuals have actively supported this work

– Joe Ivy (MSU) Richard Sheffield (MDOT-Retired)

– Brennan Anderson (MSU Alumni) Tim Cost (LafargeHolcim)

– James Williams (MDOT) Bill Barstis (MDOT)

– Mike Sullivan (MDOT) Caleb Hammons (MDOT)

• Mississippi Department of Transportation – Funded State Study 206

– http://mdot.ms.gov/portal/research.aspx

• Click Reports

• Click Interim and Final Reports

• Scroll down to locate State Study 206

Key References 1. Sullivan, W.G. (2012). Investigation of Compaction and Corresponding

Thermal Measurement Techniques for Cementitiously Stabilized Soils. MS Thesis, Mississippi State Univ. http://sun.library.msstate.edu/ETD-db/theses/available/etd-08152012-152440/unrestricted/Sullivan_Thesis_(DRAFT3).pdf

2. Howard, I.L., Sullivan, W.G., Anderson, B.K., Shannon, J., Cost, T. (2013). Design and Construction Control Guidance for Chemically Stabilized Pavement Base Layers. Report FHWA/MS-DOT-RD-13-206, Mississippi Department of Transportation, pp. 162.

3. Sullivan, W.G., Howard, I.L., Anderson, B.K. (2015). “Development of Equipment for Compacting Soil-Cement into Plastic Molds for Design and Quality Control Purposes,” Transportation Research Record, 2511, 102-111.

4. Howard, I.L., Sullivan, W.G., Anderson, B.K. (2016). “Soil-Cement Laboratory Mixture Design That Interfaces with Pavement Design and Construction Quality Control,” Transportation Research Record, 2580, 1-9.

Webinar Outline

1. Traditional approaches to soil-cement activities and pitfalls

2. Development of the Plastic Mold Compaction Device (PM Device)

3. PM Device specimen properties

4. Laboratory application of PM Device

5. Construction QC/QA application of the PM Device

6. MEPDG application of the PM Device

Webinar Learning Objectives 1. Discuss issues with traditional soil-cement approach 2. Describe design and operation of the PM Device and

tests that can be performed on specimens 3. Understand how the PM Device can be utilized as a

tool in the larger framework for design and QC/QA

Plastic Mold (PM) Device

Traditional Soil-Cement Approach • Practice review & 2012 survey of state DOTs (21 total)

– All DOTs use compressive strength as criteria. Few DOTs use durability tests. (Most states use Proctor specimens)

– Design criteria and curing protocols vary widely among all state DOTs

– Most only check field spread rate and density. Five state DOTs make field specimens or cut cores for QC.

• Portland Cement Association and US Army Corps – Both have compressive strength and durability criteria – Both utilize Proctor specimens

• Overall, not a lot of consistency

Traditional Soil-Cement Approach Pavement Design: Performed using structural coefficient or using modulus testing performed with 2:1 h/d ratio specimens Laboratory Mixture Design:

with Proctor Specimens

Check Density Check Spread Check Pulverization

Construction Quality Control / Quality Assurance

1.15:1 h/d ratio

Mississippi Soil-Cement Practices • MDOT is a heavy user of soil-cement for pavement base

and subgrade layers (243 million yd2 over past 6.5 yrs.) • Soil-Cement base layer given a structural coefficient of 0.2 • Currently, MDOT uses Proctor equipment and unconfined

compressive strength (UCS) to determine optimum cement content. Avg content is 4.5% by mass. – Base → 14 day design, 300psi (2070kPa) min. – Subgrade → 7 day design, 200psi (1380kPa) min.

• During construction check spread rate, pulverization, visual inspection, density, and prescribed curing (bituminous seal coat and minimum 7 days no traffic)

• Current procedures work but could benefit from improvements

Soil-Cement in MS Construction Process: Spread, Mix, Water, Mix, Sheep Foot Roll, Vib. Roll, Grade, Rubber Tire Roll, Seal Materials: Mainly M145 A-2-4 or USCS SM Materials (Near 100% passing No 10 sieve)

Needed Improvements: DOT’s viewpoint • Interfacing the three aspects to soil-cement practice

– Pavement Structural Design (MEPDG) – Laboratory Mixture Design – Construction QC/QA

• Pavement Design – Need material input data for MEPDG

• Lab Mix Design – Need approach to increase design reliability

• Construction QC/QA – Need an approach to verify design properties – Need a tool to reduce variability

• The PM Device is a viable solution to these needs

Design and Operation of the PM Device

• The following slides will discuss the design and operation of the PM Device. A demonstration video is available for viewing at the following link – http://www.cee.msstate.edu/cmrc/technical-documents/

• Additionally, drawings of the PM Device are available in the thesis Appendix C located at the following link – http://sun.library.msstate.edu/ETD-

db/theses/available/etd-08152012-152440/unrestricted/Sullivan_Thesis_(DRAFT3).pdf

PM Device Concept and Design • The PM Device is designed to compact soil-

cement material into a 3x6 inch plastic cylinder mold. Approximate dimensions: 11 x 10 x 9 inches

1 2

1

2 3

3

4

5 5

PM Device Concept and Design • The split mold design closely surrounds the mold

to prevent distortion while a modified Proctor hammer is utilized to compact the specimen.

• The PM Device is portable so that specimens can easily fabricated in the lab and in the field. Whole device weighs approximately 25 pounds.

Plastic Cylinder Mold Modifications • Cut hole in bottom and

keep plastic cut-out • Sand bottom smooth • Place 1/16 inch thick

aluminum plate in mold • Tape plastic

cut-out to bottom to fill gap

PM Device Operation • Tape plastic cut-out to the bottom of the plastic

mold to fill in the gap between mold bottom and aluminum plate.

PM Device Operation • Place plastic mold in PM Device, clamp device

closed with locking vise-grips, and place collar on top.

PM Device Operation • Compact specimens in three equal lifts using 5

blows per lift with a modified Proctor hammer, and scarify the surface between each lift.

Why 5 blows/lift with T180 Hammer? • Optimum moisture and density determined

according to T134 (standard cement Proctor). This is our target moisture and density for the PM Device.

• For typical A-2-4 and A4 soils found in MS, 5 blows per lift achieved target density.

• Number of blows may vary with different soils

PM Device Operation • After compaction, remove specimen from device

and strike-off the top surface. Place plastic lid on the specimen.

PM Device Specimens

Strength

Before Testing

After Testing

Mold Removal

Modulus 1.97:1 height to diameter aspect ratio as measured

Finish Compaction

PM Device Design and Operation Summary

• PM Device costs less than $1000 for a local machine shop to fabricate. Price could lessen with bulk orders.

• Plastic molds could be used multiple times. Mold assembly (with aluminum plate) costs about $1.25 a piece and take about 2 minutes to modify.

• It takes approximately 3 minutes to compact a specimen using the PM Device.

Specimen Properties: Diameter

0

10

20

30

40

50

60

76.0 76.2 76.4 76.6 76.8 77.0 77.2 77.4

Rel

ativ

e Fr

eque

ncy

(%)

Average Specimen Diameter (mm)

n = 752 Mean = 76.8 Stdev = 0.16 COV = 0.2%

Average Top n = 752

Mean = 76.4 Stdev = 0.19 COV = 0.2%

Overall Average n = 752 Mean = 76.6 Stdev = 0.15 COV = 0.2%

Average Bottom

• No two specimen diameters differed by more that 2% as per AASHTO M205

Specimen Properties: UCS Results

1600

1800

2000

2200

2400

2600

2800

3000

3200

Proctor PM Device

UC

S (k

Pa)

UCS Mean = 2290 UCS COV = 6.9% γd Mean = 98.6% γd COV = 0.85%

UCS Mean = 2090 UCS COV = 6.5% γd Mean = 97.7% γd COV = 0.98%

n = 30 n = 28

≈ 30 psi

Potential Uses and Impacts • Uses:

1. Laboratory mixture design 2. Field quality control operations 3. MEPDG material inputs (e.g. elastic modulus)

• Impacts: – There is a disconnect between these uses and the PM

device could possibly interface all three activities – Increased mix design reliability, reduced variability,

verified mix and pavement design properties during construction, improved continuity to all soil-cement activities (pavement design, material design, & construction)

UCS Variability Analysis

• Strong linear correlation between varying cement contents (R2 between 0.96 and 0.99). The same correlation was confirmed when cement contents were varied between 2 and 10%.

• Coefficient of Variation (COV) were similar for all specimen types. • Curing protocols had a significant effect on UCS (affected UCS by 11-30%). • The PM-P specimens produced very similar UCS results when compared to

Proctor specimens of the same cement content.

Not Everything Needs to Change • Part of the PM Device integration is to keep what is working well intact • Most UCS gain occurred within 56 days (≈75 to 85% of 540 day strength)

1000

1500

2000

2500

3000

3500

4000

4500

0 50 100 150 200 250 300 350 400 450 500 550

UC

S (k

Pa)

t (days)

UCS (kPa) = 378ln(t) + 1546 R² = 0.88

Average of at least 3 per test time

Pit Soil B, Proctor Specimen, Cw = 4.3%

Laboratory Mix Design • Working toward test method that ends after

specimen has been cured for 24 hr • Modified Proctor hammer • Compact to Proctor density – vary number of

blows as needed • Lab bench cure for up to 24 hr in mold –

extract from mold…(not part of the method, but recommend to cure in room at 100% humidity uncovered and not submerged thereafter for lab mix design)

Reliability Analysis

• Results are intuitive to reliability based design approach (e.g. increased replication leads to increased reliability, decreased Margin of Error (ME) requires additional replicates).

• If the number of replicates was increased to 2, then the design reliability would be 75% at ME=150kPa, 85% at ME=225kPa, and 95% at ME=300kPa.

• Recommend increasing design test specimen replicates from 1 to 2 in order to decrease UCS variability and increase overall design reliability.

Pit Cement Cw (%)

Specimen Type

Mean (kPa)

Stdev (kPa)

ME @ 85% nreps = 1 (kPa)

nreps with ME = 150 kPa

nreps with ME = 225 kPa

nreps with ME = 300 kPa

75% 85% 95% 75% 85% 95% 75% 85% 95% B TH 3.4 Proctor 1795 117 168 0.75 1.25 2.25 0.25 0.50 1.00 0.25 0.25 0.50 B TH 4.3 Proctor 2293 158 228 1.50 2.25 4.25 0.75 1.00 2.00 0.25 0.50 1.00 B TH 5.2 Proctor 2590 216 311 2.75 4.25 8.00 1.25 2.00 3.50 0.75 1.00 2.00 B1 TH 4.3 Proctor 1766 205 295 2.50 3.75 7.25 1.00 1.75 3.25 0.50 1.00 1.75 B TH 4.3 SGC 2720 168 242 1.75 2.50 4.75 0.75 1.25 2.25 0.50 0.75 1.25 B TH 4.3 PM-P 2085 135 194 1.00 1.75 3.00 0.50 0.75 1.50 0.25 0.50 0.75 B TH 4.3 PM-CF 2461 243 350 3.50 5.50 10.00 1.50 2.50 4.50 0.75 1.25 2.50 B GV 4.3 PM-CF 2831 200 288 2.25 3.75 6.75 1.00 1.75 3.00 0.50 1.00 1.75

Average from ALL Soils (A & C not shown in table) 239 1.75 2.75 5.25 0.75 1.25 2.25 0.50 0.75 1.25 1: Specimens were soaked for 5 hours prior to UCS testing

Alternative Mix Design Procedure 1. Utilize the PM-P device (other devices can be used but

the PM-P offers the most advantages)

2. Increase test replicates to 2 specimens 3. Prepare 8 UCS specimens using project cement

a) 4 at the estimated optimum cement content b) 2 at 1% above and 2 at 1 % below estimated opt.

4. After 7 days curing, test 2 specimens at the estimated optimum cement content. After 14 days curing, test the remaining specimens.

a) If subbase design is desired, reverse actions for 7 and 14 days described above

Alternative Mix Design Procedure 5. Plot average UCS as a function of cement

content. Use a linear regression trendline to determine optimum cement content that meets minimum UCS criteria.

a. Trendline slope can be transposed to test time that only has one cement content represented (i.e. 7 day USC)

6. Recommend rounding to the nearest 0.1%

Mix Design Example • Base design (minimum 14 day UCS = 2070kPa)

Construction QC/QA Applications • Working toward test method that ends after

specimen has been cured for 24 hr • Modified Proctor hammer • Compact a fixed number of blows per layer (5 to 7) –

varying density that is corrected as shown earlier to Proctor density

• Cure in mold and monitor temperature with thermocouple affixed to outside of one of the plastic molds for 24 hr (after this period, there are several options for field specimens depending on how they are used, but none of this would be part of the proposed method)

Reduced Variability From Improved QC Tool

• Nobody wants this to be their boots

Soil Cement Construction

Cement Spread In-place Mixing

Compaction Curing

Field Issues Observed in Mississippi

Structural Failure

“Soft Areas”

Shrinkage Cracking

Observed Soil-Cement Variability for MS Hwy

Early Age Coring of Soil-Cement Layers

• Soil-cement base layer cored after 7 days of curing

40 cm Air

Top Surface

Cut Face

Field Mixed UCS and Field Cores • No significant difference between PM Device and cores in

limited testing (2 projects-3 locations each)

750

1000

1250

1500

1750

2000

750 1000 1250 1500 1750 2000

Core

s (kP

a)

PM Device - Field Compacted - Lab Cured (kPa)

Can UCS Data Indicate Cement Content?

PM Device Density Correction (Current work uses 5 blows/layer in field and adjusted UCS based on density)

Lab vs. Field Curing

0

1000

2000

3000

4000

5000

0 1000 2000 3000 4000 5000

Fiel

d Cu

red

UCS

(kPa

)

Lab Cured UCS (kPa)

PM Device Testing Limitation

• Work presented is for A-2-4 (or similar) materials – larger particles would eventually require a larger mold, but same concept should apply to materials with larger particles

• A small experiment was conducted on Cold In-Place Recycled (CIR) material. Larger particle size prevented fabrication of acceptable test specimens. Scaling up the PM Device for 4x8 inch molds may help.

Mechanistic-Empirical Pavement Design Guide

• MEPDG requires a modulus input for soil-cement materials. – Level 1 – direct measurement of modulus (E) – Level 2 or 3 – E=1200(UCS) or estimated 500,000psi

• With traditional soil-cement approach, two different types of test specimens would be required – 2:1 h/d ratio specimen for modulus testing – Proctor specimen for mixture design based on

strength

MEPDG • ASTM C469 – Elastic Modulus Testing

– Directly measure modulus on the same type specimen tested for strength

Compressometer

MEPDG - Elastic Modulus

0

2

4

6

8

10

12

14

2000 2500 3000 3500 4000 4500

Ela

stic

Mod

ulus

(GPa

)

UCS (kPa)

PM Device MEPDG Input Upper/Lower Data Boundary and Linear Data Fit

MEPDG Equation E = 1200*10-6*UCS

Trendline Equation E = Ci *10-6*UCS

Vehicle to Directly Tie MEPDG to as Built Pavement Properties

• For example, an MEPDG design uses 500,000 psi as an input for a soil-cment base layer, and with that modulus determines 6 inches of asphalt are needed. During mix design, the PM Device is used and determines 4.5% cement is needed to achieve 500,000 psi at full pay in place density. During construction, a representative modulus of 400,000 psi is measured with the PM Device. The agency can use this information in several ways. One option is to re-design the needed asphalt thickness with the MEPDG (7 in for this example). The agency could then, for example, require 1 in of extra asphalt at contractor cost and leave layer in place or remove layer (this is hypothetical).

Additional Work Needed

1. Several items related to on site curing and replicating field cores (e.g. black molds may get too hot if not covered from sunlight)

2. More data/study of density correction 3. Collect more data for MEPDG interfacing 4. Collect more field data, and compare to more

in-situ properties for purposes of refining QC protocol

Summary

1. PM Device offers many advantages over more traditional approaches to soil-cement design, quality control, and pavement design

2. Project team has initiated the process of working toward a national standard test method

3. Project team encourages others to give the PM Device a try – tell us what you think!

Questions?