Coefficient of Thermal Expansion of Concrete Changes to test method will enhance pavement designs by Jussara Tanesi, Gary Crawford, Jagan Gudimettla, and Ahmad Ardani

A

new analytical model for mechanistic-empirical design of pavement structures promises to provide designers with a powerful tool for optimizing the pavement design to meet desired service lives of concrete pavements. The coefficient of thermal expansion (CTE) of pavement concrete has been found to be a very important input parameter in the model, but recent work shows that model calibrations were made with data developed using an incorrect reference CTE value to calibrate the test equipment. This article provides critical background information and instructions for ensuring that future tests and analyses are conducted using correct CTE values.

AASHTO Mechanistic-Empirical Pavement Design Guide

CTE is an empirical parameter based on change in unit length per degree of temperature change. Because the rate of thermal expansion of concrete influences curling and axial stresses in pavements, the correct evaluation of concrete CTE is essential for ensuring pavement performance and serviceability. With correct CTE values and realistic analytical tools, it’s expected that pavement designers will be able to limit early-age cracking, fatigue cracking, faulting, and joint spalling.1 The Guide for Mechanistic-Empirical Design of New and Rehabilitated Pavement Structures2 was developed under National Cooperative Highway Research Program (NCHRP) Project 1-37A and was released to the public for review and evaluation in 2004. The guide applies both theory and mechanistic principles to determine structural response and predict performance over the lifetime of a pavement structure. The Turner-Fairbank Highway Research Center (TFHRC) of the Federal Highway Administration (FHWA) has been testing cores extracted from the Long Term Pavement

Performance (LTPP) test sections since 1988 and have tested over 2900 cores to date. LTPP CTE test results were used to calibrate Version 1.0 of the model incorporated in subsequent products: the Mechanistic-Empirical Pavement Design Guide (MEPDG),3 published in 2008, and the software program DARWin-ME™,4 released in 2011. The MEPDG is believed to be the first design approach to directly incorporate the CTE as an input parameter in the design of concrete pavements. In the past 5 years alone, over 20 papers have been published on the effect of the CTE on analyses made using the MEPDG,5 indicating that CTE is one of the most important inputs for pavement design. The importance of the CTE has significantly boosted interest in CTE testing. Many state highway agencies and universities are currently in the process of characterizing their states’ materials as part of MEPDG implementation activities and the results will be included in databases of material properties. Because these will eventually be used for design of pavements using the MEPDG, it’s important that the CTE measurements from various laboratories yield accurate and comparable results to avoid any under- or overestimation in designs.6

AASHTO CTE Test Method

There are several test methods for determining the CTE of concrete. The most widely used is AASHTO T 336, “Standard Test Method for the Coefficient of Thermal Expansion of Hydraulic Cement Concrete.” Based on AASHTO TP 60-00, “Provisional Test Method for the Coefficient of Thermal Expansion of Hydraulic Cement Concrete,” T 336 was approved as a standard test method in 2009. With the exception of the Texas Department of Transportation (TxDOT), which uses a modified version of T 336, T 336 is the standard test method used by all state DOTs. The principle of AASHTO T 336 is relatively simple. The Concrete international april 2012

55

Table 1:

Precision statement for CTE evaluated using AASHTO TP 606 Pooled standard deviation, μstrain/°C (μstrain/°F) Withinlaboratory CTE

 

Betweenlaboratory CTE

Precision estimates

0.135 (0.075)

0.852 (0.473)

Acceptable range of two test results (D2S)

0.38 (0.21)

2.41 (1.34)

Basis of estimate: two replicates, two materials (two concrete mixtures and 18 CTE devices)

length change is recorded as a saturated concrete specimen in a metal frame, submerged in a water bath, is subjected to a temperature change from 10 to 50°C (50 to 122°F). The deformation of the frame is taken into account by measuring the length change of a metal calibration specimen with a known CTE.

Interlaboratory Study

The FHWA Mobile Concrete Laboratory program conducted an interlaboratory study in December 2008 to obtain an understanding of the variability of CTE measurements between different laboratories and different CTE devices. The study was performed using AASHTO TP 60 procedures. Each laboratory participating in the study was instructed to use normal operating procedures to measure CTE so the calculated variability would reflect the actual variability from the different laboratories. A total of 18 laboratories participated in the study. The test equipment included custom-made units and commercially available units from a single manufacturer.6 In this study, concrete specimens were produced using two mixtures, one known to have a low CTE and the other to have a high CTE. Calibration specimens were fabricated using SAE Type 304 stainless steel (SS). The overall withinlaboratory coefficient of variation (COV) was 1.5%. The between-laboratories COV was 10.5%. When only custommade units were evaluated, the within-laboratory COV was 1.9% and between-laboratories’ COV was 11.7%. When only the commercial units were evaluated, the within-laboratory COV was 0.6% and the between-laboratories’ COV was 8.8%.6 Table 1 shows the precision statement based on the pooled standard deviations for within- and between-laboratory CTE results.

Differences between AASHTO TP 60 and T 336-11

In 2009, TFHRC evaluated the CTEs of several reference specimens fabricated from alumina, titanium alloy, and

56

april 2012 Concrete international

stainless steel. The reference specimens provided a range of CTE values that were typically within the range of the concrete specimens that have been tested previously by TFHRC. The materials and measured CTE values are listed in Table 2. The reference material specimens were tested at TFHRC using a commercially available unit as well as custom-built units, following AASHTO TP 60 procedures. As can be observed in Table 2, the CTEs obtained for the alumina bisque, titanium alloy, and Type 410 SS were about 1 × 10–6/°C (0.6 × 10–6/°F) higher than the respective CTE values reported in the literature.5 This discrepancy led FHWA to have two independent laboratories evaluate the CTE values. Lab 1 carried out CTE tests following a modified version of ASTM E228-06, “Standard Test Method for Linear Thermal Expansion of Solid Materials With a Push Rod Dilatometer.” The modifications were made to accommodate 180 mm (7 in.) long specimens with a diameter of 80 or 100 mm (3 or 4 in.) using the 10 to 50°C (50 to 122°F) temperature range as per AASHTO TP 60. Lab 2 carried out CTE tests according to ASTM E228 procedures, using 5 x 5 x 50 mm (0.2 x 0.2 x 2 in.) coupons. Per this procedure, tests performed by Lab 2 were conducted using a temperature range of −40 to 300°C (−40 to 572°F).5 The CTE results shown in Table 2 represent the CTEs over the same temperature range as AASHTO TP 60 and do not include the full temperature range used by Lab 2. As it can be observed, the results obtained at TFHRC following AASHTO TP 60 are much higher than the ones obtained at the third-party laboratories. Moreover, other than the Type 304 SS specimens, the results reported by the thirdparty laboratories are in general agreement with the values reported in the literature. For all three Type 304 SS specimens tested, however, the CTE test results were lower than the 17.3 × 10–6/°C (9.6 × 10–6/°F) value typically listed in the literature. Over the temperature range of 20 to 300°C (70 to 572°F) evaluated by Lab 2, it was verified that the CTE for Type 304 SS was indeed 17.3 × 10–6/°C (9.6 × 10–6/°F). It’s important to emphasize that the CTE value of 17.3 × 10–6/°C (9.6 × 10–6/°F) was used to calibrate the equipment by most laboratories running AASHTO TP 60 at that time, whether a custom-built or a commercial unit was used, because AASHTO TP 60 clearly states that as the value. As a consequence of these findings and additional research conducted over the past 3 years, several improvements have been made to the test method: Calibration specimen: AASHTO T 336 now requires the use of a calibration specimen with a known CTE to determine the correction factor for the equipment. The specimen should have CTE in the range of 9 to 18 × 10−6/°C (5 to 10 × 10−6/°F) within the temperature range of 10 to 50°C (50 to 122°F). A laboratory with an ISO 9001 or equivalent accreditation should determine the

••

Table 2:

CTE values reported by the literature and test results according to AASHTO TP 60 and ASTM E2285 AASHTO TP 60

ASTM E228

ID

Material

Literaturereported CTE value × 10–6/°C (× 10–6/°F)

A

Alumina bisque

5.5 (3.1)

6.5 (3.6)

6.7 (3.7)

5.4 (3.0)

—

T

Titanium alloy

9.2 (5.1)

10.0 (5.6)

10.2 (5.7)

8.9 (4.9)

9.2 (5.1)

S

‡

Type 410 SS

10.5 (5.8)

11.8 (6.6)

11.5 (6.4)

10.4 (5.8)

10.2 (5.7)

SS743

Type 304 SS – manual unit calibration specimen

17.3 (9.6)

Not applicable§

Not applicable§

15.8 (8.8)

15.9 (8.8)

M1

Type 304 SS – manufacturer 1 calibration specimen

17.3 (9.6)

Not applicable§

Not applicable§

15.9 (8.8)

15.7 (8.7)

M2

Type 304 SS – manufacturer 2 calibration specimen

17.3 (9.6)

Not applicable§

Not applicable§

16.2 (9.0)

—

Manual unit CTE × 10–6/°C (× 10–6/°F)*

Commercial unit CTE × 10–6/°C (× 10–6/°F)*

Laboratory 1 CTE × 10–6/°C (× 10–6/°F)†

Laboratory 2 CTE × 10–6/°C (× 10–6/°F)†

Average of 2 tests Single test result. ASTM E228 precision is 0.8% for the temperature range of 25 to 400°C (77 to 752°F) ‡ SAE Type 410 SS has a weak magnetic field that could affect the LVDT during tests. Preliminary evaluations did not show any effect on the CTE units at FHWA TFHRC or the FHWA mobile concrete laboratory (MCL) § There was no need to test the SAE Type 304 SS in the manual and commercial units, as the units are calibrated with the assumed CTE literature value of 17.3 × 10–6/°C (9.6 × 10–6/°F). If Type 304 SS specimens were tested, the CTE result would match the assumed CTE value of 17.3 × 10–6/°C (9.6 × 10–6/°F) *

†

••

••

CTE of the calibration specimen according to ASTM E228-06 or E289-04, “Standard Test Method for Linear Thermal Expansion of Rigid Solids with Interferometry,” within the temperature range of 10 to 50°C (50 to 122°F) and provide a certificate of the CTE value for the calibration specimen. Verification specimen: AASHTO T 336 also now requires the use of a verification specimen (other than the calibration specimen) with a known CTE used to verify that the equipment is operating properly. The specimen should have a known thermal coefficient at least 5 × 10–6/°C (2.8 × 10–6/°F) different than the calibration specimen. As with the calibration specimen, a laboratory with an ISO 9001 or equivalent accreditation should determine the CTE of the calibration specimen according to ASTM E228-06 or E289-04 within the temperature range of 10 to 50°C (50 to 122°F) and provide a certificate of the CTE value for the verification specimen. Calibration procedure: The determination of the correction factor is now included in mandatory section of the test method.

verification: The electronic components •• Equipment in linear variable differential transformers (LVDTs) can

••

••

be adversely affected by high temperatures and contact with water. To verify the proper functioning of the LVDT and other parts of the CTE equipment, verification testing using a specimen of known CTE is required at least every month. If the CTE found in the verification test differs more than 0.3 × 10–6/°C (0.2 × 10–6/°F) from the certified value, a new frame correction factor should be determined. This change has significantly reduced the potential error associated with an LVDT that is out of calibration. LVDT zeroing: The appendix of AASHTO TP 60 stated that repositioning of the LVDT was not recommended. However, AASHTO T 336 clearly specifies that the LVDT must be adjusted so that its core is located in its midpoint or electrical zero reading before testing. This change has significantly reduced the error associated with testing concrete specimens with heights that are within ±2.5 mm (0.1 in.) of the calibration specimen. LVDT calibration: The current AASHTO T 336 requires the LVDT to be calibrated at least every 6 months. Concrete international april 2012

57

bath temperature: AASHTO T 336 requires the •• Water water bath temperature to be verified every time a

•• ••

•• ••

verification specimen is tested. Number of sensors: The number of temperature sensors in the water bath has been decreased from four to one. Water level: The water level in the controlled temperature water bath has an effect on the CTE because variations in water depth affect the length of the frame and the length of the LVDT shaft that are exposed to the ambient air. AASHTO T 336 therefore requires the water level during testing to be the same as the water level during equipment calibration. Specimens: The end condition of the concrete specimen may be a source of some test error. AASHTO T 336 requires the specimen ends to be flat and parallel, according to tolerances specified in AASHTO T 22-07, “Standard Method of Test for Compressive Strength of Cylindrical Concrete Specimens.” Number of specimens: AASHTO T 336 requires at least two specimens from each mixture be tested.

Ruggedness Test

FHWA took the lead in initiating a ruggedness study, which is scheduled to be completed by March 2012. The ruggedness test was designed according to ASTM C1067-00, “Standard Practice for Conducting A Ruggedness or Screening Program for Test Methods for Construction Materials,” and is based on AASHTO T 336-11. Two different commercially available manufacturers were included in the study, two different concrete mixtures are being evaluated, and four laboratories are participating. The ruggedness test includes seven factors at two levels (Table 3). The results obtained from the ruggedness test may point to more required changes to AASHTO T 336.

DARWin-ME Model Recalibration

Models in Version 1.0 of the MEPDG as well as the DARWin-ME software were calibrated using the LTPP database of CTE values obtained according to AASHTO TP 60. Because the values obtained according to AASHTO TP 60 and T 336 may differ considerably depending on the calibration specimen and its assumed CTE value, the CTE obtained according to AASHTO T 336 should not be used as an input in DARWin-ME, to avoid an inappropriate pavement thickness based on biased CTE data. As discussed previously, CTE is an important parameter in concrete pavement analysis and design because it’s directly proportional to the magnitude of temperaturerelated pavement deformations throughout the pavement design life. Several studies in the past few years have identified CTE as one of the most significant inputs in the MEPDG for designing rigid pavements. Aside from the improvements mentioned to reduce the testing error, the AASHTO TP 60 and T 336 test methods

58

april 2012 Concrete international

Table 3:

Ruggedness test factors Variable

Level 1

Level 2

Time at temperature

AASHTO T 336 requirement

2 hours after AASHTO T 336

Water level

Same level

5 mm (0.2 in.) above

Position of the LVDT gauge head

Aggregate

Paste

Number of segments for test

Necessary for AASHTO T 336

2 extra after AASHTO T 336

Degree of saturation

AASHTO T 336 requirement

28 days in water

Specimen length, mm (in.)

175 (6.9)

178 (7.0)

Starting temperature, °C (°F)

10 (50)

50 (122)

are very similar. The primary difference is how the CTE of the calibration specimen is determined. Starting with Standard Data Release (SDR) 24 of the LTPP database, the CTE data have been updated based on CTE values obtained using a corrected procedure for calibrating the CTE frames. Test results obtained using AASHTO T 336 can be adjusted for use in the MEPDG or in the DARWin-ME software by using the following equation in U.S. customary units CTE input = CTEpcc – CTEcalib + 9.61 × 10 –6/oF where CTEpcc is the CTE of a concrete specimen according to AASHTO T 336 or TP 60; and CTEcalib is the CTE of a Type 304 SS calibration specimen used to determine CTEpcc. It must be noted that CTE should be adjusted to reflect the CTE values used to calibrate the current models in the DARWin-ME software. CTE values listed in the LTPP SDR 24, dated January 2010 or later, have been updated with this correction. These CTE values should be increased by approximately 0.83 × 10–6/°F if they are used as a Level 3 input (input based on historical data). CTE values published in database releases prior to LTPP SDR 24 do not need to be adjusted. In October 2011, NCHRP completed Project 20-07/ Task 288, “Recalibration of DARWin-ME Rigid Pavement National Models Based on Corrected CTE Values,” to update the models based on the revised CTE values in the LTPP database as well as the lower CTE values, which result from using AASHTO T 336 compared to TP 60. A follow-up NCHRP project has been approved for 2012 funding to conduct an objective evaluation of the recalibration effort conducted in Project 20-07/Task 288, and recommend recalibrated models that use accurate Level 1 CTE values

Fig. 1: The CTE test system produced by Gilson Company, Inc.

(values measured directly) for future incorporation into the DARWin-ME program.

Quality Assurance

With the recent release of DARWin-ME pavement design software, there will be a greater emphasis on using CTE of concrete for pavement design. Since its release in April 2011, DOTs in 25 states and transportation departments in five Canadian provinces have licensed the software. Because CTE is an important element for pavement design, there is also interest in using CTE as a quality assurance test. The California DOT is the first to adopt this practice. Their pavement specifications require the contractor to test CTE from production on a daily basis for both the continuously reinforced concrete pavement (CRCP) and jointed plain concrete pavement (JPCP) sections.7 The specification for CRCP also uses a maximum CTE value, which is based on the pavement design inputs. Specifying a maximum allowable value of CTE can help prevent problems such as increased transverse cracking for JPCP due to higher-than- designed curling stresses.6 While concrete mixtures with high CTE values can provide the same level of performance as mixtures with lower CTE values, these differences need to be accounted for during the design phase and the materials selected for the project should not result in mixtures that have CTE values exceeding those used to determine the pavement design.

Concluding Remarks

As noted previously, there has been a tremendous amount of CTE research since the first version of AASHTO T 336 was published. FHWA has also conducted a lot of research during this timeframe to improve the test method and reduce its variability. FHWA plans to conduct another interlaboratory study and develop a precision statement for AASHTO T 336 in 2012, once the ruggedness study has been completed.

Fig. 2: The CTE test system produced by Pine Instruments

Currently, 11 state DOTs, 13 universities, four commercial testing laboratories, two industry laboratories, and FHWA have acquired at least one CTE device. At present, two automated CTE devices are commercially available on the market (Fig. 1 and 2). A TechBrief on CTE, FHWA-HIF-09-015 (www.fhwa.dot.gov/pavement/concrete/pubs/hif09015/) has also been published by FHWA. The TechBrief provides an overview of the CTE test method, the role of CTE on pavement performance, average CTE values based on the primary coarse aggregate from the LTPP database, and the sensitivity to variations in CTE observed with the AASHTO MEPDG model or the DARWin-ME pavement design software program. References 1. Mallela, J.; Abbas, A.; Harman, T.; Rao, C.; Liu, R.; and Darter, M. I., “Measurement and Significance of Coefficient of Thermal Expansion of Concrete in Rigid Pavement Design,” Transportation Research Record: Journal of the Transportation Research Board, No. 1919, 2005, pp. 38-46. 2. Guide for Mechanistic-Empirical Design of New and Rehabilitated Pavement Structures, Final Report for Project 1-37A, National Cooperative Highway Research Program, Transportation Research Board, National Research Council, Washington, DC, Mar. 2004. (www.trb.org/mepdg) Concrete international april 2012

59

3. Mechanistic-Empirical Pavement Design Guide, Interim Edition: A Manual of Practice, AASHTO, Washington, DC, July 2008, 212 pp. 4. DARWin-ME™ AASHTOWare® Pavement Design and Analysis System, AASHTO, Washington, DC, 2011. 5. Tanesi, J.; Crawford, G.; Nicolaescu, M.; Meininger, R.; and Gudimettla, J., “New AASHTO 336-09 Coefficient of Thermal Expansion Test Method: How Will It Affect You?” Transportation Research Record: Journal of the Transportation Research Board, No. 2164, 2010, pp. 52-57. 6. Crawford, G.; Gudimettla, J.; and Tanesi, J., “Interlaboratory Study on Measuring Coefficient of Thermal Expansion of Concrete,” Transportation Research Record: Journal of the Transportation Research Board, No. 2164, 2010, pp. 58-65. 7. State of California, Department of Transportation, “Special Provisions for Construction of State Highway in Placer and Nevada Counties in District 3 and Route 80,” Addendum 4, Oct. 2009. (www. dot.ca.gov/hq/esc/oe/project_ads_addenda/03/03-2C8604/) Note: Additional information on the AASHTO and ASTM International standards discussed in this article can be found at www.transportation.org and www.astm.org, respectively. Selected for reader interest by the editors.

ACI member Jussara Tanesi, currently with Global Consulting, is a contract Project Manager to the FHWA at Turner Fairbank Highway Research Center, McLean, VA. She is a member of ACI Committees 211, Proportioning Concrete Mixtures; 231, Properties of Concrete at Early Ages; 236, Material Science of Concrete; 238, Workability of Fresh Concrete; and 325, Concrete Pavements; and a director of the ACI National Capital Chapter. She is a member-at-large of ASTM Committee C09, Cement and Concrete Aggregates, and a member of ASTM Committee C01, Cement. She received her PhD and her BS in civil engineering from the University of Campinas, Brazil, and her MS in civil engineering from the University of São Paulo, Brazil. Gary Crawford is a Concrete Pavement Engineer with the FHWA in the Office of Pavement Technology’s Concrete Group, Washington, DC. He joined FHWA in 1983 and has spent the last 29 years promoting new technologies dealing with condition survey techniques for concrete structures, polymer concrete applications, and the use of NDT equipment for concrete. Recently, he has been involved with implementation activities involving the Mechanistic-Empirical Pavement Design Guide and managing the FHWA’s mobile concrete laboratory. He received his BS in civil engineering from Geneva College, Beaver Falls, PA. Jagan Gudimettla currently works as the Project Engineer for the FHWA’s Mobile Concrete Laboratory, Washington, DC. His areas of interest include QA/ QC, advanced material testing, NDT, and pavement design. He received his MS in pavements/materials from Auburn University, Auburn, AL, and is a licensed professional engineer in California.

www.NDTjames.com • email:[email protected]

www.NDTjames.eu • email:[email protected]

3727 N. Kedzie Ave., Chicago, IL 60618

Windmolen 22, 7609 NN Almelo, The Netherlands

800-426-6500 • 773-463-6565

60

+31 (0)548 659032

april 2012 Concrete international

ACI member Ahmad Ardani is a Research Engineer working for FHWA, conducting research and managing the Concrete Laboratories at Turner Fairbank Highway Research Center, McLean, VA. Prior to working at FHWA, he worked as Program Manager for the Research Branch of the Colorado DOT, heading up the Pavement and Geotechnical and Safety Research program. He is a licensed professional engineer in Colorado.