  |
   |
 |
 |
 |
 |
 |
 |
 |
 |
 |
July/August 2002
Taking Concrete to the Next Level
by Marcia J. Simon and Michael P. DallaireAre you building a long-lasting, effective road? Two elements contribute to the performance of concrete and the lifetime of pavements the concrete mixture itself and the construction practices engineers use to mix, place, and cure the concrete. The "perfect" mixture combined with bad practices can produce an inferior result with problems such as cracking, poor concrete durability, and premature deterioration. Likewise, a poor mixture coupled with good practices also can produce a substandard result. In either case, the outcome will be additional repair or replacement costs, user delays, and more congestion.
Turner-Fairbank Highway Research Center (above).
The Federal Highway Administration's (FHWA) Portland Cement Concrete Pavement (PCCP) team at the Turner-Fairbank Highway Research Center in McLean, VA, is studying various aspects of concrete performance, including fresh (before hardening) and hardened concrete properties. Questions they are addressing in the lab include:
- Can statistical optimization methods be used to identify the appropriate concrete mixture for projects with multiple performance requirements?
- Do the standard "rules of thumb" for an adequate air void system apply to all concretes exposed to freezing and thawing conditions?
- Is there a fast and reliable way to evaluate the susceptibility of concrete mixtures to expansion caused by alkali-silica reaction?
- How can a design engineer accurately measure the coefficient of thermal expansion for concrete?
- What better test in the field might be used in place of slump for testing workability of low-slump paving concrete?
These studies have a common emphasis—a focus on improving concrete, through better material selection and proportioning, and improving lab and field test performance predictions.
Mixture Optimization Using Statistical Methods
The advent of high-performance concrete made the process of proportioning concrete mixtures more complex. Using the right mix proportions for each project is extremely important, and designers must deal with several constraints, such as:
- Component material availability and properties varying widely throughout the country
- Climate and design constraints leading to different performance requirements for each project
- Material prices continually increasing
- Multiple engineering performance requirements (in addition to strength) needing to be met simultaneously
A means of optimizing concrete mixtures (meeting several performance requirements while minimizing costs) could result in material cost savings and more confidence that a concrete mixture will meet specifications. For example, reducing the cost of concrete by $10 per yard could yield savings of $10,000 per 765 cubic meters (1,000 cubic yards) of concrete, or about 0.8 lane-kilometer (0.5 lane-mile) of 30.5-centimeter (12-inch) concrete pavement.
Impact hammer hits the concrete to measure the internal damage in the concrete due to freezing and thawing.
Although the American Concrete Institute's (ACI) 211 Guide for Proportioning Concrete Mixtures and other procedures are good starting points for concrete proportioning, they do not provide information on the optimal proportions for meeting several performance criteria at the same time. As a result, engineers normally consider one factor at a time using a trial-and-error approach, which can be inefficient and costly, and may not produce the best combination of materials.
To replace the inefficient trial-and-error method, FHWA looked outside the transportation industry for successful optimization approaches. Response surface methods (RSMs) are a set of statistical tools that other industries use for optimizing products such as gasoline, foods, and detergents. Since these products, like concrete, are blends of various component materials, the PCCP team and researchers and statisticians from the National Institute of Standards and Technology investigated whether RSM also might be applicable to concrete. The objectives of this study were to determine whether statistical optimization methods are applicable to concrete mixture proportioning and, if so, to develop an interactive Web site tool to enable prospective users to learn about and try this approach.
In the initial portion of the project, the team performed laboratory experiments to evaluate two possible approaches: a classical mixture design and a central composite design (CCD). The materials used in both cases were Type I cement, silica fume, high-range water reducer, natural sand, and crushed limestone. The team optimized the following properties: 1-day and 28-day strength, slump, and resistance to chloride penetration as indicated by the American Association of State Highway and Transportation Officials' (AASHTO) T277 test (rapid chloride test). The experiment design (classical mixture or CCD) defined the mixture proportions for the lab experiments. Upon test completion, researchers analyzed the data, estimated models using linear regression, and employed graphical techniques, such as means plots, scatter plots, and contour plots to interpret the results.
Based on the experimental results, both classical mixture and CCD can be used in concrete mixture proportioning. However, because the CCD approach is more straightforward, the research team chose it for Web site development.
The aim of the interactive Web site, Concrete Optimization Software Tool (COST), is to introduce the procedures for using the statistically based (CCD) approach and to guide the user step-by-step through the planning and execution of a set of trial batches, and analysis of the results. The optimization procedure in COST involves six steps:
- Steps 1 and 2: The user inputs information about materials and performance requirements, and COST generates mixture proportions for a set of trial batches (according to the CCD design).
- Step 3: The user batches the concrete, fabricates specimens, and tests them.
- Step 4: The user enters the results into COST.
- Step 5: COST performs several analyses.
- Step 6: COST provides a summary of the analysis and optimal settings.
Because potential COST users may not have a statistical background, COST depicts many of the analysis results graphically and provides some guidance on interpretation. A user's guide is available online and will be published shortly. The final report for the project is slated for publication in September 2002.
Freeze-Thaw Resistance Of Concrete with Marginal Air Contents
The concrete industry has known the benefits of air entrainment (intentionally incorporating microscopic air bubbles into a concrete mixture) for more than 60 years. An adequate entrained air void system is considered necessary to prevent damage from freezing and thawing cycles, which ultimately affects the durability of concrete. The commonly accepted rules of thumb for a good air void system include: air content of 6 percent generated by a chemical admixture during mixing, specific surface greater than 600 in2/in3, and a spacing factor between air voids less than 0.02 centimeters (0.008 inches). However, some studies show that concretes with marginal air void systems (not meeting the accepted criteria) may still be resistant to freezing and thawing. Also, there is some debate in the industry about the need for air entrainment in concretes with sufficiently low water-cement ratios. The objectives of this study are to investigate the freeze-thaw durability of concrete with marginal air void systems and to suggest possible improvements to freeze-thaw testing procedures.
This screenshot shows the COST homepage at http://ciks.cbt.nist.gov/cost. COST is an online design-analysis system to assist concrete producers, engineers, and researchers in determining optimal mixture proportions for concrete.
Freeze-thaw testing by AASHTO T-161 and the American Society for Testing and Materials (ASTM) C666 exposes concrete to a temperature cycle of 54 to 32 to 54 degrees Celsius (40 to 0 to 40 degrees Fahrenheit) in 2 to 5 hours. Researchers remove specimens from the freeze-thaw machine periodically (every 10 to 30 cycles depending on condition) and assess them for visible changes, internal damage, and changes in mass. Internal damage is evaluated nondestructively using the impact method (ASTM C 215) to determine the resonant frequency of the test beam. Relative dynamic modulus (a measure of internal damage) and durability factor are calculated from the resonant frequency.
There are presently two accepted variations of the freezing and thawing procedure. Procedure A submerges the test beams in water during freezing and thawing. This procedure is not representative of field conditions due to confinement of containers used and is considered more severe than field conditions, where periodic drying can take place. Procedure B exposes test beams to air during freezing and submerges the beams in water during the thawing cycle. This method is more realistic, but concerns were raised about the possibility of excessive specimen drying during the freezing cycle. In response to these concerns, the Strategic Highway Research Program (SHRP) proposed a variation researchers wrapped test specimens in removable terry cloth covers and tested them according to Procedure B. The terry cloth protects the concrete surface from excessive drying.
Researchers in the SHRP project also identified the quality factor as a potential parameter for assessing damage to concrete specimens during testing. They found that using the quality factor as a damage measure might enable researchers to predict the freeze-thaw test result in significantly less time (100 to 150 cycles). In this study, the PCCP team is looking at the validity of the quality factor approach, which could reduce the cost and time required for freeze-thaw testing significantly.
Freeze-thaw specimens in the testing machine are labeled according to the procedure they are undergoing (Procedure A specimens are in metal containers, Procedure B no container or cover, and SHRP variation terry cloth covers).
Concretes mixtures with marginal air contents are being tested using standard AASHTO-ASTM procedures and the terry cloth procedure. Researchers are calculating the quality factor in some of the experiments to evaluate its potential for predicting the durability of concrete before 300 cycles. The experimental program has several phases.
Test Phases
Procedure A: Test beams submerged in water during freezing and thawing
Procedure B: Test beams submerged in water during freezing and air during thawing
Terry Cloth Variation: Test beams are wrapped in removable terry cloth covers and tested according to Procedure BThe first phase of research tested concretes with fresh air content levels of 2.7, 2.9, and 3.1 percent using the A, B, and terry cloth test procedures. For each air content level and test procedure, researchers tested five beams. The findings from the terry cloth procedure in most cases indicate a severity level equal to or greater than the cases tested using procedure A. There appeared to be less variability between results from procedures A and B. Mass loss was considerably greater in procedure A due to scaling.
Currently in progress, the second phase involves testing concretes with a wider air content range from 2.5 to 4.5 percent. Other variables in this phase include water-cement ratio (ranging from 0.40 to 0.50) and two different types of air entraining admixtures (AEA)vinsol resin and a synthetic AEA. In addition to performing freeze-thaw testing, the researchers will determine hardened air void system parameters for each mix and calculate and evaluate the quality factors.
Mixture-Specific Method For Prediction of ASR Expansion
Alkali-silica reactivity (ASR) occurs when alkalies in cement react with certain types of silica in concrete aggregates to form an alkali-silica gel. This expansive gel can absorb water and expand, causing pavements to crack and eventually fail. Currently, there is no rapid test method claiming to evaluate the ASR susceptibility of concrete mixtures. The ASTM C 1260 (mortar bar) test method is specified as a test to assess aggregates and not combinations of aggregates and cementitious materials (although some researchers have investigated its use for that purpose). The concrete prism test developed in Canada (ASTM C1293) is more reliable, in that it tests concrete rather than mortar, but the mixture proportions for the concrete are prescribed. Another drawback of the prism test is that it takes a year or more to perform. Other methods have been suggested or tried, but are not recommended because of limited data. The objective of this study is to identify a fast, dependable test method for assessing the ASR potential of a concrete mixture.
The approach in this study is to perform multiple prism tests varying the water-cement ratios (w/c), cement content, fly ash content, and lithium dosage for a given aggregate. A factorial experiment design will be used. Concrete prism tests will be performed at 38 degrees Celsius (100.4 degrees Fahrenheit) for 1 year (standard ASTM C1293 conditions) and at 60 degrees Celsius (140 degrees Fahrenheit) for 3 months. Tests also will be performed at 60 degrees Celsius using modified prisms with small longitudinal holes developed at the University of New Hampshire. The test results will be used to estimate a predictive model that can be used to predict expected ASR expansion anywhere within the ranges used for w/c, cement, fly ash, and lithium. Testing began in July 2002.
Thermal Coefficient of Concrete Cores from LTPP Sections
Researchers at FHWA have developed a precise method for determining concrete's coefficient of thermal expansion (CTE), a property that affects pavement design and subsequent performance. The CTE is an important element of FHWA's Long-Term Pavement Performance (LTPP) database and a critical element in the new AASHTO 2002 Design Guide for concrete pavements. The information generated by this precise method will lead to longer-lasting, smoother roads and will improve the pavement design process by better matching pavement to its environment. Before FHWA's new testing procedure, pavement designers usually assumed an average value of the concrete CTE in their designs.
A literature review revealed that several methods were developed over the years, but none were widely used within the transportation industry. Turner-Fairbank Highway Research Center (TFHRC) researchers were asked to develop a standard test for measuring CTE and to use that test to measure the CTE on cores from LTPP pavement test sites. Desired characteristics of the test method were accuracy, repeatability, ease of use, and relatively low cost.
Calculation of CTE
The CTE is calculated using the above equation: where DDL = length change (cm), L0 = initial specimen length (cm), and DT = temperature change (°C). Of the three required measurements, the most difficult to obtain is specimen length change, because it is very small (on the order of hundred-thousandths of inches). Linear variable differential transformers (LVDTs) were identified as appropriate devices for measuring this length change.
Another issue involved the moisture condition of test specimens during testing. The thermal coefficient of concrete varies with internal relative humidity (RH), with the maximum value usually occurring at 60 to 70 percent RH. The value at 100 percent RH is 20 to 25 percent less than the maximum. However, the fully saturated condition is the most practical from a testing standpoint. Furthermore, pavements in the field have an internal RH of 80 percent or more, except the region near the surface.
The test frame developed by FHWA features vertical support rods made of Invar, a very low-CTE steel, to minimize frame expansion during heating. The LVDT is centered above the specimen, which is supported by three hemispherical points. The frame is placed in a water bath to submerge the test specimen completely (but not the LVDT). The frames are calibrated (to correct for any frame expansion) using 10-centimeter diameter by 18-centimeter long (4-inch diameter by 7-inch long) A304 stainless steel cylinders.
The final test method includes the following steps: sawing the cores to a standard length (178 mm/7 in.), grinding the ends parallel, soaking the cores to reach SSD condition, mounting the core in a measuring frame with an LVDT, placing the frame in a controlled temperature water bath, varying the temperature of the bath between 10 and 50 degrees Celsius (50122 degrees Fahrenheit), and reading the length change over that temperature range.
AASHTO recently approved the new test method as a provisional test method TP60-00, Standard Test Method for the Coefficient of Thermal Expansion of Hydraulic Cement Concrete. It is included in the 2000 edition of the AASHTO Provisional Standards. The pavement team researchers are using the test in-house to measure the thermal coefficient of concrete pavement cores from around the country. The LTPP program provided about 2,500 cores for CTE testing, and tested about 600 cores on a continuing basis, with an average of 40 to 50 tests a month.
Workability of Paving Concretes
The slump test is the most widely used test for assessing concrete workability. However, slump is, at best, a tool for monitoring production variation from batch to batch. Several other test methods and devices were developed to characterize the rheological properties of concrete, but these devices will not work with the low-slump mixtures typical of paving concrete. Under an interagency agreement, the U.S. Army Corps of Engineers recently developed a workability-measuring device for FHWA. This device, called the vibrating slope apparatus (VSA), quantifies workability from measurements of mass loss as a function of time from an inclined chute at a given level of vibrator energy.
The thermal coefficient test method developed at TFHRC uses an Invar frame, an LVDT, and a 10-centimeter by 18-centimeter (4-inch by 7-inch) concrete specimen. The frame is submerged in a water bath during the test.
FHWA is performing a follow-up study on the VSA with the following objectives:
- To evaluate the operation of the VSA and the validity and interpretation of the workability parameter it provides
- To modify the device and procedure as needed, based on the results of the evaluation
- To use the VSA on a range of concrete paving mixtures to determine which factors have a significant influence on workability of paving concrete
PCCP team researchers evaluated the prototype VSA and tested it for operational characteristics, ruggedness, and ease-of-use. They also performed an initial assessment of the validity of the test result and its interpretation. Based on this evaluation, several design modifications were recommended, and three new VSAs were built using the revised design. Two of these VSAs will be taken to the field with the FHWA Mobile Concrete Laboratory for use on actual paving projects. The remaining VSA will stay at TFHRC for the next phase of laboratory research beginning in July 2002. This follow-up study will include assessment of test factors (concrete slump, chute angle, and vibration force) and evaluation of aggregate effects on workability.
Prototype version of the Vibrating Slope Apparatus (VSA) in operation. References
1. Myers, R.H. & D.C. Montgomery, Response Surface Methodology: Process and Product Optimization Using Designed Experiments. New York: Wiley, 1995.
2. ACI Committee 211, "Standard Practice for Selecting Proportions for Normal, Heavyweight, and Mass Concrete." In ACI Manual of Concrete Practice, Volume 1. Detroit: American Concrete Institute, 1995.
3. Simon, M.J., Lagergren, E.S. and K.A. Snyder, "Concrete Mixture Optimization Using Statistical Mixture Design Methods." In Proceedings of the PCI/FHWA International Symposium on High Performance Concrete, New Orleans, October, 1997, pp. 230-244.
4. Simon, M.J., Lagergren, E.S. and L.G. Wathne. "Optimizing HPC Mixtures Using Statistical Response Surface Methods." In Proceedings of the 5th International Symposium on Utilization of High Strength/High-Performance Concrete. Norwegian Concrete Association, Oslo, Norway, June, 1999, pp. 1311-1321.
5. Janssen, Donald J. and M.B. Snyder, Resistance of Concrete to Freezing and Thawing. Strategic Highway Research Program Report #SHRP-C-391. National Research Council, Washington, D.C. 1994.
6. Touma, W.E., Fowler, D.W., and R. L. Carrasquillo, Alkali-Silica Reaction in Portland Cement Concrete: Testing Methods and Mitigation Alternatives. Research Report ICAR 301-1F, International Center for Aggregates Research, Austin, Texas, 2001.
7. AASHTO TP-60-00, "Proposed Standard Test Method for the Coefficient of Thermal Expansion of Hydraulic Cement Concrete." American Association of State Highway and Transportation Officials, 2001.
8. Wong, G.S, et al, Portland Cement Concrete Rheology and Workability: Final Report. FHWA-RD-00-025, Federal Highway Administration, McLean, Virginia, 2001.Marcia Simon is a research materials engineer in the Office of Infrastructure R&D at the TFHRC. For the past 12 years, she has been involved in research on concrete materials and recycling. Her primary research areas are concrete mixture optimization and concrete performance, especially durability. She received a B.S. from MIT and an M.S. from Cornell University, and she is a registered professional engineer in Virginia.
Michael P. Dallaire is currently a concrete materials engineer with SaLUT onsite at TFHRC and oversees concrete laboratory operations. He is a graduate of the University of New Hampshire with a master's degree in civil engineering. His interests include the durability of high-performance concrete for pavements and structures with specific interests in cracking and shrinkage performance.
Other Articles in this issue:
Taking Concrete to the Next Level
Fine-Tuning Innovative Technologies
Making Roads Better and Better
Texas Tests Precast for Speed and Usability
The Biggest Bang for Your Buck
New Software Promises to Put Whitetopping on the Map
