A lot of words have been spoken about the laboratory testing for Superpave mixtures. This discussion will be about the practical issues involved with making Superpave mixes. This paper will cover the production of aggregate. The effects of production on the consensus properties of the aggregate and on mix properties. The paper will close will examples of the effects of different aggregate sources on mix properties, construction properties and pavement performance.
Superpave mixes require mixes which are usually coarser than traditional HMA mixtures. The aggregates needed to make these mixes need to be clean and uniformly graded. Enough screening capacity is needed in the crushing plant so that the screens can efficiently separate the crushed rock into separate sizes or product blends. The screens need to have a consistent rock load or depth of rock going over the screen deck for consistent gradations. Overloaded screens reduce production and increase product variability.
The loader operator is key to both stockpiling and loading. A well trained loader operator will be able to spot material problems before the aggregate is loaded into trucks or into the cold feed bins of an asphalt plant. A well trained loader operator will work the whole face of a stockpile so that material is blended and the effects of segregation are minimized during loading or stockpiling operations. The loader operator is also a traffic director, positioning trucks so that they can be loaded efficiently in a manner to reduce segregation as discussed previously.
Good sampling technique is essential for monitoring your aggregate stockpiles. Fortunately, most stockpile problems are caused by poor sampling techniques than by inconsistent material. To get an accurate sample from a stockpile a sampling board is essential. The board allows the technician to dig through the surface of the stockpile and get a representative sample of the underlying aggregate without segregated surface aggregate being mixed into the sample. Multiple samples need to be taken from the stockpile at a minimum of three random locations, keeping in mind that roughly 2/3 of the stockpile is contained in the bottom 1/3 of the pile. Remember, test results are only as good as the sample.
The effects of PG grading will be discussed by others. The characteristics of the aggregate are controlled by rock type and processing. Shape of the aggregate as previously discussed will be controlled in part by the crushing operation. Shape is important from the aggregate and HMA specification points of view. The concensus specifications limit the amount of flat and elongated particles. However, from a mix design point of view very hard aggregates such as traprock or basalt are desirable because they tend to contain elongated particles which increase the fine aggregate angularity, FAA and the mix VMA. One hundred percent crushed and cubical limestone sands produce FAA values in the low forties.
What does this mean? The concensus specifications for FAA are the consensus of a group of experts based on data available, which was mainly on natural sand FAA results. These results indicated that on average that a 45 FAA indicated a cubical natural sand. However, looking at the crushed limestone sand of 43 FAA under a microscope indicates that the particles are angular, cubical and pack well. These are the characteristics that were determined by the Aggregate ETG to be important sand attributes needed to make stable HMA mixes. What we have done in Texas is to note that the limestone manufactured sand particles are angular, cubical and thus meet the intent of the FAA specification even though the sand has a 43 FAA value. Further, The gyratory compaction slope for the Superpave mix was 10. This slope indicates that the mix has high shear resistance. This information was documented in the mix design report. The Superpave mix design with the exceptions documented was reviewed and approved by TxDOT. Then we built the road.
Five Superpave test sections later, the pavement performance indicates that a 43 FAA works for that particular quarry. I would suspect that all crushed limestone sands that meet the other aggregate specifications, look cubical and angular under the microscope and produce Superpave mixes with a high gyratory compaction slope will also provide excellent field performance. So know the philosophy behind the consensus specifications. The intend of the specification is as important as the consensus values. Thus, know the rock and document what has been done.
Another area of concern is the mix VMA. Mix design results indicate that limestone on average produce mixes with lower VMA than mixes made with basalt. This has been reported by Kandhal of NCAT as the result of the limestone aggregate crushing during compaction. Also, mixes with elongated aggregate particles produce mixes with higher VMA's than those with more cubical shapes. I think the answer to the VMA question will be answered when the NCHRP 9-7 project is completed and the results are published. NCHRP 9-7 was a study to determine the sensitivity of the mix specifications.
A study by TxDOT has determined that more consistent density and thus VMA values can be determined by using the Rice specific gravity's of lab compacted molds to calculate the effective specific gravity of a mix. Three test sections were placed on the service road for Loop 1604 in San Antonio for the NCHRP 9-7 project. What is interesting to note is that the VMA values chosen for the sensitivity study, approximately 13.6% versus the minimum Superpave specification of 14%, would have passed the existing TxDOT specification which allows rounding to the closest whole number. By the TxDOT procedure the VMA was calculated to be 0.6% higher. Two SPS-9A test sections placed on the main lanes of Loop 1604 were designed to have 14.3% by Superpave and 15% by TxDOT VMA values. Both sets of test sections have exhibited excellent performance over the last two very hot summers.
Another Superpave test section was placed last fall on Eckhert Road in San Antonio. This was a rebuild project to expand a two lane road to a four lane primary urban road. For extra durability basalt aggregate was chosen for the coarse fraction. Using a gradation similar to that used for the Loop 1604 test section a Superpave VMA of 15.5% was achieved with the basalt coarse aggregate. For all the San Antonio test sections 19mm mixes were used with a Ndsn=106.
An unpublished study comparing 19mm mixes made from 100% non-Texas limestone mixes to those made with 100% basalt indicated that the basalt mixes had VMA of 19% at Ndsn=128 and VMA of 17.8 at Nmax=208 using the same basalt used in the San Antonio test sections. The design asphalt content for the 100% basalt mixes was 5.4% compared to 5.6% for San Antonio mix which contained 20% basalt. These results are interesting from two perspectives.
The aggregates just described all meet the Consensus aggregates properties for shape and have a 100% crush count. However, the basalt aggregate fractions tend to be much more elongated than the corresponding limestone aggregate fractions which tend to be more cubical. The basalt is twice as hard as the limestone aggregate as measured by Mohs hardness and L. A. Abrasion loss. Is the higher basalt mix VMA due the aggregate particle shape or from the resistance to crushing during mixing and compaction?
An additional question, are these differences important? These are extremely important questions for the aggregate producer. On one hand, the results would indicate that a change in crushing process would improve compliance with the Superpave mix and consensus aggregate specifications. On the other, the results would also indicate to get the desired performance that the aggregate source should be changed. Changing the crushing process is many times more cost effective then importing aggregates across state lines.
The aggregate consensus specifications are that, specifications put together by a committee with working experiences using soft and hard aggregates in different parts of the country. As such, the consensus specifications are an excellent guideline to important aggregate characteristics. However, the consensus specification are not a guarantee of an excellent performing pavement. The determining test of success remains as how well is the mix performing under traffic? The practical side of the Superpave aggregate specifications are the Source Aggregate Properties. These are the test results that the committee could not agree on what average or typical values should be used. What does this all mean?
Superpave is a step forward in asphalt mix design with the goal to be able to predict pavement performance from laboratory tests. The framework has been developed. Now is the time to start filling in the details. Some of the details are hard to determine such as what is the appropriate Superpave Shear Tester (SST) result to predict the pavement distress from a given traffic level. The correlation factors in engineering units will be determined as more and more Superpave projects are built and subject to performance laboratory tests and monitored for field performance. Then engineers will have a rational method for pavement design.
This new design process had to start somewhere. So we have Consensus and Source specifications. These are baby steps in the overall picture. One must learn to crawl before walking, and walk before running. So hopefully, this discussion will help the process of learning that nothing is absolute in Superpave, ........except for the Binder specifications (note: that is supposed to be a joke). That the Superpave design procedure without performance testing is just the framework, the foundation that needs to be established before the performance testing standards are established. West Track is excellent example of what I'm trying to say. Aggregates which were used to develop the SHRP Superpave specification weren't used because they did not meet all the consensus specifications.
However, test sections paved with mixes made with aggregates which met all the Superpave aggregate and mix specifications have failed. That is good for a couple of reasons. One engineers and researchers learn from their mistakes, if nothing breaks no progress is made on design improvement. The other point is that the current state of the specification is that the current Superpave mix design is anything but a cookbook procedure. As briefly described in these Texas projects, use all the information available, check the Superpave test results against local experiences, use common sense to evaluate the results, document how decisions were made, then build the pavement and monitor the performance.