It seems like you hear about it everywhere. People are talking about noise. People living near heavily traveled roadways are becoming more vocal in their complaints about traffic noise, and they are demanding reductions. Average citizens are starting to notice that some pavements are quieter than others—and they want those pavements near their homes.
Though it may come as a surprise to some, tire/pavement noise has been shown to be a major contributor to the overall traffic noise level. Traffic noise comes from two major sources, the power train and the tires. Tire/pavement noise dominates at speeds over about 19 to 31 mph for cars and 25 to 43 mph for trucks. The noise-generating mechanisms include vibrations of the tire, air resonance and other mechanisms.
To mitigate noise complaints, barrier walls are often erected, but they can cost well over a million dollars per mile and have limited effectiveness. Experience and evidence is mounting to show that noise can be better and more economically controlled at the source by designing quieter pavement surfaces. European experience with porous asphalt pavements, for example, shows that these pavement surfaces are among those that can reduce noise generation. In addition, these asphalt surfaces can offer benefits in terms of improved rut resistance, better wet-weather traction and reduced splash and spray.
The structure of a porous asphalt surface contains interconnected voids, which help to drain away rainwater during wet weather. The porous structure also can reduce tire/pavement noise by interfering with some noise generation mechanisms. Porous pavements also have demonstrated good surface friction and decreased splash and spray when it rains.
Indiana is among the states investigating the benefits and longevity of porous asphalt surfaces. In the 1980s, the state had experimented with open-graded friction courses (OGFCs), an earlier form of porous asphalt surface. OGFC looked promising at first, but there were long-term problems with the void structure becoming clogged with road grime and the abrasives used for winter snow and ice control. The Indiana Department of Transportation (INDOT) wants to determine if the newer generation of porous asphalt surfaces, often called permeable European mix (PEM), can hold up longer under winter conditions.
The OGFCs used in Indiana and other states in the past, however, differ significantly from porous asphalt mixes. In general, OGFCs had lower void percentages (10-15%) and usually used unmodified binders, at least in Indiana. These differences could make them less durable than the newer porous asphalt mixes.
Porous asphalt mixes generally have strong gap-graded aggregate gradations to yield higher air voids (18-22%). High-quality aggregates are needed to provide good aggregate interlock and long-lasting frictional properties. Modified binders and sometimes fibers are used to hold a thick asphalt film in place, coating the aggregates. In 2003, INDOT allowed about a half-mile test section of porous friction course (PFC) to be placed on I-74 east of Indianapolis. The PFC test section was placed to allow comparison of its properties to those of an adjacent stone-matrix asphalt (SMA) pavement. SMAs are the current preferred surface for high-volume roadways in the state. A third section consisting of a conventional Superpave hot-mix asphalt (HMA) section also is being evaluated for comparison purposes. Heritage Research Group proposed the project, which was approved by INDOT and the Federal Highway Administration. Milestone Contractors LLC constructed the PFC and SMA sections in August 2003. The conventional hot-mix section was placed on U.S. 52 near West Lafayette in July 2003.
The performance of the sections has been monitored since construction by a research team from North Central Superpave Center (NCSC) and the Institute for Safe, Quiet and Durable Highways at Purdue University. The monitoring is planned to continue until at least 2008.
Comparing the threeThe main objective of the field test is to evaluate the performance of porous asphalt and SMA surfaces compared with conventional asphalt surfaces. Performance is being measured primarily in terms of tire/pavement noise generation, surface texture and pavement distress.
The primary focus is on the adjacent PFC and SMA mixes. These two mixes are very similar in terms of the component materials. The conventional mixture was not designed as part of this experiment, but was instead selected from recently constructed projects to represent typical INDOT mixes. All of the mixtures evaluated used steel-slag aggregate from the same source, though in different proportions and combined with various other aggregates and additives.
The SMA and PFC mixes had the most in common. The PFC was composed of 90% steel slag with 10% sand. The SMA consisted of 80% steel slag, 10% stone sand (from a different source than the PFC sand) and 10% mineral filler. The same binder, an SBS-modified PG 76-22, also was used in these two mixes. The PFC included 0.3% cellulose fiber and the SMA included 0.1% of the same fiber.
The conventional HMA consisted of steel-slag coarse aggregate, from the same source as the PFC and SMA, blended 50-50 with coarse dolomite. The mix also contained dolomitic manufactured sand. A PG 76-22 was included, but it was from a different source than that used in the PFC and SMA.
The PFC was designed by Milestone Contractors with assistance from Heritage Research Group. Trial mixes were prepared and compacted to 20 gyrations in a Superpave gyratory compactor. The air-void content was then measured. The target air-void content was 18-22%.
The SMA also was designed by Milestone using the same steel slag, same fiber and same binder. This mix was designed at an Ndesign level of 100 gyrations for a traffic category of 10 to 30 million ESALs. The SMA gradation was similar to the PFC in the larger sizes, but had much higher amounts passing the smaller sieve sizes. This is typical since the SMA was designed for 4% air voids versus 18-22% for the PFC. SMAs generally consist of a somewhat gap-graded aggregate structure with a high VMA to ensure good stone-on-stone contact. The space between the aggregate particles is then mostly filled with a mastic of binder (often polymer-modified, as in this case), fibers and mineral filler.
The conventional mix was designed according to Superpave mix design procedures for a design traffic level of 10 to 30 million ESALs. The Ndesign value was 100 gyrations, as it was for the SMA.
Comparing gaps
The PFC and SMA surfaces were constructed using a material transfer device (MTD) to improve ride quality and reduce segregation. The MTD transferred mix from the trucks into the hopper of a conventional paver. Mixture production of the PFC and SMA mixes proceeded smoothly. No significant problems were observed and no mixture quality penalties were assessed.
Compaction was accomplished with two steel-wheeled rollers. Only one pass with each roller was needed to seat the PFC. Due to the gap-graded nature of the PFC there is extensive stone-on-stone contact between the coarse aggregate particles with very little mastic or fine material to cushion the coarse aggregates. Relatively little compactive effort is needed to bring these coarse aggregates into contact. Over-rolling can lead to aggregate breakdown. Similar rollers and roller patterns also were used with the SMA.
The resulting surface texture of the PFC was quite porous and open, as expected. The SMA, which also is a gap-graded mix, appears much denser. In the SMA, the void space between the coarse aggregate particles, which is left quite open in the PFC, is filled with the mastic of fine aggregates, mineral filler and binder. The conventional HMA was much finer and denser in appearance.
Noise knowledge
Field testing has been conducted on the three experimental sections (PFC, SMA and HMA) to evaluate noise, surface texture and surface friction shortly after construction and again about two years after construction.
A series of measurements were performed by the National Center for Asphalt Technology (NCAT) using their close-proximity method (CPX) trailer to gather noise data for the HMA, SMA and PFC pavements before the road was opened to traffic in September 2003. This testing measures the noise generated near the interface between the tire and pavement surface. Close-proximity testing at two different speeds showed the conventional HMA produced noise levels that were 3.6 dB(A) higher than the PFC, and the SMA produced noise levels that were 4.8 dB(A) higher than the PFC. At 60 mph the PFC produced an average sound level of only 92.6 dB(A).
The Purdue research team also did pass-by testing before the road was opened to traffic using three vehicles, two passenger cars and one pickup truck. In this controlled pass-by testing, the noise is measured by a sound meter off the side of the road as the vehicles cruise by. These measurements, shortly after construction, showed the HMA produced noise levels that were 4.2 dB(A) higher than the PFC, and the SMA produced noise levels that were 5.0 dB(A) higher than the PFC at 50 mph. The pass-by noise levels were lower than the CPX measurements because they were attenuated by distance. The surface texture of the pavement between the tire and the sound meter also can affect how much sound is propagated to the roadside.
Statistical pass-by testing was performed in July 2005. This testing is similar to the controlled pass-by testing performed earlier, but it measures sound generated by random vehicles in the actual traffic stream rather than controlled vehicles. These measurements showed that the PFC was about 6 db(A) quieter than the SMA for cars and about 3 db(A) quieter for heavy trucks.
Since the decibel scale is logarithmic, these are significant differences. The noise measurements on the conventional section could not be compared due to slower traffic speeds and a scarcity of heavy trucks. The statistical pass-by measurements also cannot be compared directly to the controlled pass-by results obtained earlier because of differences in the traffic stream and speeds. The relative differences in sound level, however, can be and the sound meter also can affect how much sound is propagated to the roadside.
Statistical pass-by testing was performed in July 2005. This testing is similar to the controlled pass-by testing performed earlier, but it measures sound generated by random vehicles in the actual traffic stream rather than controlled vehicles. These measurements showed that the PFC was about 6 db(A) quieter than the SMA for cars and about 3 db(A) quieter for heavy trucks. Since the decibel scale is logarithmic, these are significant differences. The noise measurements on the conventional section could not be compared because of slower traffic speeds and a scarcity of heavy trucks. The statistical pass-by measurements also cannot be compared directly to the controlled pass-by results obtained earlier because of differences in the traffic stream and speeds. The relative differences in sound level, however, can be compared.
The surface texture of the three pavements also has been measured twice to date. NCAT did the first round of testing using a circular texture meter (CTM), which uses laser-displacement sensors to measure the surface profile. The laser sensor is mounted on an arm that rotates at a fixed distance above the pavement and measures the change in elevation of points on the surface to a vertical resolution of 3 µm (0.12 x 10-3 in.). The CTM collects data around the circumference of a circle 11.2 in. in diam. The NCSC repeated these measurements in August 2005.
This data set shows that the PFC and SMA both have significantly more texture depth than the conventional surface, with the PFC having the highest texture. The conventional surface has a more uniform gradation, which would be expected to produce a more uniform surface texture.
For the SMA, the texture is somewhat lower than that of the PFC due to the presence of the mastic of asphalt binder and fibers filling the voids between the aggregates. The texture depth of the PFC was 1.37 mm before being opened to traffic in 2003 and had increased to 1.48 mm in 2005. This increase in texture depth is likely due to traffic wearing away the thick binder coating, exposing more of the aggregate surface. The texture of the SMA was stable, averaging around 1.15 mm. The conventional surface has not been tested yet in 2005, but in 2003 its texture was only 0.30 mm.
Pavement friction has been measured using a dynamic friction tester (DFT). The DFT measures the friction between three rubber sliders on a rotating disk and the pavement surface. The DFT measures friction on the same footprint as the CTM measures texture. The DFT and CTM measurements can be combined to yield a value known as the international friction index (IFI).
In 2003, before opening the road to traffic, the PFC showed somewhat higher IFI than the SMA (0.36 and 0.28) and both were quite a bit higher than the conventional HMA (0.19). It was expected that the friction numbers would increase on the PFC and SMA after traffic had a chance to wear off the binder film on the surface. Testing in August 2005 confirmed this. The PFC and SMA both had IFI values of around 0.45. The values for these two surfaces were very similar due to the similar aggregates used. Testing on the conventional surface has not yet been completed in 2005, but it is not expected to have changed significantly; this roadway had been open to traffic over two months before the initial testing was completed.
Splash and spray was observed on I-74 during a light rain. While the amount of splash and spray on the PFC and SMA have not been measured, visual observation shows a very dramatic difference. The splash and spray is significantly less on the PFC section. A video illustrating this difference is available at the NCSC website (http://bridge.ecn.purdue.edu/~spave) under “Videos.” No visual surface distresses were noted on either the PFC or SMA sections during the August 2005 inspections.
Still looking
After two years in service, the PFC section continues to perform well in terms of noise, texture, friction and surface distress. Only two winters have passed so far, however, so further monitoring is needed to determine if the PFC can hold up to Midwestern winters. The sound levels, surface texture and friction of this experimental section will be monitored for three more years, at least, to assess the longevity of the benefits.
If the results are positive, PFC may eventually provide a more economical way to control traffic noise at one of its major sources—the tire/pavement interface.
