CVEEN 7570

Pavement Maintenance and Rehabilitation


The objective of this wiki is to summarize relevant topics related to Pavement Maintenance. Each individual in the class has been randomly assigned a topic that has been covered in class. Their job is to fill in the spaces with relevant information. The target audience will be someone looking for further information on this topic or help on a given problem. This means that you can write with the understanding that the individual reading your post has some basic understanding of pavements. The site will be open to the world wide web, so make sure everything is properly documented and that your explanations are accurate and concise. As an example, look for a topic in Wikipedia. The first post should be uploaded by March 7th; it will be discussed in class.

Pavement Maintenance


Pavement maintenance is a study of the causes and effect of distresses that affect both rigid and flexible pavements. As such, it intends to evaluate indices that can quantify the ability of the pavement to fulfill its function. Given that the need for pavement maintenance is related to its performance, the relation between material selection, traffic, load, and environment must be understood. Once these factors have been understood, the next step is to develop a condition survey that will assist decision makers in the selection of the proper rehabilitation technique.

Description of Pavement Distresses


Flexible Pavements


Environmentally Induced (Levi)


Block Cracking
Cause:
This distress is caused by shrinkage from temperature variations. Unlike thermal cracking, block cracking occurs in older roads where the pavement has aged. However, this distress is non-load associated. As a flexible pavement ages it becomes more brittle and does not allow the pavement to expand and contract as easily. This resistance results in fractures in the material both in the longitudinal and transverse directions at regular intervals.
Evaluations:
This distress is evaluated in square feet per area at each severity level. The severity levels depend on the size of the crack as well as frequency of the cracks.
Solution:
This distress can be fixed by pavement rejuvenation or a surface treatment such as a chip seal.


Block_Cracking.JPG


Thermal Cracking

Cause:
Extreme temperature drops cause the pavement to rapidly cool and contract. This contraction creates stresses in the pavement that exceed the strength of the material and result in a fracture along the length of the pavement. These cracks are typically transverse along the pavement and occur at regular intervals. This distress is non-load associated and typically is a result of a binder that is too stiff.
Evaluation:
This distress is evaluated as the number of transverse cracks per mile. Factors of crack length and severity are also observed and recorded.
Solution:
Complete reconstruction is not required for this distress. The cracks need to be maintained by sealing them with a tar mixture so moisture cannot penetrate the pavement and cause damage. This tar will allow the pavement to expand and contract with the temperature changes without causing more harm to the pavement.


Thermal_Cracking.jpg


Potholes
Cause:
This distress is caused by moisture induced raveling. When the moisture penetrates the pavement it can cause the aggregate in the mix to separate from the binder. This separation
allows the aggregates to escape the pavement causing a void. Over time, when many pieces of aggregate have been removed, it creates a very large void called a pot hole. Raveling is a load associated distress; however, the raveling begins and intensifies as a result of moisture damage. Potholes are very common during the spring. This time of year causes the moisture to freeze and thaw repeatedly which accelerates the damage that is caused by the moisture in the pavement. This distress indicates a problem with the mix design of flexible pavement.
Evaluation:
This distress is evaluated by measuring the size (area) and depth of the pothole. This area associated with depth can be related into a severity level of a particular pothole.
Solution:
This distress can be temporarily remedied by an asphalt concrete patch. This is only temporary because a similar pothole will form in or around the patch. To
permanently fix this issue, the pavement, subgrade, and sub base need to be
excavated. Then a complete rebuild of the pavement needs to be performed.


Pothole.jpg




Material Related (Fabiola)

This section represents the major flexible pavement distresses. Although the following distresses can also occur through means of the environment and/or poor construction techniques, the following focuses only on causes related to material issues.

Bleeding

Bleeding.jpg
Asphalt Bleeding

Bleeding refers to when asphalt binder fills the aggregate voids and emerges to the surface during hot weather. The asphalt binder accumulates on the surface and it usually appears as a shiny, glass-like surface. Bleeding asphalt can become very sticky and it usually causes loss of skid resistance when the surface is web.

Causes: the following are material-related causes of bleeding:
  • Excessive asphalt binder in the HMA.
  • Low HMA air voids content which do not allow enough room for the asphalt to expand during warmer seasons.
Treatments: The following measures treat the bleeding once it shows on site by eliminating or reducing the asphalt binder on the surface:
  • Minor Bleeding: use of coarse sand to sock up the asphalt binder on the surface.
  • Major bleeding: use of motor grader or heater planer to remove excess of binder.
In order to avoid the problem of bleeding, the HMA needs to be treated before is placed.

Block Cracking
Block_Cracking.JPG
Block Cracking
Group of large cracks (usually one foot or more), that breaks the pavement into rectangular pieces. Block cracking usually covers a large area of pavement and it may occur on areas where there is no traffic. It permits water infiltration in the pavement structure and also causes irregularities in the pavement surface.
Causes: Block cracking is caused by shrinkage of asphalt pavement due to temperature cycles. This is commonly associated with an ineffective selection of the asphalt binder in the mix design.
Treatment: Depends on the severity of the damage:
  • For cracks less than 0.5 inch wide, the application of crack seal is recommended.
  • For cracks more that 0.5 inch wide, removing and replacing the pavement layer is recommended.


Shoving
shoving1.jpg
Asphalt Shoving
Distortion of the pavement surface forming ripples across the pavement, perpendicular to the traffic direction. It occurs where high horizontal distresses are present such as intersections. It causes roughness on the pavement surface.
Causes: Shoving is caused by the stop and go action of traffic along with the instability of the asphalt mix. This can be caused by mix contamination, inadequate mix design, low quality of asphalt mix manufacturing, and deficiency of aeration of asphalt emulsions.
Treatment: Usually it is recommended to remove and replace damaged pavement area.

Raveling
Raveling.jpg
Raveling
Raveling is the loss of bond between the aggregate particles from the asphalt mix resulting in the wearing away of the aggregate from the HMA, beginning at the surface and progressing downwards. It causes many problems such as loose debris on the pavement surface, roughness, water accumulation which can result in hydroplaning and loss of skid resistance.
Causes: There are a number of causes including:
  • Presence of dust covering the aggregate impeding the asphalt binder to bond with the aggregate particles.
  • Gradation, including less fines and more coarse aggregate with less points of contact.
Treatment: Remove and replace damaged pavement.

Rutting
rutting2.jpg
Rutting
Is a linear surface depression in the wheelpath. Ruts can be filled with water and cause vehicle hydroplaning. There are two types of rutting: mix rutting and subgrade rutting. Mix rutting occurs when only the pavement surface shows rutting due to poor construction techniques and/or mix design issues. Subgrade rutting occurs when the subgrade shows rutting due to traffic loading.
Causes: One material-related cause for rutting is the use of an inadequate mix design or manufacture, such as exaggerate asphalt content, mineral filler or lack of angular aggregate particles.
Treatment: Ruts smaller than 1/3 inch deep generally do nott need treatment. For deeper ruts an overlay is recommended.


Stripping
Description: The loss of bond between aggregates and asphalt binder that typically begins at the bottom of the HMA layer and progresses upward. It causes rutting, shoving, raveling, or cracking.
Causes: Stripping is difficult to identify because it manifests on the pavement surface as other forms of distress to include rutting, shoving/corrugations, raveling, or cracking. It is recommended to take a core to identify stripping as a pavement distress. A material-related factor that influences in striping is the poor aggregate surface chemistry.
Treatment: A pavement that presents stripping generally needs to be removed and replaced. Also, correction of any drainage issues needs to take place before removing and replacing the pavement.

References:
Asphalt Institute: www.asphaltinstitute.org
Pavement Interactive: www.pavementinteractive.gov
Asphalt Magazine: www.asphaltmagazine.com

Construction Related


Poor construction practices can amplify all pavement distresses. If the asphalt or sub-grade layers are not compacted properly rutting may occur. If two layers are not bonded properly longitudinal cracks are more likely. When proper drainage procedures such as a slight cross slope are not implemented water may collect and promote raveling and potholes. Should the asphalt mix design is not adhered to the pavement can be expected to not perform as indicated, i.e. sorter design life, inadequate friction, excessive roughness.

Some pavement distresses are directly attributed to poor construction practices. Lack of compaction in the asphalt layer will probably result in rutting. Frequently the entire road cannot be resurfaced all at once either the road is too large to surface all at once or traffic is reduced to half lanes. Surfacing roads in stages often results in longitudinal cracking. Cracks develop along the edges of staged area/joint. In addition failure to follow design drainage guidelines will results in striping, raveling, and potholes.

An excellent reference for flexible pavement distresses is the Washington Asphalt Pavement Association (WAPA). Website is www.asphaltwa.com .


Example Construction Related Distresses

Longitudinal Cracking
Long_Cracking.jpg
(http://www.asphaltwa.com/wp-content/uploads/2010/09/longitudinal1.jpg)
Longitudinal cracks due to construction practices usually occur at joints. Typically roads cannot be paved all at once. Either the road is too wide or it is not possible to shut down all lanes at the same time. The street is paved in sections resulting in joints. Joints are weak points in the pavement structure and frequently will crack. Longitudinal cracks due to weak joints can be minimized by placing the joints outside of the anticipated wheel path. Maintenance includes crack sealing and overlays. Measured in crack length and severity


Rutting
Rutting.jpg

Rutting due to construction practices is most attributed to improperly compacted asphalt concrete. Less dense asphalt will settle under traffic loads. Measured in max depth in inches. Solution is for quality control to ensure proper compaction and adherence to design mix specifications. Excessive asphalt context is also a cause of rutting. Following design specification for amount of asphalt should reduce rutting from this source. Repair includes overlay or reconstruction

Bleeding
Bleeding2a.jpg
(http://www.asphaltwa.com/wp-content/uploads/2010/09/bleeding4_small.jpg)
Bleeding is when asphalt binder forms a shiny, glass like, reflective film on the surface; measured in square feet per area. This surface may become slick when wet or sticky when surface temperature is high. This results in less friction and appears unsightly. When bleeding is evident in a newly pave road it indicates improper mix design or excessive asphalt binder added during construction. Solutions for minor bleeding include dusting the affected surface with coarse sand. Severe bleeding requires grinding and overlay.

Improper drainage
Pothol2.jpg
If construction does not follow design guidelines for drainage raveling, potholes, shoulder heave may occur.

References:
Washington Apshalt Pavement Assicitation. www.asphaltwa.com/ Retrieved March 5, 2011

Romero, P (n.d.) "CVEEN 7570 Pavement Maintenance and Rehabiliation: Lesson 3 - Flexible Pavement Distresses". Retrived March 5, 2011, https://online.uen.org/webct/urw/lc257461621151.tp0/cobaltMainFrame.dowebct

Mallick R. B., El-Korchi., (2009). "Pavement Engineering: Principles and Practice". CRC Press


Rigid Pavements

Structural Deficiencies (Tiffany)


Structural deficiencies in rigid pavements can be determined most effectively by measuring the response of a slab to a given load. It is equally important for the response in the subgrade to also be measured. The different layers of a typical rigid pavement can be seen below in Figure 1. An inadequate subgrade is often the cause of structural problems in a rigid pavement. Similarly, lack of compaction at installation can cause structural deficiencies; the roadway will later compact under traffic loading and can cause severe distresses. Other factors which may cause structural deficiencies within the roadway are tensile stresses from traffic loading and temperature. Tensile forces result from the traffic loading as Seen in Figure 2. An optimal balance between traffic loading and pavement thickness must be achieved to have the minimal deterioration of the roadway possible. In Figure 3 it can be seen that the effect of the loading on the rigid pavement is greatly reduced as the thickness of the roadway is increased. Extreme temperature changes can cause unhealthy shrinkage and expansion of the roadway. This causes the rigid pavement to form many different types of cracks and temperature specific distresses, such as curling. Other factors which may participate in the presence of distresses in the roadway are moisture, raveling, movement of slabs, inadequate bonding, repeated loads, and the stopping and starting of traffic. While the distresses in the roadway can not be prevented, their effect can be lessened.

Figure One: Rigid Pavement Layers
Figure One: Rigid Pavement Layers




Figure 2: Load Distribution Pattern
Figure 2: Load Distribution Pattern
Figure Three: Load Distribution in Different Thicknesses
Figure Three: Load Distribution in Different Thicknesses



















Structural deficiencies result in many different types of roadway distresses, depending on the factor(s) which caused the stress. Some of these distresses include longitudinal and transverse cracking, pumping, faulting, spalling, polishing, scaling, D-cracking, corner cracking, and blowups. These deficiencies can be repaired by many means, depending on the severity and type of distress. These rehabilitations include patching, overlay, reconstruction, crack maintenance, and surface treatments. However it is important to try to prevent them from forming as much as possible when the pavement design is chosen. The testing methods mentioned above can help to lessen the effect of the causes of structural deficiency, and the tests can be performed for specified conditions and before the installation of the roadway. The tests should take into account weather conditions, traffic conditions, etc. The response of the roadway under specific loading conditions will allow for the derivation of many roadway characteristics. The results of these tests will allow for adjustments to be made which will increase the structural capacity of the roadway.



References:
Lecture #6&2
http://www.fhwa.dot.gov/engineering/geotech/pubs/05037/images/f004.gif
http://sanral.ensightnetworks.com/live/content.php?Category_ID=40
http://www.globalsecurity.org/military/library/policy/army/fm/5-430-00-1/fig5-2.gif


Environment-Induced Stresses in Rigid Pavements

Stresses Due to Temperature

Temperature variations within a rigid pavement produce dissimilarities in the length of the slab fibers, which result in the slab curling. Changes in temperature (ΔT) at a given depth in the slab result in a corresponding change of length (ΔL) at this depth; as demonstrated in equation 1. [1]

ΔL=LαtΔT (equation 1)


Where, L represents the original length of the slab, is the coefficient of thermal expansion, and ΔT is the change in temperature.

When the temperature at the upper half of the slab is lower than the temperature at the lower half, as is the case in the evening, the slab forms a concave upper shape. The opposite is true in early morning, resulting in a convex upper shape. The weight of the slab acting on these deformed shapes, and the varying reaction from the subgrade, generates stresses in the slab. When the slab has a concave upper shape, there is tension at the top and compression at the bottom. The opposite is true when the slab has a convex upper shape. [1]

Curling is often an influential factor affecting the structural and functional performance of concrete pavements.




Curling.jpg


Figure 1

Stresses Due to Subgrade Friction

Stresses produced by uniform temperature changes; due to post construction concrete shrinkage causes expansion and contraction of the slab. This expansion and contraction is only impeded by the shear stresses that are created by subsurface friction.

Subgrade friction is a function of the amount of slippage between the slab and the subgrade, which increase from the centerline of the slab toward the edges. The distribution of the shear subgrade friction stresses can be approximated by a uniform distribution of forces.

Stresses Due to Weather

The weather has a significant effect on the performance of pavement. The two main factors of concern are the presence of water/ice in pavement layers and subgrade, and the variation of temperature throughout the year. These two factors interact with each other, such as during the freezing of pore water in frost-susceptible subgrades, which result in the pavement heaving. [1]

The freezing of pore water and the melting of pore ice result in significant pavement layer volume changes, which over time and under the action of traffic loads, reduce pavement serviceability. The depth of frost penetration is a function of the type of the pavement layers and the thermal characteristics of the subgrade, as well as the duration of sub-freezing temperatures. [1]

When water freezes, it expands about 9%. As the water in moist concrete freezes, it produces pressure in the pores of the concrete. If the pressure developed exceeds the tensile strength of the concrete, the cavity will dilate and rupture. The accumulative effect of successive freeze-thaw cycles and disruption of paste and aggregate can eventually cause expansion and cracking, scaling, and crumbling of the concrete. [2] As shown in Figure 2.


image75060.gif



Figure 2



References

1. A.T. Papagiannakis and E.A. Masad, Pavement Design and Materials, 2008 John Wiley & Sons Inc.., Hoboken, New Jersey.
2. Portland Cement Association, Concrete Technology, Durability 2011 http://www.cement.org.






Evaluation of Pavement Condition


Structural Evaluation


Methods and Equipment (David)

Structural Evaluation Methods and Equipment

Engineers typically rely on the pavement layer stiffness and modulus to determine the structural integrity of a roadway. This information is used by local, state and federal highway agencies to prioritize maintenance and replacement of deteriorating rigid and flexible pavements. Many techniques and methods exist for determining a pavement section modulus. These methods can be classified primarily into two categories:

  • Destructive Testing
  • Non Destructive Testing (NDT).

Destructive Testing
Destructive testing provides more detailed data about the pavement not possible to obtain through non-destructive testing. Such detailed data include:
  • laboratory mechanical, physical, and chemical properties (obtained through coring, Shelby tubes, and trenching), and
  • visual inspection of pavement layers through coring and trenching

Trenching and Coring
Trenching and coring can be used to identify and evaluate pavement condition.

Trenching
A trenching procedure has been developed to run tests and collect samples while minimizing disturbance to the base material. The size of each trench is 3 ft. by 12 ft. Each saw cut is deep enough to penetrate all the way through the pavement layer. A strip about 6 in. wide is cut and removed by hand from one end of the trench. This creates a slot for a backhoe to reach into and lift the block a saw cut is also made across the center of the trench so the blocks can be removed without breaking.


Pic_2.JPG
Removing the AC layer


After the layer is removed and samples are collected, tests can be run, and samples collected on the base layer. The backhoe then digs out the base layer and a few inches of subgrade and the trench walls are smoothed using shovels, chisels, and brooms. On the clean trench wall the layers are highlighted with chalk or string lines. The thickness of each layer can then be measured to aid in the determination of the modulus and other properties.


Pic_1.JPG
Trench showing pavement layers


Coring
Coring involves drilling into the pavement section and extracting a representative sample for analysis and testing. To extract a core sample, a coring rig is placed at the desired location. A barrel cooled water system us used to aid in lubricating and cooling the coring bit. Dry ice can also be used to cool the core barrel. This method allows engineers to maintain a specific moisture content in the samples.


Pic_3.JPG
Coring with Dry Ice


The barrel is spun at about 500 rpm and gradually lowered through the layer. After the barrel has cut to the desired depth, it is retracted while still spinning. Then the core barrel is stopped and the location of the core is observed. The core can now be lifted out of the barrel and be transported to a laboratory for analysis. Normally, the core diameter is either 4 in. or 6 in.

Pic_4.JPG
Extracted Core Samples


Shelby Tube
Shelby tube samples have been used to determine the in situ density, moisture content, plasticity index (PI), sulfate content, and modulus of the subgrade soil. The Shelby tube is a sharpened pipe that is pushed into soil by a hydraulic ram on a truck-mounted boom. An auger can be used to remove the top layers and collect Shelby tube samples. After the top layers are removed, the Shelby tube is positioned on the subgrade and pushed into the soil. The pipe is then pulled out and placed on a rack. Aluminum foil and cardboard tubes are used to protect the Shelby tube samples. The tube is labeled with information including the location, orientation and depth of the sample.


pic_5.JPG
Extracting subgrade samples with Shelby tube

Non Destructive Testing (NDT)
NDT allows engineers to evaluate the structural integrity of a pavement with out damage or disturbance to the roadway in question. This can be a valuable tool for analyzing a pavement throughout its service life. The obtained data can then be used to develop performance models and make life cycle predictions useful in pavement design and planning. NDT is primarliy achieved by creating a measured, controlled deflection in the pavement surface. The measured deflection is then used to back-calculate the layer modulus E as shown in the pavement deflection equation below.
equ_1.JPG




Where q is the tire pressure, a is the influence radius and u is poisson's ratio for the combined layer.


Nondestructive deflection testing equipment can be separated into three categories or types.
  • Static Deflections
  • Steady State Deflections
  • Impact Load Deflections (FWD)

Static Deflections
Static deflection measurments are achieved through the use of the Benkelman Beam. The Benkelman Beam was developed at the Western Association of State Highway Organizations (WASHO) Road Test in 1952 The beam is a simple device that operates on the lever arm principle an is used with a loaded 18,000 lb (80 KN) single axle dual tire truck. The tires are inflated to 480 to 550 kPa (70 to 80 psi) and measurements are made by placing the tip of the beam between the dual tires and measuring the pavement surface rebound as the truck is moved away.


Pic_6.JPG
Benkelman Beam
Steady State Deflection
Steady State Deflection is produced by applying a dynamic generated oscillating force to the pavement surface. The resulting deflection is measures by a data collection unit containing accelerometers and transducers. The equipment can be housed in a self-contained unit or directly mounted in a specialized truck or van. The most widely used steady state deflection equipment is the Dynaflect and the Road Rater. This type of analysis is most useful when applied to thinner pavements with relatively low traffic volume.


Pic_7.JPG
Van Mounted Road Rater



Impact Load Deflections
Impact load deflection is caused by an impulse load on the pavement surface. The resulting deflection basin is measured and recorded by sensors. The test is performed by positioning a Falling Weight Deflectometer (FWD) over the pavement which releases a weight at different heights dictated by ASTM standard 2000. Sensors at the pavement surface log the deflection data produce3d by the impulse. The FWD can also be incorporated into a contained or truck mounted configuration. Impact loading is an accurate and economical method for simulating transient traffic loading.


Pic_8.JPG
Dynatest FWD Unit


Reference:

Washington Apshalt Pavemtn Assicitation.

www.asphaltwa.com/

Romero, P (n.d.) "CVEEN 7570 Pavement Maintenance and Rehabiliation: Lesson 3 - Flexible Pavement Distresses".

https://online.uen.org/webct/urw/lc257461621151.tp0/cobaltMainFrame.dowebct



Analysis and Models


Structural integrity of pavement is typically determined by comparing the current modulus to the old modulus. If the new modulus is less than the old one, then the pavement have deteriorated and is losing its structural integrity.

Since the Falling Weight Deflectometer (FWD) is the most widely used non-destructive deflection measurement device today, we will focus on analyzing the data from FWD tests.

During the FWD test deflections are measured and recorded at each of the 6 deflection transducers. Figure 1 shows the layout of a FWD test. The "curve" created by the deflections at each transducer is referred to as the deflection bowl.


cdeflectionbowl.jpg

Figure 1: FWD Layout[i]




If there were one semi-infinite, homogeneous layer of pavement the testing equipment would only require a circular load plate and one deflection transducer placed directly under the load plate. The modulus could be calculated simply by using the following equation:

equation1.jpg

Where: ω0=deflection, a=radius of the load plate, q=stress (Force/area of load plate), μ=Poisson's Ratio (0.5), and E=Poisson's Ratio (0.5), and E=modulus. First thing that needs to be done is solve for E.


equation2.jpg
Unfortunately, pavement is not placed as a semi-infinite, homogeneous layer, and even if it were, rarely is it known what is beneath the surface of the pavement. To rectify this, we need to calculate the effective modulus. To ease the backcalcualtion of layer modulus, we assume that load is spread as we go deeper in the pavement (see Figure 2). With this it is assumed that the outermost deflection transducers (sensors) correspond to the lower levels.



cbowl.jpg
Figure 2: Load Spread[ii]



To get the modulus of each layer in this manner we have to have an initial guess of what they are. The deflection is plotted as a function of radial distance from the center of the load plate. The deflections are calculated using predicted values for the height and modulus for each layer. The measured deflections are also plotted on the same graph. The modulus of the lower most layer is modified until the two lines touch at the furthest end. This process is repeated for every layer while holding the values of the previous layers constant. Once all predicted deflections match the measured the deflections we use the adjusted moduli as the measured moduli. An example of the graphs from this process is shown in Figure 3. The blue lines represent the predicted deflections and the red lines represent the measured deflections.



cgraphs.jpg
Figure 3: Graphical Analysis of Layer Modulus



Since moduli are a reflection of deflections, a second, much simpler, method of comparing pavement integrity is to calculate the area of the deflection bowl. To ease in the calculation the deflections are converted to a unitless value. This is done by dividing every deflection by the deflection at D�­0. Refer to Figure 4 for further explanation. The equation shown in Figure 4 is for 3 deflection transducers. For the setup shown in Figure 1 (6 transducers) the equation would be:


equation3.jpg
As with the modulus, the current area is compared to the previous area. If the current area is greater than the previous area, then the pavement is deteriorating.


careas.jpg
Figure 4: Area Calculation[ii]



Many software programs are available to compute layer modulus from collected FWD data. Most agencies use these software packages as opposed to using the hand calculation techniques discussed above. A few of these software programs are ELMOD6 by Dynatest[iii] and MODULUS 6.0 by the Texas Transportation Institute at Teas A&M University[iv].



References:
[i] Fontul, S. (n.d.). Structural Evaluation of Flexible Pavements Using Non Destructive Tests. Retrieved March 3, 2011, from European Conference of Transport Research Institutes: http://www.ectri.org/YRS09/Papiers/Session8/Fontul_S_Session8_Road_Transport (2).pdf
[ii] Romero, D. P. (2011, Februrary 9). CVEEN 7570 Pavement Maintenance: Lesson #7 Backcalculation of Layer Modulus. (University of Utah Spring 2011 Pavement Maintenance and Rehabilitation Class)
[iii] Dynatest. (n.d.). Dynatest FWD / HWD Test Systems. Retrieved March 7, 2011, from Dynatest: http://www.dynatest.com/structural-hwd-fwd.php?tab=structural
[iv] Anil Misra, S. K. (n.d.). Resilient Moduli and Structural Layer Coefficient of Flyash Stabilized Recyled Asphalt Base. Retrieved March 7, 2011, from Ash Library: http://www.flyash.info/2007/60misra.pdf




Safety Evaluation

Friction Theory and Applications (Jesse)

FRICTION THEORY AND APPLICATIONS

Pavement Friction
Pavement friction is a critical factor in providing safe conditions for vehicles traveling on roads. Pavement friction is the characteristic that gives drivers the ability to control and maneuver their vehicles safely in both the longitudinal and lateral directions, and is a key component for highway design. Generally, the higher the pavement friction, the more control a driver has over the vehicle (National Cooperative Highway Research Program (NCHRP), 2009).

Pavement friction is the force that resists the relative motion between a vehicle tire and the pavement surface, as shown below in Figure 1.


Figure_1_forces_on_rolling_tire.jpg
Figure 1: Illustration of Forces Acting on a Rolling Tire (NCHRP, 2009)


The Friction Force is referenced in terms of the friction factor or coefficient of friction (m) as follows:


eq_1_coefficient_of_friction.jpg
The friction factor, or coefficient of friction, can then be used to derive the Skid Number (SN), which is measured during pavement friction testing (to be discussed as a separate topic). The Skid Number (SN) is simply given as:


eq_2_skid_number.jpg
Friction force is developed in both the longitudinal and lateral directions. In the longitudinal direction, friction forces develop when operating in a free rolling or constant braked mode. The Slip Speed is referenced as the relative speed between the tire circumference and the pavement. In the free rolling mode, slip speed is zero. In the constant braked mode, the slip speed increases from zero to a potential maximum of the speed of the vehicle.


The mathematical relationship for slip speed is shown below (NCHRP, 2009):


eq_3_slip_speed.jpg
Where: S = Slip speed, mi/hr.
V = Vehicle speed, mi/hr.
Vp =Average peripheral speed of the tire, mi/hr.
w = Angular velocity of the tire, radians/sec.
r = Average radius of the tire, ft.

The Slip Ratio (SR) is defined as the ratio of the slip speed over the vehicle speed, multiplied by 100, as shown below (NCHRP, 2009):


eq_5_slip_ratio.jpg
Where: SR = Slip Ratio (%)
V = Vehicle speed, mi/hr.
S = Slip Speed, mi/hr.

From the relationships shown above, it is seen that a locked wheel state can be referred to as a 100 percent slip ratio, and a free rolling state as a zero percent slip ratio.
The coefficient of friction between a tire and the pavement changes with varying slip. The coefficient of friction increases rapidly with increasing slip to a peak value called critical slip that usually occurs between 10 and 20 percent slip. The friction then decreases to a value called the coefficient of sliding friction which occurs at 100 percent slip (NCHRP, 2009). This relationship is illustrated below in Figure 2.



Figure_2_pavement_friction_vs_tire_slip.jpg
Figure 2: Pavement Friction vs. Tire Slip (NCHRP, 2009)

When the tire is rolling, the ground force acts on the tire at a point offset from tire's vertical axis by a distance, a, creating a moment that resists the rolling motion. The value of a is a function of speed and increases with speed. The force required to overcome this moment is called the rolling resistance force (Fr). In the constant braked mode, an additional braking slip force is created. The total frictional force is then the sum of the free rolling resistance force (Fr) and the braking slip force (FB) (NCHRP, 2009). These forces are illustrated in Figure 3 below.



Figure_3_forces_and_moments_of_a_braked_wheel.jpg
Figure 3: Forces and Moments of a Constant Braked Wheel (NCHRP, 2009)


Lateral friction forces develop as a vehicle changes direction or compensates for pavement cross-slope. The relationship between the forces acting on the tire and the pavement surface is as follows (NCHRP, 2009):


eq_6_side_friction.jpg
Where FS = Side Friction
V = Vehicle speed, mi/hr.
R = Radius of the path of the vehicles center of gravity, ft.
e = Pavement super elevation, ft/ft
These terms are illustrated below in Figure 4.

It should be noted that there is a set amount of friction available at the tire - pavement interface for braking force and lateral friction forces. Therefore, a driver must often choose between braking and turning in a curve. The more friction force that is expended braking, the less lateral forces will be available for cornering, and vice versa.



Figure_4_forces_on_a_vehicle_traveling_around_a_curve.jpg
Figure 4: Forces on a Vehicle Traveling Around a Curve (NCHRP, 2009)


Friction Mechanisms
Pavement friction occurs due to the combination of adhesion and hysterisis. Adhesion is the friction that results from te small scale bonding of the vehicle tire rubber and the pavement surface, and is a function of the interface shear strength and contact area. Hysterisis results from lost energy due to deformation of the tire, which can be pictured as the enveloping of the the texture by the tire (NCHRP, 2009). These two components of friction force are illustrated below in Figure 5.



Figure_5_adhesion_and_hysteresis.jpg
Figure 5: Illustration of Adhesion and Hysteresis Components of Friction Force (NCHRP, 2009)

Both components of friction force depend largely on pavement surface characteristics, as well as temperature, speed, environmental factors. Adhesion force is developed at the pavement-tire interface and is most responsive to the micro-texture of the aggregate particules. Adhesion governs the overall friction on smooth and dry pavements. Hysteresis force is more responsive to the macro-tecture, and is the dominant component of friction force on wet and rough textured pavements (NCHRP, 2009).

Pavement Surface Texture
Pavement surface texture is defined as the deviations of the pavement surface from a true planar surface, and is categorized by wavelength (l) and peak to peak amplitude (A) of its components. The three levels of texture are (NCHRP, 2009):
- Microtexture (l< .02 in [.5 mm], A = .04 - 20 mils [1 - 500 mm]); Surface roughness quality at the sub-visible or microscopic level.
- Macro-tecture (l= .02 - 2 in [.5 - 50 mm], A = .005 - .8 in [.1 - 20 mm]); Surface roughness quality defined by the mixture properties and method of finishing of surface.
- Mega-texture (l= 2 -20 in [50 - 500 mm], A = .005 - 2 in [.1 - 50 mm]); Defined by distress, defects, or waviness on the pavement surface.

These texture levels are illustrated below in Figure 6, and the effects of each level are shown in Figure 7.



Figure_6_surface_texture_illustration.jpg
Figure 6: Simplifiled Illustration of Texture Ranges (NCHRP, 2009)


Figure_7_texture_wavelength.jpg
Figure 7: Texture Wavelength Influence on Pavement � Tire Interactions (NCHRP 2009)


There are many factors in addition to surface texture that affect the available friction on a given pavement. Some of these are represented below in Table 1.



Table_1_factors_affecting_pavement_friction.jpg
Table 1: Various Factors Affecting Available Pavement Friction

Among the critical combinations of non-operator/vehicle factors to note is the impact of wet surfaces on pavement with low friction coefficients/Skid numbers. In general, the friction resistance of most dry pavements is relatively high. However, the number of accidents on wet pavements is twice as high as on dry pavements (Romero, n.d.). The graph shown below in Figure 8 illustrates the relationship between the rate of wet pavement accidents and the skid number of various roads in Kentucky.



Figure_8_wet_surface_accidents_vs_skid_number.jpg
Figure 8: Wet Surface Accidents vs Skid Number (NCHRP, 2009)

Friction Over Time
Friction forces on a pavement's surface generally increase in the first few years following construction, then steadily decline over time. This is due to the actual aggregate particles being covered in binder and surface treatments at construction, and then becoming exposed over the first few years of use. Following the initial period, the texture at the micro-texture level is worn down, causing what is known as aggregate polishing. The graph shown below in Figure 9 illustrates the friction variance over the life of a pavement.


Figure_9_early_life_skid_resistance_changes.jpg
Figure 9: Early Life Skid Resistance Changes for AC and PCC Pavement Surfaces (Ahammed & Tighe, 2009)





References:

Ahammed, M., & Tighe, S., (2009). Early-Life, Long-Term, and Seasonal Variations in Skid Resistance in Flexible and Rigid Pavements. Transportation Research Record: Journal of the Transportation Research Board, Volume 2094
National Cooperative Highway Research Program, (2009). NCHRP Web-Only Document 108: Guide for Pavement Friction. Retrieved March 2, 2011, http://onlinepubs.trb.org/onlinepubs/nchrp/nchrp_w108.pdf
Romero, P. (n.d.) CVEEN 7570 Pavement Maintenance and Rehabilitation: Lesson #10 - Skid Resistance. Retrieved March 4, 2011, https://online.uen.org/webct/urw/lc257461621151.tp0/cobaltMainFrame.dowebct





Methods and Equipment


<!-->
Friction Testing Methods and Equipment
There are several methods currently employed to test the friction of the pavement surface. Over the years many different techniques and equipment types have been deployed to measure this important pavement characteristic. The variability in measurement techniques, procedures, data collection, and reporting can be considerable. There are several commercially available devices that can be used to study friction, they can operate at low speeds, high speeds, in the lab, and measure a variety of different friction predicting characteristics. The test tire conditions are also variable. ASTM and AASHTO have developed a set of surface characteristics measurement standards to ensure comparable texture and friction data reporting. The most common testing methods fall onto one of the following groups:

  • Locked Wheel Mode
  • Side Force (Yaw) Mode
  • Spin up (Fixed Slip)Mode
  • Surface Texture Measurement
  • Traffic Control Tests

Locked Wheel Mode
In the United States the most common type of surface friction test is the lock wheel test. The associated standard is ASTM E 264. The device is typically installed on a trailer that is towed behind a measuring vehicle at 40 MPH. Water is sprayed in front of the test tire of the trailer and a brake is applied to force the tire to lock. The drag force of the tire is measure while the tire is fully locked. The measurement can be repeated when the tire is rolling without skid again. A picture of a vehicle with a lock wheel tester is shown in figure 1 below.

Figure_1locked_wheel.jpg
Figure 1 Locked Wheel Trailer (http://www.mto.gov.on.ca/english/transtek/roadtalk/rt16-3/)


There are two types of tires that can be used for this test: A ribbed tire ASTM E 501 or a smooth tire ASTM E 524. Smooth tires are more sensitive to pavement macro texture and the ribbed tire is more sensitive to micro texture.
The advantages of this test is that it is well established, user friendly, and simple. The disadvantage is that it is can be used on only straight segments of road (where friction is usually least important) and can miss the slippery sports because of the intermittent nature of the test.

Side Force Mode
Side force measuring devices measure the pavement side friction that is needed when a vehicle is cornering or making a turn. Standard ASTM E 670 is used in this instance. The equipment required is the British Mu-Meter or the British Sideway Force Coefficient Routine Investigation Machine (SCRIM) The difference between these two devices is that the MU-Meter has the tires at a yaw of 7.5 degrees and the SCRIM has a yaw of 20 degrees. A picture of the MU-Meter is shown in figure 2 below and the SCRIM is shown in figure 3.

Fig_2_yaw.jpg
Figure 2 British MU-Meter (Romero 2011)




fig_3_SCRIM.jpg
Figure 3 SCRIM (NCHRP 2009)


In this test water is sprayed on the pavement surface and the skewed wheels are pulled over the wet pavement surface at 40 MPH. The side force, tire load, distance, and speed are recorded every 1 to 5 inches.
The advantages of this test are that it is well controlled, user friendly, simple and quick; the measurements it takes are continuous. This test is also well established as it is commonly used in Europe. The disadvantages are that is very sensitive to pavement distresses such as potholes and cracks. In the United States it is mostly only used at airports.

Spin Up Mode
The spin up or fixed slip looks very similar to the locked-wheel test in its configuration in that it usually employs a tow trailer. A picture of such a device is shown in figure 4 below. The difference is that there is a retractable wheel. The basic principle of the test is that the retractable test wheel is locked and lowered to the pavement surface while the vehicle is towing the trailer at 40MPH. The tire is then unlocked and allowed to �spin up� to the vehicle speed by a mechanism within the trailer. Wheel loads and forces are measured by sensors in the trailer. A similar method to the spin up is the �spin down� mode or variable slip mode where the retracted tire is initially allowed to rotate freely and the speed of the rotating wheel is gradually reduced and begins to skid. The Variable slip test is standardized by ASTM E 1859 and the spin up test is specified by various standards.

Figure_4_retractable.jpg
Figure 4 Friction testing device with retractable tire. (NCHRP, 2009)



This test is advantageous to the others in that the test tire will not wear out as quickly as it is not locked for as long a time or always skidding. Also, no expensive force measurement apparatus is necessary. The force can be calculated by knowing physical properties of the test wheel. A disadvantage is that fixed slip testers only take readings at their specified fixed speed. The slip speed does not always coincide with the critical slip speed value.

Surface Texture
Because the friction of a pavement surface is so closely tied to the texture of the pavement surface, there are methods that test the pavement�s texture then compare it with skid measurements. Three of the most common methods are the Sand Patch test, ASTM C 956 and ISO 10844; the outflow meter (OFM), ASTM E 2380; and Circular Texture Meter (CTM) ASTM E 2157. The sand patch test is simply done by spreading a known volume of glass beads (sand) over a surface in a circular pattern then measuring the diameter of the circle. The outflow meter is related to hydroplaning, it measures the water drainage rate through the surface. The circular test meter uses a laser to measure the surface texture in a circular profile at regular intervals.

Traffic Control Friction Tests
All of the surface friction tests that have been mentioned so far can be done without traffic control. The following three tests require traffic control and include (1) Stopping Distance Measurement, ASTM E 445, (2) Deceleration Rate Measurement, ASTM E 2101, and (3) Portable Testers.
The stopping distance measurement is performed by driving a vehicle at 40 MPH over a wet pavement surface then applying the maximum braking and measuring the distance it takes to stop the vehicle.
The Deceleration Rate measurement is typically done in winter conditions. The vehicle is driven at 20-30MPH then the maximum braking is applied and the deceleration rate is measured.
Portable testers are used to measure friction properties of the pavement surfaces in the field or in the laboratory. The most common of these portable testers are the British Pendulum Tester (BPT) and the Dynamic Friction Tester. Pictures of both devices are shown in figures 5 and 6 respectively.
The British Pendulum Tester (ASTM E 303) is a simple device that measures the loss of energy of the pendulum as skids past the pavement surface by determining the loss of height of the swing of the pendulum. The Dynamic Friction Tester (ASTM E 1911) measures the torque necessary to rotate three small rubber pads over the pavement surface at a given speed.

Figure_5pendlum1.jpg
Figure 5 British Pendulum Tester (http://www.highwaysmaintenance.com/skidtext.htm)




Figure_6_dynamic_friction_tester.jpg
Figure 6 Dynamic Friction Tester (http://www.pti.psu.edu/allFacilities)




References:

National Cooperative Highway Research Program, (2009). �NCHRP Web-Only Document 108: Guide for Pavement Friction�. Retrieved March 6, 2011, http://onlinepubs.trb.org/onlinepubs/nchrp/nchrp_w108.pdf
Romero, P. (n.d.) �CVEEN 7570 Pavement Maintenance and Rehabilitation: Lesson #10 - Skid Resistance�. Retrieved March 6, 2011, https://online.uen.org/webct/urw/lc257461621151.tp0/cobaltMainFrame.dowebct




Analysis and Models (Augusto)

PAVEMENT SAFETY EVALUATION

Introduction.
To maintain the safety condition of our roads is a high priority thing. There are thousands of miles of paved roads in the United States that are traveled daily by millions of cars and trucks transporting people and goods. For this reason, the condition of the Nation's roadways plays a very important role in everyday life. Billions of dollars are spent to build, maintain, and improve roadways.

The cost of building a new roadway or rehabilitating an existing pavement can be considerable. If these roads are not repaired, poor pavement conditions can be just as costly to the driving public. Rough pavements can decrease speeds of traffic flow, cause damage to vehicles, and increase the number of traffic accidents. These costs, defined as social costs, are difficult to quantify and unfortunately, are born by the public at large.
To address these concerns, the Federal Highway Administration has developed guidelines for developing a Pavement System.
Development of reliable pavement deterioration prediction models is a challenge to developers. Accurate pavement deterioration prediction models can be a valuable tool to the DOTS to achieve a more efficient highway management.

Due to the challenges of modeling the behavior of pavements, current pavement management's strength depends upon the measurement of existing pavement conditions rather than predicting future conditions of pavements.
Projected roughness trends are a big factor used for evaluation, since pavement roughness is a good indicator of its future performance.

There is also the need for modeling different types of pavements. Portland cement concrete pavements are solid structures (i.e. rigid pavements). Most deterioration models for these structures are fairly accurate because their failure follows a more typical structural pattern.
On the other hand, the deterioration of asphalt pavements is more difficult to predict due to the visco-elastic characteristic of the asphalt.




Accident.gif
Accident.gif

Even though modeling the behavior of asphalt material can be easily done in a pavement laboratory, there are various external conditions that can be impossible to mimic. By including the many variables a roadway pavement endures, such as construction techniques, weathering or aging, the modeling effort becomes even more difficult.



PAVEMENT ROUGHNESS AND SKID RESISTANCE

One of the main objectives of our transportation system is to provide a comfortable ride for users. Roadway roughness is a good indicator of whether this criteria will be fulfilled. A brief look at the historical development of this indicator can be useful. In the 1940�s the roadway longitudinal profiles were measured using an in/mile scale, which was the popular basic unit of measurement. The in/mile scale represents the change in elevation over a given interval. In the 40's the devices used were simple and not as sophisticated and efficient as those used in recent years.

Pavement roughness is generally defined as an expression of irregularities in the pavement surface that adversely affect the ride quality of a vehicle (and thus the user). Roughness is an important pavement characteristic because it affects not only ride quality but also vehicle delay costs, fuel consumption and maintenance costs. Roughness is also referred to as "smoothness" although both terms refer to the same pavement qualities




picture.jpg


Roughness of a pavement surface is commonly correlated to its serviceability. On the other hand, on many occasions, investigators have attributed pavement roughness to inadequate skid resistance (friction) as well. However, current pavement friction evaluation and standardization models have yet to incorporate effects of pavement roughness.

Measurement

Today, roughness is typically quantified using form of either present serviceability rating (PSR), international roughness index (IRI) or other index with IRI being the most prevalent.

Skid resistance is the force developed when a tired is that is prevented from rotating slides along the pavement surface. It is an important pavement evaluation parameter, because the inadequate, or lack of skid resistance could lead to a higher incidents and accidents. The government agencies have the obligation to provide a reasonably safe roadway.


Factors_influencing_skid_resistance.png
Factors_influencing_skid_resistance.png



Measurement

Skid resistance is generally quantified using some form of friction measurement such as a friction factor or skid number (SN). Hence a study was conducted to investigate and quantify the effects of pavement roughness on the skid number (SN) (or 100 coefficient of friction). First, an experimental program was executed to evaluate SN measured from a locked wheel tester (LWT) on pavement sections with similar micro- and macrotexture conditions but different levels of roughness. The measured average SN was seen to be significantly lower on relatively rougher pavement sections. To explain the above observations, a second set of experiments was conducted to study the effect of the normal load on the LWT tire on SN. Statistical analysis including regression and ANOVA was used to validate the nonlinear reciprocal relationship found between SN and the normal load which contradicts the general perception of constant SN with respect to the normal load. Then, a one-dimensional two-degrees-of-freedom vibration model was formulated to incorporate the significant dynamic fluctuations of the normal load of the LWT induced by pavement roughness and the vehicle speed. The variation of the normal load and its nonlinear relation to SN was used to explain lower SN values measured on relatively rougher surfaces. The feasibility of using the international roughness index and the dynamic load coefficient as predictors of the reduction in SN due to pavement roughness was also investigated. Assurance of adequate skid resistance is a vital factor considered in allocating pavement rehabilitation funds at the network level. Since excessively rough pavements also create skid hazards, it is concluded that roughness effects must be considered in pavement management systems not only for serviceability purposes, but also in safety evaluations.

PAVEMENT QUALITY MODELS:
International Roughness Index (IRI):
The IRI is a statistic index that summarizes the surface deviations for just one wheel track. This mathematical simulation uses the quarter car system to generate an imaginary profile. As shown in figure 1, the quarter car system is composed of two parts: a sprung mass representing the vehicle body (where the user is seated) and an unsprung mass representing the set of wheel/tire and half axle/suspension. The sprung mass is connected to the unsprung mass by the suspension, which is simulated by a damper and a spring. The sprung mass is in contact with the real pavement surface by another spring. Lower values represent a smoother ride while higher values indicate rougher one. IRI is easy to collect, also is reliable and repeatable.However limited knowledge as to IRI's relationship with other measures distress



external image fig1.gif



Figure 1. Quarter car simulation.
During the simulation, the quarter car system runs over the longitudinal profile, measured in the field at a constant speed of 80 kilometers per hour (km/h). The roughness over this surface induces dynamic excitation to the quarter car system, generating different vertical speeds ( and ) or accelerations ( and ) in the sprung and unsprung masses. As a result, a relative movement is produced between the chassis and the axle of the imaginary vehicle. The IRI value for a given section length (e.g., 100 m) is computed according to equation 1.
Where:
IRI = International Roughness Index (in mm/m or m/km).
L = length of the section (m).
x = longitudinal distance (m).
V = speed of the quarter-car model (m/s).
x/V = time the model takes to run a certain distance x.
dt = time increment.
= vertical speed of the sprung mass.
= vertical speed of the unsprung mass.
The IRI represents the rectified average slope, or the absolute sum of the relative vertical displacement experienced by the user when driving a fictitious model car over a section (L) of the road at a constant speed of 80 km/h.
A perfectly smooth road results in an IRI value of 0, roads with moderate roughness give IRI values of around 6 meters per kilometer (m/km), and in extreme cases a very bumpy unpaved road can result in IRI values up to 20 m/km.(6) Maintenance intervention threshold varies according to the country, road type, etc. For instance, limit values of 2.7 m/km in the United States, 3.5 m/km in Brazil, 4.0 m/km in Chile, Uruguay, and Spain, and 6.0 m/km in Honduras have been reported. Figure 2 illustrates these threshold differences.


external image fig2.gif

Figure 2. IRI thresholds adopted in different countries.

Standard Deviation of Longitudinal Roughness (σ)
In Japan, the standard deviation of longitudinal roughness (σ) is used to summarize surface deviations. Initially, as illustrates figure 3, the relative difference height �di� (roughness) is measured every 1.5 m, considering an imaginary reference line.


external image fig3.gif

Figure 3. Roughness measurement.

The surface roughness can be measured using a 3-m-long straightedge profilometer (figure 4). Nowadays, laser profilometers are the most used devices. The laser readings are coincident with the desired intervals.

external image fig4.gif

Figure 4. Three-meter-long straightedge profilometer.

During the measurement, hundreds of elevations are registered, thus the heights �di� can be computed according to equation 2.
external image 2.gif (2)
Where:
di = Registered profile heights.
hi, hi-1, hi+1 = Surface elevations.
According to figure 3a, after measuring the height �di� at point B, the beam is displaced to point D, where the heights hi, hi-1, hi+1 are measured at points B, C, and D, respectively. The reference line becomes the line linking points B, C, and D. Using the equation 2 once more, the height �di� at point C can be calculated. When a pavement section is measured continuously, a profile with positive and negative elevations is obtained. These elevations are the relative difference height �di� considering the reference line every 1.5 m.
The longitudinal roughness (σ) is computed through the standard deviation of �di� values, as shown in equation 3.

external image 3.gif (3)

Where:
σ = Standard deviation of longitudinal roughness (mm).
di = Registered profile heights.
nr = Number of registered data.
The Japan Road Association recommends pavement sections 100 m long to calculate σ, whereas the Japan Highway Public Corporation suggests pavement sections 150 m long.

Neural network techniques.-
According to Anderson, the neural network concept started with the work of McCulloch and Pitts in 1943. They developed computational elements based on the physiological properties of biological neurons. Later, Widrow-Hoff developed a linear model called ADALINE (ADAptive LINear Element), which was then generalized for multiple layers and called MADALINE (Multiple ADALINE). The next important development came in 1950 with the work of Rosenblatt, who proposed neural networks known as perceptrons. In this model, the network can learn when fed with examples and the responses can assume continuous values, whereas the original neurons of McCulloch-Pitts operated with only binary numbers.

Artificial neural networks (NN) can be understood as a computational technique that helps to develop nonparametric mathematical models. Different from usual statistics techniques, the models do not explicitly exhibit a set of fitting coefficients or parameters (although they are somehow embedded in the model).
The term �neural network� is a little presumptuous and derives from the fact that earlier models were inspired in the neuronal structure of intelligent organisms. However, the technique is rather simple and easy to use. From a mathematical point of view, a neural network is a simple set of points, called nodes or neurons, arranged in a few consecutive layers. There is necessarily an input layer, one or more intermediate layers, and an output layer. The number of input and output neurons depends on the available data and type of problem, whereas the number of intermediate layers and nodes (the NN architecture) is generally a matter of empirical investigation for each case under study. However, there are recent developments in adaptive NN in which the network architecture is iteratively modified during the learning stage.

The neurons of a given layer are generally linked to all neurons of the next layer, although some connections (called synapses) may be disabled and layers may be bypassed. Information is generally processed from the input layer to the output layer in the feed-forward process. A node or neuron (i) of a given layer (t+1) performs very simple mathematical operations. It multiplies each entry Sj(t) from a neuron (j) of the previous (t) layer by some coefficient (wij) and computes the sum (S) of all entries thus modified.

external image 4.gif

The coefficients (wij) are known as synaptic weights and store the main model characteristics during the learning process.
Later, the previous sum (S) is compared to limit value (�i), known as threshold, for each layer.

external image 5.gif

The value x above will be the argument of function, f(x), known as activation function. The output of this function will be the entry of the neuron for the next layer.
external image 6.gif

Different activation functions may be tried during the development of a model. The most common are the step function and the sigmoid function.
The development of an NN model involves two stages: learning and validation. For that purpose, a sufficiently large number of experimental data must be available. The data with known input and output values are divided into two sets. The first and larger set is used to train the NN and the validation set is used to test the generalization capacity of the trained NN.
The learning stage consists of finding the appropriate synaptic weights (wij) to reproduce the desired output values. The weights are initialized randomly and the computed output values are compared to the desired output values. A root-mean square (RMS) error between computed and desired output values is calculated. Neural networks generally use a learning algorithm known as backpropagation of error. In the backpropagation method, the weights are recalculated from the output layer to the input layer to minimize the RMS error. The process is repeated during several iterations, until a specified error is achieved. This algorithm, also known as the generalized delta rule, is a modification of Widrow-Hoff�s ADALINE, but considers nonlinear activation functions and the entries may assume continuous values. It was developed by Rummerlhart, Hinton, and Williams, and is extremely efficient in minimizing the quadratic error, RMS. Other algorithms are also available.

Once an NN is trained, the validation data set, which was not used in the learning stage, is used to test the forecasting capacity of the model developed. An NN has the property of generalization.
Neural networks are easy to implement, robust even when treating data with some noise, and very efficient, especially when dealing with problems for which a specific knowledge of the underlying mechanisms is not totally available and when analytical formulations are too complicated to be obtained. The use of NN depends on the ability to adapt it to the desired problem by means of appropriate changes in its synaptic weights to enhance its efficiency.
Several commercial and academic programs are available to help develop neural network models. Basically, the user prepares the appropriate input and output files, decides the appropriate NN architecture, and defines a few other analysis parameters. For the analyses in this paper, the authors used a multilayered neural network program called Qnet. The program uses a backpropagation algorithm.


Lineal Regression:

The linear regression method is a favorite method utilized by researchers due to its simple form and ease of implementation. No matter what kind of prediction model is used, the accurate predictions are of importance.
A typical procedure for the modeling of the pavement condition prediction using the regression method includes the following main steps:
� Identify pavement families. Pavements can be classified into families according to factors such as pavement type (flexible, rigid and composite), climate, and geographic region. Pavements within a family are assumed to follow a similar deterioration pattern.
� Identify dependent (response) and potential independent (explanatory) variables. The response variable is the indicator of the pavement condition, such as the pavement condition index and roughness. The potential explanatory variables may include pavement age, thickness of pavement, traffic loading, etc.
� Select an appropriate prediction function form. Linear function is the simplest form. However, other function forms, such as polynomial, S-shaped curve, and power curve, can be used.
� Identify the significant explanatory variables in the model by using the stepwise regression method. The variables are added or removed from the model according to the F values associated with the F-test for the hypothesis.
� Develop and analyze the final model. Statistics associated with the model can be used to justify the accuracy of the model. Validation of the model is necessary before its application to the pavement condition predictions.
� Make predictions for the individual pavements. If the pavement age is the only independent variable used in the prediction functions, the condition of an individual pavement at any time can be predicted by a curve which is adjusted from the family curve and passes through the known pavement condition-age point the linear relationships between roughness and cracking, between roughness and rutting are small for the statistical significance of influence the rate of progression. The extent of rutting influence roughness has more sensitivity than cracking.




Neural_network_model.gif
Neural network model

It is confirmed that the linear relationships between roughness and cracking, between roughness and rutting are small for the statistical significance of influence the rate of progression. The extent of rutting influence roughness has more sensitivity than cracking.

References:

  • Carlos, F., et al., Artificial Neural Network-Based Methodologies for Rational Assessment of Remaining Life of Existing Pavements. 1999, Texas Department of Transportation.
  • Cook, W.D., and A. Kazakov. Pavement Performance Prediction and Risk Modeling in Rehabilitation Budget
  • Planning: A Markovian Approach. Proceedings, Second North American Conference on Managing Pavements,Vol. 2, Toronto, Ontario, Canada, 1987, pp.2.63-2.75.
  • Pavement Design and Rehabilitation Manual, the Ohio Department of Transportation, Columbus, OH, 1999.
  • Spath, H. Clusterwise Linear Regression. Computing, 22 (4), 1979, pp. 367-373.



Ride Evaluation

Methods and Equipment (Jasmin)

Methods and Equipment

Pavement Roughness
Pavement Roughness is defined according to AASHTO PP 37-04 "The deviation of a surface from a true planar surface with characteristic dimensions that affect vehicle dynamics and ride quality". During construction or network inspection a typical road will always experience some sort of irregularities due to poor construction, environmental conditions and heavy traffic flow. Depending on the phase of the project or the age of a network there are procedures and methods to survey the pavement roughness.
Two most common methods of measuring pavement roughness are by taking road surface profile measurements and by vehicle response measurements. Profile measurements are preferably taken on large road networks by vehicular equipment that record elevations while maintaining inertial and or fixed reference. The profile data can also be obtained on a project level by equipment such as a Dipstick® and profilograph. Vehicle response measurement is more focused in determining the quality of pavement drivability and is focused on the user comfort known as response-type road roughness meters (RTRRM). Data obtained from profile measuring equipment is typically interpreted in terms of the International Roughness Index (IRI). The obtained data is used as input in a computer algorithm known as the quarter car simulator. Developed by the World Bank in the 1980's IRI is a popular index made part of ASTM standards such as ASTM E1926-08, ASTM E1364-95 and many others. Figure 1 shows a range values obtained from various road conditions.

fig1.jpg
Figure 1: IRI scale (Romero, 2011)


Dipstick profiler
The Dipstick profile is a one man operated device which measures and records longitudinal road contours in terms of elevation over a given length. The device is relatively inexpensive and easy to operate. From Figure 2, the operator is turning the entire device 180º from one leg to another (legs at the base are 12" apart) making his way along the road segment at an operating speed of 0.2 mph. The onboard computer records data from the internal inclinometer that measures elevation difference between the two legs. Its small size makes it relatively accurate, although the accumulated errors may skew the overall result. The device is intended for road sections with small survey units, not practical for a large survey network. The Intended use of the equipment is at project level.

fig2.jpg
Figure 2: Operator using a Dipstick® (pavemaintenace.org)


Profilograph

Profilograph is used during the construction period of a road to measure acceptable pavement roughness. It is manually operated where the operator is walking the profilograph along the road segment. The apparatus is low cost equipment and currently there are two types of profilographs in use today, California truss-type and Rainhart type. The main difference between the two types, as it can be seen from Figure 3 & 4, is that the Rainhart model uses twelve wheels arranged in four groups of three.


fig3.jpg
Figure 3: Typical California truss-type profilograph (FHWA)


The basic functional outline of the profilograph is the use of smaller support wheels on the outside establishing a reference plane while the road variations are measured by the larger center wheel. The center wheel is allowed to pivot with the road irregularities and the data is recorded and stored to a onboard computer device. The device is not intended for measuring roughness on a large network and lacks precision. The Intended equipment use is at project level.

fig4.jpg
Figure 4: Typical Rainhart type profilograph (FHWA)


Profilers

A vehicle equipped with profile metering hardware design to measure road roughness at large network levels. Depending on the manufacturer the costs can range from $50,000 to $220,000. The system requires two operators, one to drive the vehicle and the other operates system functions. The onboard systems are able to record data at speeds anywhere from 10 to 70 mph. Depending on the setup a typical vehicle is equipped with profilometers, accelerometers, a non-contact measuring system (laser or acoustic), a GPS unit and a computer for data storage (see Figure 6).

fig5.jpg
Figure 5: Profilometer van (FHWA)


The inertial profilometers measure the road surface profile and rut depth. The accelerometers reference the plane by accounting for the vertical displacement of the vehicle. The light or acoustic devices used are to measure relative displacement between the accelerometers and the road surface. The road profile measurements are taken at 1, 2 or 6 inches over a 12 inch interval. Rut depths are taken every 3 ft and average is recorded at 100 ft intervals. The vehicle setup is versatile and can be configured to record various roughness indices, wheel path rutting cross-slope, and other viable parameters. Profile data and roughness index are computed and available immediately after the section has been surveyed. The equipment is not intended for at project level basis.


fig6.jpg
Figure 6: Schematic of Profilometer van (FHWA)


Response-type Road Roughness Meter

The response-type road roughness meter is designed to measure the ride comfort of the driver. The equipment on the vehicle is designed to measure the bounce response of the vehicle due to road roughness.


fig71.jpg
Figure 7: RTRRMS vehicle with Mays Meter (FHWA)


Onboard of the vehicle are devices used to measure relative axle body motion and acceleration. Shown in Figure 7 is a most commonly used Mays meter which typically comes equipped with the rotary transducer, the pavement condition recorder, and the distance measuring instrument. The rotary transducer converts the axle/body movement to an electrical signal (FHWA). The distance measuring instrument is an electronic odometer (FHWA). The pavement condition recorder is a microprocessor, which accepts input from the rotary transducer, the distance measuring instrument, and the keyboard processes the various signals into an output (FHWA). This output is commonly in the form of accumulated inches of relative motion over a distance (FHWA). Roughness measurements vary from one system to another and prove to be inconsistent with time. The system price tag ranges from $8,000 to $10,000 depending on manufacturer and it has been in use for the last 50 years on a network level.





References:


Romero, P. "CVEEN 7570 Pavement Maintenance and Rehabilitation: Lesson #12-Roughness" Spring 2011, https://online.uen.org/webct/urw/lc5986969246411.tp5986969268411/displayContentPage.dowebct?pageID=7081236601331&resetBreadcrumb=false&displayBCInsideFrame=true


Joseph Budras P.E. - August 2001 "A Synopsis on the Current Equipment Used for Measuring Pavement Smoothness" Federal Highway Administration Record, http://www.fhwa.dot.gov/pavement/smoothness/rough.cfm


Pavement Interactive Organization, http://www.pavementinteractive.org


FHWA-HRT-05-068 "Achieving a High Level of Smoothness in Concrete Pavements Without Sacrificing Long-Term Performance CHAPTER 2. PAVEMENT SMOOTHNESS MEASUREMENTS" Federal Highway Administration publication, http://www.fhwa.dot.gov/pavement/pccp/pubs/05068/02chapter2.cfm





RIDE EVALUATION - ANALYSIS AND MODELS
By Vernon M. Black

In order to evaluate the ride quality of a given pavement, it is necessary to somehow quantify and measure the principal component of the motorist's experience - which is the roughness of the pavement surface.
Roughness is generally defined as an expression of the irregularities in the pavement surface that affect the quality of the ride inside a vehicle. Several different methods have been determined which are used to measure the roughness of a pavement. All of these methods can be grouped into one of two categories of measurement type: profile or response.
Profile measurements typically consist of a device which measures the ups and downs of the pavement using some fixed or inertial reference point as a datum. A profilograph is the generic term used to describe one type of instrument which is used to measure the profile of a pavement surface. Straight edges, inertial reference systems, and laser measurement systems can all be used to measure the level of roughness. Each different measuring device has its own quirks, and gives different values, which can make it difficult to directly compare roughness levels in different areas of the world. Therefore, in the 1980s, the World Bank backed an effort to develop what is known today as the International Roughness Index, or IRI. The IRI is an internationally-accepted measure of roughness, for which there are a number of correlation equations which relate the roughness measured by any given instrument to an IRI number.
Response measurements typically quantify how a given vehicle or user responds to the roughness. The Present Serviceability Rating, or PSR, is an example of a response measurement. This particular measurement is based on individual observations � a motorist gives the pavement ride quality a subjective score from 0 (impassable) to 5 (excellent), based upon their experience driving on it. The main disadvantages with PSR ratings is that they are very subjective

IRI Analysis and Model
In order to perform an analysis of the profile roughness of a given pavement, it is first necessary to take field measurements of the profile of the surface for which the profile is desired. Numerically, there are three ingredients to any roughness/profile measurement:
1. A Reference Elevation
2. A Height Relative to the Reference
3. Longitudinal Distance
There are several methods that can be utilized to obtain this information, of which the following are but a few examples:
  • Use of a rod and a level - typical surveying equipment. Although it may be possible to obtain very accurate information, this is a very laborious procedure for this application.

Figure_#1.png



  • Use of the Dipstick - a trademarked device that can be "walked" along the line being profiled, which includes an onboard computer which automatically records the profile measurements. Although far less laborious than the rod and level approach, this device still requires the measurements to be made at a "pedestrian" pace.

Figure_#2.png



  • Use of an inertial profiler - a device that utilizes laser, infrared or ultrasonic sensors, accelerometers and a computer mounted in a vehicle in order to measure the profile at higher speeds. This is far more practical at the network level.

Figure_#3.png




The overall profile measured by an inertial profiler may be significantly different from that measured by the dipstick. However, the IRI values that are obtained from each measuring device are remarkably similar.


References:
  • Sayers, Michael W., et al., The Little Book of Profiling. 1998. University of Michigan.
  • Finn, Fred, Pavement Management Systems. Past, Present and Future. 1998. U.S. Department of Transportation. Federal Highway Administration, Washington D.C.
  • Romero, Pedro, Lesson #12. Roughness. 2011. University of Utah.
  • International Roughness Index. 2011. Wikipedia.org. Permanent Link = <http://en.wikipedia.org/w/index.php?title=International_Roughness_Index&oldid=411970033>.




Pavement Inventory and Forecasting


Once the data for the assessment of the condition for a given pavement network has been determined, the next step is to present the data in a logical manner so that those interested can make the proper decision. Obviously the way the data is presented will depend on the intended use; namely Network or Project Level decisions.

Network versus Project Level Decisions (Dave)



In order to effectively manage a pavement network, DOT's, local agencies and engineers must first develop a comprehensive Pavement Management System (PMS) which encompass all the required pavement planning aspects. On a large scale this can include the design, construction, pavement maintenance and utilities maintenance of an entire network. Evaluation and analysis at this level can be difficult due to the uncertain interrelated nature of these variables. To simplify the analysis it is required to establish a management and decision making criteria at both the network and project levels. Software packages such as MicroPave, Excell and Acccess are usefull in managing the large amount of usable data included in such a network.

This pavement management system is analyzed at two general management levels, the network level and the project level. A PMS is based on input data from both management levels can be more effective in meeting the entire network needs, providing a well-manage, optimized system.

Network Level


At the network level, pavement management relies on a combination of procedures, data, software, policies and decisions to produce solutions that are optimized for the entire pavement network. In essence, a network-level approach uses aggregate data such as traffic loading, safety, inventory and pavement condition to identify the best overall solution. Then, project-level decisions are made using target strategies identified at the network level. In order to accomplish this, the network-level approach requires large amounts of data, accurate aggregation, computer models and trained individuals.

Network-level approaches are very powerful and can produce optimum solutions for the entire defined system. Key elements in a network-level approach are:

  • System definition. The network-level approach optimizes solutions for the defined system. If the system is incorrectly defined, solutions will not be optimal. For instance, many consider the appropriate system for pavement management to be the larger transportation infrastructure rather than the more limited pavement network.
  • Network model. Network-level decisions, and thus all decisions, are based on outputs from a complex simulation model. Thus, these decisions are only as good as the model. Inputs, accuracy, sensitivity, assumptions and calibration must be known and considered when selecting an appropriate model.

The network-level approach is powerful and sophisticated, but requires large amounts of data/resources and attention to detail in order to function in a meaningful manner.



The project-level approach to pavement management uses a bottom-up methodology to combine methods, procedures, data, software, policies and decisions to produce network solutions. In essence, a project-level approach first uses similar data classes as compared with the network level but in individual sections as opposed to aggregate sections. Because the initial decisions are made at lower levels they tend to drive the overall network solution, which then may or may not be optimized for the entire pavement network. Thus, network priorities are enforced by the inclusion or exclusion of projects or by relying on the compatibility of project-level decisions with network-level goals.
Project-level approaches can be very useful and constitute the majority of pavement management systems in operation today. The key elements in a project-level approach are:

  • Project-level vs. network-level goals. Since decisions are made first at the project level, the project-level approach requires more effort to coordinate the anticipated or promulgated network-level priorities down to the project level.
  • Project ranking. This will determine included/excluded projects based on network-level goals.

Many State and local pavement management systems can be classified as project-level. Although less capable at producing optimum solutions and conditional scenarios, the project-level approach is advantageous because it maintains the detailed project-level information needed to make fully informed project-level decisions.



Pavement management systems assist decision makers in determining cost effective strategies for maintaining, upgrading and operating a network of pavements. There are two general levels to pavement management:

  1. The network level. Deals with the pavement network as a whole. Structure is built to support high-level decisions relating to network-wide planning, policy and budget.
  2. The project level. Deals with smaller constituent sections within the network. Structure is based on lower-level decisions relating to condition; maintenance, reconstruction and rehabilitation.

Both the network-level and the project-level approaches have significant advantages and disadvantages. In general, the advantages offered by the network-level approach, such as network optimization, conditional scenarios, consistency and versatility are most suitable to a federal, state or large municipal pavement management system. These agencies have large enough networks that manual or ad hoc methods of network optimization, conditional scenario generation and maintaining consistency are prohibitively difficult. Additionally, these larger organizations are better able to handle the large data and resource requirements of a more sophisticated system. Conversely, the advantages offered by the project-level approach, such as simplicity, economy and familiarity are more suitable for smaller agencies such as local municipalities who generally have small networks, few resources and minimal information requirements.


Reference:

Washington Apshalt Pavemtn Assicitation.

www.asphaltwa.com/



Pavement Condition Indices (Eric)
What are the common indices used to summarize all data into one number.

Pavement Condition Indices
Pavement maintenance and rehabilitation is a field of study that is subjected to demands and constraints from engineers, politicians, the public, scientific principles, data (too much or too little) budgets, weather, traffic, and judgments. As pavement maintenance consultants or experts there is often a lot that is lost in communication when trying to convey an idea to a politician or a member of the public or even another engineer when trying to tell them to spend a dollar on pavement maintenance rather than some other public works venture. It is not enough to tell a layman that the deflection of the FWD test indicates that the pavement is weakening and will need maintenance dollars. We need to tell them something that is quantifiable and comprehensive; we need to use an index.
Simply put, an index is an arbitrary score given to a pavement section based on a compilation or a summarization of several different inputs. For Example, it is easier to relate and understand that the condition of a new pavement is 95 it and when it reaches 70 it will need some maintenance that cost $1 but if we wait until it reaches 50 it will cost $100.
An index is not only used to translate complex technical details into a compiled number it is also used to make sense of the data as it translates to the use of the road pavement in regards to the smoothness of the ride and the perceived quality of the road by the users or drivers. Indices make the measured data more usable. Indices make the decision of what projects to undertake clearer. Also with the historical use of indices it helps estimate the costs of maintenance based on when it is performed.
There are three commonly used indices in pavement evaluation, they are the International Roughness Index (IRI), the International Friction Index, and the Pavement Condition Index (PCI).

International Roughness Index (IRI)
The IRI is a measure of the roughness of the ride as perceived by the riders in a vehicle. The lower the roughness index the smoother the ride. The IRI procedure is described above. The IRI is useful because the roughness of the ride will most likely be the source of most of the complaints from the public. If the IRI is high for a certain part of the network then it can be expected that the public is not happy with that stretch of pavement. The roughness of a road does not necessarily have to do with the structural integrity of the pavement. A pavement section can have good structure but a poor ride or vice-versa. It is common that the ride quality is directly related to the structural condition of the road, however, because the conditions that lead to a rough ride also indicate a failing pavement structure. Figure 1 shows an ARAN (Automatic Road ANalyzer) profilometer produced by ROADWARE Corp. of Canada. The ARAN device scans the road and outputs the IRI.
ARAN.jpg

Figure 1 ARAN (source http://dcnonl.com/article/id26632)


The relationship equation for IRI is shown below and as can be seen it is dependent on many factors such as rutting, alligator cracking, transverse cracking, frost, swell, and soils.

IRI = INI_IRI + 40.0*MRUT + 0.4*CRACK + 0.008*TRANS_CK + 0.015*SF

Or from the MEPDG:
IRI = INI_IRI + 40.8*MRUT + 0.575*CRACK + 0.0014*TRANS_CK + 0.00825*SF

Where:
INI_IRI = initial IRI, in/mi (use MEPDG default of 63.4 in/mi)
MRUT = total rutting, in
CRACK = alligator cracking, percent lane area
TRANS_CK = transverse cracking, ft/mile
SF = FROSTH + SWELLP*AGE1.5
FROSTH = LN([PRECIP+1]*FINES*[FI+1])
SWELLP = LN([PRECIP+1]*CLAY*[PI+1])

International Friction Index (IFI)
International Friction Index (IFI) is an index that indicates the friction available between the tire and the road. The physics of this friction is explained above. The IFI provides a standard to reporting how the sliding friction of a tire on the pavement is reported. The importance in this index lies in the safety of the pavement. A pavement that lacks friction is not as safe. If testing on a network level indicates sections of pavement that have low friction it should be a priority to restore the friction as soon as possible.
The IFI was produces as a result of various types of friction tests including locked-wheel, fixed-slip, ABS, variable-slip, side-force, pendulum, and some prototype devices. Surface texture was measured by means of the sand patch, laser profilometers (using the triangulation method), an optical system (using the light sectioning method), and outflow meters. The IFI is composed of two numbers—F(60) and SP—and the designation and reporting of this index is IFI(F(60), Sp). The F(60) is the friction reported at 60 km/hr and the Sp is a measure of macro-texture influence on surface friction. Both the F(60) and Sp are measure d using various techniques. The IFI is typically given by the following equations and shown in figure 2 and the graph in figure 3.
figure2.png

Figure 2 IFI equations (source: http://onlinepubs.trb.org/onlinepubs/nchrp/nchrp_w108.pdf)

figure3.png

Figure 3 IFI Friction Model (source: http://onlinepubs.trb.org/onlinepubs/nchrp/nchrp_w108.pdf)

Pavement Condition Index (PCI)

The pavement Condition Index (PCI) is a visual distress survey. This means that the pavement is visually inspected for distresses and each distress has a deduct value based on its type and severity. A perfect road with no distresses will have a PCI of 100. Each distress observed takes points away from the value. The PCI is not simply a measure of how good the pavement looks. Each of the distresses may be an eyesore, yes, but its purpose is to translate the visual distresses into an aggregate index of overall condition-structural condition, roughness condition, and visual condition. Figure 4 shows the PCI as related to maintenance operations.
PCI_curve.jpg

Figure 4 PCI over time (source: http://www.cecer.army.mil/paver/Paver.htm#top)

Pavement Condition Index (PCI) developed by the Army Corps of Engineers and used by MicroPAVER and other similar programs. Figure 5 shows an example screen-shot of MicroPAVER software. MicroPAVER was originally was developed in the late 1970s to help the Department of Defense (DOD) manage maintenance and rehabilitation for its vast inventory of pavements. It uses inspection data and a pavement condition index (PCI) for consistently describing a pavement's condition and for predicting its M&R needs many years into the future. The PCI for roads and parking lots became an ASTM standard in 1999 (D6433-09).

figure5.png
Figure 5 MicroPAVER interface example (source: http://www.cecer.army.mil/paver/Documents/AboutPave)

References:

National Cooperative Highway Research Program, (2009). �NCHRP Web-Only Document 108: Guide for Pavement Friction�. Retrieved March 6, 2011, http://onlinepubs.trb.org/onlinepubs/nchrp/nchrp_w108.pdf

Romero, P. (n.d.) �CVEEN 7570 Pavement Maintenance and Rehabilitation: Lesson #17 - Condition Assessment. Retrieved April 20, 2011 https://online.uen.org/webct/urw/lc257461621151.tp0/cobaltMainFrame.dowebct

Romero, P. (n.d.) �CVEEN 7570 Pavement Maintenance and Rehabilitation: Lesson #7 - Performance Models. Retrieved April 20, 2011 https://online.uen.org/webct/urw/lc257461621151.tp0/cobaltMainFrame.dowebct

Prediction Models

Once the present data is summarize, it needs to be extrapolated to a future date when the actual maintenance needs to be carried out. The models used to forecast the pavement conditions are divided into stochastic (i.e., statistics based) and deterministic (mechanics based)

Stochastic Prediction Models (Fabiola)

Planning and managing activities for a large network of transportation infrastructure is a daunting task. Many projects and interests compete for the limited resources allocated to a transportation agency and infrastructure management is only one of such competing interests. How much resources to allocate to transportation infrastructure and how to get the best value for the allocated resources have received high priority by top management officials of these agencies. The decision makers who have to make these types of choices often do so based on a number of criteria. Such criteria include limited budget for capital and recurrent expenditure, the need to keep the transportation network open at an acceptable level of service, etc

Probability theory has been used for assessing life-cycle costs for infrastructures. Piyatrapoomi et al. note that there are several researches done in this area (Kong and Frangopol; Zayed et al.; Kong and Frangopol; Liu and Frangopol; Noortwijk and Frangopol; and Novick, as cited in Piyatrapoomi, 2004). However it is evident from the literature that there is very limited information on the methodology that uses the stochastic characteristics of asset condition data for assessing costs for pavement maintenance. Salem et al. (2003) have used a risk-based approach using probability theory and data input modeling to predict probabilities of occurrence of different life cycle costs associated with the construction/rehabilitation of an infrastructure unit. Markovian Decision Process has been a popular tool among various researchers modeling asset management decision problems because of its ability to include stochastic nature of pavement deterioration (Golabi et al., 1982; Kostuk, 2003). It has been widely accepted now that pavement deterioration must be represented probabilistically and not deterministically because of the following reasons as outlined by Kostuk (2003).

The mechanistic causes of pavement distress are not well known; one can model stress, strain, and deflection but not fatigue, cracking, rutting and other pavement characteristics. Pavement is under a continuous influence of random variables such as traffic loads, utilization and weather conditions. Pavement being a heterogeneous material performs differently at different sections.




MODELS:


MARKOV THEORY APPLIED TO PAVEMENT DETERIORATION MODELING
Essentially, a stochastic process is a collection of random variables. In statistical terms, a random variable is one of the possible outcomes of an experiment, together with its associated probability of occurrence.
When dealing with pavement deterioration the requirement may be to determine the condition at a particular location on the road network. The random variable would then be the pavement condition at that location coupled with the associated probability of finding it in that condition. The main components of the stochastic process are, therefore, states (condition) and transition probabilities. The transition probabilities specify the likelihood that the pavement will move from one state to another. However, at the network level, the probability that a pavement is in a given condition is interpreted as the expected proportion of pavements in that condition, thereby allowing the proportion of the network expected to be in a certain condition to be calculated:
The Markov prediction model is a specific type of stochastic process and is governed by three ‘restrictions’. It is possible to show that the Markov process may be used in the determination of pavement deterioration as it approximates to these three restrictions as follows:

(i) The stochastic process should be discrete in time. Although pavement deterioration is ‘continuous’ in time, it is possible to consider it as being discrete in time, as it is common to analyse road network condition at specific points in time, usually annually.

(ii) The stochastic process should have a countable or finite state space. Although the state space (that is, the theoretical number of possible outcomes) is infinite in pavement deterioration, this is overcome by expressing the range of possible outcomes as a set of discrete states. That is, the state space is defined as a relatively small number of fixed bands of condition for the particular defect under consideration.

(iii) The stochastic process should satisfy the ‘Markov property’, where this means that the future state of the process depends on its present, but not past, state. In pavement deterioration it is assumed that the Markov property holds.

Furthermore, by applying the Markov model over a series of years, it is possible to predict the proportion of the network in any state in any future year. Such a series of Markov predictions is called a Markov chain.
In addition to these three restrictions, a discrete-time Markov chain is said to be ‘stationary’, or homogeneous, in time if the probability of going from one state to another is independent of the time at which the step is being made4. In this case it is considered that the road network will deteriorate following the transition probabilities of a single transition matrix. However, if the pattern of deterioration of a particular road network is likely to change at a certain point in time, t, the deterioration process may be modelled by a ‘nonstationary’ chain. This implies the use of a different transition matrix before and after t. In this case, the distribution of condition at t will become the starting distribution for the second chain, which will operate with a different transition matrix. This type of arrangement may be performed as many times as required, thereby allowing account to be taken, for example, of changing traffic patterns.




STOCHASTIC, SINGLE-OBJECTIVE NETWORK-LEVEL PMS
A stochastic single-objective network-level PMS model is a dynamic model where the pavement deterioration process is stochastic (semi-Markovian) in nature. This is a more appealing method of modeling pavement deterioration since there are many factors, ranging from weather condition, to traffic flow levels, to material characteristics, that make the process of pavement deterioration probabilistic in nature. The deterioration process is therefore, represented by transition probabilities. In general terms a transition probability, Pij, represents the probability that a pavement section will deteriorate from condition (e.g., distress or roughness level) i to condition j in one year. Table 1 shows an example of such probabilities, grouped in a matrix. The matrix in table 1 is one of the pavement distress deterioration models that were developed for New York State highway system [11]. Based on such a transition matrix, an average distress deterioration curve such as that shown in Figure 2 can be derived.

Table 1. One-Year Markov Transition Probability Matrix for a Flexible Pavement:
(Low trafficked asphalt concrete road after 2.5-3" overlay & preventive maintenance)

matrix1.jpg


Grafica_1.png

Figure 2: Average Performance Curve for a Flexible Pavement in New York State

Figure 3 gives a rough idea of the distress scale (pavement surface rating – PSR, ranging from 1 to 10) used in New York State and its qualitative rating. Historical data from this distress scale were used to develop the stochastic deterioration model. Detailed discussions of how to estimate such transition probabilities from historical data are presented in Mbwana [8] and Mbwana/Meyburg [11].


graf_3.png

The following is an example of a single-objective, network-level (or long-term) stochastic PMS model. This is a distress-based model that minimizes agency-cost (single-objective) subject to a number of constraints. In this model the pavement condition modeled by the dynamic (time-variant) stochastic process is distress

example.png




This policy states that, in the long run Dtcia percent of road segments of category c, carrying
traffic level t will be in distress level i and repair action a will be applied. Another way of 345looking at these policies is that, whenever road segments of category c, carrying traffic level t are in distress level i apply repair action a Dtcia percent of the time. Please note that since the
transition probabilities are modeling pavement deterioration in distress units, the resulting policies are distress-based. This means the optimal maintenance policies will use pavement distress levels as trigger values for repair actions.
For the sake of comparison, let us look at a roughness-based PMS model. The single objective used in this case is user-cost. Since roughness (rather than distress) is a better determinant of user-cost, it is preferable to use pavement distress as the dynamic process that drives the model. Therefore, the pavement deterioration matrix, P, is in roughness scale.

example_2.png


The optimal solution Ytcla from this model will generate long-term roughness-based policies Rtcla where:


Optimal maintenance policy Rtcla = Ytcla 100 ..............................16 Ytclb

The roughness-based policy Rtcla has similar interpretations as the distress-based policy Dtcia .
The only difference is that the former is used pavement roughness levels to trigger repair actions while the latter used pavement distress levels to trigger actions.
As discussed earlier, if one wishes to find pavement maintenance policies that minimize both objectives (i.e., agency-cost and user-cost) the best way would be to develop a multi- objective model that will generate an efficient frontier of non-dominated policies. These non- dominated policies can then be viewed as alternative policies from which different decision makers may choose differently, based on their respective criteria. Later, the paper discusses a technique by which the decision makers can rationally choose from a set of non-dominated policies generated by the multi-objective PMS.



STOCHASTIC, MULTI-OBJECTIVE NETWORK-LEVEL PMS
This section presents a framework for developing a stochastic multi-objective PMS model. To illustrate this framework, an example formulation is used in which two objectives (agency-cost and user-cost) are to be minimized. It is assumed (as it is indeed the case) that pavement roughness and traffic volume will be the major determinants of user-cost while pavement surface distress and repair type will be the determinants of agency-cost. Since there is no clear relationship between pavement distress and pavement roughness, the model will include deterioration models for both conditions as two separate stochastic processes. Even though distress and roughness will be presented as separate processes, it should be noted that these processes are being driven by the same set of decision factors, namely the repair actions applied. The two objectives will be denoted as O1 (agency-cost) and O2 (user- cost), respectively.





ex_4.png


ex_5.png


Given the above multi-objective network-level PMS model one can generate a number of non-dominated (Pareto-optimal) solutions that form the efficient frontier. Techniques such as the weighting method or the constraint method, which are outlined in several multi-objective analysis texts, like that by Goicoechea et al. [6], can be used to accomplish this task. The main focus of this paper however, is not to discuss the generation of the efficient frontier, but rather to explore the implications of:
  • a dual policy (distress-based, Dtcia , and roughness-based, Rtcla ) existence for each non-dominated solution to the above problem, and

  • rational choice by decision makers from various non-dominated policies resulting from the non-dominated solutions generated by the above PM model.


REFERENCES
  • Transportation Research Board
  • Journal of Transportation Engineering @ ASCE
  • Reliability-based Model for Estimating Long Term Pavement Maintenance Contracts Under Performance Specifications.


Deterministic Prediction Models


(Jesse Ralphs)

Models are used as part of a pavement maintenance strategy to forecast the condition of the pavement. Models allow us to evaluate trends and predict future conditions so that sound decisions can be made with regards to maintenance actions and budget development. A good model will allow us to conduct ‘what if’ scenarios so that the effect of different alternatives can be evaluated (Romero, n.d.).

Deterministic models of pavement performance are relationships composed of the variables understood or assumed to influence pavement performance (Martin, 1996). The future pavement condition is predicted as a precise value using mathematical functions that are derived from observed or measured pavement deterioration using mechanistic, regression, or mechanistic-empirical methods (Abraza, 2004). In contrast to the stochastic prediction models discussed above, deterministic prediction models take no account of the probabilistic nature of pavement performance.

Deterministic models are defined on the basis of their derivation, and are classified into three main model types: Mechanistic, Mechanistic-Empirical, and Empirical (Martin, 1996):
  • Mechanistic models are based on a fundamental and primary response approach to predicting pavement performance, such as elastic theory.
  • Mechanistic-empirical models are based on theoretical postulations about pavement performance, but are calibrated by observational data using regression analyses. These models must adhere to known boundary conditions and physical limits. These models can incorporate interactive forms of distress near the end of pavement life, such as the interaction of rutting with cracking, when these interactions are well understood. If these models are theoretically sound and correctly calibrated, they may be applied beyond the range of data from which they were developed.
  • Empirical models are developed from regression analyses of experimental or observed data. These models are useful when the mechanism of pavement performance is not understood. Empirical models should not be used beyond the range of data from which the model was developed.

PSI
The most common indices describing the condition of the pavement are friction (IFI), roughness (IRI), structural condition (modulus), and visual distresses (PCI). An alternative to developing a separate model for each of these indices is to combine each index into a single value. An example of this is the present serviceability index (PSI) which was developed during the AASHO road test in the 1950’s (Romero, n.d.).

ralphs_EQ_PSI.jpg

Where:
SV = slope variance (a measure of roughness)
C = cracking in linear feet per 1000 ft2
P = patching in area per 1000 ft2
RD = mean rut depth

RSL
A more current approach to predicting future pavement performance is the Remaining Service Life Method (RSL). Service life of a pavement is the period over which a pavement section adequately performs its desired function or performs to a desired level of services. For an existing pavement, RSL is simply the amount of service left (FHWA, n.d.).

Traditionally pavement condition has been defined based on only the structural and functional condition of the pavement—the period of time under specified site conditions during which a pavement’s structural or functional condition is expected to remain within stated limits, provided that appropriate routine and preventive maintenance are carried out. Examples of stated design and site conditions include:
  • Demands actually placed on the pavement while in use (e.g., traffic applications)
  • Environmental conditions (e.g., harshness of the environment [freeze-thaw or wet-dry])
  • Planned routine, preventive, and corrective maintenance and other preservation activities
  • End condition level of the pavement
Figure 1 shows an example of how future pavement performance is forecast and used to estimate pavement RSL.
ralphs_Example_of_Pavement_RSL_FHA.jpg
Figure 1: Example of pavement RSL (based on IRI only) (FHA, n.d.)


Current practice is to apply a multi-condition approach for determining pavement RSL. Commonly used pavement performance measures/indicators used to characterize pavement structural and functional condition are presented in Table 1 (FHWA, n.d.). With these performance measures, pavement RSL can be determined to satisfy varying objectives such as remaining functional life, remaining structural life, remaining service life (overall), etc.

ralphs_table1.jpg

Using multiple performance measures, RSL is determined first for each individual pavement performance measure as follows:

ralphs_EQ_RSL_General.jpg
Where

RSLPMX = RSL based on performance measure X
AgeTD = Forecasted pavement age at threshold value of performance measure X
AgeNOW = Current pavement age.

Overall pavement RSL for a given section can be determined using the two methods presented below.

Method 1:
Overall pavement section RSL is basically the minimum of all the individually computed RSLs for each performance measure considered. It is defined mathematically as follows:
ralphs_EQ_Overall_RSL.jpg
Method 2
Overall pavement section RSL is basically the weighted average of all the individually computed RSLs for each performance measure considered. It is defined mathematically as follows:
ralphs_EQ_RSL_Method_2.jpg
Where:
Wi = Weight assigned to a given performance measure i
RSLPMi = RSL computed for given performance measure i.

A summary of individual performance measures and associated models used to characterize pavement condition is shown below in Table 2. For more details and specific equations please visit http://www.fhwa.dot.gov/pavement/healthtrack/pubs/technical/pht03.cfm.
ralphs_table2.jpg

The specific steps that used to determine RSL for a given pavement section are summarized below:
1. Select analysis type and parameters
a. Analysis type:
i. Minimum RSL
ii. Weighted average
b. Analysis Parameters:
i. Assign weighting factors to each distress/IRI to be used to determine overall RSL
2. Extract from the Highway Performance Monitoring System (HPMS) program and default databases all relevant data for the given pavement section.
3. Determine current age and condition (distress/IRI)
4. Use relevant pavement condition (distress/IRI) models to forecast future pavement condition (distress/IRI). The predicted future distress/IRI is adjusted using actual distress/IRI data if available in HPMS. This ensures that current pavement condition is projected into the future, minimizing potential errors in future condition predictions.
5. Determine pavement age at which each individual distress/IRI threshold value is attained.
6. Estimate RSL for each individual performance measure (distress/IRI) of interest.
7. Determine overall pavement section RSL (note that computed RSL cannot be greater than the maximum pavement service life).
a. Minimum RSL
b. Weighted Value
8. Generate reports and graphs as needed
a. Histograms
b. GIS Plots

Techniques for Model Development
When it becomes necessary to develop a new model, the general rule is to use the simplest model that explains the data.

Straight Line:
The straight line technique is the simplest method and consists of a straight line. At least two points are required: initial condition and current conditions. The rate of change is assumed to be constant.
A more sophisticated approach consists of transforming the data so that a linear trend can be developed but applied to a more complex behavior. Such equations look like this:

ralphs_EQ_straight_line_eq.jpg
Polynomial:
Depending on the amount of data available, a polynomial equation might be used to fit the data. A polynomial takes the following form:
ralphs_EQ_polynomial_1.jpg
Under certain circumstances the polynomial approach may result in a positive slope, meaning that the conditions are improving. To prevent this from happening constraints can be added to the equation such as always having negative slope.
ralphs_EQ_polynomial_2.jpg
Both the linear and polynomial models are based on least squares regression and do not take into consideration the expected behavior of the material.

S-shaped Curve:
An alternative to purely regression models is to incorporate the expected behavior of the material into a predetermined function. In the case of a decreasing value at an increasing rate, the most common one being the S-shaped curve. For example:
ralphs_EQ_S_shape.png
Where ρ, α and β are regression constants.

Once a model has been verified, it can be used to predict the time interval to reach the trigger value for whatever action is appropriate. In general, the trigger value should be selected to result in the lowest possible maintenance cost (Romero, n.d.).


References:
Abraza, K (2004) Deterministic Performance Prediction Model for Rehabilitation and Management of Flexible Pavement. The International Journal of Pavement Engineering, Vol. 5.

Martin, T (1996) A Review of Existing Pavement Performance Relationships. Research Report ARR 282. Retrieved April 18, 2011, http://www.arrb.com.au/admin/file/content13/c6/ARR%20282%20Review%20of%20existing%20pavement.pdf

Romero, P (n.d.) CVEEN 7570, Lesson #23: Adapted from FHWA Pavement Health Track (PHT), Remaining Service Life (RSL) Forecasting Models, Technical Information: Forecasting Models. Retrieved April 19, 2011, https://online.uen.org/webct/urw/lc257461621151.tp0/cobaltMainFrame.dowebct

US Department of Transportation: Federal Highway Administration (FHWA) (n.d.) Pavement Health Track Remaining Service Life (RSL) Forecasting Models, Technical Information. Retrieved April 19, 2011, http://www.fhwa.dot.gov/pavement/healthtrack/pubs/technical/pht03.cfm


Maintenance Strategies


Once the present and future pavement conditions are known, the next step in the evaluation is to select the proper maintenance strategy.

Asphalt Pavements



Design Methodology (Levi)


An overlay is a maintenance strategy that consists of placing a layer on top of the existing pavement structure. Overlay can be used to strengthen existing pavement, improve ride quality, reduce safety hazards, and improve profile. Overlays can be divided into two groups: structural and non-structural.

Non- structural overlays are dense graded asphalt mixtures, usually less than 2-inches thick. This is the highest type of maintenance treatment available. It can even be considered a rehabilitation strategy. A non-structural overlay will protect the pavement structure and replace to top surface when rutting has occurred.

Structural overlays are used, as the name implies, to correct for structural deficiencies in the pavement. They are used to correct for fatigue cracking, rutting (attributed to the structure), allow for excess traffic, etc.

Steps in Overlay Design
1- Structural Evaluation
2- Traffic Evaluation
3- Thickness Design
4- Surface Preparation
5- Construction

Structural Evaluation:
The most effective way to evaluate the structural condition of a pavement is by using a falling weight deflectometer (FWD) deflections to backcalculate effective layer modulus. This process consists in finding a set of layer modulus that will deflect in a manner consistent with the measured values. The modulus is then used in mechanistic design and related to specific failure criteria. This process requires use of a layered design and related to specific failure criteria. This process requires use of a layered analysis program similar to the ones used in pavement design.

If no layered analysis program is available, the area method can be used. The area method provides an index of structural condition of the pavement. However, it does not provide enough information to evaluate the condition of individual pavement layers.

FWD_Deflection_Graph.PNG

The area is calculated as 6*(D0 + 2D1 + 2D2 + 3D3) / D0

Thus, an area of 36 indicates a structurally intact pavement. A criterion can be set based on previous data and its relationship to 36. The Asphalt Institute provides the following typical values to use as a guideline for deciding structural adequacy.

Area_Table.PNG

An area < 15 indicates a very weak pavement.

Traffic Evaluation:

Traffic can be evaluated based on existing traffic counts adjusted for growth over the design period. Traffic counts must be adjusted based on percent of trucks, truck weight, and lane distribution. These numbers are projected 10 or even 20 years into the future. Historically, traffic counts have been extremely low.

An alternative is to use traffic based on road classification. The asphalt Insititute provides the following estimates.
Traffic_Table.PNG

Thickness Design:

Deflection Method
The deflection method models the pavement as a homogeneous half-space regardless of the number of layers. The deflection caused by a uniformly distributed load q, with a radius a, can be calculated using Boussinesq’s equation.


Equation1.PNG


Using the modules, E2 is calculated based on the deflections we can model the pavement as a 2-layer system. The equation that relates deflection to loading in a 2-layer system was developed by Burmeister and is as follows:
Equation2.PNG


In this equation h1 represents the overlay thickness. E1 represents the modulus of the overlay, assumed as 500,000 psi.



Equation3.PNG

This equation is used to determine the allowable number of load repetitions (ESAL’s) for a given overlay thickness based on the deflection prediction using the previous equation.
Effective Thickness method

The effective thickness method is similar to the deflection method in that they transform the pavement into a single-layer system regardless of how many actual layers there are present.

The effective thickness method is based on the following equation.

Equation4.PNG

Where hoverlay is the thickness of the overlay, hneeded is the thickness for a full depth pavement and heffective I sthe effective thickness of the existing layers.

Equation5.PNG

Ci is the factor that converts the pavement thickness into an effective thickness based on its condition. Values of Ci are included in most design manuals.

To determine the total thickness needed, use the full depth pavements design charts. As shown below, these charts have ESAL’s and subgrade resilient modulus as an input.


Asphalt_overlay_chart.PNG




Reflective Cracking Strategies (including Crack and Seat)


The term ‘reflective cracking’ is very synonymous with its meaning; the term refers to previously existing cracks in lower layers of the pavement reflecting back up onto the pavement surface as traffic loading is applied as seen in Figure 1. This type of crack occurs in both hot mix asphalt concrete and Portland cement concrete overlays. However, in this summary, only the reflective cracking which is found in asphalt will be taken into account. Preventative maintenance must be implemented to minimize this problem. By minimize it is meant that preventative maintenance will delay reflective cracking from occurring again, it will not prevent it. Reflective cracking will only be eliminated through reconstruction.

Figure 1: Reflective Craking from Traffic Loading
Figure 1: Reflective Craking from Traffic Loading

Reflective cracking does not act the same way in the two different types of pavement. This is because hot mix asphalt and Portland cement concrete have dissimilar stiffness properties. Thus, the worst case of reflective cracking is found when hot mix asphalt is placed over Portland cement concrete. This is the reason why this case was chosen. Three results of loading can occur to induce this type of crack; thermal contraction, downward movement of both slabs, or one slab moving upward and one moving downward. Identifying the movement of slabs as the problem which causes reflective cracking, we now move into finding a solution.

In solving the problem of reflective cracking, the slab must first be stabilized and made rigid. Possible problems which may cause the slab to move are: insufficient compaction, insufficient strength in the subgrade, insufficient drainage design. Once these problems are cared for, a plan of action (or two, depending on the severity of the reflective cracking) may then be formulated to rehabilitate the roadway in the places where reflective cracking is found. Possible solutions that a roadway engineer might decide to act upon include: using fabric, installing a thicker overlay, inserting a membrane which will absorb stress, or applying crack and seat methods.

The crack and seat solution is often seen as most effective by engineers. This is because the lifetime of the pavement is extended much further with this method and the maintenance costs were reduced most with this method. However, it is theorized that the strength of the roadway is reduced as the slabs are broken into smaller sections. This can be somewhat avoided by increasing the thickness of the overlay. Full-depth cracking has also been observed in some cases where this method has been used.

The crack and seat procedure begins with cracking the underlying Portland cement concrete into pieces to minimize movement resulting from temperature. Typical cracking methods include drop hammers (most commonly used) as seen in Figure 2, guillotines as seen in Figure 3, pile drivers as seen in Figure 4, or whip hammers. The drop hammers work based on gravity and the weight of the hammer, and are continually dropped as the name implies. One must be careful to not crack the slab into too small of pieces, approximately four square feet pieces are desired.


Figure 2: Drop Hammer
Figure 2: Drop Hammer
Figure 3: Guillotine
Figure 3: Guillotine
Figure 4: Pile Driver
Figure 4: Pile Driver

The seat method is then induced by compressing the segments with a heavy roller as seen in Figure 5. Seating the pavement will rid of any voids underneath the slab, will create a smooth grade to pave on top of, and will allow pavement engineers to find places in the subgrade which are weak and need reform. It is very important in this process, however, to not excessively roll the pavement. At this point, special treatments can be applied if needed. Finally, the pavement is overlain with hot mix asphalt. This completes the crack and seat method.


Figure 5: Roller Performing Seating
Figure 5: Roller Performing Seating


References


Lesson # 17 From Lecture
Study on advantages and disadvantages of using crack and seat:
http://www.dot.state.il.us/materials/research/pdf/104.pdf
http://www.intrans.iastate.edu/reports/rubblization.pdf
Drop Hammer Figure:
http://www.shelbycountytourism.org/attractions/pof/PublishingImages/ShelbyIronWorks/Original/SIW%20Drop%20hammer.png
Guillotine Figure:
http://www.antigoconstruction.com/images/p-sp-t8600-1.JPG
Pile Driver Figure:
http://img1.topfreebiz.com/o2010-8/Hydraulic-Static-Pile-Driver-TYY500--2915209634.jpg
Reflective Cracking from Traffic Loading Figure:
http://www.fixasphalt.com/cracks.php
Roller Performing Seating Figure:
http://www.forconstructionpros.com/print/Asphalt-Contractor/Features/HMA-Overlay-on-Crack-and-Seat-Provides-Cost-Effective-Solution-to-County-Roads/2FCP2805




Cold In-Place Recycling (Brent)

Background

The current concept of Cold In-place Recycling (CIR) of bituminous pavement was introduced in Eastern Canada in 1989. The benefits associated with the CIR process are significant when compared with traditional pavement rehabilitation methods from both cost and performance perspectives. The CIR method is an environmentally friendly process as it reuses in-place bitumben and aggregates which reduces the overall fuel consumption while minimizing hazardous emissions [1].

Process

CIR is based on the principle that the in-place bituminous pavement is a source of materials that may be used to build a new bituminous layer. The process is carried out in-place and reuses the existing bituminous material to a depth between 65 and 125 mm (2.5 and 5 inches). The existing bituminous pavement is reclaimed, transformed into a bituminous aggregate, which is then mixed with an emulsion, laid down and compacted to the specified density [1].

Cold deep in place recycling is carried out using specialized recycling machines, the heart of which is a milling drum equipped with a large number of hardened steel picks. The drum rotates upwards, milling the material in the existing road, as shown below.

injectionill.jpg
Figure 1: Milling Processinjectionill.jpg


As the milling process is taking place water from a water tanker is delivered through a flexible hose and sprayed into the mixing chamber. The water, which is metered accurately by the microprocessor controlled pumping system, is mixed together with the milled material to bring it up to its optimum compaction moisture content. The milled material and the bitumen emulsion are thoroughly mixed together in the recyclers mixing chamber before being discharged from the rear of the recycler and profiled using a motor grader. The new layer is then compacted in the same way as the cement treated layer [2].


Project Selection

CIR may be considered wherever cracking is present. It may also be considered when permanent deformation and/or loss of integrity in the existing bituminous pavement occurs. Structurally sound and well-drained pavements are the best candidates, as shown below.

before_and_after.jpg
Figure 2: Before and After Cold In-Place Recyclingroad.jpg


Economics
CIR is a cost effective rehabilitation alternative to traditional methods. Based on the life cycle cost of pavement rehabilitation, Municipalities have reported that the annual cost of CIR projects may be as low as 80 % of a traditional method [1].

Full Depth Reclamation (Brent)

Full Depth Reclamation is distinguished form other rehabilitation techniques like Cold Planning (CP), Cold In-Place Recycling (CIR) of Hot In-Place Recycling (HIR), by the fact the rotor or cutting head always penetrates completely through the asphalt section into the underlying base layers, thereby erasing deep pavement crack patterns and eliminating the potential of reflective cracking. With today's innovative equipment and vast range of stabilizing additives, FDR can be utilized to depths exceeding 12", although it is typically performed at 6" to 9". The pulverized layers and additives become a homogenous well-graded material (as shown in figure 3) with an improved structural characteristic [3].


soil_cement_diagram.jpg


Figure 3: Homogenous Layers Createdsoil_cement_diagram.jpg


Full Depth Reclamation provides the ideal opportunity to add stabilizing additives to the new base, either solely or in a variety of combinations, to further enhance the characteristics of the reclaimed material. For example, bituminous stabilizing agents, in the form of asphalt emulsion or foamed asphalt, may be added to achieve a flexible bound material with high fatigue resistance. For added compressive strength, Portland cement or fly ash is sometimes blended with the new base. In areas where the freeze/thaw cycle is severe, liquid calcium chloride is often added to new base. To reduce plasticity and improve load-bearing characteristics, the addition of lime may be specified. If necessary granular material can be added for improving gradation and/or lessening the asphaltic binder content, thereby increasing the structural characteristics of the base [3].

Due to the wide array of stabilizers and equipment available for Full Depth Reclamation, the process is genreally broken down into four primary disciplines: Pulverization, Mechanical Stabilization, Bituminous Stabilization and Chenmical Stabilization. Although all of these disciplines utilize the same “core” equipment and procedure (i.e.-reclaimer, motor grader, compactors and water trucks) [3]. See reclamation diagram procedures below.





full_depth_reclamation.jpg


full_depth_reclamation_2.jpg



Figure 4: Full Depth Reclamation ProceduresReclaimOne.jpgReclaimTwo.jpg



References

1. The Miller Group, Cold In-Place Recycling http://www.millergroup.ca/pavement/cir.html
2. ASN Lewis and DC Collings, Cold In Place Recycling: A Relevant Process for Road Rehabilitation and Upgrading
3. Asphalt Recycling & Reclaiming Association, Full Depth Reclamation, www.arra.org



Surface Treatments

Chips Seal

Background


Chip seal is a surface treatment intended to improve pavement condition, increase friction, treat minor fatigue cracking, seal out moisture, increase usable pavement life, and improve overall pavement appearance. Use of chip seals began in the 1920s originally for low volume lower speed roads. With improvements in polymer modified asphalt emulsion technologies and industry practice its use has expanded to higher volume roads and in some cases it is used as part of new construction. It is intended as a preventative maintenance. If the road is severely distressed it is not a good candidate for a chip seal. This is a widespread surface treatment as it is quick, easy, and has relatively low costs.

There are some noticeable drawbacks to chip seals. They do not correct structural deficiencies. Only very minor cracks can be arrested or significantly retarded by this treatment. At higher operating speed more traffic noise can be generated. This is of little concern for low speed/volume roads. If a chip seal is used on a high volume high speed road, which is becoming more common, noise can be a concern. Increased friction can be caused by a chip seal. However the improvement in roughness due to correcting surface defects by the chip seal is often a net improvement regardless of and roughness caused by the chips.


800px-Chipnseal.jpg
Figure 1. Typical chip seal http://en.wikipedia.org/wiki/File:Chipnseal.jpg


Chip_Seal_close_up.jpg
Figure 2. Close up of typical chip seal http://www.asphaltwa.com/?p=957



Process
This treatment consists of a thin application of asphalt binder or rapid setting asphalt emulsion. Common thicknesses are 1 inch. Once the binder has set a thin layer of aggregate is applied and then pressed into the asphalt by a roller. The aggregate is small and uniform size and shape. Common size is 3/8 inch or 10-mm. Preferred shaped is cubical. Cubical chips promote aggregate interlock and do not depend on how they land on the road. Figure 3 shows aggregate being applied. After the binder with chips has set often the pavement is swept to remove loose chips. A slurry seal or other sealant method may then be applied to the road to further seal out moisture and retain aggregate. The road may be prepped by minor grinding or texturing, selected patching, pothole repair, and crack sealing. Or the road may be left as is. For best results pre-work should be completed months before the chip seal is placed.

Chip_seal_spreader.jpg
 Figure 3. Aggregate chips being
applied to asphalt binder www.hannibal.net



DesignThis treatment is often done empirically however there is an applicable design approach. Empirically a binder of the appropriate performance grade is applied to the pavement of about one inch, then aggregate chips are applied to the binder. The design approach consist of evaluation the existing pavement surface texture, analyze traffic conditions, evaluate climate and seasonal characteristics, determine type of chip seal, type of aggregate, determine proper chips seal type i.e. single layer, sandwich seal, racked in seal, etc, then determine proper construction techniques. There are numerous ship seals for different pavement conditions and budget restraints. Single Chip: This is the most common chip seal and is used mostly in the empirical approach or when no special considerations are needed.
Single_chip.png
Figue 4. SingleChip
Double Chip: This type produces less traffic noise, better waterproofing, and altogether a more robust seal.

Double_Chip.png
Figue 5. Double Chip Racked in seal: A sort of sacrificial layer is added dry to allow the chip seal to completely cure before any of the primary aggregate is dislodge. Helpful in a road with a high level of turns.
Racked-in_seal.png
Figure 6. Racked in seal
Cape Seal: Developed in South Africa. Basically a slurry seal applied to a chip seal. This is a very strong surface treatment and provides additional shear resistance.
Cape_Seal.png
Figure 7. Cape seal Inverted Seal: This chip seal has larger aggregate over smaller sized aggregate. This seal is used to repair or correct a pavement that is bleeding.


Inverted_seal.png
Figure 8. Inverted Seal
Sandwich Seal: A seal involving one binder application between two separate aggregate applications. Useful for restoring surface texture for pavement experiencing severe raveling.

Sandwich_seal.png
Figure 9. Sandwich seal
Geotextile-Reinforced Seal: An advanced chip seal. This is intended for extremely oxidizer and/or thermal cracked pavements. The geotextile is rolled over a tack coat followed by a single chip seal.
Geo_2.png
Figure 10. Geotextile-Reinforced seal

Referances
Washington Apshalt Pavement Assicitation. www.asphaltwa.com/ Retrieved April 18, 2011

Mallick R. B., El-Korchi., (2009). "Pavement Engineering: Principles and Practice". CRC Press

Chip Seal Best Practices. (2005). Transportation Research Board

Romero, P (n.d.) "CVEEN 7570 Pavement Maintenance and Rehabiliation: Lesson 19 - Asphalt Overlay". Retrived April 18, 2011, https://online.uen.org/webct/urw/lc257461621151.tp0/cobaltMainFrame.dowebct

Fog, Slurry seals, etc. (Augusto)




Slurry Seal.

Slurry Seal is a cold mix blend of high quality crushed aggregates, asphalt emulsion, water and mineral fillers, mixed together according to a pre-determined mix design from a laboratory.
It is applied to an existing surface, such as surface treatment or pavements, that are still in fair to good condition, as a means of a cost effective preventative maintenance. It reduces deterioration by sealing, prevents further oxidization, corrects raveling and provides or replaces a high degree of skid resistance.
Slurry Seal is applied with a spreader box, which is connected to the slurry mixing unit, as shown on the right. The box is the width of a single lane, allowing a uniform spread of material in a single pass.
Crews ahead of the unit set up traffic control, and sweep the surface before applying the slurry seal. Operators of the units monitor the mixing and application. After a short curing time traffic will be allowed to resume use of the freshly rehabilitated lane.

external image New-Picture-2-300x145.png

There are three types of Slurry seal commonly used, each using a different size of aggregate;
  • Type I- Slurry Seal (Fine Aggregate,1/8” max) Used for lower traffic volumes and maximum crack penetration
    Recommended application rates are 4.3-6.5 kg/m².
  • Type II-Slurry Seal ( General Aggregate,1/4” max) Most commonly used slurry. Good for moderate to high volume roads
    Recommended application rates are 6.5-10.8 kg/m².
  • Type III- Slurry Seal ( Coarse Aggregate ,3/8”) Used for high volume roads and heavier traffic.
    Recommended application rates are 9.8- 16.3 kg/m²

external image slurry_seal_04.jpg


Slurry seals will fill small surface cracks, stop raveling, and improve the skid resistance of the pavement.

Advantages
The surface of a slurry-seal treatment is smoother than a chip seal treatment. The slurry-seal treatment is, therefore, more “surface friendly” than a chip-seal treatment in areas such as campgrounds. A person is able to rollerblade on a slurry-seal treatment.

Disadvantages
Equipment to apply a slurry seal is not as common as the equipment for a chip-seal application. Many counties, which are partners with the U.S. Department of Agriculture, Forest Service, as well as local contractors, own equipment for chip seal applications but not slurry-seal equipment.

Slurry Seal Specifications

The following specification is for basic standard unmodified slurry seal for use in areas where the longer life cycles of polymer-modified and black aggregate are not required. Before using this specification the user must reduce the aggregate types listed under (c) to only one. Type I may be used for parking lot resurfacing but should only be specified if it is readily available. Type II may also be used for parking lots, streets and arterials. Type III is used for arterials and highways.


MATERIALS
The materials for slurry seal immediately prior to mixing shall conform to the following requirements:

(a) Asphalt Emulsion
Asphalt emulsion shall be a cationic quick-setting type conforming to the requirements for CQS-1H grade under CalTrans Standard Specifications, July 1992, Section 94, Table 4, requirements for "Quick Setting Asphalt Emulsion".

(b) Water
Water shall be potable, free of harmful salts and shall be of such quality that the asphalt will not separate from the emulsion before the slurry seal is in place in the work.

(c) Aggregate
Aggregate shall consist of sound, durable, crushed stone or crushed gravel and approved mineral filler. The material shall be free from vegetable matter and other deleterious substances. Aggregates shall be 100% crushed with no rounded particles. The percentage composition by weight of the aggregate shall conform to one of the following gradings:


PERCENTAGE PASSING
SIEVE SIZES
TYPE I
TYPE II
TYPE III
3/8"
--
100
100
No. 4
100
90-100
70-90
No. 8
90-100
65-90
45-70
No. 16
60-90
40-70
28-50
No. 30
40-65
25-50
19-34
No. 200
10-20
5-15
5-15
Theoretical asphalt
content, % dry
aggregate
10-12
7.5-13.5
6.5-12
Approx. application
rate (pound per
square yard)
8-12
10-15
15-25


The aggregate shall also conform to the following quality requirements:

TESTS
CALIF. TEST
ASTM TEST
REQUIREMENTS
Sand Equivalent
217
D2419
55 Min.

PROPORTIONING
Asphalt emulsion shall be added at a rate of from 11 to 25%. A job mix design shall be submitted by the Contractor for approval by the Engineer that conforms to the specification limits, and that is suitable for the traffic, climate conditions, curing conditions and final use.

The Slurry Seal mixture shall be proportioned by the operation of a single start/stop switch or lever which automatically sequences the introduction of aggregate, emulsified asphalt, admixtures, if used, and water to the pugmill.

Calibrated flowmeters shall be provided to measure both the addition of water and liquid admixtures to the pugmill. If necessary for workability, a retarding agent, that will not adversely affect the seal, may be used.

Water, and retarder if used, shall be added to ensure proper workability and (a) permit uncontrolled traffic on the slurry seal no more than three (3) hours after placement without the occurrence of bleeding, raveling, separation or other distress; and (b) prevent development of bleeding, raveling, separation or other distress within seven (7) days after placing the slurry seal.

Uniformity of distribution of asphalt will be determined by extraction tests in accordance with California Test 310. The average bitumen ratio (pounds of asphalt per 100 pounds of dry aggregates) shall not vary more than five (5) percent above or below the amount designated by the Engineer. This requirement shall apply to samples taken from any location or operation designated by the Engineer.

MIXING
The Slurry Seal shall be mixed in a self-propelled mixing machine equipped with a continuous flow pugmill capable of accurately delivering and automatically proportioning the aggregate, emulsified asphalt, water and admixtures to a double shafted, multiblade pugmill mixer capable of minimum speeds of 200 revolutions per minute.

A minimum of two mixing machines shall be maintained on each project of a 12 cubic yard or larger capacity. The slurry seal retention time in the pugmill shall be less than three seconds. The mixing machine shall have sufficient storage capacity of aggregate, emulsified asphalt, and water to maintain an adequate supply to the proportioning controls.

The mixing machine shall be equipped with hydraulic controls for proportioning the material by volume to the mix. Each material control device shall be calibrated, properly marked, preset and lockable at the direction of the Engineer. The mixing machine shall be equipped with a water pressure system and nozzle type spray bars to provide a water spray immediately ahead of the spreader box.

The mixing machine shall be equipped with an approved fines feeder that provides a uniform, positive, accurately metered, pre-determined amount of a mineral filler, if used, at the same time and location that the aggregate is fed.

SPREADING EQUIPMENT
The slurry mixture shall be uniformly spread by means of a controlled spreader box conforming to the following requirements:

The spreader shall be capable of spreading a traffic lane width and shall have strips of flexible rubber belting or similar material on each side of the spreader box and in contact with the pavement to prevent loss of slurry from the box and the box shall have baffles, or other suitable means, to insure uniform application on super-elevated sections and shoulder slopes.

The rear flexible strike-off blade shall make close contact with the pavement and shall be capable of being adjusted to the various crown shapes so as to apply a uniform seal coat.

Slurry mixture, to be spread in areas inaccessible to the controlled spreader box, may be spread by other approved methods.

PLACING
The slurry seal shall not be placed if either the pavement or the air temperature is below 13 degrees C (55F) and falling, but may be applied when both the air and pavement temperature is 7 degrees C (45F) or above and rising. The mixture shall not be applied if high relative humidity prolongs the curing beyond a reasonable time.

Before placing the slurry seal, the pavement surface shall be cleaned by sweeping, flushing or other means necessary to remove all loose particles of paving, all dirt and all other extraneous material.

Prior to 24 hours before beginning slurry seal operations, the contractor shall notify all residents, businesses and agencies by an approved written notice detailing streets and limits of work to be done and the hours of work. The contractor shall also 24 hour post all streets that are to be worked upon with temporary "No Parking - Tow Away" signs at 100 foot intervals. These signs shall also state the day of the week and hours of no parking.

Immediately before commencing the slurry seal operations, all surface metal utility covers (including survey monuments) shall be protected by thoroughly covering the surface with an appropriate adhesive and oiled or plastic paper. No adhesive material shall be permitted to cover, seal or fill the joint between the frame and cover of the structure. Covers are to be uncovered and cleaned of slurry material by the end of the same work day.

Hand tools shall be available in order to remove spillage. Ridges or bumps in the finished surface will not be permitted. The mixture shall be uniform and homogeneous after spreading on the surface and shall not show separation of the emulsion and aggregate after setting.

Adequate means shall be provided to protect the slurry seal from damage from traffic until such time that the mixture has cured sufficiently so that the slurry seal will not adhere to and be picked up by the tires of the vehicles.

MEASUREMENT (BY SQUARE YARD)
Slurry seal will be measured and paid for by the square yard for the actual surface area covered.

PAYMENT (BY SQUARE YARD)
The contract price paid per square yard for slurry seal shall include full compensation for furnishing all labor, materials, tools, equipment and incidentals and for doing all the work involved in the furnishing and placing the slurry seal complete in place, including cleaning the surface and protecting the slurry seal until it has set, all as shown on the plans, as specified in these specifications and as directed by the Engineer.

Application
The slurry mix is applied at the thickness of the largest aggregate in the mix. The amount of aggregate, filler, additives, and water is based on the mix design, depending on the component materials, environmental conditions, and existing road surface.

Cost and life expectancy
The typical cost is $1.20/square yard in 1999. The life expectancy is 4 to 6 years.


Fog Seal

A fog seal is an application of asphalt emulsion sprayed onto a pavement surface with or without a sand cover (figure 3). The emulsion is diluted to the proper consistency in order to get complete coverage on the roadway but not be too thick to cause a slippery surface. A fog seal works better on a coarse aggregate surface where the asphalt emulsion has room to pond between the aggregate particles. On a smooth aggregate surface, the asphalt rests on the surface covering the top of aggregate particles, creating a slippery surface for the vehicles. If the fog seal was not properly applied and a slippery surface exists, a dry choke cover is applied to the surface. The choke is usually clean sand or aggregate that is less than 0.25 inches in diameter.

Fog seals are used to delay weathering of the pavement, waterproof the pavement surface, improve the pavement’s ability to keep water from penetrating the base course or subgrade, and reduce raveling.


external image fog_2.jpg


Asphalt emulsions with rejuvenating properties such as GSB-Emulsion Sealer and Rejuvenator supplied by Asphalt Systems, Inc. or Reclamite from Witco Corporation can be used to penetrate, rejuvenate, and seal the surface of asphalt pavements. Reclamite requires a thin layer of sand (1 to 2 pounds/square yard) to be applied before allowing traffic onto the roadway.


Advantages
Fog seals are inexpensive compared to other surface treatments. Only a distributor truck is required to apply the fog seal in most cases.

Disadvantages
The expected life of the fog seal is generally shorter than other surface treatments. If applied too heavily, the fog seal could be slippery and hazardous for the road users.

Fog Seal Specifications.

Site Conditions

To be effective, fog seals need to break quickly (revert to solid asphalt) and cure completely (lose water to form a cohesive film). This should be at a rate that allows traffic to be accommodated without the binder being picked up by vehicle tires. To achieve this behavior, the film forming properties of the binder must be adequate (i.e., the binder must be able to coalesce into a continuous film prior to allowing traffic on the new seal). Asphalt films do not form well at low temperatures in the absence of low viscosity diluents. Thus, warm conditions with little to no chance of rain are necessary to ensure successful applications. Fog seals should not to be applied when the atmospheric temperature is below 10°C (50°F), and pavement temperature below 15°C (59°F).
If unexpected rain occurs, prior to the emulsion breaking, the emulsion may wash out of the pores of the pavement and break on the surface of the pavement creating a slippery surface.

Surface Preparation

Immediately before applying a fog seal, the pavement surface must be cleaned with a road sweeper, power broom, or flushed with a water pump-unit to remove dust, dirt, and debris. The pavement surface must be clean and dry before applying the fog seal. If flushing is required, it should be completed 24 hours prior to the application of the fog seal to allow for adequate drying.

Materials Preparation

Although asphalt emulsions (original emulsions) contain up to 43% water, they must be diluted further before use. This additional dilution reduces viscosity (see Figure 8) and allows the application of small amounts of residual binder to be adequately controlled. Generally, the supplier will dilute the original emulsion, in the field or at the plant. A dilution rate of 50% (1:1) (equal parts water to equal parts emulsion) is recommended. Dilution water must be potable and free from detectable solids or incompatible soluble salts (hard water).


A chart shows a decrease in viscosity with increase in percent dilution.
A chart shows a decrease in viscosity with increase in percent dilution.


Water can be checked for compatibility with the emulsion by mixing a small amount of the emulsion in a can (approximately 1 liter). The materials are mixed for 2 to 3 minutes with a stirrer and the resulting mixture is poured through a pre-wetted 150 mm sieve. If more than 1% by weight of material is retained on the sieve, the water is not compatible and clogging in spray jets may result.

Simple Water Compatibility Test Method
Simple Water Compatibility Test Method


Incompatible water may be treated with 0.5 to 1.0% of a compatible emulsifier solution (the emulsion manufacturer can provide advice regarding compatible solutions). The emulsifier solution should be added to the water tanker and circulated for 10 to 15 minutes via pump before adding to the emulsion. If a water treatment is used, the compatibility test should be repeated using the treated water to ensure compatibility.
The emulsion should be diluted no more than 24 hours before its intended use to avoid settlement of the diluted emulsion. (6) Water is always added to the emulsion and not the other way around. The emulsion may be circulated using a centrifugal or other suitable pump to ensure uniformity (6).

Application Rates and Spraying

Properly calibrated distributor trucks should be used to apply the emulsion (see Figure 2). Spray nozzles with 4 to 5 mm (1/8” to 3/16”) openings are recommended (6). The emulsion may be heated to 50°C (122 °F) maximum, although, generally the emulsion is sprayed at ambient temperature (6). The emulsion is sprayed at a rate determined by the surface conditions (see Table 1). A test section representative of the entire surface should be chosen to approximate application rates (see Section 4.5). Typical application rates for diluted emulsion (1:1) range from 0.15 to 1.0 l/m2 (0.03 to 0.22 gal/yd2) depending on the surface conditions (5). A 1:1 diluted emulsion is an original emulsion that has been subsequently diluted with equal parts water.
Table 1: AEMA Recommendations for Application Rates (5) ||~ % Original Emulsion ||~ Dilution Rate
Tight Surface*
Open Surface**
(l/m2)
(gal/yd2)
(l/m2)
(gal/yd2)
50
1:1
0.15 - 0.5
0.03 – 0.11
0.4 - 1.0
0.09 – 0.22

Ideally, half the application should be sprayed in each direction to prevent build-up on one side of the stones (this is particularly important in the case of chip seals) and rough surfaces. Build-up on one side can result in a slippery surface and is inadequate binder to fully enrich the surface or hold the stone.

Estimating Application Rates

To estimate the application rate, take a one-liter can of diluted emulsion (usually 1:1 dilution rate) and pour evenly over an area of 1 m2. This represents a diluted application rate of 1 l/m2. If the emulsion is not absorbed into the surface after 2-3 minutes, decrease the application rate of the emulsion and apply to a new 1 m2 area and repeat until the approximate application rate is found. If, after the first test, the surface looks like it can absorb more emulsion, increase the application rate of the emulsion and spread it over a new 1 m2 area. Repeat until the approximate application rate is found. This same procedure can be followed using gallons and square yards to determine application rate.

Traffic Control

Place traffic control before work forces and equipment enters the roadway or work zone. Traffic control is required both for the safety of the traveling public and the personnel performing the work. Traffic control includes construction signs, construction cones and/or barricades, flag personnel, and pilot cars to direct traffic clear of the construction operation. Traffic control details should conform to agency requirements.
Traffic control is also required to protect the integrity of the application. The curing time for the fog seal material will vary depending on the pavement surface conditions and the weather conditions at the time of application. Under ideal conditions, including increasing air and surface temperatures, it is suggested that traffic be kept off the fog seal material for at least two hours and until acceptable skid test values are achieved.

Safety (Personal Protective Equipment–PPE)

All employees should wear and use the safety gear required for a fog seal operation. This includes, but is not limited to, items such as hard hats, approved shirts, safety vests, earplugs, gloves, and safety glasses (8, 9).

Quality Control

Quality control and workmanship are critical to the performance and life of a fog seal treatment. There must be a cooperative effort between the Agency’s representative and the contractor’s representative to conduct inspections of all project equipment before and during the project. The primary piece of equipment for a fog seal operation is the asphalt distributor. It is critical that it is functioning as required by the project specifications. The spray bar must be set to the appropriate height (distance) from the pavement surface and the nozzles must be set at the proper angle to assure a uniform application of material (1). The material temperatures should also be measured for quality control purposes.
The emulsion must be certified to specification according to established sampling and testing procedures (6). Excess emulsion can create slick pavements.
It is recommended that project inspections be conducted so that any deficiencies in workmanship or materials are addressed and corrected. This process will also assist the department in identifying the performance of various fog seal materials; how they are performing on various surface conditions and how they are performing in various climatic zones.

Post Treatment

Sand blotters may be used at approximately 1 kg/m2 (1.8 lb/yd2) to allow early opening to traffic. Sweeping may be required. Agency personnel will assess this after application and opening to traffic. Even with sand cover, traffic control may be required to keep speeds down.
Skid resistance (a coefficient of friction) can be measured using ASTM E 274. It is recommended that this be done after the application has cured to ensure the proper value is measured. The final surface shall yield a coefficient of friction not less than 0.30 as determined by ASTM E 274. A treated pavement shall not be opened to traffic until an acceptable value is recorded. If a treated pavement does not produce an acceptable coefficient of friction, see Table 3 for corrective action. Permeability may be monitored by Agency procedures to ensure that an effective seal has been achieved. This should be done at the discretion of the Resident Engineer.

Equipment
The equipment needed for a fog seal is a distributor truck to dispense the asphalt emulsion and a sand spreader if sand is applied.

Application
Fog seals are applied at 0.05 to 0.15 gallons/square yard of the diluted asphalt emulsion.

Cost and life expectancy
Typical costs are $0.45/square yard. The expected life of a fog seal is 1 to 3 years.


Sand Seal


A sand seal is a sprayed application of asphalt emulsion followed by a covering of clean sand or fine aggregate (figure 4). A pneumatic-tire roller is often used after applying the sand. Excess sand is removed from the road surface after rolling.

Sand seals enrich weathered pavements and fills fine cracks in the pavement surface. The sand can provide additional skid resistance to the pavement while also inhibiting raveling.

As in fog seals, emulsions with rejuvenating properties can be used. The additional expense of the rejuvenating emulsion could be cost effective if there are many small cracks in the pavement. CRF Restorative Seal of Witco Corporation and GSB-88 of Asphalt Systems, Inc. are products that are designed to penetrate and restore aged pavements.


CONSTRUCTION METHODS


a. Streets to be Treated

The Contractor and the Director shall mutually determine the streets which shall receive sand seal treatment. Measurements of streets to be treated shall be made by the Contractor and the Director or his/her Designee, and the Contractor shall
prepare a cost estimate for each street prior to beginning work.
b. Surface Preparation
Surface preparation, which may include pothole patching, truing and leveling, adjusting of street irons (valve covers, manhole covers, drop inlet gratings), etc., will be the responsibility of the awarding authority and will be completed before
the contractor moves onto the job.
Immediately prior to the application of asphalt materials Highway Department personnel shall remove small branches and other debris, and use a mechanical street sweeper to clean any loose material from the pavement surface.
The Highway Department shall protect manhole covers, drop inlets, catch basins, curbs, and any other structures within the shoulder areas against the application of the surface treatment materials.





















c. Weather Limitations

Work will not be done unless the road surface is dry. No work shall be done during rain or foggy periods. No work shall be done if the ambient temperature is elow 10 °C, (50°F).

d. Spreading Asphalt and Sand
Prior to application of asphalt material on any street, sufficient quantities of materials to cover the entire street at the specified rates shall be on the site and ready for application. The awarding authority shall be responsible for providing
the Contractor with an aggregate storage area near the job site. The asphalt material shall not be applied more than 90 meters, (300 feet), in advance of the self-propelled aggregate spreader. AT NO TIME SHALL ANY ASPHALT MATERIAL BE ON ANY ROAD SURFACE FOR MORE THAN FIFTEEN MINUTES BEFORE IT IS COVERED WITH SAND.
e. Rolling
Initial rolling shall be done immediately following the application of sand. Rollers shall be operated at a speed that will not displace aggregate.

800px-Road_resurfacing.jpg


Advantages
The sand seal generally provides a thicker coating on the pavement surface than the fog seal, resulting in a longer life expectancy. The sand seal on polished aggregate surfaces can provide additional skid resistance.

Disadvantages
Only fine cracks are filled and larger cracks tend to reappear within a year.

Equipment
The equipment needed for a sand seal is a distributor truck to dispense the asphalt emulsion and a sand spreader to add the sand cover. A pneumatic-tire roller can be used but is not required. A broom is used to remove the excess sand.

Application
Emulsified asphalts are applied from 0.10 to 0.25 gallons/square yard. The sand is applied at 18 to 25 pounds/square yard yielding a 3/16-inch-thick new layer over the existing pavement.

Cost and life expectancy
The typical costs are $0.70/square yard in 1999. The expected life of a sand seal is 3 to 4 years.


Crack Sealing

Crack sealing and filling prevent the intrusion of water and incompressible materials into cracks. The methods vary in the amount of crack preparation required and the types of sealant materials that are used.
Crack sealing is the placement of materials into working cracks. Crack sealing requires thorough crack preparation and often requires the use of specialized high quality materials placed either into or above working cracks to prevent the intrusion of water and incompressible materials. Crack sealing is generally considered to be a longer-term treatment than crack filling.

Due to the moving nature of working cracks a suitable crack sealant must be capable of:
• Remaining adhered to the walls of the crack,
• Elongating to the maximum opening of the crack and recovering to the original dimensions without rupture,
• Expanding and contracting over a range of service temperatures without rupture or delamination from the crack walls, and
• Resisting abrasion and damage caused by traffic.

Materials for Crack Sealing
Crack sealing materials are designed to adhere to the walls of the crack, stretch with the movement of the crack over the range of conditions and loads associated with the crack location, and resist abrasion and damage caused by traffic. For sealing working cracks, the preferred sealant is usually elastomeric.



crack.jpg



This means the sealant has a low modulus of elasticity and will stretch easily and to high elongations (usually around 10 times its non strained dimensions) without fracture. Such sealants also recover over time to close to their original dimensions. The sealants are usually applied at elevated temperatures due to their high viscosity at ambient temperatures and they set or cure by cooling and reforming into complex structures. This is called thermoplastic. Thermoset is sometimes used to describe these materials, however this is incorrect. A thermoset is a material that undergoes a chemical cross-linking when heated. This structure is retained as it cools and is not reversible by reheating. Thermoplastics form physical structures on cooling but this process is reversible with reheating. Hot application ensures good adhesive bond to the crack walls. In California most of the hot pour materials are rubbermodified asphalt. These materials have excellent abrasion resistance and are useful for trafficked surfaces.

Cold pour materials for crack sealing in California are usually silicone based and often used prior to paving. These materials cure either by exposure to moisture in the air or by mixing a hardening agent with the base silicone. These materials often have poor abrasion resistance and should not be used in trafficked areas. Other materials such as epoxies and polyurethanes are almost always cured by addition of a second chemical.

Micro-Surfacing


Micro-surfacing is a thin surfacing, and can be laid at two to three times the thickness of the largest stone in the grading. The emulsion in the system is always polymer modified and special additives are used to create a chemical break that is largely independent of weather conditions. In breaking, the emulsion forces water from the aggregate surface. Such systems can often be opened to traffic within 1 hour or less of its application under a range of conditions (Holleran, 2001a).
Micro-surfacing can be used for the same applications as slurry seals. However, micro-surfacing uses better quality aggregates and a fast setting emulsion of higher stiffness allowing thicker layers to be placed.
These aspects create the following extended performance characteristics and applications for micro-surfacing:
  1. • Correction of Minor Surface Profile Irregularities
  2. • Rut Filling
  3. • Higher Durability
  4. • Ability to be Placed at Night or in Cooler Temperatures

Micro-surfacing, like slurry seal, is not intended as a crack treatment and will not prevent cracks in the underlying pavement from reflecting through to the surface. Micro-surfacing does not add any structural capacity to an existing pavement; it is applied as a maintenance treatment to improve the functional characteristics of the pavement surface.

APPLICATIONS
AGGREGATE TYPE
II
III
Void Filling

Wearing Course (AADT) < 100

Wearing Course (AADT) 100 – 1,000


Wearing Course (AADT) 1,000 – 20,000


Minor Shape Correction (0.4-0.8 in [10-20mm])

Application Rates in pounds of dry aggregate per square yard
10 – 15
20 – 25


The main use of micro-surfacing materials is for pavement preservation as a part of a program of periodic surfacing before distresses appear. The main criteria for project selection are:


  • Sound and well drained bases, surfaces, and shoulders.
  • Free of distresses, including potholes and cracking. These must be repaired before slurry application. Potholes should be filled and compacted several weeks prior to slurry surfacing. Emulsion crack filling should be done several months prior to slurry surfacing.



Distress modes that can be addressed using micro-surfacing include:
  1. Raveling: Loose surfaces or surfaces losing aggregate may be resurfaced using slurry seals or micro-surfacing.
  2. Oxidized pavement with hairline cracks: These surfaces may be resurfaced using slurry seals or micro-surfacing.
  3. Rutted pavements: Deformation resulting from consolidation of the surfacing only. Rutting due to base failure of significant plastic deformation of the HMA cannot be treated except as a temporary measure.
  4. Rough pavements with short wavelength: These irregularities may be treated with micro-surfacing, provided the frequency of the irregularities is shorter than the spreader box width.



Distress modes that cannot be addressed using micro-surfacing include:
  1. Cracking (including reflection cracking)
  2. Base Failures of any kind
  3. HMA Layers that exhibit plastic shear deformation

Micro-surfacing will not alleviate the cause of these distresses. As a result, the distresses will continue
to form despite the application of a slurry surfacing.

Application Conditions

The application conditions required are addressed in detail in the Caltrans "Micro-surfacing Pilot Study 2001", Appendix A (Caltrans, 2002). The basic requirement for success is that the emulsion must be able to break and form continuous films, as it is the only way a slurry mixture can become cohesive. As a result, humidity, wind conditions, and air and surface temperature are important and need to be considered. Modifications to additives should be made according to the changing environment during application. Because micro-surfacing slurry systems use a chemical break, they can be placed at night.
Micro-surfacing shall only be placed when the ambient temperature is 8°C (46°F) and rising and the high temperature for the day is expected to be at least 20°C. Micro-surfacing shall not be placed if rain is imminent or if the ambient temperature is expected to fall below 2°C within 24 hours after placement. Slurry surfacing systems will typically resist rain induced damage after as little as one hour but typically require at least three hours to cure to a fully waterproof state. Additionally, breaking time for a slurry system is affected by ambient temperature. Figure 9-14 shows the effect of temperature on the breaking rate of emulsion.
REFERENCES:

1.- Holleran, G, 2001b. Micro-surfacing, Bitumen Asia 2001, Singapore, Asia,2001

2.- Asphalt seal coat treatments http://www.fs.fed.us/eng/pubs/html/99771201/99771201.htm

3.- Diferent types of asphalt and sealcoating http://www.sealcoatingsolutions.com/articles/different-types-of-asphalt-sealcoating

4.- Keys to successful slurry seal and/or micro-surfacing projects http://www.techtransfer.berkeley.edu/pavementpres07downloads/metcalf.pdf



Concrete Pavements


Overlays


Design Methodology (Jasmin)



The intent of bonded overlays is to improve existing pavement structural capacity where as unbounded overlays are generally designed with a service life of a new pavement. Bonded overlays placed over distressed underlying pavements range from 2 to 5 inches thick acting as one system and unbounded overlays range from 4 to 11 inches thick separating the existing pavement from the new overlay. Existing pavement surface plays a big role when designing concrete overlay. The concrete overly relies heavily on load transfers to the distressed pavement layer and the pavement needs to be properly characterized in order to meet the design serviceability. Figure 1 shows design factors that affect each other.

figure_1.JPG

Figure 1: Overlay design factors.
Depending on the design service life of the new overlay cost effective repairs need to be considered and made to the existing pavement. If done properly both methods are used as a preventative measure to stop reflective cracking from surfacing. Figures 2 shows overlay possibilities for concrete used in bonded and unbounded designs.


figure_2.JPG


Figure 2: Bonded and Unbonded Concrete Overlays


Design considerations for Bonded and Unbonded concrete overlays:
·
Bonded Overlays of Concrete Pavements
  • Designed using AASHTO Guide 1993 & 1998 considering design deficiency or remaining life approach.
  • Concrete thickness determined based on anticipated traffic load.
  • Design assumes no stresses between the two layers at the bond plane.
  • Transverse joints cut at full depth of the overlay plus another 0.5 inches.
  • Longitudinal joints are cut at half the depth of transverse joints.
  • Preoveraly repairs need to be considered.
  • Life expectancy of 15 years if done properly.

· Bonded Overlays of Asphalt Pavements
  • Design based on AASHTO 1998 Guide and ACPA 2004 based on corner breaking mode failure.
  • Bond layer between the two surfaces is analyzed and horizontal bond stresses are modeled and compared to ACPA 2004 fatigue model.
  • Joint spacing is optimized and reduced to mitigate wrapping and curling.
  • Preoveraly repairs need to be considered.
  • Life expectancy of 15 – 25 years if done properly.

· Bonded Overlays of Composite Pavements
  • Designed using AASHTO Guide 1993 & 1998 where serviceability load failure is assumed.
  • Interface bond between the two surfaces is ignored.
  • Joint spacing is a design consideration.
  • K value is calculated at the top of the asphalt layer.
  • Preoveraly repairs need to be considered.
  • Overall considerations made result in an overly conservative design.

· Unbonded Overlays of Concrete Pavements
  • Designed using AASHTO Guide 1993 & 1998 where the underlying pavement is considered a base.
  • Existing pavement is used as base although it is made sure that the base will provides adequate support.
  • A one (1) inch stress relief layer is designed so that it prevents cracks from surfacing.
  • Shorter joint spacing elevates temperature curling and load combinations.
  • Drainage design consideration help reduce pumping faulting and cracking.
  • Design life is usually expected to be 20-30 years.

· Unbonded Overlays of Asphalt Pavements
  • Designed using AASHTO Guide 1993 & 1998 where the underlying pavement is considered a base.
  • Used over crown and rut distresses, good for ruts up to 2 inches of depth.
  • No friction is assumed in the design between the two surfaces.

· Unbonded Overlays of Composite Pavements
  • Designed using AASHTO Guide 1993 & 1998 where the underlying pavement is considered a base.
  • Stress relief layer already considered in a form of the composite asphalt layer.





Reference:

Guide to Concrete Overlays "Sustainability Solutions for Resurfacing and Rehabilitating Existing Pavements" Second Edition September 2008, National Concrete Pavement Technology Center







Bonded or Unbonded (Vernon)


BONDED OR UNBONDED CONCRETE OVERLAYS
By Vernon M. Black


There are two principal types of concrete overlays that are used: bonded or unbonded. The decision of which type of overlay will be used is based on two primary variables: the existing condition of the pavement, and the required increase in structural capacity. Although the discussions below focus on the application of concrete overlays over concrete pavements, they are equally applicable to the use of concrete overlays over asphalt pavements.

Bonded Concrete Overlays


Figure_#3-1.png



Figure 1 – Bonded Concrete Overlays of Concrete Pavements

Bonded concrete overlays of concrete pavements are primarily used to increase pavement structural capacity. They consist of a thin concrete layer (4 inches or less) bonded to the top of the existing concrete surface to form a monolithic or composite section. Typically, pavements that have very little deterioration but are too thin for an increasing traffic volume are good candidates for a bonded overlay. Bonded concrete overlays are not recommended when the existing pavement is badly deteriorated and a substantial amount of removal and replacement of existing layers is necessary during rehabilitation. Bonded concrete overlays are also not appropriate if there is significant deterioration of the existing pavement from a material durability problem like "D" cracking or alkali-silica reaction.

Bonded concrete overlays over existing concrete pavements require existing-surface preparation in order to enhance the bonding to the overlay. This involves any minor repairs that may be necessary of existing distresses in order to provide an even surface, as well as grinding to increase the bonding surface area and airblasting to remove any debris from the surface, as demonstrated in figures 2 and 3 below.


Figure_#3-2.jpg



Figure 2 – Teeth of a Rotomill


Figure_#3-3.jpg



Figure 3 – Airblasting to clean the concrete surface

Unbonded Concrete Overlays


Figure_#3-4.png



Figure 4 – Unbonded Concrete Overlays of Concrete Pavements

Unbonded concrete overlays consist of a relatively thick concrete layer (5 inches or greater) on top of an existing concrete pavement. Unbonded overlays are generally most cost effective when an existing concrete pavement is badly deteriorated and removal of existing pavement layers is not desirable. Unbonded overlays react structurally as if built on a strong, non-erodable base course.

Unbonded overlays do not require much pre-overlay repair before placement because of a separating layer used between the overlay and old pavement. The separation interlayer is usually a thin asphalt layer of about 0.5-1.5 inches thick. The layer is sometimes called a debonding layer or stress relief layer. The purpose of the interlayer is to separate the old and new layers so that they may act independently of each other through temperature cycles and load deflection. The separation interlayer prevents distresses in the old pavement from reflecting through into the overlay.

Unbonded concrete overlays are a much better option for deteriorated concrete pavements than rubbleization and an asphalt overlay.



References:
· Texas Department of Transportation, Pavement Design Guide. January 2011. Accessed Online – April 15, 2011. <http://onlinemanuals.txdot.gov/txdotmanuals/pdm/pdm.pdf>.
· National Concrete Pavement Technology Center, et al. Summary of Concrete Overlays. Poster.




Ultra Thin White Topping (Chris)


Ultra Thin Whitetopping (UTW) is a bonded concrete overlay. Being such, it is constructed on top of an existing asphalt pavement. Before application, the asphalt pavement is ground to give a nice, even surface for the overlay to bind to. UTW if considered ultra thin because it is less than 4 inches thick. Since it is so thin, the bond between new and old pavements is required. Also, slabs are much smaller with a UTW than a typical overlay. Standard overlays typically have a slab size of about 10 feet. With UTW, slabs can be anywhere from 6 feet to as small as 2 feet. This is to prevent curling.

joint_spacing_cm.jpg
Figure 1: Joint Spacing on Ultra Thin whitetopping[i]

UTW is known to significantly extend the life of the pavement. Since UTW is a rigid overlay, it is especially effective in areas that exhibit high levels of rutting and shoving, such as intersections. Another benefit to UTW is that it is easy to work around utilities when placing the overlay.


joints_cm.jpg
Figure 2: Joint placement around utilities[ii]

One other benefit of whitetopping is that it can be used to enhance the aesthetics of an area. This can be done by adding color, stamping, and/or forming.


stamped_concrete_cm.gif
Figure 3: Examples of Stamped and Colored Concrete[iii]

The most important thing to consider when debating on using UTW is: Will there be enough thickness left in the asphalt layer after grinding to support the UTW structurally? The next most important issue is to remove whatever conditions caused the current deterioration of the pavement. If this is not done, it will not be long before the same creep back into the overlay. Other issues to consider are: impacts to the corridor, cost, time, traffic loading and contractor experience.

Using tables similar to the ones shown in Figure 4, the thickness and slab size (joint spacing) can be determined, given the flexural strength and thickness of the asphalt. The flexural strength and asphalt thickness can be determined in the field, if they are not known.


Tables_cm.jpg
Figure 4: Table for UTW Thickness and Size[ii]


References:
[i] CPTP Status Report - Task 65 Engineering ETG Review Copy: Chapter 2 - CPTP Focus Areas: Focus Area 1: Advanced designs. (n.d.). Retrieved April 18, 2011, from U. S. Department of Transportation: Federal Highway Administration: http://www.fhwa.dot.gov/pavement/concrete/sr04ch2.cfm
[ii] Romero, D. P. (2011, April 6). CVEEN 7570 Pavement Maintenance: Lesson #18 White Topping. (U. Class, Interviewer)
[iii] Decking. (n.d.). Retrieved April 18, 2011, from Riverton Pool and Garden Center: http://www.rivertonpoolandgarden.net/catalog/cc/3/ccc/11/decking.php