IMPLICATION OF TIME-DEPENDENT TEXTURE-DEGRADATION ON PAVEMENT ON- BOARD-SOUND-INTENSITY PATTERNS IN MNROAD TEST CELLS

Transcription

1 IMPLICATION OF TIME-DEPENDENT TEXTURE-DEGRADATION ON PAVEMENT ON- BOARD-SOUND-INTENSITY PATTERNS IN MNROAD TEST CELLS Lev Khazanovich, PhD Associate Professor Department of Civil Engineering University of Minnesota Bernard I. Izevbekhai (Corresponding Author) Research Operations Engineer, Minnesota Department of Transportation PhD Candidate Department of Civil Engineering University of Minnesota Proceedings of the International Noise Conference, Dearborn Michigan July 2008 ABSTRACT Pavement texture is an important parameter in tire-pavement-interaction-noise (TPIN). As pavements carry traffic load over the years measurable degradation occurs in texture. As the pavement is exposed to environmental and traffic elements, changes occur in ride quality measured by the International Roughness index (IRI) as well as the Surface Rating (SR). Research investigated the correlation between TPIN measured with the On-Board Sound intensity (OBSI) Protocol, Estimated Single Axle load (ESAL); SR and age of pavement in MnROAD test cells. General observation shows decrease of OBSI with respect to traffic evident in the relative values of driving and passing lane in the rigid pavements and conversely an increase in OBSI in the bituminous segments with time. An initial overall model of all the cells was not feasible as various pavement types exhibited unique characteristic residuals. This led to the development of individual models for each surface type. Although hysteretic effects are implicated in both parameters, no tenable relationship between Friction Number (FN) and TPIN was yet established. TPIN exhibited non-linear characteristics with respect to the Pavement age ESALS IRI and SR using the universal Levenberg Marquardt hybrid of steepest-descent and least-squares non-linear model fitting technique. INTRODUCTION The Minnesota Department of Transportation (Mn/DOT) constructed the Minnesota Road Research Project (MnROAD) between 1990 and MnROAD is located along Interstate 94 forty miles northwest of Minneapolis/St.Paul and is an extensive pavement research facility consisting of two separate roadway segments containing 51 distinct test cells. Each MnROAD test cell is approximately 500 feet long. Subgrade, aggregate base, and surface materials, as well as, roadbed structure and drainage methods vary from cell to cell. All data presented herein, as well as historical sampling, testing, and construction information, can be found in the MnROAD database and in various publications. Layout and designs used for the Mainline and Low Volume Road can also be found on its web site at

2 Izevbekhai & Khazanovich Page 2 of 19 Low Volume Road Mainline (I-94) (Interstate 94) Fig 1.1 Aerial view of MnROAD Low Volume Road and Mainline Parallel and adjacent to Interstate 94 Mainline (ML) is the Low Volume Road (LVR). The LVR is a 2-lane, 2½-mile closed loop that contains 20 test cells. Traffic on the LVR is restricted to a MnROAD operated vehicle, which is an 18-wheel, 5-axle, tractor/trailer with two different loading configurations. The "heavy" load configuration results in a gross vehicle weight of 102 kips (102K configuration). The legal load configuration has a gross vehicle weight of 80 kips (80K configuration). On Wednesdays, the tractor/trailer operates in the 102K configuration and travels in the outside lane of the LVR loop. The tractor/trailer travels on the inside lane of the LVR loop in the 80K configuration on all other weekdays. This results in a similar number of ESALs being delivered to both lanes. ESALs on the LVR are determined by the number of laps (80 per day on average) for each day and are entered into the MnROAD database. These are adjusted for down times during construction as well as unusual traffic events. The mainline consists of a 3.5-mile 2-lane interstate roadway carrying live traffic. The Mainline consists of both 5-year and 10-year pavement designs. The 5-year cells were completed in 1992 and the 10-year cells were completed in Originally, a total of 23 cells were constructed consisting of 14 Hot Mixed Asphalt (HMA) cells and 9 Portland Cement Concrete (PCC) test cells. Traffic on the mainline comes from the traveling public on westbound I-94. Typically the mainline traffic is switched to the old I-94 westbound lanes once a month for three days to allow MnROAD researchers to safely collect data. The mainline ESALs are determined from an IRD hydraulic load scale was installed in 1989 and a Kistler Quartz sensor installed in The mainline has received roughly 5 million flexible Equivalent Single Axle Loads (ESALS) and 7.8 million Rigid ESALS as of December 31, MnROAD Instrumentation and Surface Data Collection Strength shrinkage, deflection and temperature data collection at MnROAD is accomplished with a variety of methods that describe the layers, the pavement response to loads and the environment as well as actual pavement performance. Layer data is collected from more than 4500 sensors of various types located at various depths. Data flows from these sensors to several roadside cabinets, which are connected by a fiber optic network that is fed into the MnROAD database for storage and analysis. Data can be requested from the MnROAD database for each sensor along with the performance data that is measured throughout the year. Performance data includes Noise, Texture, ride and friction measurements are conducted seasonally. Surface characteristics data is stored in temporary files from which they are subsequently transcribed to the MnROAD database. Such data is not yet directly fed to the database and is thus more expositive to collect, reformat and export. Texture measurements are done with the ASTME Texture Meter as well as the ASTM E-956 sand volumetric technique. Ride measurements are done seasonally with the Mn/DOT pathways van that records an International roughness Index (IRI) value. The vehicle makes one sweep through each of the 4 lanes. The lanes include the Low Volume Road (LVR) inside and outside loop, the Mainline (ML) driving lane and Passing Lane. For each run, a unique File of IRI is recorded and an ERD file is preserved for further analysis. Friction measurement is done according to ASTM E-274 for smooth and ribbed tire measurement. Sound intensity measurements are conducted with a Mn/DOT set-up that records the tire pavement sound intensity without the bias of aerodynamic noise or stack noise. In the near future, On-Board-Sound-Intensity (OBSI) data from MnROAD will be retrievable from the database..

3 Izevbekhai & Khazanovich Page 3 of 19 Figure 1.2: OBSI Set Up with intensity meters and communication cables leading to in-vehicle front end & Computer 1.2 Paper Objectives This paper examines results of TPIN from many cells in MnROAD measured with the On-Board Sound Intensity Method. It investigates any trend between the TPIN and other surface characteristics that include International Roughness Index (IRI), Age in Years, Pavement or surface Type, Surface rating and friction Number. The process consequently seeks to validate the original hypothesis of this research. It is believed that since texture degrades with respect to time and Surface rating and IRI are also time dependent if the ESALs or Load spectrum is maintained through the service period, There should be a correlation between TPIN and the independent variables Literature Review According to Izevbekhai (1) Texture affects ride quality vehicle delay costs, fuel consumption and maintenance costs. The requirements for quiet pavements moderates excessive texturing against skid resistance and hydroplaning.. Hessian drag techniques and longitudinal texturing produce the quietest concrete pavements. The Marquette University Pooled Fund Study (2) study showed that astro-drag favorably compared with many other finishing types, including the bituminous surface, in quietness. Some agencies also use the longitudinal broom drag for surface finishing. When corrective action is required to remove bumps or restore texture in a poorly textured surface, diamond grinding or grooving is used for corrective action. Another report by Hanson et al (3) on their noise and friction testing in Minnesota in 2003/2004 showed that astro-drag surfaces were quieter than tined and diamond ground surfaces within the auditory spectrum. Hanson et al (4) proposed a correlation between noise and texture for all surface finish and texture types (4). Such a correlation agglomerates a group of textures of differing frequencies, pavement stiffness, porosity and absorption and corresponding noise characteristics in one equation. That algorithm underestimates the effect of tine spacing, stiffness, porosity and grinding geometries and other uniform texture patterns that affect the noise spectrum. Wu and Nagi (5), categorized textures into various degrees ranging from micro-texture to mega -texture or giga-texture. Sandberg and Ejsmont (6) identified the surface characteristics influenced primarily by these distinct texture ranges. In the reference manual (6), Sandberg et al attributed wet weather skid resistance

4 Izevbekhai & Khazanovich Page 4 of 19 to surface textures ranging from.07mm to 50mm. The chart also suggested that higher textures may be responsible for increased tire pavement interaction noise. According to Izevbekhai (6, 7& 8) a recent diamond grinding research initiative was performed in 3 test cells in MnROAD in The 3 cells contained the conventional grind, the innovative grind and the pre-existing random transverse tine for control. Texture measurements ranges from 0.3 mm to 0.5 mm prior to grinding. In Cell 8 shoulder texture measurements indicated that original textures 0.8 mm had been original textures were maintained over time. Texture improved in the conventional grind to range of 1.3mm to 1.8 mm. A Texas DOT study (9) examined the effect of a traffic load spectrum or traffic types on OBSI and SPPB noise. Although no models were generated, they (9) indicate the frequency range where the influence of various vehicle types are dominant and on various pavement types. Average tire-pavement noise values ranged from 100 dba on the low end to over 111 dba on the high end. The texture and noise variability observed within section for transverse tining was generally higher than the standard deviation for drag or ground textures. A slight variation in texture geometry appears to have an appreciable influence on noise at the tire-pavement interface. Other variables (particularly texture wavebands) had much more of an influence on noise. With texture and noise variability observed both within-section and between sections of identical nominal texture, only measurements of texture itself during construction remain as a means to consistently control the potential for noise. Texture measurements during construction operations will probably have to go beyond sand patch if texture and noise variability is to be effectively controlled. Only with revised measures of texture will the concrete pavement industry be able to more adequately ensure that a quieter pavement is being constructed. According to Izevbekhai (1) there are at least 7 commonly used surface finishing textures. Dragging an inverted astro-turf behind a concrete paver creates Astro-Turf-drag. Asphaltic-concrete does not require any special finish unless the oil content is excessive and friction is considerably reduced. The Bituminous pavement exhibits a natural texture that is characteristic of the aggregate type and mix design. Burlap Drag is achieved by dragging burlap fabric behind a paver or with a work bridge over a newly paved surface. Some contractors use a broom finish that is applied from a work bridge behind the paver. Diamond grinding is performed with the aid of a diamond-grinding tool that corrects bumps detected by a profiler. A typical diamond grinding is characterized with parallel 1/8 X 1/8 grooves at 1/8 interval. Exposed Aggregate finish is obtained by jetting the desired segment at a convenient time after placement to ensure removal of the surface mortar without impacting the aggregate. It is facilitated with the use of a surface retarder. Etching is a process of controlled chemical blasting of the surface to erode the aggregate and mortar to a desired mean texture depth. Tining is achieved by dragging a rake longitudinally or transversely in the fresh pavement, create ridges at longer intervals but at greater depths than would be obtained by diamond grinding. Texture-planing is the acceptable method of imparting texture without necessarily correcting bumps and dips as well as surfaces poorly textured surfaces service. Such corrective actions are achieved by diamond grinding where the dominant cut is the bump removal and the grooving or unacceptable ride (below the disincentive range) in newly completed projects. Public agencies ensure that the public rides safely on pavements. Consequently, a pavement is expected to satisfy minimum safety requirements against skidding particularly in wet conditions where friction is reduced. 2.0 DATA COLLECTION PROTOCOL This section discusses the test types conducted on the cells and how the various data was collected. The parameters measured in addition to TPIN have been introduced. They include Surface Rating, Age of pavements, IRI and Friction. For the purpose of this exercise friction was not added to the model. It was used to model hysteresis phenomena in other studies on the same cells.

5 Izevbekhai & Khazanovich Page 5 of 19 Ride data was collected using the Mn/DOT Pathways van that is equipped to provide a video log, measure Ride quality and perform some level of distress counting. The pathways van performs IRI measurements at least 3 times a year in each of the MnROAD Lanes. The same profile run made by the pathways van also produces surface data that facilitates computation of the surface rating SR numbers for each test cell. Software crops the data for each cell. The OBSI data was measured with the Mn/DOT OBSI Set up. It consists of a 2006 Chevrolet Impala, a Bruel and Kjaer front end, a Lap top, 4 communications cables and 8 BRC Intensity meters. A Standard Reference Test Tire (SRTT) is used The OBSI set up is shown in figure 2. The Intensity meters are amounted on a rig system that allows tire rotation while ensuring fixity in position and direction of each intensity meter. The Rig system is held on to the wheel of the right side of the vehicle via a system of long bolts which when fastened securely ensures that the Intensity meters are 4 inches above the ground and that each of 2 sets of microphones coincide directionally with the front or rear end of the tire pavement contact patch. To perform a run in a MnROAD Cell, the intensity meters were calibrated according to the procedure set forth by the Inventor Paul Donavan. Testing was done at 60 miles an hour for 5 seconds. To activate the frontend, operator ensures a lead distance to facilitate a cruising speed of 60 miles per hour. It was therefore difficult to measure the TPIN of cells at the transition from loop to straight course in the LVR. A measuring run covers approximately the length of a cell. 2.2 Texture Types in MnROAD Texturing of Pavements in MnROAD followed the historical standards to which concrete pavements were constructed and finished in accordance with the Mn/DOT construction and Specific and Standard for Construction. The newer cells including Mainline Cells 60 to 63 built in 2004 as well as Low volume road cells 32, 52 and 54 built in 2000, 2000 and 2004 respectively were textured to current standard of dragging an inverted turf or broom and ensuring that a texture measured by the sand Volumetric Technique (ASTM E ) of 1mm was obtained. Cell 54 was built in 2004 with taconite aggregate. To study the effect of texture degradation pattern and its implication on ride quality, an untextured strip was created on cell 54 while all other concrete cells in the low volume road were finished with the nonuniform transverse tines that were prevalent at the time of initial MnROAD cells construction in 1993/1994. New textures including the exposed aggregate surface and transverse broom will be created in the 2008 construction. Other proposed innovative surfaces include the pervious concrete full depth pavement, the porous Asphalt full depth pavement and the porous overlay. Concrete textures as at 1994 were built to a standard of 0.8mm Mean Texture Depth (MTD). This was validated in cells 7 and 8 that included 12 foot wide inside shoulder that received no traffic. An average texture of 8mm was consistently measured in the 500 ft long shoulder built at the same time as the Passing and Driving lanes. In 14 years, the cells driving lane textures had degraded texture ranging from 0.4 to 0.8-mm textures. The current texture measurements for cells 37, 7 and 8 are shown as appendix table 3.7. Concrete textures as a 1994 were built to a standard of 0.8mm Mean Texture depth (MTD). This was validated in cells 7 and 8 that included 12 foot wide inside shoulder that received no traffic. An average texture of.8mm was consistently measured in the 500-ft long shoulder built at the same time as the Passing and Driving lanes. In 14 years, the driving lane textures had degraded to texture ranging from 0.4 to 0.8 mm. The current texture measurements for cells 37,7 and 8 are shown in table 3.7 Although annual measurements of texture was not conducted on the Astro turf drag cells. A characteristic decay was observed. Initial texture measurement on cell 32 as at 1999 was 0.8mm and in Cell 54 initial texture was 0.8 mm and Cells 60 to 63 were measured 2.3 Principle of the OBSI. The OBSI agglomerates the component frequencies ranging from 500 hertz to 5000 hertz of the measured section. The component frequencies are logarithmically summed to give a unique OBSI value for that section.

6 Izevbekhai & Khazanovich Page 6 of 19 Based on the logarithmic summation, a reduction of 3 dba is equivalent to a reduction of the reference traffic volume by 50 percent all things being equal. Two equal sources will increase the OBSI value by 3dbA and 3 equal sources will increase the OBSI value by 4.7 dba. Similarly, a reduction of 4.7 DbA is equivalent to a 60 in percent reduction in traffic. This can be observed by substituting the OBSI difference as shown in equations 2.1a to 2.1c. The logarithmic addition though the OBSI values typically appear clustered from 94 to 110dBA, a difference of 0.5 is a significant research level deviation and a difference of 3 is usually felt by the driver or passenger. 10( I OBSI level L s =10 * log10( 1 /10) (I /10)+ (I /10 (I / ).+ 10 n ) 1 L + L = 10 log10 (n X 10L)= (10 log 10 n) + L 2 (Δ L/10) Δ L = 10 log 10 n or n = 10 3 Where n is the number of equivalent sources L is the OBSI level and Δ L is the Change in OBSI due to the increase from a unit source. This is the A-weighted scale that is simply explained by the following expatiation. If n similar sources generate a noise level i db (A), the total Noise level is given by 4 Consequently, if there are 2 sources with the same sound intensity, the cumulative intensity is thus 3 db (A) higher than the individual intensity as shown below. This implies that a reduction of the sound intensity by 3dba is equivalent in effect to a traffic reduction to 50 % of original traffic. DB(A) I2 = 10*log[10 (db(a)i/10) +10 (db)ai/10) ]for 2 sources 5 = 10 log [2 *10 (db (A)/10) ] for 2 sources = 10*( Log 2 + db(a)i) DB( A ) = (i + 3) db(a) 6 I 2 Similarly in the A-weighted metric a reduction of 4.7 d B(A) is equivalent to a 67% reduction in the overall noise level. 2.4 On Board Sound Interaction (OBSI) Collection Process OBSI equipment consists of a Chevrolet Impala, 4 intensity meters, connected via 4 communication cables to a Bruel and Kjaer Frontend Collector connected to a dell laptop computer. The intensity meters are mounted on a rig system attached to a Standard reference test tire that is installed at the rear left side of the vehicle and maintained at a temperature of 30 degrees. After recording temperature, 4 intensity meters were plugged in to the B &K Front End Unit, as well as 12v power supply and Ethernet (computer) cable. With this arrangement, the unit is capable of measuring repeatable tire pavement interaction noise of the tire pavement contact patch at a speed of 60 miles an hour, thus measuring approximately 440 ft within 6 seconds. Mounting the rig on a non-dedicated vehicle and calibration of the microphones as well as durometer evaluation of the tire prior to measurement were mandatory operational procedure, prior to data collection. The existing Excel programs and Fourier transform programs that facilitate analysis of the noise measured simplify subsequent data analysis. Implicit in the Transforms are the window functions that facilitate the program. The Mn/DOT Collection Template is saved the collection laptop s desktop for use in collection OBSI. Pulse software program is used to collect data. A texture noise degradation algorithm is being validated. It is believed that texture degradation results in a reduction of TPIN in concrete pavements while ageing of flexible pavements result in increase of tire pavement interaction noise. While texture degradation of rigid pavements Causes a redistribution of friction into its constituents: hysteresis and adhesion the effect of the tire pavement noise suction component is attenuated by the averaging the reception of two sets of microphone 4 each on the trailing edge and the leading edge of the contact patch.

7 Izevbekhai & Khazanovich Page 7 of 19 Intensity, dba Sound Intensity, 1/3 Octave Bands Leading Edge Trailing Edge Average Frequency, Hz Intensity, dba Sound Intensity, Narrow Band Frequency, Hz Leading Edge Trailing Edge Average Figure 2.1 One third Octave and Narrow band Frequency Respond of 2 sets of Microphones The Mn/DOT Standard equipment consists of a Bruel and Kjaer front end, a rig system developed by Illingworth and Rodkin, the Inventor of the OBSI method and a set of sophisticated microphones supplied by BRC Engineering. The Rig is installed on the Standard Reference Test Tire and connected to the frontend through a series of communication cables. The front-end is also linked to a laptop that allows direct analysis of the noise data equipped with software in convenient logarithm scale is used to mimic the human hearing spectrum. The Mn/DOT equipment is installed on a Chevrolet Impala. The entire MnROAD Cells except those in the transitional curves from loops to straight courses at the LVR cells were measured. It was not practicable to safely attain 60mph measuring speed in the transitional curves and so it was not possible to measure cells 54, and 40 in the LVR during the fall of In Spring of 2008, another measurement strategy was studied. That strategy accommodated measurement of subcells that are less than 500-feet. 3 DISCUSSION AND ANALYSIS OF RESULTS Data sources for other independent variables in this study deserve mention. The age of the cells were easily deduced from the year of construction up to In the case of cells where texture treatments were done on existing cells, the age of the texturing was recorded. However, with flexible pavements, the age of texturing and the age of pavement were considered as reasonable independent variables since bituminous pavements exhibit OBSI changes irrespective of the traffic volume. Generally, bituminous slurry seals on a new pavement will not necessarily produces the same degradation functions as a slurry seal on an old pavement. OBSI levels in Passing lane and Driving lanes were compared. In all cases the Driving lane concrete pavement were quieter than the Passing lane Concrete pavements. Conversely, the bituminous pavements in the passing lane were quieter than their driving lane counterparts. It is worth mentioning again that the LVR is loaded 80 times a day with an 18 Kip 5-axle semi 4 times a week in the inside lane and with 80-trips of the 102-Kip 5-axle Semi once a week. IRI is measured seasonally in all the Cells in MnROAD. A single Loop or run is made in each lane (Mainline) or Loop (LVR). The IRI values obtained in the fall of 2007 were rated against the OBSI data from fall of The loudest pavement in MnROAD was the set of Cells 92 to 97. These cells were badly faulted curled warped and cracked. Moreover the IRI values in these cells were correspondingly high. The effect of IRI on OBSI is dependent on the frequency regimes at which the response of the Quarter Car is chaotic. It appeared that the 12 ft frequency characteristic of the cells in question would bring about lower frequency issues that the OBSI responded to. Moreover, the existence of spalled joints in these cells as well as cracks formed in 6 by 6 blocks influenced ride quality as well as noise. Consequently, the loudest concrete pavement in MnROAD had the worst ride quality. Based on fall 2007test data, cells 60 to 63 were the quietest cells although they were the newest concrete cells built in There are no spalled joints in these cells and even the effect of numerous joints implicit in a 6 by 6 panel did not producer any measurable effect on the TPIN.

8 Izevbekhai & Khazanovich Page 8 of 19 The Original modeling approach was to seek a global TPIN correlation to Age, ESALs and IRI while concealing the identity of each cell type. The Levenberg Marquardt model parameters did not converge and when the initial parameters were chosen from observation of the transverse tined cells preponderant in the data set, each pavement type exhibited unique characteristic residual. That original model stayed for the transverse tined pavement as the sum of absolute residuals converged to zero. Interestingly the residuals grouped the pavements distinguishably into the following groups; bituminous pavements, transverse tined concrete pavements, transverse tined white topping, and turf drag pavements. Another general observation was that there was very insignificant difference between the inside and outside lanes of the low volume loop compared to the mainline. The loading sequence of the 2 lanes is different but the computed damage was conceptualized in initial (1994) design to be similar. In terms of OBSI, the similarity of damage effect on the low volume cells appears to be validated by the fact that the OBSI values are similar. 3.1 Texture Noise Degradation Models for concrete cells In the preferred model, the Function was based on the following function Abs (OBSI) measured = ESAL α Age β IRI μ Texture degradation and IRI for AstroTurf Drag (ATD)Pavements OBSI IRI L m Age ESALS E E E E E E E E E E+05 The linear model OBSI ATD = IRI Age E-07 ESALS OBSI ATD = IRI Age E- 07 ESALS When the ATD cells were separated into a group a linear model was initially investigated. This model looked reasonable but was rejected after father analysis. The model equation 3.1 shows that it is not an actual decrease in the age component of the equation that reduces noise with time. The model imposed a negative factor on the IRI, which increases with time. The implication of this model is that as the pavement degrades it gets quieter. This was not validated in other cells where chaotic levels of IRI were observed and the resulting bump-at-the-joints were transformed into noise. Typical examples observed were cells 92 to 97. These were randomly transverse tined concrete cells that were built over bituminous pavements. Because of the degradation of the joints and the creation of cracks at every six feet in the 12 ft panels, the ride quality was exacerbated. This led to a return to the Multiplicative Exponential Model for where the model parameters may change. This model took into account the Traffic ESALs representing the load input on the pavement, the IRI and the age of the pavement. The Model for the correlation of OBSI to IRI, Age and ESALs For Astro Turf Drag Finish is 3.2 OBSI(cracked White topping) = ESAL ( ) *Age (2.208) *(IRI) ( )

9 Izevbekhai & Khazanovich Page 9 of 19 Cells finished with the Turf drag are few in the LVR and mainline. This is due to the age of the pavements and the standard of finishing that prevailed at the time most of them were built. Noise data for the few Astro turf drag cells could not be modeled along with the transverse tined cells and in spite of the few ATD cells they were influential to the overall absolute residual sum. Removing this data set enabled the modeling of the transverse tined section and the creation of a separate model for the turf drag. Although the sum of absolute residuals SAR of this model is less than what was obtained in equation 3.1 this model is more tenable. Primarily the relationship is non-linear given the parameters in question. Furthermore, the rate of decrease in ESAL contribution is evident in the exponent. Moreover the negative exponent of the IRI contribution explains how higher IRI results in higher noise. Although the model does not directly ascribe decrease in noise to age, this relationship is implicit in the ESAL and IRI function. Where porosity is constant, the preponderant noise feature is the texture component. The loss of texture is brought about by ESALs and that is partially responsible for the noise reduction with respect to time. Another factor of relevance is the ride component that is ordinarily a reflection of the degree of damage brought about by ESAL and Environmental factors. Because all distresses are not traffic volume related, the IRI and the ESALs are not necessarily dependent variables. For a data test the Surface rating was included in the model to study the effect. Table 3.1 Multiplicative exponential model For ATD Cell Number ZETA = ETHER = THETA = Mean OBSI ESAL AGE (YRS) IRI CROPPED OBSI Cal Abs residual E-05 Texture NWT ATD NWT ATD k ATD 32 80k ATD 32 80k ATD k ATD k ATD k ATD Sigma Absolute Residual NWT ATD NWT ATD NWT ATD 3.3 White Topping With Transverse Tine Cells 92 to 97 consist of 4 inch and 5 inch concrete built over bituminous surfaces. In that White topping design, joints were not mainly designed according to the optimum known today. 12ft by 12 ft cells were maintained in order to ascertain what natural crack patterns would dictate where the joints are cut. Panels cracked into 6 X 6 dimension and the joints so formed were irregular. The ride quality in these cells was poor and so was the quietness. When these Whitetopping cells were Included in the overall concrete model, they exhibited influential data characteristics. It was not possible to fit a model of Noise as a function of Ride quality, ESAL and Age for all concrete cells when the white topping cells were included. A plausible explanation resides in the composite nature of the pavement. Concrete over bituminous may have a non-surficial noise effect that is introduced by the bituminous layer. The isolated model parameters are shown in table 3.3.

10 TABLE 3.3 WHITE TOPPING WITH TRANSVERSE TINE Izevbekhai & Khazanovich Page 10 of 19 ZETAT = ETHERT = THETAT= Cell Number Mean OBSI ESAL AGE (YRS) IRI CROPPED OBSI CAL abs residual WT RT WT RT WT RT E-06 WT RT WT RT WT RT OBSI(cracked White topping) = ESAL ( ) *Age (2.208) *(IRI) ( ) Bituminous Multiplicative Model Bituminous cells exhibited an increase in TPIN with respect to time. Passing lane TPIN was generally lower than driving lane TPIN whereas the Weigh -in -Motion (WIM) data showed ESAL values of 9E +06 in and 3 E+06 in the respective lanes. Further investigation of that approach was beyond the present scope. The surface rating (SR) parameter was introduced into the model after an unsuccessful attempt use two age parameters (the age of the current surface and the age of the pavement) to fit the curve. Exponent of ESAL, (Leta) = 0.188; Exponent of IRI (in/mile) (Nita) = 0.041; Exponent of Age (Muta) =1.245; Exponent of SR (Peta) = OBSI (MnROAD Bituminous) = ESAL ( 0.188).) *Age (1.245) *(IRI) (0.041) *(SR)^ Bituminous pavements respond to noise differently form Solid and pervious concrete. Tire pavement interaction noise was already observed in the white topping cell where the measured OBSI values did not fit the general model for full dept transverse tined concrete and of course was different. The exponents of the model indicate the preponderance of age over the other parameters in their bituminous degradation model. The effect of ESALs is more pronounced than all concrete surfaces investigated. This observation is in line with comparative life cycles of concrete and bituminous. The surface rating component (SR) was introduced in the model to accommodate the surface defects that do not necessarily affect Ride quality. The Model parameters indicate that each of the component parameters is important. Respective exponents of ESALs and Age parameters indicate the preponderance of these 2 parameters over IRI and SR in the OBSI model. Due to the order of magnitude of ESALs the 0.18 exponent attenuates it to the typical early 3 digit value of OBSI. The 1.24 exponent of age is an amplifying function that is meaningful within the age range of pavements and service life of pavements measured. 3.5 Multiplicative Exponential Model Model For Transverse Tined Pavements alpha = ; beta = 1.767; mu = OBSI (Transverse tine) = ESAL ( ) *Age (1.7678) *(IRI) ( )

11 Izevbekhai & Khazanovich Page 11 of 19 The preponderant texture in MnROAD is the Transverse tined pavement. This was the state of the art of texturing when the original MnROAD Cells were constructed. And currently form the majority of concrete pavement textures in MnROAD. Since these are in various degrees of degradation, the surface rating was suggested as an n additional parameter for modeling the surface Characteristics. The transverse tine model does not readily describe the original OBSI value at the inception of service. The model also desensitizes ESAL with the insensitive equation (ESAL^ ) that tends to 1 irrespective of the order of magnitude of A and amplifies the effect of age. Observation in MnROAD test cells show that degradation of transverse tine textures in general is related to mix design and texture type. Transverse textures are not easily susceptible to degradation as the actual indentation on the pavements are less frequent and far between when compared to other texture types. Random transverse textures are quieter than conventional uniform transverse textures transverse because the former is characterized by a unique frequency tone. This tone is chaotic when the interval is inch and the vehicle moves at 60 miles per hour (11) Table 3.4 OBSI Data for Transverse Tined Concrete

12 Izevbekhai & Khazanovich Page 12 of 19 Table 3.5 MnROAD TEST Cells and Corresponding Textures ML C e ll Const Year of Recent Surface Texture Type Current Textures Age LVR C e lls Const Year of Recent Surface Texture Type Current Textures Age Bit Bit (14) Bit Bit (14) Bit Micro Bit Bit (4) Bit Bit Bit Bit Bit Micro Oil Gravel TT+BD TT+BD (15) Bit TT+BD TT+BD (15) Bit Bit (1) TT+BD TT+BD (15) Oil Gravel TT+BD TT+BD (15) Bit TT+BD TT+BD (15) Oil O il G r. (7) Gravel TT+BD TT+BD (15) Oil Gravel TT+BD TT+BD (15) Bit TT+BD TT+BD (15) Micro Micro (4) TT+BD TT+BD (15) Micro Micro (4) Bit Micro (4) /2004 Bit/Tac Bit/Tac (3) Bit Micro (3) Various Various (7) B it M icro (3) Bit Bit (3) B it M icro (3) Bit Bit (8) Bit Micro (4) Bit Bit (8) B it M icro (4) TT+BD TT+BD (13) B it M icro (4) TT+BD TT+BD (13) B it M icro (4) TT+BD TT+BD (13) B it M icro (4) TT+BD TT+BD (13) Bit Micro (3) TT+BD TT+BD (13) TT+BD ATD (3) ATD ATD (7) TT+BD ATD (3) ATD ATD (7) TT+BD ATD (3) ATD ATD (7) TT+BD ATD (3) TT+BD ATD (3) TT+BD TT+BD (10) TT+BD TT+BD (10) TT+BD TT+BD (10) TT+BD TT+BD (10) TT+BD TT+BD (10) TT+BD TT+BD (10)

13 Izevbekhai & Khazanovich Page 13 of 19 Table 3.6 MnROAD Historical Friction Measurements on a Typical Cell C ell 7 Driving Ribbed RTF PASSING Ribbed RTF D lane R TF Plane R TF M L-Driving-RL 23-Jun Ribbed ML-Passing-LL 23-Jun Ribbed M L-Driving-RL 29-Oct Ribbed ML-Passing-LL 23-Jun Ribbed M L-Driving-RL 20-Oct Ribbed ML-Passing-LL 20-Sep Ribbed M L-Driving-RL 31-Oct Ribbed ML-Passing-LL 4-May Ribbed M L-Driving-RL 3-Nov Ribbed ML-Passing-LL 20-Jun Ribbed M L-Driving-RL 24-May Ribbed ML-Passing-LL 29-Oct Ribbed M L-Driving-RL 19-Apr Ribbed ML-Passing-LL 20-Oct Ribbed M L-Driving-RL 24-Oct Ribbed ML-Passing-LL 31-Oct Ribbed ML-Driving-RL Ribbed ML-Passing-LL 3-Nov Ribbed ML-Driving-RL Ribbed ML-Passing-LL 24-May Ribbed ML-Driving-RL Ribbed ML-Passing-LL 19-Apr Ribbed ML-Driving-RL Ribbed ML-Passing-LL 24-Oct Ribbed ML-Passing-LL 14-Oct Ribbed Driving Smooth STF Passing Smooth STF D Lane STF P Lane STF M L-Driving-RL 14-Oct Smooth 14-Oct Smooth M L-Driving-RL 20-Oct Smooth ML-Passing-LL 20-Oct Smooth M L-Driving-RL 19-Apr Smooth ML-Passing-LL 3-Nov Smooth M L-Driving-RL 24-Oct Smooth ML-Passing-LL 19-Apr Smooth C ell 8 Driving Ribbed RTF PASSING Ribbed RTF D lane R TF P Lane R TF M L-Driving-RL 23-Jun Ribbed ML-Passing-LL 23-Jun Ribbed M L-Driving-RL 20-Sep Ribbed ML-Passing-LL 20-Sep Ribbed M L-Driving-RL 4-May Ribbed ML-Passing-LL 4-May Ribbed M L-Driving-RL 20-Jun Ribbed ML-Passing-LL 20-Jun Ribbed M L-Driving-RL 29-Oct Ribbed ML-Passing-LL 29-Oct Ribbed M L-Driving-RL 14-Oct Ribbed ML-Passing-LL 14-Oct Ribbed M L-Driving-RL 20-Oct Ribbed ML-Passing-LL 20-Oct Ribbed M L-Driving-RL 31-Oct Ribbed ML-Passing-LL 31-Oct Ribbed M L-Driving-RL 3-Nov Ribbed ML-Passing-LL 3-Nov Ribbed M L-Driving-RL 24-May Ribbed ML-Passing-LL 24-May Ribbed M L-Driving-RL 19-Apr Ribbed ML-Passing-LL 19-Apr Ribbed M L-Driving-RL 24-Oct Ribbed ML-Passing-LL 24-Oct Ribbed Driving Smooth STF Passing Smooth STF D Lane STF Plane STF M L-Driving-RL 14-Oct Smooth ML-Passing-LL 14-Oct Smooth M L-Driving-RL 20-Oct Smooth ML-Passing-LL 20-Oct Smooth M L-Driving-RL 19-Apr Smooth ML-Passing-LL 3-Nov Smooth M L-Driving-RL 19-Apr Smooth ML-Passing-LL 19-Apr Smooth M L-Driving-RL 24-Oct Smooth

14 Izevbekhai & Khazanovich Page 14 of 19 Table 3.7 A Sand Patch Test result on Monitored Locations in MnROAD Test Cells. Spots Provide Texture Degradation Data M easured By Bernard Izevbekhai Sand Patch ASTM E-965 A ra s h M o in C e ll 3 7 Date 6/19/07 Tim e 10:00am Tem p 65 Deg F S ta tio n D IA 1 D IA 2 D IA 3 R U N 1 R U N 2 R U N 3 A verage V o l m m 3 T ex tu re (m m ) TX9 TS TX9 TS TX8TS TX7 TS TX6 TS TX5 TS TX4 TS TX3 TS TX2 TS TX 1TS Arash M oin POST GRINDING Date 6/22/07 Sand Patch Test Tim e 11:10 AM C e ll 3 7 S ta tio n D IA 1 D IA 2 D IA 3 R U N 1 R U N 2 R U N 3 A verage V o l m m 3 T ex tu re (m m ) TX9 TS TX8 TS TX7 TS TX6 TS TX5 TS TX4 TS TX3 TS TX2 TS TX1 TS S ta tio n D IA 1 D IA 2 D IA 3 R U N 1 R U N 2 R U N 3 A verage V o l m m 3 T ex tu re (m m ) TX9 TS TX8 TS TX7 TS TX6 TS TX5 TS TX4 TS TX3 TS TX2 TS TX1 TS S ta tio n D IA 1 D IA 2 D IA 3 R U N 1 R U N 2 R U N 3 A verage V o l m m 3 T ex tu re (m m ) TX9 TS TX8 TS TX7 TS TX6 TS TX5 TS TX4 TS TX3 TS TX2 TS TX1 TS S ta tio n D IA 1 D IA 2 D IA 3 R U N 1 R U N 2 R U N 3 A verage V o l m m 3 T ex tu re (m m ) TX9 TS TX8 TS TX7 YS TX6 TS TX5 TS TX4 TS TX3 TS TX2 TS TX1 TS

15 Izevbekhai & Khazanovich Page 15 of 19 Table 3.8 Friction Data from Innovative Grinding Cells CELL LANE DAY TIME FN SPEED AIR_TEMP TIRE_TYPE EQUIPMENT STA DATE_UPDA Texture E- 274 Type 7 Driving Tuesday 10:32 AM Ribbed 7 Driving Tuesday 10:32 AM Ribbed 7 Driving Tuesday 10:32 AM Smooth 7 Driving Tuesday 10:32 AM Smooth 7 Passing Tuesday 11:08 AM Ribbed 7 Passing Tuesday 11:08 AM Ribbed 7 Passing Tuesday 11:08 AM Smooth 7 Passing Tuesday 11:08 AM Smooth 8 Driving Tuesday 10:32 AM Ribbed 8 Driving Tuesday 10:32 AM Ribbed 8 Driving Tuesday 10:32 AM Smooth 8 Driving Tuesday 10:32 AM Smooth 8 Passing Tuesday 11:08 AM Ribbed 8 Passing Tuesday 11:08 AM Ribbed 8 Passing Tuesday 11:08 AM Smooth 8 Passing Tuesday 11:08 AM Smooth Friction 22-Oct Innovative Friction 22-Oct Innovative Friction 22-Oct Innovative Friction 22-Oct Innovative Friction 22-Oct Innovative Friction 22-Oct Innovative Friction 22-Oct Innovative Friction 22-Oct Innovative Friction 22-Oct Conventional Friction 22-Oct Conventional Friction 22-Oct Conventional Friction 22-Oct Conventional Friction 22-Oct Conventional Friction 22-Oct Conventional Friction 22-Oct Conventional Friction 22-Oct Conventional

16 Izevbekhai & Khazanovich Page 16 of 19 Table 3.8 Unadjusted Traffic ESALs from MnROAD LVR ESALS FOR MNROAD LVR HMA Test Cells (24-31) PCC Test Cells (36-40) 80K Lane 102K Lane 80K Lane 102K Lane Year Total Total Total Total Total Total Total Total Laps ESALs Laps ESALs Laps ESALs Laps ESALs ,245 7, ,760 3,245 12, , ,979 25,059 3,830 28,962 10,979 40,999 3,830 44, ,577 44,680 6,406 48,496 19,577 73,123 6,406 74, ,451 60,379 8,578 64,981 26,451 98,808 8,578 99, ,384 83,074 11,046 83,713 36, ,930 11, , ,070 93,781 12,861 97,487 41, ,434 12, , , ,173 15, ,188 50, ,832 15, , , ,594 17, ,052 57, ,324 17, , , ,912 19, ,141 67, ,723 19, , , ,482 21, ,466 72, ,328 20, , , ,433 22, ,969 76, ,020 22, , , ,676 24, ,749 82, ,746 24, ,787

17 Izevbekhai & Khazanovich Page 17 of 19 Table 3.9 Historical (14 year) Texture degradation. More Texture Observed in the Passing Lane PREGRIND Measured By Bernard Izevbekhai Sand Patch ASTM E-965 Cell 7 &8 10/15/07, 10/16/07 Time 12:00pm Temp 55 Deg F LOCATION WHEELPATHRUN 1 RUN 2 RUN 3 Average V ol mm3 Texture (mastm E-27CTM Check Cell 8 BX1 RR Cell 8 BX1 RL Cell 8 BX1 LR Cell 8 BX1 LL Cell 8 BX2 RR Cell 8 BX2 RL Cell 8 BX2 LR Cell 8 BX2 LL Cell 8 BX3 RR Cell 8 BX3 RL Cell 8 BX3 LR Cell 8 BX3 LL Cell 8 BX4 RR Cell 8 BX4 RL Cell 8 BX4 LR Cell 8 BX4 LL Cell 8 BX5 RR Cell 8 BX5 RL Cell 8 BX5 LR Cell 8 BX5 LL Cell 8 BX6 RR Cell 8 BX6 RL Cell 8 BX6 LR Cell 8 BX6 LL Cell 8 Bx7 RR Cell 8 Bx7 RL Cell 8 Bx7 LR Cell 8 Bx7 LL Cell 7 BX8 RR Cell 7 BX8 RL Cell 7 BX8 LR Cell 7 BX8 LL Cell 7 BX9 RR Cell 7 BX9 RL Cell 7 BX9 LR Cell 7 BX9 LL Cell 7 BX10 RR Cell 7 BX10 RL Cell 7 BX10 LR Cell 7 BX10 LL Cell 7 BX11 RR Cell 7 BX11 RL Cell 7 BX11 LR Cell 7 BX11 LL Cell 7 BX12 RR Cell 7 BX12 RL Cell 7 BX12 LR Cell 7 BX12 LL Cell 7 Bx13 RR Cell 7 Bx13 RL Cell 7 Bx13 LR Cell 7 Bx13 LL Cell 8SH Bx14 RR Cell 8SH Bx14 RL Cell 8SH Bx15 RR Cell 8SH Bx15 RL Cell 8SH Bx16 RR Cell 8SH Bx16 RL Cell 8SH BX17 RR Cell 8SH BX17 RL Cell 8SH Bx18 RR Cell 8SH Bx18 RL Cell 8SH BX19 RR Cell 8SH BX19 RL Cell 8SH Bx20 RR Cell 8SH Bx20 RL Cell 8SH BX21 RR Cell 8SH BX21 RL