Evaluation of mechanical parameters of an asphalt mixture under the environmental conditions of Bogotá D.C

The present paper introduces the evolution of resilient modulus, permanent strain and fatigue life of an asphaltic concrete mixture exposed under weather conditions in Bogota D.C. (Colombia) during 47-48 months. Mixtures were elaborated by using two asphaltic cements (CA) made in Colombia: CA 80100 and CA 60-70. Mixtures generally tend to increase their stiffness values after an exposure period, mainly due to hardening and ageing processes undergone by asphaltic binder. A stiffness gain leads to increase endurance against rutting phenomenon and fatigue life of the analyzed mixture employed to lay thick asphaltic layers on the pavement. Based on the resilient modulus evolution due to exposure under environmental conditions, an equation is proposed to approximately predict the evolution of fatigue life endured by a mixture.

Evaluación de los parámetros mecánicos de una mezcla asfáltica/Evaluation of mechanical parameters of an asphalt mixture The main degradation mechanisms intended to be controlled by empirical and mechanistic methodologies of pavement design are fatigue and the excess of permanent deformation on vertical road direction (rutting) (e.g., Shell, 1978;TAI, 1982;AUSTROADS, 1992;AASHTO, 1993, TRL, 1993;HMSO, 1994; Urban Development Institute and Universidad de Los Andes, 2002;MEPDG, 2004).Fatigue phenomenon caused by cyclic vehicular loads take place inside bonded layers.In the case of flexible structures, this phenomenon takes place because of high deformation values due to tensile stress in the lowest area of an asphaltic layer.The continuous vehicle traffic makes the asphaltic layer bending down thus generating tensile stresses in its lower area.
Such repeated load stress causes a stiffness loss of the asphaltic mixture and provokes plastic deformations due to tensile stress, which in turn lead to micro-crackings.Because of continuous and repeated load stress and stiffness loss, micro-cracking coalescence takes place leading to the occurrence of a visible cracking (macrocracking).If the fatigue phenomenon occurs due to load stress, macro-crackings take place in the lower area of the asphaltic layer (which is the area where the tensile stress is higher).Macro-crackings are spread upwards to the road pavement surface where they are reflected as parallel crackings to the lengthwise carriage way direction, which under repeated traffic loads are spread creating angular pieces that develop a map-cracking appearance.If this phenomenon takes place due to temperature conditions, the crackings spreading effect occurs in the opposite way, thus micro-crackings are provoked from the grade line and they are spread towards the inner fiber in the asphaltic layer.On the other hand, rutting is the residual vertical deformation accumulated due to traffic load, which may cause structural or functional failures to the road pavement.In an asphaltic layer, this phenomenon is mainly controlled by means of its own mixture stiffness (dynamic or resilient modulus).
In Colombia there are few researches developed on asphaltic mixtures fatigue phenomena mainly because of the lack of appropriate equipment to conduct field and laboratory measurements and also due to a small budget for research purposes.Such situation explains the lack of knowledge on this subject and the lack of an appropriate model characterizing fatigue behavior from material properties (mechanical, chemical, rheological and physical point of view), environmental conditions and traffic behavior.
Generally, fatigue tests on asphaltic mixtures consist in placing a specific geometric sample under repeated and identical load stresses; in order to determine the number of load cycles the sample is able to endure (Di Benedetto and De la Roche, 2005).This number of load cycles endured before failure is known as fatiguelife.According to Epps and Monismith (1972), Rao Tangella et al. (1990) and Di Benedetto et al. (2004) Di Benedetto andDe la Roche (2005), the fatigue response from asphaltic mixtures is mainly affected by: Evaluación de los parámetros mecánicos de una mezcla asfáltica/Evaluation of mechanical parameters of an asphalt mixture Sample compactation methods.Loading mode and the type of test applied.Mixture design parameters.Variants associated with the environment.
A detailed reference about fatigue phenomenon on asphaltic mixtures may be consulted from González and Vásquez (2009).
On the other hand, the main factors affecting asphaltic mixtures durability, assuming they are properly constructed, are: ageing rate, and moisture damage (Airey, 2003).Therefore, an asphaltic mixture must be designed and constructed not only to endure traffic loads but also environmental conditions.Evaluating the influence of environmental conditions on this type of material behavior is not an easy task.Nowadays it is carried out by isolating every single component (water, temperature, ultra-violet rays, among others) to separately evaluate their influence on binder and mixtures (from a mechanical and chemical point of view) (Kemp and Predoehl, 1981;Welborn, 1984;Kim et al., 1987;Shiau et al., 1991;Bishara et al., 2000;Bocci and Cerni, 2000;Brown and Scholz, 2000;Khalid and Walsh, 2000;Khalid, 2002;Airey, 2003;Said, 2005;Shen et al., 2006).Additionally and generally, these researches employ test methods and equipment (e.g.thin-film-oven test TFOT, rolling-thin-film-oven test RTFOT, microwave ovens, automatic pressure aging vessels PAV), which can not fully reproduce the influence of each environmental component on mechanical and rheological properties of mixtures and asphaltic binders (Jemison et al., 1991;Choquet and Verhasselt, 1992;Kuppens et al., 1997;Verhasselt, 1997).Furthermore, a significant amount of researches tend to combine these tests and equipment to evaluate material properties changes under real environmental conditions (Jemison et al., 1991;Migliori and Corté, 1998;Montepara, 1999;Montepara and Giuliani, 2000;Airey, 2003).As a general conclusion these test methodologies are unable to reproduce real conditions endured by asphaltic mixtures in situ.A method to measure the environmental influence on the behavior of asphaltic mixtures is by developing segments test or tracks test at scale units.Hugo Alexander Rondón Quintana, Fredy Alberto Reyes Lizcano The limitation of these tests is their high cost and they can not directly or separately measure the influence of traffic loads and environmental conditions.
Consequently, this paper introduces a study that intends to evaluate the influence of environmental conditions in Bogota D.C (Colombia) on stiffness and fatigue strength of a dense hot asphaltic mixture type MDC-2 (as per specifications by INVIAS, 2007 -Roads National Institute).It was decided to analyze this type of mixture as it is commonly used to build tread layers, which are directly exposed to environmental conditions.In the same way, climatic conditions of Bogotá D.C. were chosen because the city is located in an area where the following environmental conditions are present during a given day: predominantly cold with minimum and maximum average temperatures of 5°C and 19°C, respectively and; periodical rains at any moment in the same day.In order to accurately understand the environmental influence on the behavior of asphaltic mixtures, further researches will analyze different climate conditions.This paper introduces the results from the first 47-48 months of the project, which evaluates mixtures elaborated with the two CAs produced in Colombia: CA 60-70 and CA 80-100.Regarding climatic conditions, these binders are employed provided that yearly average temperature is higher and lower than 24°C, respectively in the area where the asphaltic layer will be built (INVIAS, 2007).
In order to evaluate the environmental influence on the mixture mechanical parameters, dynamic characterization tests were developed, such as resilient modulus, permanent deformation strength and fatigue law.Fatigue tests were conducted under controlled stress.Table 1 summarizes the main differences observed between tests under controlled stress and controlled deformation.
Tabla 1. Comparación de los ensayos bajo esfuerzo controlado y deformación controlada (Epps y Monismith, 1972;Di Benedetto y De la Roche, 2005) Table 1.Comparison between tests under controlled stress and controlled deformation (Epps and Monismith, 1972;Di Benedetto and De la Roche, 2005) Evaluación de los parámetros mecánicos de una mezcla asfáltica/Evaluation of mechanical parameters of an asphalt mixture Presently the investigation team, under a doctorate level study, is evaluating the effect of each component of Bogota environmental conditions and the loads on the mixture chosen for this research.For this team it was important to get acquainted with the behavior of every environmental component, in order to clearly describe and understand the results provided by micro level studies (isolated components).

Methodology
Design of asphaltic mixtures and characterization of materials.The quarry rock aggregate used for the elaboration of asphaltic mixtures to be employed in Marshall Test (called briquettes) comes from "Subachoque" quarry (Cundinamarca, Colombia).The following tests were conducted on such materials, in accordance with specifications by the Roads National Institute  In order to characterize mixtures the following standard tests were conducted on asphaltic cements CA 80-100 and CA 60-70 according to (INVIAS 2007) specifications: penetration (ASTM D-5), dynamic viscosity (ASTM D-4402), ductility (ASTM D-113), solubility in trichloroethylene (ASTM D-2042), water content (ASTM D-95), softening point (INV.E-712), residual analysis after rolling-thin-film-oven test (RTFOT) ASTM D-2872).These tests results are presented in Tables 3 and 4.
After analyzing tests conducted on quarry rock aggregates and on asphaltic cements, five briquettes were elaborated (compacted with 75 strikes per face) for each asphalt percentage between 4.5% and 6.5%, in order to develop Marshal Design (INV.E-748, INVIAS, 2007a), so as to determine the optimum asphalt content of the mixture.In order to meet INVIAS (2007) specifications and to elaborate asphalt mixtures type MDC-2, the original aggregates grading was modified by using as a reference the percentage average values from grading range, established by the specification for briquettes elaboration for Marshall Test.   5 and 6.The optimum percentages of asphaltic cements, according to data in tables 5 and 6, are 5.3% and 5.6% for mixtures elaborated with CA 80-100 and CA 60-70, respectively.These percentages meet the minimum requirements established by INVIAS ( 2007) specification for MDC-2 and traffic types NT1 and/or NT2 (low and medium traffic volumes).
Tabla 5. Resumen del ensayo Marshall para mezcla asfáltica MDC-2 con CA 80-100  (10, 20 and 30°C), and load frequencies (2.5, 5.0 and 10.0 Hz) by using the Nottingham Asphalt Tester (NAT).The permanent deformation resistance test, under repeated load cycles, was conducted under 100 kPa stress at 3600 load cycles in accordance with standard procedure by EN 12697-22 (CEN, 2005).By using these tests results, an empirical equation was developed in order to predict the strength evolution of the analyzed mixture, in relation to the exposure time period to environmental conditions.Tabla 7. Variables para simular la rigidez de la mezcla MDC-2 en función de T y F Table 7. Variants used to simulate MDC-2 mixture stiffness in function of T and F Hugo Alexander Rondón Quintana, Fredy Alberto Reyes Lizcano These variants do not deliver a defined physical meaning and they mainly change with the type of binder, grading and type of rock aggregate employed in the elaboration of mixtures.Load frequency F in the Equation (1) may be translated as the vehicle speed (V, in cm/s) by means of equation ( 2) proposed by Lin (1989Lin ( , 1989a)).
t c is loading time expressed in seconds for an in-motion vehicle at different depths (h in cms and measured from the pavement structure surface) and L is the wheel rim length in contact with the road pavement (usually 30 cms). ( (2) Evaluación de los parámetros mecánicos de una mezcla asfáltica/Evaluation of mechanical parameters of an asphalt mixture Figure 2 indicates the evolution of resilient modulus for asphaltic mixtures throughout time.Mixtures fabricated from CA 80-100 and CA 60-70 (Figures 2a and  b, respectively) show a typical modulus increase when load frequency is incremented and test temperature is decreased.In mixtures fabricated from CA 80-100, the modulus tends to increase throughout exposure time under environmental conditions (see Figure 2a) and, therefore, permanent deformation strength increases (Figure 3).Such phenomenon takes place mainly due to oxidative ageing experimented by asphaltic cements due to temperature and ultra-violet (UV) radiation exposure effects.

Vertical permanent deformation
Hugo Alexander Rondón Quintana, Fredy Alberto Reyes Lizcano Mixtures fabricated from CA 60-70 experiment a different behavior in comparison to those fabricated from CA 80-100 (see Figures 2b and 3).During the first exposure months, the mixtures decrease their modulus by increasing permanent deformations.Afterwards such modulus is stabilized and finally there is an increasing tendency to decrease deformation values (especially when the test is developed at temperatures ranging from 20 to 30°C).Probably the modulus decay is produced by micro-cracking generation, which appears at low temperatures when mixtures are too rigid (fragile behavior).Then those micro-cracking are stabilized and the stiffness increases due to an ageing and asphalt oxidative phenomenon similar to the one taking place in CA 80-100 mixtures.Micro-cracking phenomenon at low temperatures has been widely studied by several researchers (e.g., Kliewer et al., 1996;Sebaaly et al., 2002;Nesnas and Nunn, 2006).
Water effect is not clear from observed results, since mixture humidity produces de-bonding between rock aggregate and asphalt; probably provoking a modulus decrease and mechanical resistance drop.Additional researchers in this field shall be conducted.
Figure 4 presents the relation (E/E o ) between resilient modulus obtained after exposing briquettes under environmental conditions during different periods of time (E) and, the initial mixtures resilient modulus for a period of t=0 months (E o ).In the case of mixtures elaborated from CA 80-100, the modulus obtained from t=48 increases from 48 to 176%, in relation to the initial modulus (depending on the temperature and frequency of the test).The most significant increment is observed as long as temperature increases in the test.CA 60-70 mixtures experiment a maximum modulus decrease of 35% during the first five months of exposure.However, after t=47 months, these mixtures increase and experiment modulus that slightly supersede the initial value.These results can be mathematically expressed by means of equations (3a) and (3b) for CA 80-100 and CA 60-70 asphaltic mixtures.This equation can only be employed for the analyzed mixture in the period of time under analysis (47-48 months), because the mixture behavior might change as long as t increases.As well as in equation ( 1), in equation (3) the status variants k1-k3 are obtained by means of regressions and do not provide a defined physical meaning (see Table 8).
Figure 5a presents the results from fatigue testing developed on 360 samples of asphaltic concrete type MDC-1, MDC-3 and MDC-3.There is amplitude value decrease max (deformation amplitude for material failure when N f load cycles are applied) as long as resilient modulus is incremented in mixtures (E), which is equivalent to the status where the material is changing its behavior from ductile to fragile due to stiffening.These results can be mathematically represented by means of Equation ( 5).This equation was chosen taking into consideration that for a very low deformation amplitude applied on the material, it can endure such amplitude infinite times without rupturing (concept of fatigue endurance limit stated by Carpenter et al., 2003 andReyes, 2003).Status variant values in the equation ( 5) were determined by means of regression: k 4 =6.24x10-6,k 5 =0.48 and k 6 =15.79.Similar results as the ones found in this research were reported by Schmidt and Santucci (1969).Although material allows lower tensile deformations as long as it stiffens, fatigue-life or the number of cycles required for material failure (N f ) increases, as described by equation ( 6) (see figure 6).Such results are in accordance with those reported by the referred biography.Evaluación de los parámetros mecánicos de una mezcla asfáltica/Evaluation of mechanical parameters of an asphalt mixture If equations (4-5) are fused, the evolution experimented by ?max with the evolution of mixtures stiffness exposed to environmental conditions in Bogota D.C. are obtained.By using these equations it is possible to predict that acceptable tensile deformation of the mixture will decrease from t inicial =0 to t=29 months, from 55 to 64%, only due to the effect of increased stiffness during exposure to environmental conditions in Bogota D.C.In the same way, during the same period of time the equation ( 6) reports an average increase of fatigue-life 6.85 times higher, due to the mixture stiffening effect.From t=30 up to t=48 months, stiffening effect decreases and the mixture will experiment a fatigue-life increase of 10 times, due to stiffening from t inicial =0 up to t inicial =48 months.
The analysis of fatigue-life for CA 60-70 mixtures is far more complex, since during the first exposure months under environmental conditions, stiffness drops in an average of 12.8 % causing a reduction of 33 % in the mixture fatigue-life.However, between t=5 and t final = 47 months, it reappears and increases stiffness, thus generating at the end of exposure time a slight average increment of fatigue-life of 38 % in comparison to t inicial = 0 months.
(2007a): grading analysis on coarse and fine aggregates (INV.E-213), fine aggregates absorption and specific weight (INV.E-222), coarse aggregates absorption and specific weight (INV.E-223), aggregates abrasion resistance (sizes lower than 3/4") by employing the Los Angeles apparatus (INV.E-218); Micro-Deval abrasion test (INV.E-238), strength loss test (INV.E-220), fractured particles (INV.E -227), del modulo de rigidez Stiffness modulus increase Efecto del tiempo de receso Effect of recess time Duración de la propagación de las macro-fisuras Macro-cracking spreading duration lo que ocurre in situ More representative than in situ Tabla 2. Caracterización de los agregados asfalto original Tests developed on original asphalt Ensayos sobre el residuo luego del RTFOT Test developed on residue after RTFOT Calculations obtained from Marshall Test are registered in Tables

Figura 6 .
Figura 6. Evolución del número de ciclos de falla con /E Figure 6.Evolution of number of failure cycles with /E

Table 2 .
Aggregates Characterization Hugo Alexander Rondón Quintana, Fredy Alberto Reyes Lizcano test to measure sand content (INV.E -133) and elongation -flattening indexes (INV.E-230).These test results are presented in table 2, which values meet the minimum quality standards established by INVIAS (2007) specifications for the elaboration of MDC-2 mixtures used for tread layers construction.

Table 4 .
General Characteristics of CA 60-70 asphaltic cement Evaluación de los parámetros mecánicos de una mezcla asfáltica/Evaluation of mechanical parameters of an asphalt mixture

Table 5 .
Summary of Marshall Test for MDC-2 asphaltic mixture containing CA 80-100