Construction and Building Materials

Finite element-based framework to study the response of bituminous concrete

pavements under different conditions
--Manuscript Draft--

Manuscript Number:

Article Type: Research Paper

Keywords: effect of air voids; Indian traffic conditions; generalised Kelvin model; creep
compliance; resilient modulus

Corresponding Author: Abhinav Kumar, B.Tech.

VARANASI, UTTAR PRADESH INDIA

First Author: Abhinav Kumar, B.Tech.

Order of Authors: Abhinav Kumar, B.Tech.

Ankit Gupta

Kumar Anupam

Vivek Pratap Wagh, M.Tech

Abstract: HIGHLIGHTS

• Linear viscoelastic material characterization of BC-2 is presented.

• Lab evaluation of creep compliance has been highlighted.
• Effect of temperature and air void on viscoelastic properties of BC-2 is identified.
• A direct approach to find Prony series constants has been discussed.
• Pavement response based on linear viscoelastic properties of BC-2 has been
predicted.
• Effect of contact pressure and load on pavement response analysed statistically.

Suggested Reviewers: Nikhil Saboo, Ph.D

Assistant Professor, Indian Institute of Technology Roorkee
[email protected]
His research domain belongs to Viscoelastic Characteristics of Asphalt Binders and
Mixes, Pavement Evaluation, Design and Analysis, Pavement Materials and Testing.
These domains are similar to research area covered in our paper.

Prabin Kumar Ashish, Ph.D

Assistant Professor, Indian Institute of Technology Kanpur
[email protected]
His research domain belongs to Performance based mechanistic characterization of
asphalt binder and mixes which is similar to research area covered in our paper.

Denis Jelagin, Ph.D

Associate Professor, KTH Royal Institute of Technology
[email protected]
His research area are mix performance, response of flexible pavement, framework for
modelling which are similar to our work

Sudhir Varma, Ph.D

Assistant Professor, Indian Institute of Technology Patna
[email protected]
His research domain belongs to Characterization of pavement materials, constitutive
modeling of pavements, asphalt mix design which are similar to our research work

Benjamin Underwood, Ph.D

Professor, NC State University
[email protected]
His research domain belongs to performance of asphalt mix, mastics, rheological
behaviour of binders which are similar to our research work

Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation
Cover Letter

To, Date: 26/08/2023

Editor In Chief
Construction and Building Materials

Sub: Submission of Manuscript for Possible Publication

Respected Sir,

Manuscript titled “Finite element-based framework to study the response of bituminous concrete pavements
under different conditions” by Abhinav Kumar, Ankit Gupta, Kumar Anupam, and Vivek Pratap Wagh is
submitted for possible publication in your journal, Construction and Building Materials.

In most of developing countries across the world, pavement design is still based on an empirical approach that may
result in premature failure or overdesigned pavements. A shift from an empirical to a semi-mechanistic or
mechanistic approach is the need of modern time. In this regard, computational tools such as finite element (FE) are
being successfully utilized to gain deeper insights because such tools have allowed researchers to study the complex
behaviour of bituminous concrete (BC) materials. It is well recognized that BC material typically exhibits
viscoelastic/visco-elasto-plastic behaviour depending on applied loading (including temperature) conditions.
However, due to the complexity of the whole procedure yet many pavement design tools consider them as pure
elastic material. The aim of this research is to develop FEM based simple and practical framework to evaluate the
structural response of BC material with viscoelastic material characterization which can be an effective tool to
predict field behaviour with commonly available pavement material tests. Such a framework will be helpful in
analysing variations in the critical response of BC pavement with varied traffic loads and ambient temperatures. The
framework provides a relatively simple procedure to obtain the viscoelastic parameters of BC mix with a creep
compliance test conducted at different temperatures. It was concluded that Creep compliance data if pre-smoothened
by the Power law model reduces mathematical optimization issues to some extent. Furthermore, with the obtained
parameters, a 3-dimensional FE model was developed to obtain sensitivity to critical stresses, strains, and
deformations at desired conditions. Material characterization of unbound granular layers was evaluated through
resilient modulus based on empirical relations. Analysis was carried out taking into consideration the traffic load,
contact pressure, mix type, air-void, and temperature variation.

Kindly consider this manuscript for the review and possible publication in the Journal.

Yours Sincerely

Dr. Kumar Anupam

Assistant Professor
Faculty of Civil Engineering & Geosciences
Delft University of Technology
Delft, The Netherlands
[email protected]
Manuscript (For revision, remove the original version and upload Click here to view linked References
clean version of revised paper)

Finite element-based framework to study the response of bituminous

1
2
concrete pavements under different conditions
3
4
5
6
7
8 Abhinav Kumar1, Ankit Gupta1, Kumar Anupam2, Vivek Pratap Wagh1
9 1
10 Department of Civil Engineering, IIT BHU, Varanasi, India
11 2
12
Department of Civil Engineering & Geoscience, Delft University of Technology, Delft, The Netherlands
13
14
15
16
17 HIGHLIGHTS
18
19
20  Linear viscoelastic material characterization of BC-2 is presented.
21
22  Lab evaluation of creep compliance has been highlighted.
23  Effect of temperature and air void on viscoelastic properties of BC-2 is identified.
24  A direct approach to find Prony series constants has been discussed.
25  Pavement response based on linear viscoelastic properties of BC-2 has been predicted.
26  Effect of contact pressure and load on pavement response analysed statistically.
27
28
29 ABSTRACT
30
31 In most of developing countries across the world, pavement design is still based on an empirical approach that
32 may result in premature failure or overdesigned pavements. A shift from an empirical to a semi-mechanistic or
33 mechanistic approach is the need of modern time. In this regard, computational tools such as finite element (FE)
34 are being successfully utilized to gain deeper insights because such tools have allowed researchers to study the
35
36
complex behaviour of bituminous concrete (BC) materials. It is well recognized that BC material typically
37 exhibits viscoelastic/visco-elasto-plastic behaviour depending on applied loading (including temperature)
38 conditions. However, due to the complexity of the whole procedure yet many pavement design tools consider
39 them as pure elastic material. The aim of this research is to develop FEM based simple and practical framework
40 to evaluate the structural response of BC material with viscoelastic material characterization which can be an
41 effective tool to predict field behaviour with commonly available pavement material tests. Such a framework will
42 be helpful in analysing variations in the critical response of BC pavement with varied traffic loads and ambient
43
44 temperatures. The framework provides a relatively simple procedure to obtain the viscoelastic parameters of BC
45 mix with a creep compliance test conducted at different temperatures. It was concluded that Creep compliance
46 data if pre-smoothened by the Power law model reduces mathematical optimization issues to some extent.
47 Furthermore, with the obtained parameters, a 3-dimensional FE model was developed to obtain sensitivity to
48 critical stresses, strains, and deformations at desired conditions. Material characterization of unbound granular
49 layers was evaluated through resilient modulus based on empirical relations. Analysis was carried out taking into
50
consideration the traffic load, contact pressure, mix type, air-void, and temperature variation.
51
52
53
54
55
Keywords: effect of air voids, Indian traffic conditions, generalised Kelvin model, creep compliance, resilient
56
57 modulus.
58
59
60
61
62
63 1
64
65
 1. Introduction
1
2 Direct/indirect empirical approaches in the current pavement design procedures may result in premature
3 pavement failure or over-designed pavements [1]. However, these empirical approaches are still being practiced
4 in many developing countries around the world. In modern days, a lot of developing countries such as India are
5 identifying an urgent need to shift their design practices from empirical to semi-mechanistic/mechanistic
6
7
pavement design suited to their local requirements. In such cases, different material characterization tests will be
8 required. One of the important aspects that need to be incorporated is the time and temperature dependency of
9 the pavement materials. Whereas, for the characterization of time/temperature dependency of binding materials,
10 Dynamic shear rheometer (DSR) tests are often used, however, for mixtures either no such tests are carried out
11 or they are scarcely used. In the authors’ opinion, the characterization of temperature dependency of mixture is
12 important for countries like India where huge temperature variation across the region and seasons is observed.
13
14 In India, bituminous concrete (BC) pavements which are multilayer structures with different material properties
15 [2] are becoming popular [3]. Depending on the aggregate gradations, the Ministry of Road Transport &
16 Highways [4], India has categorised BC as BC-1 and BC-2. The nominal aggregate size of BC-1 is 19 mm
17 whereas that of BC-2 is 13.2 mm. BC-2 is composed of a higher percentage of fine aggregates (62% aggregates
18
19
passing 4.75 mm IS sieve) than BC-1 (50% aggregates passing 4.75 IS sieve). So, the minimum binder content
20 specified by [4] for BC-2 is higher (5.4%) than BC-1 (5.2%). Since BC-2 consists of a higher percentage of fine
21 aggregates, the prepared mix provides a smoother surface than BC-1, which helps in fixing deformation
22 measuring transducers and motivated to consider in the present study as the top layer. Depending on material
23 choices and their production processes, they exhibit different mechanical characteristics resulting in different
24 performances.
25
26 As discussed in the previous paragraph, understanding the mix behaviour is important for reaching the expected
27 lifetime. Therefore, uncertainties related to the mix design phase will result in a significant reduction in its
28 lifetime. To minimise such risks, road agencies in India mandate the testing of these mix samples under controlled
29 laboratory conditions, for example, the Marshall stability test, resilient modulus test, etc. However, tests
30
31
measuring the viscoelastic characterization of the mixture are not posed as a requirement. It is well known that
32 depending upon loading and ambient (temperature) conditions, asphaltic materials such as BC can exhibit
33 viscoelastic/visco-elasto-plastic behaviour [5]. Hence, many researchers in the past have studied the viscoelastic
34 behaviour of binder [6][7] and asphalt mixes [8][9][10][11][12][13] for the analysis of flexible pavement. It is
35 important to measure the properties of BC at a wide range of frequencies and temperatures to obtain viscoelastic
36 behaviour. At high temperatures or under slow-moving loads, it may exhibit a purely viscous flow reflecting its
37 propensity towards rutting. However, at relatively low temperatures and fast-moving loads, it becomes
38
39 progressively harder and eventually brittle, which makes it vulnerable to non-load-associated distresses like low-
40 temperature cracking. Moreover, it is susceptible to fatigue-related problems at normal temperatures as most of
41 the vehicular load is applied at these temperatures [14].
42
43
The characterization of viscoelastic properties is often done by measuring creep behaviour [15]. The creep-related
44 tests are popular because these tests make it possible to determine and separate the time-independent (elastic
45 strain) and time-dependent (viscoelastic and plastic strains) components of the strain response [16] in a simplified
46 way. In addition, parameters obtained from creep tests at low temperatures (m-value) are used to predict thermal
47 cracking development and propagation, and those at high temperatures are used to predict rutting in the HMA
48 [17]. Once creep compliance is obtained, viscoelastic behaviour is obtained through the time-temperature
49 superposition principle [15].
50
51 Even though, as highlighted in the previous section, viscoelastic characteristics are important to be included in
52 the design of pavements, most of the present mechanistic design standards [18][10][20] assume the BC layer as
53 elastic during multilayer pavement analysis. Some of the modern-day tools (ABAQUS, ANSYS, 3-D Move)
54
55 allow for the inclusion of viscoelastic characteristics for pavement design. However, in India, generally, elastic
56 modulus (E) and Poisson ratio (υ) are used to characterize materials as linear elastic and subsequently used as
57 input parameters in different design tools [21]. However, researchers [22] concluded that the elastic theory cannot
58 represent permanent deformation or delayed recovery, which is an important characteristic of BC materials. BC
59 shows elastic behaviour at low temperatures only and thus the use of elastic theory to explain material’s response
60 is limited to low-temperature analysis [23]. For intermediate and high temperatures, elastic theory often
61
underestimates these responses [23].
62
63 2
64
65
 As explained in the previous paragraph, since BC exhibits viscoelastic behaviour, controlling and conducting
experiments for all different types of combinations to represent field conditions is nearly impossible. In such
1
2 cases, analytical and computational tools are readily used. Increasing computational capabilities have allowed
3 pavement scientists to study complex material behaviour using tools like finite element (FE), discrete element
4 (DE), etc. [5]. To include viscoelastic characteristics in such tools, often a generalised Kelvin viscoelastic model
5 (GKM) or Prony series model [24][22] is used. Furthermore, Zhang et al. [25] showed that the generalized
6 Maxwell model is well suited for the prediction of asphalt concrete deformation under repeated load.
7
8 State-of-the-art equipment such as dynamic modulus test, bending beam rheometer test, and creep compliance
9 test is recommended to be used in developed nations. However, in developing nations, often limited resources
10 are available and therefore conducting different types of tests is not always possible. In such cases, a simple test
11 is required to obtain the desired behaviour. Creep compliance test facility is readily available in most of the
12
research institutions in India but to the best of the author’s knowledge, a framework to obtain GKM parameters
13
14 from this test is not available. Creep compliance test measures material deformation with time at different
15 temperatures which is defined as time-dependent strain per unit stress [26]. In addition to temperature, creep
16 compliance varies with the constituent’s properties and their proportion in the bituminous mix. Compacted mix
17 properties like air void also affect creep compliance results, hence, affecting GKM parameters. Since India is a
18 region with a tropical climate and temperature varies from -2o to 42o C throughout the year, it is important to
19 study mix behaviour at different temperatures. Also, the lack of quality control in mix compaction results in a
20
21
wide range of air voids. Under these conditions, a lot of uncertainties in the calculation of GKM are expected
22 depending upon region and season. At the end, these uncertainties will result in unreliable design and a significant
23 reduction in their expected lifetime. This article aims to study some of these effects for designing a pavement
24 with BC-2 materials as surface course including their viscoelastic characteristics.
25
26 In this study, material properties of various layers were characterized to evaluate the structural response of typical
27 pavements. Material characterization of unbound granular layers is based on resilient modulus and Poisson’s
28 ratio as recommended by IRC:37 in India. A continuum-based finite layer approach, 3-D Move [27], has been
29 used in this study as a reference tool to predict and validate the structural response of BC pavement. it can identify
30 the elastic and viscoelastic moduli of BC pavement [28]. A 3-dimensional FE model of BC pavement has been
31 created in ABAQUS (FE based analysis tool). The present study uses ABAQUS for predicting the sensitivity of
32
33
the structural response of bituminous concrete pavement at various air voids in BC-2 mix, temperature, tire load,
34 and contact pressure.
35 1.1. Framework of the paper
36
37
38 For easy understanding and quick review, the manuscript has been arranged into different sections and
39 subsections. The first section provides various literature available on the viscoelastic material characterization of
40 bituminous concrete, pavement modelling, structural response prediction, the objective of the study, and the
41 methodology adopted. The second section presents lab test results of the physical properties of materials used in
42 BC-2, mix design, and creep compliance test method. The third section discusses how creep compliance data has
43
been fitted to the generalised Kelvin model. Evaluation of Prony series constants and conversion of creep
44
45 compliance data to relaxation modulus has also been discussed. Material characterization of unbound granular
46 layers has been discussed as per the guidelines of IRC:37-2018 in the last part of this section. The fourth section
47 presents finite element modelling of bituminous concrete pavement, boundary conditions, and pavement
48 structure. The fifth section discusses the results of the creep compliance test, structural response of the pavement,
49 and statistical analysis using 2-way ANOVA for understanding the significance of contact pressure and tire load
50 on pavement response. The sixth and seventh sections provide the critical conclusion and future scope of the
51
study respectively.
52
53 1.2. Objective and scope
54
55
56 The objective of this paper is to predict the sensitivity of the structural response of bituminous concrete pavement
57 subjected to various load classes, contact pressure, air void in BC mix, and temperature for Indian roads. A
58 constitutive framework for evaluating Prony series parameters of the BC-2 mix for Indian conditions has been
59 discussed. Test temperature has been selected corresponding to average temperature variation in summer and
60 winter in India. Tire load has been selected which corresponds to light to heavy vehicular load passes on Indian
61 roads.
62
63 3
64
65
 For linear viscoelastic material characterization of bituminous concrete, a creep compliance test has been
employed for the determination of creep compliance as a function of deformation in the mix with time. Since the
1
2 present study aims to evaluate the influence of air voids on the creep behaviour and relaxation modulus, the
3 compaction efforts were varied at optimum binder content (OBC) to obtain different air voids in compacted
4 Marshall specimens. The present study employs a dynamic testing system for the experimental determination of
5 creep compliance. For modelling the viscoelastic properties of BC-2, a generalized Kelvin model has been used.
6
7 For characterizing materials used in unbound granular layers, the resilient modulus and Poisson’s ratio have been
8 used considering the elastic response of materials as suggested by IRC:37-2018 [29], code of practice for the
9 design of flexible pavement in India. Resilient modulus has been evaluated using empirical relations based on
10 CBR values of supporting layers.
11
12 A 3-dimensional bituminous concrete pavement has been modelled in ABAQUS (FE-based software) to identify
13 the response to material properties of the bituminous mix. The analysis was carried out at three different contact
14 pressures of 420, 560, and 740 kPa and three different single-wheel loads of 20, 25, and 30 kN. The effect of air
15 void, temperature, tire load, and contact pressure on asphalt concrete Pavement response in terms of vertical
16
stress and strain has been noted and discussed.
17
18
19 2. Materials
20
21 Locally available silty sand has been used as subgrade material. The free swelling index was found 12% which
22 is within the specified limit of a maximum of 50% by IS:2720-part 40 [30]. Maximum dry unit weight when
23
24 tested as per IS:2720-part 8 was found to be 17.91 kN/m3 which is higher than the minimum specified value of
25 17.5 kN/m3. The physical requirements of aggregates to be used in the granular subbase and base layer and binder
26 for the BC layer have been discussed in the next subsections.
27
28 2.1. Aggregate
29
30 In the present study, Calcareous aggregate (Dolomite) was used to prepare bituminous mixtures. These are
31 common types of aggregate used for the construction of flexible pavement in India. It is basic in nature and
32 provides good bonding with bitumen compared to acidic aggregates [31]. To check the quality of aggregates, the
33 following test was conducted as per MoRTH [4] specifications followed in India as shown in Table 1.
34
35 Table 1. Physical properties of aggregate.
36
37 Los Angeles abrasion Aggregate impact
38 Properties Combined flakiness and elongation index
value value
39
40 Specification limit Max 35% Max 40% Max 30%
41
42
Methodology IS 2386-Part I [32] IS 2386-Part IV [33] IS 2386-Part IV [33]
43
44
45 Value obtained 18% 20% 16.7%
46
47
48
49 2.2. Asphalt binder
50
51 A viscosity-graded binder, VG30 was used. This is the most common binder type used in road construction in
52 India. Physical tests were performed on the binder in accordance with IS 73:2013 and ASTM standards [34], as
53
shown in Table 2. In India, binder gradation is based on viscosity so test protocol and specifications for high-
54
55 temperature continuous grade and performance grading are not available. High-temperature continuous grade
56 and performance grade tests using a dynamic shear rheometer were conducted in accordance with ASTM D6373.
57 Table 2 presents some of the physical properties of the asphalt binder, the test protocol used for their evaluation,
58 and the limits in which these results should fall to be used as an asphalt mix binding material.
59
60
61
62
63 4
64
65
 Table 2. Physical properties of asphalt binder.

1
2
Properties Value obtained Methodology Specification limit
3
4
5 Penetration value at 25o C, 0.1mm 60 IS 1203 [35] Min 45
6
7
Absolute viscosity at 60o C, Poises 2900 IS 1206-Part 2 [36] 2400-3600
8
9
10 Softening point, ⁰ C 51.5 IS 1205 [37] Min 47
11
12 High-temperature continuous grade
83.65 ASTM D6373 -
13 (°C)
14 High-temperature
15 PG82-XX ASTM D6373 -
PG (°C)
16
17
18 2.3. Mix design
19
20 The bituminous concrete (BC-2) gradation in accordance with the Ministry of Road Transport and Highways,
21 India [4] with a nominal maximum size of aggregate of 13.2 mm was used for the preparation of the asphalt
22 mixture. The obtained gradation and specification limits are shown in Figure 1. In India, gyratory compaction
23
24 machine is available at only a few research institutes, and all the design and mix performance test are still based
25 on Marshall mix samples. In this study, the Marshall mix design was carried out in accordance with the Asphalt
26 Institute specification, MS-2 [38]. The OBC was evaluated for samples prepared using VG 30.
27
28 100
29 Lower Limit
90
30
31 Upper Limit
80
32 Blend Achieved
33 70
% Passing

34 60
35
36 50
37 40
38
39 30
40 20
41
42 10
43
0
44
0.01 0.1 1 10 100
45
46 Sieve Size, mm
47
48 Figure 1. Aggregate gradation adopted in the study.
49
50 Table 3 shows the volumetric properties of Marshall mix samples for BC-2 gradation with VG30 asphalt binder.
51 The compaction efforts required to achieve particular air voids were determined from the relationship between
52 air voids and compaction by interpolation method.
53
54 Table 3. Summary of volumetric properties of Marshall mix sample.
55
56 OBC Air void VMA VFB Stability Flow
57 Binder type Gmb Gmm
(%) (%) (%) (%) (kN) (mm)
58
59 VG 30 5.8 2.400 2.501 4.038 13.1 71 18.8 3.8
60
61 Specification limit Min 5.4 - - 3-5 Min 12 65 - 75 >9 2-4
62
63 5
64
65
 2.4. Creep compliance test
1
2 The test sample was prepared with all the volumetrics as discussed in section 2.3 (see Table 3). Creep compliance
3 test procedures apply to samples having a maximum aggregate size of 38 mm or less. As per AASHTO T-322,
4 the sample should be 38 to 50 mm high and 150 ± 9 mm in diameter. Both ends of the sample were cut by 6 mm
5 to provide smooth and parallel surfaces for better mounting of the deformation measurement transducer (LVDT).
6 Four brass gauge points were attached to each flat face of the sample. The gauge distance was kept at 50 mm.
7
Two 0.1 mm LVDTs were connected on each face, one in the axial direction of loading and the other in the lateral
8
9 direction. The specimen was allowed to remain at test temperature for 3 hours prior to testing for conditioning of
10 the sample. A constant load of 2 kN for 100 sec was applied on the diametrical axis of the sample. Creep load
11 was selected to keep the strain within the linear viscoelastic range i.e., produces a horizontal deformation of
12 0.00125 mm to 0.0190 mm. Based on horizontal and vertical deformations, normalised deformations and trimmed
13 mean were calculated as per AASHTO T-322 for evaluating the creep compliance of the mix. The test setup for
14 creep compliance is shown in Figure 2.
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37 (a) (b)
38
39 Figure 2. (a) Specimen loading frame (AASHTO T-322) (b) BC-2 sample and deformation measurement transducers.
40
41 3. Material modelling
42
43
44 3.1. Material characterization of bituminous concrete
45
46 The generalized Kelvin model (GKM) and the generalized Maxwell model (GMM) are the two widely used
47 models for characterizing linear viscoelastic properties of asphalt mix [39]. These models best fit the relaxation
48 modulus and creep compliance as a series of decaying exponentials also called as Dirichlet or Prony series [39].
49
50
Viscoelastic functions represented using the Prony series can be mathematically converted from the time domain
51 to the frequency domain [39]. The GMM for a viscoelastic solid consists of a spring and several Maxwell
52 elements assembled in parallel. Maxwell elements are a combination of spring and dashpot representing elastic
53 and viscous components of the mix. The GMM is convenient for analysing the relaxation behaviour of linear
54 viscoelastic materials. The relaxation modulus of GMM has the form of the Prony series as shown below:
55 𝑛
56 𝑡
57 𝐸(𝑡) = 𝐸Ꝏ + ∑ 𝐸𝑘 exp (− ) (1)
𝜌𝑘
58 𝑘=1
59
60 Where EꝎ is the equilibrium modulus, ρk, and Ek are the relaxation time and the stiffness of the kth Maxwell
61 element respectively. The complex modulus of GMM in the frequency domain is given by Eq. 2 [40].
62
63 6
64
65
 𝑛
∗ (𝜔)
𝑖𝜔𝜌𝑘 𝐸𝑘
𝐸 = 𝐸Ꝏ + ∑ (2)
1 1 + 𝑖𝜔𝜌𝑘
𝑘=1
2
3 where ω is the angular frequency. However, the GKM consisting of a spring and several Voigt elements
4 connected in series is convenient for analysing the creep behaviour of viscoelastic material [39]. The complex
5
6
compliance (D*) and creep compliance (D(t)) of GKM is given by [40] as follows:
7 𝑛
𝐷𝑘
8 𝐷 ∗ (𝜔) = 𝐷𝑜 + ∑ (3)
9 1 + 𝑖𝜔𝜏𝑘
𝑘=1
10
𝑛
11 𝑡
12 𝐷(𝑡) = 𝐷𝑜 + ∑ 𝐷𝑘 [1 − exp (− )] (4)
13 𝜏𝑘
𝑘=1
14
15 where Do is instantaneous creep compliance, τk, and Dk are the retardation time and creep compliance of the kth
16 Voigt element respectively. Model parameters τk and Dk are determined by a series of constraint optimization
17 processes using the least square approach [41].
18
19 In this study, the Prony series and power law series have been used to model linear viscoelastic material properties
20 of BC-2. Power law series is a simple power function of time in which evaluation of only instantaneous creep
21 compliance and time exponent is required [42]. A power law model if used with multiple power law terms can
22 present a smooth and reliable viscoelastic response of asphalt mix with minimal impact from local variance in
23
24 data. However, from the computational point of view, a Prony series representation is preferable to a power law
25 model because of its efficiency and better exponential fitting of material response [43]. In the present study, the
26 five-term Prony series and single-term Power series are used to model the viscoelastic properties of BC-2. It is
27 noted that the Power law series has been kept simple just for the pre-smoothing and comparison purposes (see
28 Eq. 5). Pavement response analysis has been carried out based on outcomes of the Prony series model only.
29
30 𝐷(𝑡) = 𝐷𝑜 + 𝐷1 (𝑡)𝑛 (5)
31
32 where Do is instantaneous creep compliance, D1 is the model parameter and n is the time exponent. In both models,
33 the square of the error between the experimental creep compliance value and model predicted value is minimized
34 to evaluate model constants [41].
35
36 The Prony series model was also used to convert creep compliance to relaxation modulus E(t). The
37 interconversion technique is based on an approximate solution developed by Schapery and Park [44]. The
38 relaxation modulus and creep compliance are related by a convolution integral, as follows:
39 𝑡
40 ∫ 𝐸(𝑡 − 𝜏)𝐷(𝜏)𝑑𝜏 = 𝑡 𝑓𝑜𝑟 𝑡 > 0 (6)
41 0
42
43 A common method for the numerical solution requires the decomposition of the integral into several intervals as
44 the spread of the function may extend to a few decades [43]. However, this may render inaccurate results and
45 cause computational difficulties if intervals are not carefully selected [43]. Therefore, another approach [44] in
46 which data within each small interval are locally fitted to the pure power law model has been used. The relaxation
47
modulus within one interval was predicted using the following equation as proposed by [44]:
48
49 sin(𝑛𝜋)
50 𝐸(𝑡) = (7)
51
𝑛𝜋𝐷(𝑡)
52 where n is the time exponent of local power law expression.
53
54 3.2. Material characterization of unbound granular layers
55
56
57 The material properties of unbound granular layers were obtained using empirical relations as provided by
58 IRC:37-2018 [29]. These relations are based on the CBR value of subgrade soil and are given by:
59 𝑀𝑅𝑆 = 10 × 𝐶𝐵𝑅 𝑓𝑜𝑟 𝐶𝐵𝑅 ≤ 5% (8)
60
61 𝑀𝑅𝑆 = 17.6 × (𝐶𝐵𝑅)0.64 𝑓𝑜𝑟 𝐶𝐵𝑅 > 5% (9)
62
63 7
64
65
 where MRS is the resilient modulus of subgrade soil (MPa), and CBR is California bearing ratio of subgrade soil
(%). The granular base and granular subbase layer are considered as a single layer for the analysis; therefore, a
1
2 single modulus value is assigned to the combined layer. the resilient modulus of the granular layer from its
3 combined thickness and modulus of the supporting layer is estimated as:
4
𝑀𝑅(𝑔𝑟𝑎𝑛𝑢𝑙𝑎𝑟) = 0.2 (ℎ)0.45 × 𝑀𝑅(𝑠𝑢𝑝𝑝𝑜𝑟𝑡) (10)
5
6
where MR(granular) is the resilient modulus of the granular layer (MPa), h is the combined thickness of base and
7
8 subbase layer, and MR(support) is the resilient modulus of subgrade as supporting layer. This is one of the major
9 simplifications made in the Indian standard design code that takes into account equal material modulus of both
10 the base and subbase layer. These materials shall be tested separately and pavement analysis and design shall be
11 based on their evaluated resilient modulus. Material characterization of the unbound granular layer was solely
12 based on CBR values of subgrade soil as resilient modulus tests using repeated tri-axial equipment are usually
13 expensive. CBR tests on unsoaked local granular soil were performed. Table 4 presents the resilient modulus and
14
Poisson’s ratio values used for unbound granular layers in pavement analysis.
15
16
17 Table 4. Material properties of unbound granular layers.
18 S No. Pavement layers Thickness (mm) Resilient modulus (MPa) Poisson’s ratio
19
20
21 1 Subgrade 500 87.25 0.35
22
23 2 Granular sub base 350 321.81 0.35
24 3 Granular base 300 321.81 0.35
25
26
27 Since resilient modulus takes into account only the elastic deformation of the specimen and often underestimates
28 material stiffness at higher temperatures. Therefore, the process of material stiffness modulus evaluation as
29
30 suggested by flexible pavement design codes in India [29] is on the conservative side. Also, in the absence of
31 triaxial testing equipment, the code suggests empirical equations that further question the reliability of evaluated
32 material properties. Extensive studies beyond the elastic response of these materials at different test temperatures
33 need to be performed for a better analysis of the stiffness modulus.
34
35 4. FE modelling of asphalt concrete pavement
36
37
38 FE modelling of asphalt concrete pavement is not a new technique for pavement response prediction. Many
39 researchers [1], [45]–[55] have developed similar models in the past and reported pavement response subjected
40 to various loads, contact pressures, and material characteristics. Most of these models have been developed in
41 European countries suited to load class, contact pressure, and material properties of their origin. Limited studies
42 [1], [50] are available in developing countries like India, Bangladesh, and Sri Lanka where mixed traffic offers a
43
44 wide range of load classes and contact pressure. Also, material properties used in various layers changes
45 significantly with change in source and need characterization and modelling for pavement response prediction.
46 A three-dimensional FE model of asphalt concrete pavement has been developed in this study using ABAQUS.
47
48
The conventional geometric design consists of four different layers namely, bituminous concrete (including dense
49 bituminous macadam), base layer, granular subbase layer, and compacted soil subgrade layer. The present study
50 takes into consideration the elastic properties of granular unbound aggregate layers and subgrade layers while it
51 considers linear viscoelastic properties of bituminous concrete layers based on creep compliance test data. The
52 accuracy of the model depends on several parameters including material properties, number of elements, degree
53 of freedom, contact stress distribution, and boundary conditions [51]. With the increase in several elements and
54 degree of freedom, computational complexities increase. However, with the advent of modern high-speed
55
56 computers with increased computational efficiency, dealing with a high number of structural elements is no more
57 a significant constraint. In India, FE modelling of asphalt concrete pavement is relatively a new topic of research
58 and gaining interest with the success of consecutive modelling efforts. For the prediction of the structural
59 response of bituminous concrete pavement using FE analysis, one needs to develop a pavement model, tire model,
60 and material model. In developing nations like India, researchers have gained success in pavement modelling,
61
62
63 8
64
65
 however, material characterization and 3-dimensional modelling of test tire is relatively unexplored and has a lot
to be done.
1
2 4.1. Pavement structure
3
4
5
In this study, stress has been given to LVE material characterization of asphalt mix for sensitivity analysis of the
6 pavement response. In India, pavement design is based on linear elastic theory as selected in the guidelines for
7 the design of flexible pavement (IRC: 37-2018).
8
9
The pavement structure consists of a 150 mm BC layer (including dense bituminous macadam), 300 mm
10 aggregate base layer, 350 mm granular subbase layer, and 500 mm compacted subgrade layer.
11 The top surface of 150 mm BC was used to model the contact area for tire loading. The 20 kN tire load was
12
assumed to be uniformly distributed over the contact area at the tire pavement interface. The contact area was
13
14 estimated using a standard contact pressure of 560 kPa [29]. The contact area can be represented using a rectangle
15 and two semi-circles at the ends as shown in Figure 3 (a). Further, this shape of the loading area is converted to
16 an equivalent rectangle as suggested by Huang [56] having an area of 0.5227 L2 and a width of 0.6 L. Dimensions
17 of the contact area can be evaluated using tire load and contact pressure. The contact area was loaded using 560
18 kPa of contact pressure uniformly distributed over the entire area for validation of FE model.
19
20 4.2. Loading and boundary conditions
21
22 Sufficient depth below the subgrade layer (7000 mm) has been provided to subside the pavement structural
23 distresses to zero. Longitudinal and lateral movement of pavement elements has been restricted as it is infinite in
24
25 length compared to small sections considered for analysis and laterally supported by shoulder or earthen soil.
26 Only vertical movement of integration points and elements is allowed as shown in Figure 3 (b). All nodes at the
27 bottom plane have been restricted from the translational or rotational degrees of freedom.
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44 (a) (b)
45 Figure 3. (a) Contact area dimension (mm) for tire loading (b) Meshing in various layers and boundary conditions.
46
47 The bituminous concrete pavement model was meshed after the evaluation of contact area dimensions and
48
49 applying boundary conditions. The hexahedral elements type was used to mesh all the layers in the pavement. A
50 finer mesh close to the loading area and a relatively coarser mesh size were used for the region away from the
51 loading area.
52
53 5. Results and discussion
54
55
56 5.1. Creep compliance
57
58 A creep compliance test was conducted for the determination of the time-dependent response of BC-2 at three
59 different air voids of 4, 5, and 6%. Samples have been tested at three different test temperatures of 5o, 15o, and
60 25o C against each air void. Horizontal and vertical deformations of the sample with temperature have been noted
61
(see Figure 4). It can be noted that as loading time increases, both horizontal and vertical deformation also
62
63 9
64
65
 increases. The gradient of the deformation-time graph is steeper at the beginning and gets flatter at the end. These
deformations were also found to increase with air void and temperature. With the increase in temperature, the
1
2 stiffness of the material decreases thus deformation is found to increase. Further, with the increase in air void in
3 the mix, the density of the mix decreases which allows constituent particles to settle more easily into these voids,
4 and deformations in the axial as well as lateral direction increases. Percentage change in deformation at each time
5 step has been evaluated and shown for the temperature of 25o C compared to 15o C in Figure 4.
6
7
8 0.12 Air void = 4% 0.14 Air void = 4% 184%
9 262% 196%
0.1 0.12
10 25⁰ C 15⁰ C 5⁰ C
25⁰ C 15⁰ C 5⁰ C
Horizontal Def (mm)

11 0.1

Vertical Def (mm)

236% 163%
0.08
12
13 0.08 148%
0.06 223%
14 0.06 127%
15 0.04 188% 112%
16 172% 0.04 107%
147%
17 0.02 144% 0.02
18
19 0 0
20 1 2 5 10 20 50 100 1 2 5 10 20 50 100
21
22 Time (s) Time (s)
23
(a) (b)
24
25 0.12 Air void = 5% 0.16 Air void = 5% 175%
26 217%
0.14 176%
0.1 25⁰ C 15⁰ C 5⁰ C
Vertical Def (mm)

27 25⁰ C 15⁰ C 5⁰ C
Horizontal Def (mm)

28 0.12
199% 146%
29 0.08
0.1
30 140%
0.06 189% 0.08
31 120%
32 163% 0.06 103% 110%
0.04 162%
33
34 123% 145% 0.04
0.02
35 0.02
36
0 0
37 1 2 5 10 20 50 100 1 2 5 10 20 50 100
38
39 Time (s) Time (s)
40 (c) (d)
41
42
43 Air void = 6% 0.18 Air void = 6%
0.14
44 236% 157%
45 0.16
0.12 152%
Horizontal Def (mm)

25⁰ C 15⁰ C 5⁰ C
Vertical Def (mm)

46 25⁰ C 15⁰ C 5⁰ C 218% 0.14

47 0.1 123%
0.12
48
0.08 218% 0.1 108%
49
50 187% 0.08 93%
0.06 98%
51 175% 0.06 92%
0.04
52 120% 132% 0.04
53
0.02 0.02
54
55 0 0
56 1 2 5 10 20 50 100 1 2 5 10 20 50 100
57 Time (s)
58 Time (s)
59 (e) (f)
60
61 Figure 4. Variation of horizontal and vertical deformations with time at different temperatures and air voids.
62
63 10
64
65
 The average annual temperature in India hovers around 25o C, so analysis is more focused on this temperature.
Also, to maintain clarity in the figures, the percentage change in deformation at 15o C compared to 5o C has not
1
2 been shown here. It can be seen from Figure 4 that, the percentage change in deformation is continuously
3 increasing with time. It shows, higher material stiffness at lower temperatures, and consecutively, the deformation
4 plot flattens more quickly than at higher temperatures. A similar trend was not observed for 15 o C temperature
5 compared to 5o C. Percentage changes in deformation were relatively stable and constant. This may be due to at
6 lower temperatures, change in material stiffness is relatively constant, and elastic deformation is more dominant
7 than plastic deformation. These deformation data were used to find creep compliance of the material as per
8
9
AASHTO T:322-07 [57]. Variation of creep compliance as a function of time at an air void of 4,5, and 6% and a
10 temperature of 5o, 15o, and 25o C is shown in Figure 5.
11
12
13 0.006 Air Void - 4% 0.006 Air Void - 5%
14

Creep Compliance (MPa-1)

Creep compliance (MPa-1)

15
0.005 0.005
16
17
0.004 0.004
18
19
20 0.003 0.003
21
22 0.002 0.002
23
24 0.001 0.001
25
26 0 0
27 0 20 40 60 80 100 0 20 40 60 80 100
28
Time (s) Time (s)
29
30
31 (a) (b)
32
33 Legends used
34 0.006 Air Void - 6%
Creep Compliance (MPa-1)

35
36 ◊ - Predicted creep compliance at 5o C
0.005
37
38 ӿ - Experimental creep compliance at 5o C
0.004
39
40 Δ - Predicted creep compliance at 15o C
41 0.003
42 □ - Experimental creep compliance at 15o C
43 0.002
44
45 0.001
○ - Predicted creep compliance at 25o C
46
47 0 × - Experimental creep compliance at 25o C
48 0 20 40 60 80 100
49
50 Time (s)
51
52 (c)
53
54 Figure 5. Variation of creep compliance (experimental and model predicted) with time at (a) 4% air void (b) 5% air void
55 (c) 6% air void.
56
57 It can be seen from Figures 4 and 5 that trends of creep compliance and deformation plots are simillar as they
58
59 hold a proportionality relationship among them. Creep compliance data predicted from the Prony series model
60 shows good agreement with lab-calculated data at lower temperatures and variation increases at higher
61 temperatures. It can be noted that, a model which predicts compliance data with great accuracy at lower
62
63 11
64
65
 temperatures may not fit well at higher temperatures. In this study, the Prony series model was fitted to consider
the elastic stiffness of the material and hence it is not that accurate to consider plastic flow behaviour at a higher
1
2 temperature.
3 Creep test results were also used to study the material stiffness using relaxation modulus. This relaxation modulus
4
was used to further study the shear and bulk modulus of BC-2 material. Power law and the Prony series model
5
6 were used to fit the relaxation modulus variation with time. Then, the time-dependent variation of shear modulus
7 G(t) and bulk modulus K(t) were estimated assuming constant Poisson’s ratio given by the following relation:
8 𝐸(𝑡)
9 𝐺(𝑡) = (11)
10 2(1 + 𝜐)
11
𝐸(𝑡)
12 𝐾(𝑡) = (12)
13 3(1 − 2𝜐)
14
15 Creep compliance fitted by Power law and Prony series is presented in Figures 6 (a) and (b) respectively.
16 Relaxation modulus, shear modulus, and bulk modulus as predicted from equations (7), (11), and (12) were
17 plotted against reduced time and shown in Figures 6 (c) and (d).
18
19 Air Void 0.1 Air Void
0.1
20 4%
Creep Compliance (MPa-1)

4%
Creep Compliance (MPa-1)
21 5%
5%
22 0.01 6%
0.01 6%
23
24
25 0.001 0.001
26
27
28 0.0001 0.0001
29
30
31 0.00001 0.00001
32 0.1 10 1000 0.1 10 1000
33 TR (Sec) TR (Sec)
34
35 (a) (b)
36
37 Air Void Air Void
38 1000 1000
4% G (4%)
Relaxation Modulus (MPa)

39 5%
40 G (5%)
6% G (6%)
41 1
G(t) & K(t)

100 100 1 K (4%)

42
43 K (5%)
44 K (6%)
45 3
10 2 10 3
46 2
47
48
49 1 1
50 0.001 0.1 10 1000 0.001 0.1 10 1000
51
TR (Sec) TR (Sec)
52
53
(c) (d)
54
55
56 Figure 6. Variation of creep compliance with time fitting (a) Power law (b) Prony series (c) Variation of relaxation
57 modulus with time fitting Prony series (d) Variation of shear modulus and bulk modulus with time fitting Prony series.
58
59 The Power law model as shown in Figure 6 (a), is not a good fit of creep compliance data, as for a long period,
60 there is almost no change in creep compliance and thereafter starts increasing with a steep gradient. It can be used
61 to pre-smoothen the creep compliance data before fitting it to a Prony series model. Trends of creep compliance
62
63 12
64
65
 data predicted using the Prony series model were found simillar as evaluated in the lab. So, the Prony series
model was found to be a good fit for creep compliance data and can predict it well as a function of time as shown
1
2 in Figure 6 (b). Further, relaxation, shear, and bulk modulus were predicted. Relaxation modulus was found to
3 follow the inverse trend of relaxation modulus as shown in Figure 6 (c). Since shear and bulk modulus hold a
4 proportional relation with relaxation modulus, their plot also follows a simillar trend as that of relaxation modulus
5 as shown in Figure 6 (d). Plots of relaxation modulus, shear modulus, and bulk modulus can be classified into
6 three zones. Zone 1 and 3 are identical in nature and reflects very small variation in modulus values with reduced
7 time. Plots are relatively flatter in these zones. In zone 1, the material offers the highest resistance to the applied
8
9
load and modulus values mobilise to maximum as constituent particles are intact in the mix. As time passes,
10 particles start slipping from their original position and the modulus keeps on reducing in zone 2. At the end of
11 zone 2, when the resilience of the material is negligible against applied load, a relatively flat plot is obtained in
12 zone 3.
13
14 5.2. Asphalt concrete pavement response
15
16 After the evaluation of Prony series coefficients and resilient modulus of unbound granular layers, asphalt
17 concrete pavement was modelled for structural response analysis. Pavement response as obtained from FE-based
18 tool ABAQUS was first validated using 3D Move Analysis, a continuum-based finite layer approach software
19
developed by the University of Nevada, Reno. FE model was validated considering the elastic properties of the
20
21 asphalt mix. A validated model was used to predict pavement response in terms of stress, strain, and deformation.
22 Different load class of 20, 25, and 30 kN has been considered for pavement response sensitivity analysis. This
23 load class corresponds to a single wheel load. Considering dual wheel assembly, the load shared by a single wheel
24
25 is 20 kN, and the equivalent standard axle load considered for pavement design in India is 80 kN [29]. Considering
26 the heavy truck tire loads which are frequent in India, wheel load of 25 kN and 30 kN has also been considered
27 for analysis. State highways and national highways cater mixed traffic in India and allow movement of all
28 categories of vehicles on the same lane where multiple lanes are not provided. Mixed traffic offers a wide range
29 of wheel bases for different vehicles and thus contact pressure also changes. For pavement design in India, the
30 standard contact pressure of 560 kPa is used. Contact stress distribution was analysed for 20 kN tire load. It was
31
observed that the maximum contact pressure developed in the mid-section of the tire-pavement contact patch was
32
33 740 kPa and had an average value of 420 kPa. Therefore, three different contact pressures of 420, 560, and 740
34 kPa were selected for pavement analysis.
35
The model uses the linear viscoelastic properties of asphalt concrete mix and resilient modulus properties of
36
37 unbound granular layers. Pavement response was analysed to evaluate the effect of LVE properties of bituminous
38 concrete and sensitivity with changes in air void, vertical load, contact pressure, and temperature. The plot of
39 deformation and stress against pavement depth at 4, 5, and 6% air void has been presented in Figure 7. Vertical
40 stress has been plotted at log scale as the stress gradient is steep and reduces significantly in a small vertical depth
41 domain.
42
43 0.25 1
44 Air Void - 4% Air Void - 4%
45 0.2
Vertical stress (Mpa)

Air Void - 5% Air Void - 5%

Deformation (mm)

46 0.1
Air Void - 6% Air Void - 6%
47 0.15
48
0.01
49
50 0.1
51
0.001
52 0.05
53
54 0 0.0001
55 0 2000 4000 6000 0 2000 4000 6000
56
57 Depth (mm) Depth (mm)
58 (a) (b)
59
60 Figure 7. (a) Pavement deformation vs. vertical depth (b) Vertical stress vs. pavement depth.
61
62
63 13
64
65
 The rate at which vertical stress subsides initially against pavement depth is steeper than deformation. The stress
reduction of 65.74, 58.68, and 55.53% was observed at air voids of 4, 5, and 6% respectively in the BC-2 layer.
1
2 This is because the top layer of BC-2 has the highest stiffness among all the layers and is much higher than its
3 subsequent base layer. So, most of the stress is taken by the bituminous concrete layer. It was also interesting to
4 note that at a higher air void in the mix, the percent reduction in stress in the BC-2 layer was lesser than at a lower
5 air void. This is due to the higher stiffness of the mix at lower air void giving a more compact structure and thus
6 stress dissipates more quickly within the layer. Whereas, the opposite trend prevails in the case of deformation.
7 As the material stiffness of the upper layers is higher than lower layers, so, the rate of subsidence of deformation
8
9
is much lesser in BC-2 and base layers. Only 33.83, 28.10, and 23.27% of stress were reduced corresponding to
10 4, 5, and 6% of air voids in these layers, and most of the surface deformation was found to subside in the natural
11 subgrade and compacted subgrade layers.
12
The sensitivity of pavement response was further analyzed at different vertical loads and contact pressure. Load
13
14 sensitivity analysis was performed corresponding to a vertical load of 20, 25, and 30 kN keeping other parameters
15 of the FE model constant. Pavement response in terms of normal stress and vertical microstrain for the BC-2, and
16 granular subbase (GSB) layer is shown in Figure 9 for a contact pressure of 740 kPa. Ideally, it is expected to
17 have maximum stress at the surface, and subsequently, it should reduce as it moves down the pavement. However,
18 this is not the actual case in the top layer of bituminous pavement. It was found that stress in the BC-2 layer first
19 increases to a certain depth (approx 85 mm) and then starts decreasing. A simillar pattern is observed in the case
20
21
of vertical strain with a depth of BC-2 layer. Initially, the strain was found negative (up to 70 mm) representing
22 compression, and then with increasing depth, it is positive representing tension. This can be justified by the fact
23 that when a thin wearing course of the bituminous layer is provided over granular layers, there is compressive
24 bending strain due to a wheel load at the bottom of the bituminous layer which decreases with the increase in
25 thickness. It becomes tensile at a higher thickness. When thickness reaches about 50 mm, the reduction in tensile
26 strain starts with a further increase in thickness as shown in Figure 8 [29].
27
28
29
30
Positive

31
32
Tensile strain

33
34
35
36 50 mm
37
Negative

38
39 Thickness of bituminous layer
40
41
42
43 Figure 8. Tensile strain in a thin bituminous wearing course (regenerated from IRC:37-2012).
44
45 Based on the above justification, stress and strain (micro) plots against pavement depth were noted as shown in
46 Figure 9. It is important to note that, compressive stress at lower load (20 kN) is higher than compressive stress
47 at higher loads (25 and 30 kN) till it becomes tensile (85 mm depth). After that, tensile stress starts decreasing
48 with the depth of the pavement and the reverse is true, i.e.,. tensile stress at higher load (30 kN) is higher than
49 tensile stress at lower loads (20 and 25 kN). Response in other lower layers like granular base, subbase, and
50
subgrade layers was found unidirectional. Normal stress and vertical microstrain decreases throughout the depth
51
52 of the layer. It can be concluded from Figure 9 that, variations in stress and strain in lower granular layers at
53 different load classes are more distinct than in upper bituminous concrete layer as the stiffness of granular layers
54 is lesser. Although stresses coming to lower layers also reduces with depth, however, these are more susceptible
55 to visible changes with different load class.
56
57 FE model of asphalt concrete pavement used for pressure sensitivity analysis at three different contact pressures
58 of 740, 560, and 420 kPa at a constant vertical load of 20 kN as shown in Figure 10. Here, variation has been
59 shown for asphalt concrete and granular subbase layer only. For pressure sensitivity analysis, it is interesting to
60 note that, unlike load sensitive response, variation of normal stress and vertical strain in asphalt concrete layer is
61 more distinct and shows more vertical gap between consecutive curves. However, the reverse is the case with
62
63 14
64
65
 lower granular layers. Response at different contact pressures in granular layers seems much closer and
overlapping. It can be concluded that the BC-2 layer is more sensitive to contact pressure than vertical load.
1
2 Variations of stress and strain in BC-2 and lower granular layers are the same as in the case of load sensitivity
3 analysis.
4
5 Asphalt concrete layer Asphalt concrete layer
6
7 0.1
8 0.09 50 20 kN 25 kN 30 kN
9

Vertical micro-strain
0.08
10 0.07 30
11
Stress (MPa)

0.06
12 10
0.05
13
0.04
14 -10 0 30 60 90 120 150
0.03
15
0.02 20 kN 25 kN 30 kN
16 -30
17 0.01
18 0
-50
19 0 30 60 90 120 150
20
21 Depth (mm) Depth (mm)
22
23 (a) (b)
24
Sub-base layer Sub-base layer
25
26 120
0.03
27 20 kN 25 kN 30 kN 110 20 kN 25 kN 30 kN
Vertical micro-strain

28 0.025 100
Stress (MPa)

29
30 90
0.02
31 80
32 0.015 70
33
60
34 0.01
35 50
36 0.005 40
37 450 500 550 600 650 700 750 800 450 500 550 600 650 700 750 800
38 Depth (mm) Depth (mm)
39 (c) (d)
40
41 Figure 9. Pavement response as normal stress and vertical compressive strain for asphalt concrete and granular subbase
42 layer at a vertical load of 20, 25, and 30 kN.
43
44
45
46 Asphalt concrete layer Asphalt concrete layer
47
0.09 60
48 740 kPa
49 0.08 50
560 kPa
Vertical micro-strain

50 0.07 40
Stress (MPa)

51 0.06 30 420 kPa

52 0.05 20
53 0.04 740 kPa 10
54 0.03 0
55 560 kPa
0.02 -10 0 50 100 150
56 0.01 420 kPa -20
57 0 -30
58 0 50 100 150 -40
59 Depth (mm)
Depth (mm)
60
61 (a) (b)
62
63 15
64
65
 Sub-base layer Sub-base layer
1 0.019
2 740 kPa 560 kPa 420 kPa 85
0.017 740 kPa 560 kPa 420 kPa
3

Vertical micro-strain
75
Stress (MPa)

4 0.015
5 0.013 65
6
7 0.011
55
8 0.009
9 45
0.007
10
0.005 35
11
450 500 550 600 650 700 750 800 450 500 550 600 650 700 750 800
12
13 Depth (mm) Depth (mm)
14 (c) (d)
15
16 Figure 10. Pavement response as normal stress and vertical compressive strain for asphalt concrete and granular subbase
17 layer at a contact pressure of 740, 560, and 420 kPa.
18
19 In previous sections, the effect of tire load, contact pressure, and air void on the structural response of the pavement layers
20 has been discussed. A comparative analysis of these responses among various layers has been presented in Table 5 and Table
21 6.
22
23 Table 5. Effect of load and contact pressure on the structural response of various layers of BC pavement.
24
25 Contact pressure (kPa) Pavement layer Stress (kPa) % Micro-strain % change
26 decrease
27 BC 40.22 1.12
28 420 Base 32.23 19.86 77.19 6791.96
29 GSB 9.98 69.03 52.63 -31.81
30 Subgrade 4.39 56.01 58.22 10.62
31 BC 54.54 3.99
32 560 Base 36.68 32.74 88.26 2112.03
33 GSB 10.44 71.53 54.78 -37.93
34 Subgrade 4.47 57.18 58.99 7.68
35 BC 73.29 8.14
36 740 Base 40.88 44.22 98.77 1113.39
37 GSB 10.80 73.58 56.37 -42.92
38 Subgrade 4.53 58.05 59.39 5.35
39 Load (kN) Pavement layer Stress (kPa) % Micro-strain % change
40 decrease
41
BC 69.35 8.13
42
20 Base 40.88 41.05 98.77 1114.88
43
GSB 10.80 73.58 56.36 -42.93
44
Subgrade 4.53 58.05 59.39 5.37
45
BC 68.06 5.87
46
25 Base 46.91 31.07 112.97 1824.53
47
GSB 13.15 71.96 68.92 -38.99
48
Subgrade 5.60 57.41 73.87 7.18
49
50 BC 67.10 3.86
51 30 Base 52.09 22.37 125.09 3140.67
52 GSB 15.37 70.49 80.84 -35.37
53 Subgrade 6.65 56.73 88.05 8.91
54
55 It was found that vertical stress and micro-strain in the BC layer are relatively smaller than expected. This is
56 because initially stress in the BC layer first increases and then decreases as explained earlier. Micro-strain is
57 initially compressive (negative) and with the increase in depth, it becomes tensile (positive) so most of the strains
58
cancel out and effective magnitude diminishes to a smaller total. It was also noted that the effect of varying load
59
60 and contact pressure is least significant for lower layers of GSB and subgrade layers as most of the stress is being
61 taken by upper layers having higher stiffness. The effect of air void in BC mix and test temperature was also
62
63 16
64
65
 studied and presented in Table 6. Creep compliance test for BC mix was conducted at 5, 15, and 25o C only
however looking into the temperature variation over the year in India, creep compliance data with time was also
1
2 evaluated for a temperature of 40o C using Williams-Landel-Ferry (WLF) equation [58]. The WLF equation
3 allows for the estimation of material properties beyond test data.
4
5 Table 6. Effect of air void and temperature on the structural response of various layers of BC pavement.
6
7 Air void (%) Pavement layer Stress (kPa) % decrease Vertical % decrease
8 displacement (mm)
9 BC 54.54 0.175
10 4 Base 36.68 32.74 0.149 14.85
11 GSB 10.44 71.53 0.121 18.79
12 Subgrade 4.47 57.18 0.089 26.86
13 BC 60.86 0.193
14 5 Base 41.28 32.17 0.169 12.43
15 GSB 13.22 67.97 0.142 15.97
16 Subgrade 4.63 64.97 0.104 26.76
17 BC 67.12 0.212
18 6 Base 47.30 29.52 0.186 12.26
19 GSB 15.82 66.55 0.163 12.36
20 Subgrade 4.98 68.52 0.112 31.28
21 Temperature Pavement layer Stress (kPa) % decrease Vertical % decrease
22 (oC) displacement (mm)
23 BC 53.94 0.176
24 5 Base 36.12 33.03 0.151 14.20
25 GSB 10.19 71.78 0.120 20.53
26 Subgrade 4.38 57.01 0.085 29.16
27 BC 54.54 0.198
28 15 Base 36.68 32.74 0.152 23.23
29 GSB 10.44 71.53 0.123 19.07
30 Subgrade 4.47 57.18 0.088 28.45
31 BC 54.73 0.232
32 25 Base 36.88 32.61 0.154 33.62
33 GSB 10.57 71.33 0.126 18.18
34 Subgrade 4.52 57.23 0.092 26.98
35 BC 55.05 0.268
36 40 Base 37.23 32.37 0.156 41.79
37 GSB 10.79 71.01 0.128 17.94
38 Subgrade 4.54 57.92 0.094 26.56
39
40
41 It can be seen from Table 6 that the effect of air void on the structural response of the BC mix is significant.
42
Stresses in the BC layer are 11.58% higher when compacted with 5% air void as compared to 4%. This increase
43
44 in stress rises to 23.06% when compacted with 6% air void. Vertical displacement in the BC layer observed was
45 also significant. A rise of 21.14% was found in the displacement of the BC layer when the air void in the mix
46 increased from 4% to 6%. The effect of temperature on the vertical displacement of pavement layers especially
47 the BC layer was found significant. An increase in displacement of nearly 52.27% was found in the BC layer
48
when the temperature rises from 5o C to 40o C.
49
50
51
5.3. Statistical analysis
52
53 Two-way Analysis of Variance (ANOVA) was carried out to study the effect of load and contact pressure on the
54 stresses and strains at different pavement depths. Load and contact pressure were considered as the independent
55 variables whereas stresses and strains at different depths were taken as dependent variables. Table 7 presents the
56 results of the statistical analysis at a 95% confidence level (α = 0.05). It is noted that the effect of independent
57 variables is significant when the p-value is less than 0.05. Both independent variables i.e., loads and contact
58
59
pressure have a significant impact on dependent variables (see Table 7). This shows the importance of these
60 parameters in governing the design of asphalt concrete pavement.
61
62
63 17
64
65
 Table 7. Summary of Two-way ANOVA at 95% confidence level.

1 Variables Source of variation F calculated P-value F critical

2
Contact pressure 12543.81 3E-08 6.94
3 Stress at 100 mm
4 Load 0.71 0.542 6.94
5
Contact pressure 158.88 0.00015 6.94
6 Stress at 150 mm
7 Load 19.44 0.01 6.94
8
9 Contact pressure 51.49 0.0014 6.94
Stress at 300 mm
10 Load 57.68 0.0011 6.94
11
12 Contact pressure 26.21 0.0050 6.94
13
14 Stress at 800 mm Load 1611.58 0.000002 6.94
15 Contact pressure 9130.2 5E-08 6.94
16 Strain at 100 mm
17 Load 278.3 5E-05 6.94
18 Contact pressure 294.03 0.0004 6.94
19 Strain at 150 mm
20 Load 220.1 0.039 6.94
21 Contact pressure
22 52.6 0.0013 6.94
Strain at 300 mm
23 Load 51.1 0.0014 6.94
24 Contact pressure
25 14.21 0.015229 6.94
Strain at 800 mm
26 Load 973.17 0.000004 6.94
27
28
29 It can be seen that contact pressure has a significant effect on stress and strain at each depth whereas load has a
30 significant effect on these parameters except at 100 mm depth (as p-value > 0.05). This is mainly due to the
31 presence of a viscoelastic asphalt layer up to 150 mm depth in which tensile strain starts building nearly 50 mm.
32 Additionally, this might be due to the load levels within the linear viscoelastic regime for the asphalt layer. Table
33 5 shows that contact pressure has a more dominant effect than load as the p-value in most of the cases for contact
34 pressure is much less when compared with the load.
35
36
37
To better understand the significance of these variables on stresses and strains, main effect plots were plotted at
38 a 95% confidence level. The results for other depths are not shown here for brevity. Figure 11 shows the change
39 in dependent variables with the variation in the attributes of independent variables.
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55 (a) (b)
56
57
58
59
60
61
62
63 18
64
65
 1
2
3
4
5
6
7
8
9
10
11
12
13 (c) (d)
14
15 Figure 11. Main effect plots for stress and strain at 100 mm and 300 mm depth.
16
17
18 6. Conclusions
19
20
The correct characterization of mechanical properties of BC mixes is important for predicting the accurate
21
22 lifetime expectancy of the flexible pavements particularly, under Indian design specifications. Current standard
23 specifications are primarily based on characterizing the BC material as elastic. In order to obtain these mechanical
24 behaviours, resilient modulus tests are recommended. Several studies reported that characterizing BC material
25 with only elastic properties induces uncertainty in the prediction of an expected lifetime. Although modern-day
26 computational tools allow input of complex material characterization such as viscoelastic/visco-elasto-plastic etc.
27 but they generally require specific input parameters. To the best of the authors’ knowledge, there are very few
28
29
studies that tried to develop a framework to obtain viscoelastic characterization of BC mixes under Indian
30 conditions. The present study proposed a framework to obtain and evaluate Prony series parameters for the BC
31 mixes on the basis of creep compliance tests performed for different compositions of mixes (air voids: 4, 5, and
32 6%) and temperatures (5, 15, and 25o C).
33
34 As expected, it was observed that the initial gradient of deformation as well as creep compliance were much
35 steeper and flattened quickly after 30-40 seconds. Also, with the increase in temperature, these gradients were
36 further apart and became more distinct, however at lower temperatures, these curves were found to be relatively
37 flatter. This could be the result of the material becoming softer at higher temperatures and therefore deforms
38 continuously with time.
39
40 In this study, it is proposed that at first creep compliance data should be pre-smoothen by using the Power law
41 model which to an extent minimizes mathematical nonlinear optimization issues. After pre-smoothing nonlinear
42 optimization tools were used to obtain viscoelastic characteristics of material by means of the Prony series. It is
43 also noted that the Power law model was found to be a good tool for pre-smoothing however fails to provide a
44
45
good representation of the viscoelastic characterization of BC materials. Power law reflects almost no variation
46 in creep compliance for a longer duration of time, hence, it is not recommended. On the other hand, the Prony
47 series model was found to be a good representation of LVE characterization.
48
As a next step 3-dimensional FE models were developed to obtain sensitivity of critical stresses and strains in
49
50 different layers with the variation of load types (20, 25, and 30 kN), contact pressures (420, 560, 740 kPa), air
51 voids (4, 5, and 6%) and temperatures (5, 15, 25, and 40o C). Key conclusions are highlighted below:
52
- From load and pressure sensitivity analysis, it was found that the effect of contact pressure variation on the
53
54 structural response of pavement layers is more prominent than load. An increase in vertical stress of 82.22% was
55 found when contact pressure changed from 420 kPa to 740 kPa. This highlights the importance of considering
56 vehicle tire type and stress distribution at tire-pavement contact during analysis.
57
58 -Change in temperature was found to have a significant influence on vertical deformation in the BC layer. For
59 example, vertical displacement was found to increase by 52.27% in the BC layer when temperature was changed
60 from 5o C to 40o C. This highlights the importance of consideration of region (based on temperature) in the design
61 process.
62
63 19
64
65
 -Change in air voids in the BC mix also plays a vital role in the pavement design stage itself as it affects the
critical structural response of the BC layer considerably. It was found that an increase in air void from 4% to 6%
1
2 increases vertical displacement by 21.14% and vertical compressive stress by 23.06%. It shows the importance
3 of BC mix compaction in the field. Good construction practices such as adequate compaction play a significant
4 role in achieving the desired lifetime.
5
6 -It was also noted that the variation in structural response was found to be in agreement with past literatures i.e.,
7 the least for lower layers of GSB and subgrade.
8 In the last step, Two-way ANOVA was carried out to see the relative effects of contact pressure and vertical
9
10 loading on the structural response of the pavement. It was observed that pavement response is more sensitive to
11 contact pressure than vertical loading. Further, the effect of load at the mid-section of the bituminous concrete
12 section was not found significant however, the same is not true for contact pressure.
13
14
In general, this research is an attempt to present a FEM-based simple and practical framework to evaluate the
15 structural response of BC material with viscoelastic material characterization which can be an effective tool to
16 predict field behaviour with commonly available pavement material tests. Such a framework could be a very
17 useful tool for practitioners and researchers.
18
19 7. Future scope of the study
20
21
22 To make the study more inclusive and comprehensive, other ranges of temperature and air void shall be included.
23 It will make analysis more effective in the sense of understanding material properties with more clarity.
24 Additionally, the tire footprint considered for loading uniformly distributes contact stresses which is not the actual
25 condition in the field. In a further study, we will simulate actual tire contact and contact stress distribution. Also,
26 the resilient modulus approach used for unbound granular layers underestimates layer stiffness at higher
27
28
temperatures. Future studies should explore the material nonlinearity effect on pavement response.
29
30
31 References
32
33 [1] M. S. Ranadive and A. B. Tapase, “Parameter sensitive analysis of flexible pavement,” International Journal of Pavement Research and
Technology, vol. 9, no. 6, pp. 466–472, Nov. 2016, doi: 10.1016/j.ijprt.2016.12.001.
34
35 [2] Á. Holanda, E. P. Junior, and T. Araújo, “Finite element modeling of flexible pavements,” no. September, 2006, [Online]. Available:
36 http://repositorio.ufc.br:8080/ri/handle/123456789/1380
37
38 [3] A. Ghosh, A. Padmarekha, and J. M. Krishnan, “Implementation and Proof-checking of Mechanistic-empirical Pavement Design for Indian
39 Highways Using AASHTOWARE Pavement ME Design Software,” Procedia Soc Behav Sci, vol. 104, pp. 119–128, Dec. 2013, doi:
40 10.1016/j.sbspro.2013.11.104.
41 [4] “MoRTH, Specifications for road and bridge work (Fifth revision),” New Delhi, 2013.
42
43 [5] M. J. Williamson, “Finite element analysis of hot-mix asphalt layer interface bonding,” 2015.
44
[6] Y.-R. Kim and D. N. Little, “Linear Viscoelastic Analysis of Asphalt Mastics,” Journal of Materials in Civil Engineering, vol. 16, no. 2, pp. 122–
45
132, Apr. 2004, doi: 10.1061/(ASCE)0899-1561(2004)16:2(122).
46
47 [7] C. Y. Cheung and D. Cebon, “Experimental study of pure bitumens in tension, compression, and shear,” J Rheol (N Y N Y), vol. 41, no. 1, pp.
48 45–74, Jan. 1997, doi: 10.1122/1.550858.
49
50 [8] Y. T. Chou and H. Larew, “Stresses and displacements in viscoelastic pavement systems under a moving load,” Transp Res Rec, pp. 25–40,
1969.
51
52 [9] J. F. Elliott and F. Moavenzadeh, “Analysis of stresses and displacements in three-layer viscoelastic systems,” Highway Research Record, 1971.
53
54 [10] Y. H. Huang, “Stresses and strains in viscoelastic multilayer systems subjected to moving loads,” Highway Research Record, no. 457, pp. 60–
55 71, 1973.
56
[11] S. W. Park and Y. R. Kim, “Analysis of Layered Viscoelastic System with Transient Temperatures,” J Eng Mech, vol. 124, no. 2, pp. 223–231,
57 1998, doi: 10.1061/(asce)0733-9399(1998)124:2(223).
58
59 [12] R. A. Schapery, “A method of viscoelastic stress analysis using elastic solutions,” J Franklin Inst, vol. 279, no. 4, pp. 268–289, Apr. 1965, doi:
60 10.1016/0016-0032(65)90339-X.
61
62
63 20
64
65
 [13] P. Mackiewicz and A. Szydło, “Viscoelastic parameters of asphalt mixtures identified in static and dynamic tests,” Materials, vol. 12, no. 13,
2019, doi: 10.3390/ma12132084.
1
2 [14] M. A. Elseifi, I. L. Al-Qadi, and P. J. Yoo, “Viscoelastic Modeling and Field Validation of Flexible Pavements,” J Eng Mech, vol. 132, no. 2, pp.
172–178, Feb. 2006, doi: 10.1061/(ASCE)0733-9399(2006)132:2(172).
3
4 [15] L. Zhang, X. Zhang, X. Liu, and Y. Luo, “Viscoelastic Model of Asphalt Mixtures under Repeated Load,” Journal of Materials in Civil Engineering,
5 vol. 27, no. 10, p. 04015007, 2015, doi: 10.1061/(asce)mt.1943-5533.0001256.
6
7 [16] M. W. Witczak, T. K. Pellinen, and M. M. El-Basyouny, “Pursuit of the simple performance test for asphalt mixture rutting,” in Asphalt Paving
8 Technology 2002, Colorado: Association of Asphalt Paving Technologists (AAPT), 2002, pp. 767–778.
9 [17] M. Jeong, “Comparison of Creep Compliance Master Curve Models for Hot Mix Asphalt,” 2005.
10
11 [18] Shell International Petroleum Company, “Shell pavement design manual: Asphalt pavements and overlays for road traffic,” 1978.
12
[19] “Thickness design-Asphalt pavements for highways and streets, 9th Ed,” Lexington, KY, 1999.
13
14 [20] H. L. Theyse, M. De Beer, and F. C. Rust, “Overview of South African Mechanistic Pavement Design Method,” Transportation Research Record:
15 Journal of the Transportation Research Board, vol. 1539, no. 1, pp. 6–17, Jan. 1996, doi: 10.1177/0361198196153900102.
16
17 [21] A. Gupta and A. Kumar, “Comparative Structural Analysis of Flexible Pavements Using Finite Element Method,” International Journal on
18 Pavement Engineering & Asphalt Technology, vol. 15, no. 1, pp. 11–19, May 2015, doi: 10.2478/ijpeat-2013-0005.
19 [22] M. A. Elseifi, I. L. Al-Qadi, and P. J. Yoo, “Viscoelastic Modeling and Field Validation of Flexible Pavements,” J Eng Mech, vol. 132, no. 2, pp.
20 172–178, Feb. 2006, doi: 10.1061/(ASCE)0733-9399(2006)132:2(172).
21
22 [23] I. L. Al-Qadi, M. Elseifi, and P. J. Yoo, “In-situ validation of mechanistic pavement finite element modeling,” in International Conference on
23 Accelerated Pavement Testing, Minneapolis, Minnesota, USA, 2004.
24
[24] S. W. Park and R. A. Schapery, “Methods of interconversion between linear viscoelastic material functions. Part I—a numerical method based
25
on Prony series,” Int J Solids Struct, vol. 36, no. 11, pp. 1653–1675, Apr. 1999, doi: 10.1016/S0020-7683(98)00055-9.
26
27 [25] L. Zhang, X. Zhang, and C. Hu, “Deformation Prediction of Asphalt Mixtures under Repeated Load Base on Viscoelastic Mechanical Model,” in
28 Paving Materials and Pavement Analysis, Reston, VA: American Society of Civil Engineers, May 2010, pp. 116–125. doi:
29 10.1061/41104(377)15.
30
[26] D. N. Richardson and S. M. Lusher, “Determination of Creep Compliance and Tensile Strength of Hot-Mix Asphalt for Wearing Courses in
31
Missouri,” Apr. 2008.
32
33 [27] M. Nasimifar, S. Thyagarajan, and N. Sivaneswaran, “Backcalculation of Flexible Pavement Layer Moduli from Traffic Speed Deflectometer
34 Data,” Transportation Research Record: Journal of the Transportation Research Board, vol. 2641, no. 1, pp. 66–74, Jan. 2017, doi:
35 10.3141/2641-09.
36
37 [28] Z. Sun, C. Kasbergen, K. N. van Dalen, K. Anupam, A. Skarpas, and S. M. J. G. Erkens, “A parameter identification technique for traffic speed
deflectometer tests of pavements,” Road Materials and Pavement Design, vol. 24, no. 4, pp. 1065–1087, 2023, doi:
38
10.1080/14680629.2022.2060125.
39
40 [29] “IRC:37-2018, Guidelines for the design of flexible pavements (Fourth Revision),” New Delhi, Nov. 2018.
41
42 [30] IS:2720 - Part 40, Determination of free swell index of soils. India, 1997.
43 [31] Preethi Shailaja, Ranjitha Tangadagi, Manjunath Maddikeari, and Bharath Ashwath, “Sustainable Effect of Chemically Treated Aggregates on
44 Bond Strength of Bitumen,” Journal of Green Engineering , vol. 10, no. 9, pp. 5076–5089, Sep. 2020.
45
46 [32] B. of Indian Standards, “IS 2386-1 (1963): Methods of Test for Aggregates for Concrete, Part I: Particle Size and Shape.”
47
[33] B. of Indian Standards, “IS 2386-4 (1963): Methods of test for aggregates for concrete, Part 4: Mechanical properties.”
48
49 [34] “ASTM: D6373, Standard Specification for Performance-Graded Asphalt Binder,” West Conshohocken.
50
51 [35] IS: 1203-2022, Methods for Testing Tar and Bituminous Materials — Determination of Penetration. India, 2022, pp. 1–10.
52
[36] IS:1206 (Part 2), Methods for Testing Tar and Bituminous Materials — Determination of Viscosity. India, 2022, pp. 1–12.
53
54 [37] IS:1205-2022, Methods for Testing Tar and Bituminous Materials — Determination of Softening Point — Ring and Ball Apparatus. India, 2022,
55 pp. 1–10.
56
57 [38] Asphalt Institute, MS-2: Asphalt mix design methods, 7th ed. Asphalt Institute, 2014.
58 [39] Y. Zhao, Y. Ni, and W. Zeng, “A consistent approach for characterising asphalt concrete based on generalised Maxwell or Kelvin model,” Road
59 Materials and Pavement Design, vol. 15, no. 3, pp. 674–690, 2014, doi: 10.1080/14680629.2014.889030.
60
61 [40] N. W. Tschoegl, The phenomenological theory of linear viscoelastic behavior: an introduction. Springer, Science & Business Media, 2012.
62
63 21
64
65
 [41] S. K. Srirangam, Numerical simulation of tyre pavement interaction, vol. 53, no. 9. 2013.

1 [42] Samer W. Katicha, “Analysis of Hot Mix Asphalt (HMA) Linear Viscoelastic and Bimodular Properties Using Uniaxial Compression and Indirect
2 Tension (IDT) Tests,” Virginia Polytechnic Institute and State University, Blacksburg, Virginia, 2007.
3 [43] Y. R. Kim, J. S. Daniel, and H. Wen, “Fatigue performance evaluation of WesTrack asphalt mixtures using viscoelastic continuum damage
4 approach,” Apr. 2002.
5
6 [44] A. [ Schapery and \ S[ W[ Park, “Methods of interconversion between linear viscoelastic material functions[ Part II*an approximate analytical
7 method.”
8 [45] B. Sukumaran et al., “THREE DIMENSIONAL FINITE ELEMENT MODELING OF FLEXIBLE PAVEMENTS,” 2004.
9
10 [46] B. Sukumaran, M. Willis, and N. Chamala, “Three Dimensional Finite Element Modeling of Flexible Pavements,” 2005.
11
12 [47] I. L. Howard, M. Asce, and K. A. Warren, “Finite-Element Modeling of Instrumented Flexible Pavements under Stationary Transient Loading”,
doi: 10.1061/ASCE0733-947X2009135:253.
13
14 [48] Y.-H. Cho, B. F. Mccullough, and J. Weissmann, “Considerations on Finite-Element Method Application in Pavement Structural Analysis.”
15
16 [49] B. Sam Helwany, J. Dyer, and J. Leidy, “FINITE-ELEMENT ANALYSES OF FLEXIBLE PAVEMENTS.”
17
[50] Rahman MT, Mahmud K, and Ahsan S, “StressStrain characteristics of flexible pavement using Finite Element Analysis,” 2011.
18
19 [51] M. N. S. Hadi and B. C. Bodhinayake, “Non-linear finite element analysis of flexible pavements,” in Advances in Engineering Software, Elsevier
20 Ltd, 2003, pp. 657–662. doi: 10.1016/S0965-9978(03)00109-1.
21
22 [52] J. M. Duncan, C. L. Monismith, and E. L. Wilson, “Finite Element Analyses of Pavements.”
23 [53] F. Tang, T. Ma, Y. Guan, and Z. Zhang, “Parametric modeling and structure verification of asphalt pavement based on BIM-ABAQUS,” Autom
24 Constr, vol. 111, Mar. 2020, doi: 10.1016/j.autcon.2019.103066.
25
26 [54] P. Liu, Q. Xing, Y. Dong, D. Wang, M. Oeser, and S. Yuan, “Application of Finite Layer Method in Pavement Structural Analysis,” Applied
27 Sciences, vol. 7, no. 6, p. 611, Jun. 2017, doi: 10.3390/app7060611.
28
[55] E. Parente Junior, J. B. Soares, Á. Silva de Holanda, E. Parente Junior Teresa Denyse Pereira de Araújo, and L. Tadeu Barroso de Melo Francisco
29
Evangelista Junior Jorge Barbosa Soares, “FINITE ELEMENT MODELING OF FLEXIBLE PAVEMENTS,” 2006. [Online]. Available:
30 https://www.researchgate.net/publication/281974814
31
32 [56] Y. H. Huang, Pavement analysis and design, Second. Upper Saddle River, NJ: Pearson Prentice Hall, 1993.
33
34 [57] AASHTO : T 322-03, “Determining the Creep Compliance and Strength of Hot-Mix Asphalt (HMA) Using the Indirect Tensile Test Device,”
American Association of State Highway and Transportation Officials, Washington D.C., USA, pp. 1–11, 2005.
35
36 [58] K. Anupam, T. Tang, C. Kasbergen, A. Scarpas, and S. Erkens, “3-D Thermomechanical Tire–Pavement Interaction Model for Evaluation of
37 Pavement Skid Resistance,” Transp Res Rec, vol. 2675, no. 3, pp. 65–80, Jan. 2020, doi: 10.1177/0361198120963101.
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63 22
64
65
Declaration of Interest Statement Click here to view linked References

Declaration of interests

☐The authors declare that they have no known competing financial interests or personal relationships
that could have appeared to influence the work reported in this paper.

☒The authors declare the following financial interests/personal relationships which may be considered
as potential competing interests:

Abhinav Kumar reports financial support was provided by MHRD, India. Abhinav Kumar reports a
relationship with Indian Institute of Technology BHU Varanasi that includes: non-financial support.