INTRODUCTION
Pavement Design is a complex process, since it involves many variable factors, such as dynamic loading, nonlinear material properties and environmental conditions. Usually the design factors for flexible pavement design are divided into four broad categories (Huang, 2003) that are traffic and loading, environment, material and failure criteria.
During the development of the 1986 AASHTO guide, it was recognized that future
design procedure would be based on MechanisticEmpirical principles (National
Cooperative Highway Research Program, 2004). Over the past 20 years there
has been a tendency for road agencies to direct their efforts toward MechanisticEmpirical
(ME) method (Prozzi and Madanet, 2002). Probably this
was the reason thus AASHTO replaced its 1993 empirical design method with the
more reliable MechanisticEmpirical (ME) design method in 2004.
MechanisticEmpirical (ME) design combines the elements of mechanical modeling and performance observation in determining the required pavement thickness for a set of design condition.
The mechanical model is based on elementary physics and determines pavement
reactions to wheel loads in terms of stress, strain and displacement. The empirical
part of design uses the pavement responses to predict the life of the pavement
on the basis of actual field performance. Specific advantages of ME design
over traditional empirical procedures are that as follow (Timm
et al., 1998):
• 
Consideration of changing load types 
• 
Better utilization and characterization of available materials 
• 
Improved definition of the role of construction by identifying the parameters
that are most influential over pavement performance 
• 
Relation of material properties to actual pavement performance 
• 
Better definition of the existing pavement layer properties and 
• 
Accommodation of environmental and aging effects of materials 
At the present time, AASHTO method (American Association
of State Highway And Transportation Officials, 1993) is the most creditable
method, which is used in Iran for design of pavement structure. The major drawback
of this method of flexible pavement design is its dependency on the specific
conditions for which they are derived.
Any change in material, loadings, environmental conditions and assumptions
would reduce the accuracy and increase the error. The objective of this study
is to present ME design process in Iran. This process has been defined for
pavement design since 2004 (Khavandi, 2004).
DEVELOPMENT OF MECHANISTICEMPIRICAL PROCEDURE
The pavement design in this study, is based on valid and creditable scientificmathematical
theories in which factors such as exiting facilities particularly laboratory
and other necessary data effective in pavement design such as CBR, temperature,
etc have been defined. Figure 1 shows the process proposed
for the proposed ME design.

Fig. 1: 
Axle proposed ME design flowchart 
One advantage of the process, common to most ME
design processes is iterative loops.
Material properties: Since the phenomenological behavior of any real
material is extremely complex, therefore certain idealization of material behavior
are inevitable. It is well known that pavement materials exhibit nonlinear
viscoelasticplastic behavior and are generally depended on loading time, stress
level, temperature and moisture content. It is presently not possible to account
for every factor influencing the responses of inservice pavements. Therefore
simplifications are necessary.
The most significant material properties in the proposed ME method, is elastic modulus and poison ratio. Until recently in Iran, traditional CBR test was the only laboratory test performed on base, subbase and subgrade materials. In order to utilize these data in ME design, the correlation charts for treated and untreated granular base and subbase materials presented in the AASHTO (1993) guide used for granular materials.
In the proposed method, the air temperature data are used to account for the
effect of temperature on moduli of asphalt mixture. The relationship between
mean pavement temperature (Tp) and mean monthly air temperature (Ta) is based
on the depth below pavement surface (z) as presented by Eq. 1
(Asphalt Institute, 1997). The elastic modulus of asphalt
layer is calculated using Eq. 2 (Timm
et al., 1998).
Where:
T_{p} 
= 
The average Asphalt layer temperature (°C) 
T_{a} 
= 
The average air temperature (°C) 
Z 
= 
The average Asphalt layer thickness (cm) 
Where:
E_{AC} 
= 
Modulus of hot mix Asphalt concrete (kg m^{2}) 
T_{p} 
= 
The average Asphalt layer temperature 
Annual average elastic modulus of Asphalt layer is determined based on the
weighted average modulus of Asphalt layer for each month.
The information of Poisson ratio is also one of the other properties of materials that is required in the proposed ME method. However, because Poisson ratio has a relatively small effect on pavement responses it is customary to assume a reasonable value for use in design, rather than to determine it from actual tests. In the present study, the Poisson ratio for base and subbase layers is considered 0.4 and for subgrade, 0.5.
Conducted researches and investigations have demonstrated that Poisson ratio
of the Asphalt is affected by the pavement temperature. The values of Poisson
ratio for the Asphalt based on the changes in temperature are shown in Table
1 (Timm et al., 1998).
Traffic: Proper consideration of traffic loading in pavement design requires knowledge of full axle load distribution by the main axle types, including single, tandem and tridem axles. Although the equivalent single axle load concept has been used since the 1960s foe empirical pavement design, the new mechanisticbased pavement design procedure most likely require the use of the axle distribution.
Conducted research has shown that majority of heavy vehicle traveling on paved roads in Iran have axles loads and wheels configurations shown in Fig. 2.
The distance of axles and the distance of wheels from each other are also among
the effective factors in the pavement performance which leave some impact on
the degree and intensity of road destruction.

Fig. 2: 
Axle loads and wheel configurations of vehicle traveling on paved roads
in Iran 
Table 2: 
Distance of axels and wheels 

Table 2 shows the distance of axles and the distance between
the double wheels for different trucks in Iran. In the analyses and graphs presented
in this article, the distance between axles and also the distance between the
double wheels are considered 140 and 35 cm, respectively (Fig.
2). It is worth mentioning that in the case of Tridem 24 t axle with single
wheels, the distance between axles is 136 cm which should also be considered
in proposed pavement design (Fig. 2). Tire pressure also considering
different types of tires in Iran, the size of the wheels and load on each axle
is variable.
MECHANISTIC COMPUTER MODEL
As described earlier and shown in Fig. 1, material properties
and load configurations are entered into a mechanisticbased loaddeformation
computer model.
The mechanisticbased loaddeformation model is the heart of ME design process and determines pavement response to applied loads. There are many models available, including linear layeredelastic, nonlinear layeredelastic, elastoplastic, viscoelastic and viscoplastic.
In proposed ME method, using the computer software KENLAYER whose validity and credibility is proved by Huang (2003) all pavement are analyzed.
The KENLAYER computer program can be used to analyze a multilayer elastic pavement structure by cumulative damage techniques for a single, dual and multiplewheel system.
Transfer functions: The empirical component of ME design is pavement
life equation, known as a transfer function. Transfer function use pavement
responses calculated by the mechanistic model and predict the life of pavement
in terms of fatigue cracking or rutting. In fact, Transfer functions act as
a chain between the pavement reactions and appeared damages in the pavements.
Transfer functions presented in this study are (Asphalt Institute,
1999; Khavandi, 2008):
N_{f} = 6x10^{10} (ε_{t})^{4.59} 
(3) 
N_{d} = 1.365x10 ^{9}
(ε_{c})^{4.477} 
(4) 
Where:
ε_{t} 
= 
The maximum tensile horizontal strain at bottom of Asphalt layer 
E_{t} 
= 
Asphalt layer elastic module 
ε_{c} 
= 
The maximum compressive vertical strain at the top of the subgrade 
N_{f} 
= 
Number of allowed loads until fatigue cracking occurs 
N_{d} 
= 
Number of allowed loads until rutting occurs 
After the numbers of applied and allowable loads have been determined, miner
hypothesis is used to quantify accumulating damage in terms of rutting or fatigue,
over the life of the pavement.
When damage exceeds unity, the pavement has been under designed and thickness will increase. If the damage is much less than unity, the pavement has been over designed and thicknesses are decreased. An optimum design is achieved when the damage is near but not exceeding 1.
Miner’s hypothesis:
Table 3: 
Variables and their values 

_{*}Tire pressure for 24 ton tridem Ax 
Where:
D 
= 
Total fatigue or rutting damage 
n_{i} 
= 
Applied number of i axle 
N_{i} 
= 
Allowable number of i axle 
PROPOSED EQUATIONS FOR ME DESIGN
The computational techniques of KENLAYER computer program are complex and cumbersome
to be used for routine design. To incorporate KENLAYER structural model into
a ME method, simplified analytical equations that reliably predict KENLAYER
response solutions for typical flexible pavements are needed.
The first step in developing theses equations is to create a data base. A 3*3*2*2*3*(2, 1)*(5, 1) full factorial data base totaling 1836 cases was created. Table 3 shows the specific values of pavement variables. Multiple regression was then applied to the response data to develop best fit equations. The results of the regression resulted in the following equations:
Proposed equations for 6 t singleaxle with single wheel:
Proposed equations for 8.2,10 And 13 t singleaxle with double wheels:
Proposed equations for 21 t tandem axle with double wheels:
Proposed equations for 24 t tridem axle with single wheel:
Where:
T_{asphalt} 
= 
Aphalt layer thickness (cm) 
E_{asphalt} 
= 
Aphalt elastic module (psi) 
T_{base} 
= 
Bse layer thickness (cm) 
E_{base} 
= 
Base elastic module (psi) 
T_{subbase} 
= 
Subbase layer thickness (cm) 
E_{subbase} 
= 
Subbase elastic module (psi) 
E_{subgrade} 
= 
Subgrade elastic module (psi) 
T_{pressure} 
= 
Tire pressure (psi) 
W_{load} 
= 
Axle weight, tons (this is defined for single axles with double wheels) 
ε_{t} 
= 
The maximum tensile horizontal strain at bottom of Asphalt layer 
ε_{c} 
= 
The maximum compressive vertical strain at the top of the subgrade 
N_{f} 
= 
Allowable number of loads for fatigue control 
N_{d} 
= 
Allowable number of loads for rutting control 



Fig. 3: 
Design graph for 8. 2, 10 and 13 t single axle with double wheels, (Design
criterion: rutting) 
DESIGN DIAGRAMS BY ME METHOD
Some diagrams suggested for proposed equations. A typical diagrams of
Eq. 7, is presented in Fig. 3. For using of this diagrams,
At first the minimum functional thickness should be selected by determining
the pavement layer technical characteristics, in accordance with management
and planning organization, second, with existing Asphalt elastic module and
selected thickness of Asphalt layer and base layer , from horizontal axis of
asphalt elastic module (Fig. 3a), a line is drawn to intersect
the minimum asphalt thickness line. Then, a horizontal line in the same direction
with diagonal base layer thickness lines is drawn to intersect the minimum selected
thickness of this layer from this point a vertical line in the direction of
horizontal axis of parameter A is drawn and at last the quantity of this parameter
will be determined. All these steps are applied for subgrade elastic module
of horizontal axis (Fig. 3b), a vertical line is drawn to
intersect the minimum subbase layer thickness curve . A horizontal line is drawn
in the direction of base layer elastic module curve to intersect the subbase
elastic module curve. From this point a vertical line in the direction of parameter
B axis is drawn and the value of this parameter will be obtained.
The maximum compressive vertical strain at the top of the subgrade is determined by using the parameters A and B values are shown in Fig. 3c.
CONCLUSION
This study presented the results of a study to develop a mechanisticempirical
process for pavement design in Iran. The process offers a flexible, comprehensive
and simple framework for pavement design. The process starts with load and climatic
data and material properties for each layer in the pavement structure. The input
parameters of the ME method proposed in this study, are based on the existing
data and facilities particularly environmental data and laboratory equipments
in Iran. the structural analysis usually involve a linear elastic, static analysis
of the multilayer system, resulting in the pavement response to the loading
condition expressed in terms of strains at critical positions in the pavement
structure. The pavement response serves as input of the transfer functions.
In proposed ME process, fatigue and rutting transfer functions are considered.
This study also presented relationships and diagrams for types of axles loads
based on effective variable on pavement design to facilitate design process.
One of the most important advantages of the present method in compare to other
pavement designs is that, there is no need to use the equivalent load factors
for converting the different axels to equivalent single axel load.