Page 83 - SAPEM-Chapter-10-2nd-edition-2014

South African Pavement Engineering Manual
Chapter 10: Pavement Design

(iii) Pavement Response Models
The pavement response model uses layered elastic analysis to determine the displacements, strains and stresses
induced in the pavement by the loading. This is done through the pavement response model, which attempts to
model the resilient response of the individual pavement layers and the whole pavement system. The response model
therefore includes the material models and the system model:
• The material model describes the stress-strain behaviour of the material in each of the pavement layers when
viewed in isolation.
• The system model combines the material models of the individual pavement layers, the interaction between
pavement layers, the external loading and the boundary conditions of the problem to model the response of the
complete system.

The continuum mechanics model used in classical ME models for pavement materials is the homogenous, isotropic,
linear-elastic model. See Chapter 2: 3 and 4.2 for a description of this model and further discussion. This solution is
available in a number of software packages, the oldest being BISAR, ELSYM5, CHEV15 and WESLEA, some of which
are part of Cyrano, ME-Pads and Rubicon Toolbox, which are more commonly used now. See Sections 7.10, 8.4
and 9.6.

The pavement response model provides stress and strain results at any location within the pavement system using
the multi-layered linear-elastic system. The damage in pavement layers is determined by the stress or strain induced
at specific locations in the pavement structure. The location and stress or strain parameter is determined by the
material type and associated expected distress mechanism. These stresses and strains at specific locations in the
pavement are referred to as critical parameters and serve as the primary, load related, input to the damage model.

See the green side-box below for a discussion on the complexities of the response models.

(iv) Damage Models
The permanent response of the pavement to loading is captured in the damage models, also referred to as ‘transfer
functions’ or failure criteria. The transfer functions are material specific and are calibrated for the dominant
mechanical modes of distress, permanent deformation and fatigue in flexible pavements and shattered slabs,
pumping and faulting for rigid pavements. The calibration of the transfer functions is done from observed damage
data and is thus the empirical component of the process. The damage models often include some measure of
material strength. Although the strength properties are associated with the damage models, they are also part of
the material input parameters. Often these are not known explicitly and typical properties associated with the
material type and class are used, termed “default” properties. Examples of the types of strength and materials
properties included are the shear strength, density and saturation for granular materials and strain-at-break for
lightly cemented materials.

The structural capacity of a pavement also depends on the conditions in which the pavement is operating. Wet
conditions, for example, reduce the structural capacity of a pavement. Classical ME design methods assume that a
consistent set of conditions apply for the duration of the structural life of the pavement.

Most mechanistic-empirical (ME) methods were only calibrated for predefined terminal conditions. These methods
estimate the structural capacity (N) from the initial condition to a predefined terminal level of distress. No
information is provided on how damage is accumulated and how the terminal condition is reached.
(v) Recursive Analysis Methods
Recent developments in ME design methods involve modelling the incremental damage that occurs within periods of
similar conditions in terms of traffic loading, environmental conditions and pavement characteristics. Methods based
on this approach are referred to as recursive ME-design methods. The incremental damage that occurs within a
period of similar conditions may be modelled using a linear incremental damage model, based on Miner’s Law, or a
non-linear incremental damage model. For linear recursive methods, the damage models from classical ME design
methods may be used, although they were not calibrated for this use. Non-linear recursive methods require more
advanced mathematical formulations of the damage models to capture the non-linear accumulation of damage.
Research is underway to incorporate recursive methods into the South African Mechanistic-Empirical Method. See
Section 4.1.3.3 for a brief discussion on recursive methods.

Section 6: Structural Capacity Estimation
Page 72