PERS® utilizes damage predictions by incremental-recursive methods combined with the mechanistic derived pavement responses in terms of stresses and strains calculated for each layer interface.

The method is based on dividing the lifetime into minor time increments, typically a year or a season.
At the beginning of each time increment the present condition is used as input for calculating the damage over the time increment considering the applied traffic loading.
A new condition can then be calculated as input for the next time increment.

In the widely used empirical relationships for predicting damage only the terminal condition is predicted as a function of the total number of load repetitions over the period.
Parameters in the relationship would normally be the present condition in terms of strain/stress level and elastic modulus as expressed in the equation:
N = a * strainb * Ec,
where N is number of load repetitions causing a defined damage, E the elastic modulus and a, b, c are constants calibrated for the specific material under the local conditions regarding climatic and traffic characteristics.
Miner’s law is then assumed, allowing to sum damages linearly over the design period.

In a pavement management system where timing of the activities is important in order to minimize the total lifetime costs for the society, it would not be realistic to operate with linear damage models.
If damage e.g. were defined as fatigue cracking, you would not expect a linear development of the amount of cracked surface with time. It is more reasonable to assume that the relative decrease of the elastic moduli is a nonlinear function of the past number of load repetitions as illustrated in figure 2.

Figure 2 (left): Nonlinear damage modelling. (Ullidtz, 2005). MNp is the accumulated number of load repetitions, Ei is the initial elastic modulus and dE is the decrease in modulus. The damage function can then be expressed as in equation 1 Equation 1: Damage model used for incremental-recursive performance prediction.

In PERS® damage with respect to relative decrease in elastic modulus, to IRI (roughness) and rutting can be entered according to this model for each material type defined by the user.
A, Rref, Eref, k1, k2, k3 are constants defined by the user.
RT is response type, which can be either stress or strain and Rref is a reference response, and Eref the reference modulus.

During the analysis PERS® will calculate the predicted damages layer by layer in the pavement structure, and combine them to get the total predicted performance of the structure.
It is also possible to set up models for predicting surface wear and friction.
These models are entered as a function of tire contact pressure and traffic. Effects of ageing and water penetration can be defined as well.

All of the mechanistic modelling is based on the input of elastic modulus.
The elastic moduli are usually derived through backcalculation of deflection bowls measured with the Falling Weight Deflectometer.
The method for calculation of the responses can be selected by the user as either Odemark-Boussinesq (MET) or Linear Elastic Theory (LET) using Waterways Experiment Station’s program (WESLEA)

For practical and financial reasons field testing with the Falling Weight Deflectometer may only be carried out on a fraction of the road network.

For the part of the network where the structural information is not present, PERS® offers the facility to use pure empirical relationships as a function of traffic or age of the pavement.
These relationships can also be used in combination with the mechanistic models to secure a minimum decrease in performance caused by other factors than the load repetitions.
Chapter 6 describes the use of this alternative.