Partner und Internationale Organisationen|
Belgian Road Research Centre,
Technical University of Wien,
Centro de Estudios y Experimentacion de Obras Publicas, National Technical University of Athens, Swedish National Road and Transport Research Institute, Technical Research Centre of Finland,
Druzba za Drzavne Ceste, d.o.o., Institute for Transport Sciences ltd, Transport Research Laboratory, Laboratorio Nacional de Engenharia Civil, Technical University of Denmark,
Bergische Universität, Gesamthochschule Wuppertal
The structural design of a pavement and the prediction of its long term performance are two complementary and closely linked tasks. They indeed are both based on similar models (empirical or mechanistic). Such models now in use in several countries are considered as essential tools in the construction and maintenance of a road network.
The AMADEUS research project is a major step in the development of a European advanced pavement design method. It concerns the review and evaluation of pavement design models that are already in use for practical applications. The purpose of this activity is to propose guidelines for potential users of advanced models and set up the plans for a design procedure, already outlined in the frame of the COST333 action. The new procedure will integrate the strong points of existing methods while considering the long-term evolution of the pavement properties in an incremental way.
A first part of this activity consisted in making the inventory of existing design models or software packages and selecting those that are most relevant to our purpose.
In a second part of the work, a detailed program was proposed in order to organise the evaluation of these software's by different teams formed among the partners of the AMADEUS consortium. It was decided that the evaluation would be carried out in 3 phases of increasing complexity.
Phase 1 to evaluate 'simple' response models (calculation of stress, strain and displacement).
Phase 2 to compare calculated responses with measurements carried out on accelerated tests managed by three partners (DTU, LAVOC and CEDEX).
Phase 3 to compare the long term damage predictions made by means of the models with the actual behaviour observed on German road sections monitored by BASt over long periods of time (up to 20 years).
The work was divided between four teams with the team leaders being responsible for summarising their team's assessment of the models. These were combined to produce a detailed assessment report and information from this report has been used to compile a users guide. This guide will help potential users in choosing design tools suited to their specific problems.
This co-operation finally resulted in a plan for developing a more comprehensive harmonised pavement design method.
The inventory of existing tools reveals a wide variety of products in terms of response models and type of damage they address, but none of them are presently accounting for all damage types nor take their mutual interactions into account.
Models based on the multi-layer elastic theory are easy to use and they give generally similar results. Some of these models, however, have limitations in precision or produce wrong results in stress/strain analysis at particular locations of the structure (particularly in the vicinity of the applied loads and in the subgrade).
Multi-layer elastic models can be used for non-linear elastic materials (with stress dependent stiffness) provided they are included in an iteration loop.
Applications made during the AMADEUS evaluation task has shown that linear visco-elastic models better describe the shape of the stress and strain waves generated by moving loads. Their advantage over linear elastic approaches. The applicability of such models remains limited by the lack of reliable rheological input data. Their practical advantage depends mainly on the magnitude of the errors made by using the more simple linear elastic approach.
Models based on finite element approaches are much more difficult to apply unless users have a thorough knowledge of basic principles. Unlike analytical multi-layer methods FE methods need the definition of a system with horizontal and vertical limits in space. Hence their implementation needs:
1. a precise definition of the boundary conditions.
2. a decomposition in discrete elements allowing an accurate evaluation of the stresses and strains.
A distinction should be made between plane stress/strain two-dimensional methods (2D), axi-symmetrical methods and three-dimensional methods (3D).
In the context of pavement analysis, 2D methods are irrelevant and axi-symmetrical methods have little advantage over multi-layer elastic models.
True 3D methods are really worth to be applied for situations where particular boundary conditions and local discontinuities must be modelled, which is the case of crack propagation and reflective cracking.
Long term performance prediction must consider the actual history of a pavement structure, including the evolution of all the external (loads, temperatures, etc…) and internal (geometry, material properties) factors. This requires an iterative procedure, called 'incremental', composed of a response model surrounded by different design elements providing time dependent input data.
The work performed by COST 333 and AMADEUS resulted in an outline harmonised pavement design method, taking into account the present knowledge in the field of pavement design.
The incremental procedure recommended for the new design method will be able to perform life-cycle analysis throughout the whole pavement life, including not only the initial stage, but also future maintenance and rehabilitation strategies.
A staged approach was proposed, whereby the design method will be based on current best practice, and will be latter improved with the incorporation of new developments in this field.
In fact, in order to be able to describe more accurately the deterioration mechanisms that pavement components undergo during their life cycle, future research is needed on several related topics, in order to reach a fundamental pavement design. These aspects can later be incorporated in the procedure, so that it will be able to take the following into account:
Material and structure changes occurring during the life-cycle of the pavement, such as asphalt ageing.
More detailed traffic input, taking into account different axle configurations, axle weights, etc.
Most relevant forms of deterioration observed in actual pavements, including, not only the classical forms - fatigue cracking starting from the bottom of the bound layers, and permanent deformation originating in the subgrade -, but also:
- Cracking initiated at the surface;
- Rutting initiated at the surface layers;
- Wear due to studded tyres, frost heave and temperature cracking, in cold climates.