01373673

Component

The present paper aims to incorporate an artificial neural network (ANN) as an innovative technique for modeling the pavement structural condition in pavement management systems (PMS). The initial task is to set criteria for strains assessment in order to characterize the structural condition of flexible pavements in regards to fatigue failure. This is followed by the second task to develop an ANN model for the prediction of strains not by using synthetic databases but based on field data. For this purpose falling weight deflectometer (FWD) measurements were conducted on a highways network where required pavement thickness data was evaluated mainly based on ground penetrating radar technique. The FWD data (i.e. deflections) were back-analyzed to assess strains that were used as output data in the process of developing the ANN model. A paper exercise demonstrated how the developed ANN model in conjunction with the suggested conceptual approach to assess strains for characterizing pavement structural condition could make provisions for pavement management activities, categorizing network pavement sections according to the need for maintenance or rehabilitation. It is evidence that the ANN technique could assist decision makers in finding optimum strategies for planning maintenance of pavement infrastructure.

01362476

12-1377

English

16p

2012

Transportation Research Board 91st Annual Meeting

Digital/other

Figures; References; Tables

Condition surveys; Falling weight deflectometers; Flexible pavements; Neural networks; Pavement management systems; Pavement performance; Strain (Mechanics)

Design; Highways; Maintenance and Preservation; Pavements; I22: Design of Pavements, Railways and Guideways; I60: Maintenance

Transportation Research Board Annual Meeting 2012 Paper #12-1377

TRIS, TRB

Feb 8 2012 5:01PM