Desenvolvimento de modelos de irregularidade longitudinal e afundamento em trilhas de roda para um trecho da BR 285

Abstract

The age of brazilian pavements, along the intense traffic and the limited investiment in infrastructure has resulted in a poor road network. Restricted budgets require a management system to decide on investiments for each section of a pavement, trough the use of performance prediction models. The goal of this research was to develop performance prediction models for part of a road with asphaltic pavement at the northern region of Rio Grande do Sul, aiming their application on network level. The models were developed to predict wheel track rutting and roughness, through the international roughness index (IRI), two important measures of pavement performance. The age of the pavement was set by the year that showed deffect value reduction larger than the minimum established. It was observed that both rutting and IRI evolved nonlinearly and logarithmic equations were developed. The equations were developed through regression analysis between the dependent variable (rutting or IRI) and the independent variable (pavement age and initial pavement condition), determining the coefficients b1, b2 and b3. To improve regression results was applied the statistical technique bootstrapping. The equation created to estimate IRI evolution showed rmse equal to 0,6772 and the equation for rutting evolution, rmse equal to 5,2577. The difficulty to acess detailed and current information about material and building systems made the whole pavement be considered homogeneous and made it impossible to create models more suited to each part of the pavement characteristics. It was concluded that the IRI model was adequate and, even as a simple model, represents the pavement better than other known models. The rutting model was not adequate and needs a more suitable relation to regression analysis. It was possible to determine, through these models, which sections needs maintanence, predicting, then, the priorities and the spendings with maintanence in each year.