Comparison of Pavement Condition Assessment and Prediction Models on Road Section and Network Level

Periodic condition assessments of flexible and rigid pavements together with condition predictions are the basis for investment decisions in every systematic pavement management system (PMS). Typical approaches include periodic surveys of several failure types every 3-6 years with the results being analysed, rated and combined to condition indices e.g. with regard to road safety or structural health. In every advanced PMS in addition condition prediction models are utilized allowing a comparison of different maintenance options and an optimization of investment strategies. The paper presents an overview on current condition survey and rating approaches on highway and regional road level in Germany, Switzerland and Austria. Currently utilized deterministic pavement performance functions and origins of defined failure criteria for all relevant failure types are analysed as well. However, the main emphasis of the paper is a comparison of common deterministic condition prediction models with discrete stochastic approaches mainly from scientific literature and prediction models based on advanced regression techniques. All prediction models are applied to real world data from condition surveys in Austria and the Long Term Pavement Performance – Database LTPP (USA) both on road section and network level. The results of these models are then compared to an innovative stochastic continuous time and continuous state space process (HOFFMANN – Process). The results of the analysis show that simple deterministic prediction approaches based on shifting and scaling of performance functions (master) fall short e.g. due to the stochastic nature of pavement performance leading to substantial bias in condition distribution and remaining service life. Discrete stochastic Markov-chain approaches are highly praised in literature, but fall short regarding modelling of non-linear condition development, account for data censoring and are neglecting the fact of changing transition probabilities with increasing age. Applying common bivariate and multiple regression techniques may also lead to certain bias due to collinearity effects and specification bias. Almost always occurring data censoring due to a lack of information on already failed sections is neglected leading to an overestimation of service lives and underestimation of maintenance costs. The research provides substantial mathematical evidence on ways to overcome these shortcomings based on the presented innovative stochastic process even on the basis of current condition survey data. With these results a higher reliability in condition assessment, rating and accuracy of condition prediction may be achieved leading to lower risks for road users and higher efficiency of invested funds.


  • English

Media Info

  • Media Type: Digital/other
  • Features: Figures; Photos; References; Tables;
  • Pagination: 16p
  • Monograph Title: Proceedings of the 25th World Road Congress - Seoul 2015: Roads and Mobility - Creating New Value from Transport

Subject/Index Terms

Filing Info

  • Accession Number: 01709909
  • Record Type: Publication
  • ISBN: 9782840604235
  • Files: TRIS
  • Created Date: Jun 20 2019 10:07AM