International Roughness Index Model for Jointed Plain Concrete Highway Pavements: An Artificial Neural Network Application

An efficient and safe road network secures the nation’s economy and prosperity by providing public mobility and freight transport. Maintenance and rehabilitation of the road network cost billions of dollars annually. If maintenance and rehabilitation are not done promptly, the damages to the road caused by heavy traffic and extreme climate may lead to life-threatening conditions for road users. The literature review to date indicates that the Maintenance and Rehabilitation (M&R) history was not considered in the concrete pavement condition models. The hypothesis testing in this study demonstrated that it is imperative to use M&R history of the pavement in the development of the International Roughness Index (IRI) prediction model. Therefore, this study utilized Construction Number (CN) for developing IRI prediction models for Jointed Plain Concrete Pavements (JPCP). The models were developed using Long Term Pavement Performance (LTPP) database. A total of 590 data points was used to develop the IRI prediction using Artificial Neural Network (ANN) modeling technique. Three ANN models were developed using variables such as initial IRI, pavement age, concrete pavement thickness, equivalent single axle load (ESAL), climatic region, and CN. The best model was found to be the one with an R² value of 0.87. The IRI prediction model developed in this study can successfully estimate the increase of IRI values with time and decrease of IRI value after maintenance and rehabilitation. The developed IRI models can be utilized by the local and state agencies to prepare M&R programs and budget allocations.

Language

  • English

Media Info

  • Media Type: Digital/other
  • Features: Figures; References; Tables;
  • Pagination: 19p

Subject/Index Terms

Filing Info

  • Accession Number: 01764420
  • Record Type: Publication
  • Report/Paper Numbers: TRBAM-21-04227
  • Files: TRIS, TRB, ATRI
  • Created Date: Dec 23 2020 11:04AM