Pavement Performance : Approaches Using Predictive Analytics

Details

• Creators:
  Duckworth, William ; Nath, Ravi
• Corporate Creators:
  Creighton University. Department of Business Intelligence & Analytics
• Corporate Contributors:
  Midwest Transportation Center ; United States. Dept. of Transportation. Office of the Assistant Secretary for Research and Technology
• Subject/TRT Terms:
  Analytics Mathematical Models Pavement Condition Index Pavement Distress Pavement Management Systems Pavement Performance Prediction Regression Regression Analysis
• Resource Type:
  Tech Report
• Geographical Coverage:
  Iowa ; United States
• Corporate Publisher:
  Midwest Transportation Center
• Abstract:
  Acceptable pavement condition is paramount to road safety. Using predictive analytics techniques, this project attempted to develop models that provide an assessment of pavement condition based on an array of indictors that include pavement distress, pavement type, traffic load, structural data, and pavement repair history. Data collected by the Iowa Department of Transportation (DOT) regarding road conditions across the state of Iowa were used to model pavement condition index (PCI). All data were from calendar year 2013 and consisted of nearly 4,000 observations. Various distress indicators were used to model PCI. These distress measures quantify a variety of cracks (types of cracks, severity of cracks, and amount of cracking) as well as joint spalling (severity and amount) and the condition of previous patching (condition and amount). Twenty-three distress measures were considered as possible model inputs. In addition to distress measures, nine descriptive variables were tested as potential model inputs for improving the overall fit of the model to the data. These descriptive variables included traffic, load, speed limit, number of lanes, pavement thickness, and pavement age. Series of multiple regression models were developed for different pavement types and for combined data (when all pavement types were aggregated). The results reveal that a number of distress variables and descriptive variables have a statistically significant relationship with PCI. The efficacies of the derived models, as measured by R2 values, range from 44% to 86%. The results of further analyses show that the introduction of the quadratic effects of certain variables on PCI improves model efficacy. Therefore, it is concluded that linear predictive models that involve distress and descriptive characteristics of road conditions provide a reasonable basis for estimating PCI. However, these models can be further improved by examining nonlinear effects.
• Format:
  PDF
• Funding:
  Part of DTRT13-G-UTC37
• Collection(s):
  US Transportation Collection
• Main Document Checksum:
  urn:sha256:192e65ea2f93476037787c157f530afa9f1fc819a4514e48e31a511cc5ac2ec3
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/35692/dot_35692_DS1.pdf
• File Type:
  [PDF - 938.39 KB ]