Development of an AI-Powered IDEAL Fatigue Test with a Low-Cost Loading Frame

Fatigue is a fundamental property of asphalt mixes and a key input to pavement mechanistic-empirical (ME) designs. Fatigue failure of asphalt pavements makes state departments of transportation (DOTs) spend a hefty sum annually in their rehabilitation and reconstruction. Such failure is affected by asphalt mixes, but it is essentially a pavement structural design issue. Currently, more and more state DOTs are designing their pavements using ME design programs that require input of fatigue properties of asphalt mixes determined by a fatigue test. However, bending beam fatigue (BBF) test (AASHTO T321) is seldom performed by state DOTs, asphalt industry, or academia because of at least three limitations: (1) expensive compaction and test equipment (> $100,000), (2) tedious specimen preparation and cutting time, and (3) long testing time (weeks). Without fatigue testing, default fatigue property values (or model parameters) are generally used, making the resulting pavement designs questionable. For example, reclaimed asphalt pavement (RAP) often stiffens asphalt mixes. The increased modulus leads to a thinner pavement structure from the ME design programs if the actual fatigue property of asphalt mixes containing RAP is not used in the design process. With emphasis on sustainability and use of more recycled materials in asphalt mixes, it becomes rather urgent to have a simple and expensive fatigue test for routine use by state DOTs and the asphalt industry. For NCHRP 20-30/IDEA 265, the research team will develop an artificial intelligence (AI)-powered IDEAL fatigue test to make it simple and easy for state DOTs to determine fatigue of their asphalt mixes and optimize their pavement structural designs more accurately. The test will use a low-cost loading frame that most DOTs or contractors already have for balanced mix design. An AI-powered learning algorithm will be developed that will take the measured displacement-force data to determine the fatigue model parameters. The algorithm will be trained with historical Flexural BBF test datasets and validated by comparing the fatigue model parameters of at least 10 asphalt mixes determined by the BBF test with those using the AI-powered algorithm. Using test results, the system’s hardware and software will be refined and finalized. A test procedure will be drafted for evaluation by partner state DOTs. Eight state DOTs (Texas, California, Michigan, Minnesota, Mississippi, Massachusetts, South Carolina, and Virginia) are collaborating on this effort. These partner states will also help develop a plan to demonstrate the new AI-powered fatigue test to other state DOTs.

• Record URL:
  • http://apps.trb.org/cmsfeed/TRBNetProjectDisplay.asp?ProjectID=5941

Language

• English

Project

• Status: Active
• Funding: $139,000.00
• Contract Numbers:

  Project 20-30, IDEA 265

• Sponsor Organizations:

  National Cooperative Highway Research Program

  Transportation Research Board
  500 Fifth Street, NW
  Washington, DC  United States  20001

  American Association of State Highway and Transportation Officials (AASHTO)

  444 North Capitol Street, NW
  Washington, DC  United States  20001

  Federal Highway Administration

  1200 New Jersey Avenue, SE
  Washington, DC  United States  20590
• Project Managers:

  Jawed, Inam

• Performing Organizations:

  Texas A&M Transportation Institute

  ,    
• Principal Investigators:

  Zhou, Fujie

• Start Date: 20240101
• Expected Completion Date: 0
• Actual Completion Date: 0

Subject/Index Terms

• TRT Terms: Artificial intelligence; Asphalt mixtures; Fatigue tests; Mechanistic-empirical pavement design
• Subject Areas: Design; Highways; Pavements;

Filing Info

• Accession Number: 01993225
• Record Type: Research project
• Source Agency: Transportation Research Board
• Contract Numbers: Project 20-30, IDEA 265
• Files: TRB, RIP
• Created Date: Jun 23 2026 1:37PM