Dynamic load responses of perpetual pavement test roads on US 23: Full-scale instrumentation and evaluation

Perpetual pavement experimental roads of varying asphalt concrete thicknesses (28 cm, 33 cm & 38 cm) were constructed and instrumented on U.S. Route 23 in Delaware County, Ohio. Controlled Vehicle Load (CVL) testing was performed to obtain dynamic load responses under vehicular load and to evaluate influence of vehicle speed, tire pressure and axle configuration. CVL tests were conducted in November 2012 and July 2013 to evaluate temperature effects on pavement responses. A linear elastic finite element model program, OpenPave was used to obtain layered elastic solutions of load responses with time history to validate measured results. Paired t-tests were performed to compare mean difference between measured and calculated strain responses. Factorial ANOVA test was performed to evaluate significance of truck speed and tire pressure on measured strain responses. It was found that the calculated strain responses were overall in good agreement with the measure values. A uniquely designed strain gage rosette was installed to obtain vertical strain in the pavement and was found to be effective. However, significant differences existed between measured and calculated strain responses in summer test due to unaccountability of the FE program for viscoelasticity of asphalt concrete materials. 33 cm section with stabilized subgrade performed the best with lowest strain responses. Tandem axle truck produced strains lower than single axle truck even though total weight was heavier. Transverse strain tends to increase as lateral wheel offset increases.

Automated summary

Experimental roads with different thicknesses of asphalt concrete were constructed to study the effects of vehicle load, temperature, tire pressure, and axle configuration. The results showed that the calculated strain responses were generally in good agreement with the measured values. A strain gage rosette was effective in measuring vertical strain in the pavement. However, there were significant differences between measured and calculated strain responses in the summer test due to the limitations of the finite element program used.