Long-Term Pavement Performance Automated Faulting Measurement

CHAPTER 5. CONCLUSIONS AND RECOMMENDATIONS

This study was conducted to develop a new LTPP AFM algorithm to detect transverse joint locations and compute joint faulting and to compare the new method with the existing AASHTO R 36-12 AFM methods (including the ProVAL AFM (AASHTO Method-A) and the FDOT PaveSuite AFM (AASHTO Method-B)). LTPP profiler longitudinal elevation profiles at 25-mm sampling intervals and the FDOT profiler at 20.7-mm sampling intervals were used. The joint detection results from the six selected LTPP sections show that the LTPP AFM algorithm was more effective than the ProVAL AFM routine. The JDRs from the ProVAL AFM for the six selected sections ranged from 58 to 99.4 percent, whereas the JDRs for the same six LTPP test sections using the LTPP AFM ranged from 95 to 100 percent. The average section biases computed for the ProVAL AFM and the manual GFM for test sections 364018, 370201, 421606, and 493011 were greater than 1 mm, as were the average section biases for the LTPP AFM and the manual GFM for test sections 313018, 364018, and 421606. These results could be because the joint faulting measurement surveys using the GFM and the LTPP profilers were not conducted on the same wheelpaths, at the same time of the day, or under the same temperature conditions.

The average of absolute differences between faulting estimated by the LTPP AFM and the manual GFM (average section bias) for the six LTPP sections ranged from 0.44 to 5.07 mm. Similarly, the average of absolute differences between faulting estimated by the ProVAL AFM and the manual GFM (average section bias) for the same six LTPP sections ranged from 0.88 to 8.75 mm. FDOT used an HSIP profiler to collect elevation profile data with a 20.7‑mm sampling interval on the same wheelpaths (left and right) where manual GFM measurements were collected, at same time of the day, and under the same temperature conditions. The JDR for the one FDOT test section using both the LTPP AFM and the FDOT AFM was 96 percent. The averages of absolute differences in faulting estimated by the FDOT PaveSuite AFM and the LTPP AFM with the measured manual GFM were 1.05 and 1.14 mm, respectively. From this study, it appears that the newly developed LTPP AFM is relatively reliable in detecting transverse joints and computing joint faulting.

RECOMMENDATIONS

• A profiler sampling interval of less than 25.4 mm may yield better results for both transverse joint detection and faulting measurements.
  
  
• Faulting measurements surveys conducted using the manual GFM and the LTPP profiler on the same wheelpaths, at the same time of the day, and under the same temperature conditions may generate better AFM results.
  
  
• Further research is needed to understand the bias of the AFMs and manual GFM measurements.