Caltrans Senior Engineer Sean Shepard (I-5 Red Bluff job) and Caltrans Resident Engineer John Bailey (I-5 Weed job) share lessons learned on the Long-Life Asphalt Pavement Projects on Interstate 5.
Editor’s Note: This is a web-exclusive piece to complement “Long Live Asphalt Pavements,” the National Asphalt Pavement Association (NAPA) Special Section in the Better Roads October 2013 print edition (pp. 18-27).
In gathering research for this article, Bob Humer had the following e-mail exchanges with Caltrans Senior Engineer Sean Shepard (I-5 Red Bluff job) and Caltrans Resident Engineer John Bailey (I-5 Weed job) regarding lessons learned on the Long-life Asphalt Pavement Projects on Interstate 5.
Bob Humer: Please elaborate on mix production, traffic control and other logistical issues:
Sean Shepard: Long-life HMA is produced using the same general process as conventional HMA, in a continuous-flow plant having adequate controls on the materials. There are somewhat tighter tolerances on long-life binder content than with conventional “Type A” HMA and this created some challenges with the QC/QA (Quality Control/Quality Assurance) process.
For the “North Red Bluff Rehab” project, traffic control was one of the biggest construction challenges. Interstate 5 is a four-lane facility within the “long-life” post-miles, and traffic volumes required us to keep both lanes open in each direction during the daytime.
The staging necessary to accomplish this work used an arrangement not frequently seen on Interstates — the traffic lanes in one direction were split, with one of the lanes crossing the median to run against traffic in the other direction, separated by a temporary concrete wall.
With one lane removed from service in a given direction, we were able to shift the remaining lane on to the shoulder enough to install another temporary concrete wall (Type K barrier rail) to protect traffic and construction crews, then roto-mill the full depth of the old Asphalt Concrete (AC) in the other lane, exposing the cement treated base (CTB) underneath.
In both directions, the CTB under the truck lane was found to be heavily cracked and in need of replacement prior to repaving with long-life HMA. The CTB was removed and replaced before the long life paving continued.
John Bailey: We have not encountered any challenges regarding the mix since the Job Mix Formula. All production/transport of the product has gone without issues.
BH: Can you describe the construction, phasing, and time needed for this project?
SS: Because of the narrow shoulders in many areas of the project, most of the long-life paving was done at night, when we were able to close the traffic lane that did not cross the median. During these times all traffic in that direction used the lane that did cross over the median.
Haul trucks carrying long life HMA entered the construction zone using the closed, non-crossed-over lane adjacent to the lane being repaved. Short windrows were placed in front of a Material Transfer Vehicle, which promptly picked up the windrow, reheated and re-mixed it, then used its conveyor belt to carry the material over the K-rail directly into the paver’s hopper.
The long-life mix goes down a lot like normal Type A, but rolling patterns and temperatures were even more important than usual because of the tighter compaction windows specified. Tullis was able to place 2000+ tons per night with this process. Each of four construction stages (one for each lane, basically) took about a month, working five or six 10- to 14-hour shifts/week. In total, we paved about 100,000 tons of long-life mix between July 4th and October 20th last summer.
The key to maintaining schedule was to keep the paver moving, so Tullis developed some hybrid staging plans that kept both the paver and traffic moving without losing any time between stages. These improvements to the contract plans were key to completing all of the long-life paving this past season.
JB: Our project has two mix designs. A bottom layer of HMA 25 percent RAP with a rubberized chip seal sandwiched between two lifts, and a top layer of HMA 15 percent RAP. All items associated with this pavement strategy (i.e., cold planing and chip sealing) are going well, and we have experienced no challenges.
I found the option of crossing traffic over the median to give the contractor full access to the entire northbound or southbound side to be very useful and cut construction times considerably.
BH: Can you describe in more detail the QC/QA and the challenges the variability in performance tests encountered, and the solutions implemented to address them?
SS: The QC/QA process for this long-life job uses the same structure as a normal Caltrans QC/QA job. The differences appear both in the target values and tolerances of our QC/QA performance measures.
For example, though we typically have a +/- .45 percent range on the binder content, long-life performance requires +/- .3 percent. Greater compactions are required as well: the normal 92 percent-96 percent compaction for Type A QCQA is replaced with 94 percent-97 percent for the 15 percent RAP and 25 percent RAP mixes, and the rich bottom required 97 percent-100 percent!
In addition to the more stringent QC/QA specs, new HMA performance requirements were introduced. Tullis was required to show that each of their three long-life mixes, when produced from the plant, met the design parameters for permanent deformation, beam stiffness, fatigue and moisture sensitivity.
Our project experienced some delays at the beginning of the construction season because multiple iterations were required, making adjustments to the mix design each time, before the desired performance was achieved. During this process, there was fantastic cooperation between Pavement Engineering, Inc. (Tullis’ QCQA sub), Tullis, Inc., the University of California Pavement Research Center, Caltrans HQ Pavement program personnel and Caltrans District 02 personnel.
We found that mix performance was fairly sensitive to changes in binder content and fine gradings. With multiple testing iterations and feedback from UCPRC, private testing lab PEI, Tullis collaboratively refined their initial long life mix designs to bring their performance in compliance with our specifications.
During mainline production, the most challenging part of maintaining mix quality had to do with controlling the total binder content. The 25 percent RAP long life HMA was particularly sensitive to variations in binder content of the RAP used. Total oil content was checked frequently and the new oil quantity was adjusted accordingly, in order to maintain total binder content within the tighter specifications. Rolling patterns and temperatures were fine-tuned to achieve the desired in-place compaction.
JB: All QC/QA testing has gone relatively well. I can’t associate the challenges we encountered to the LLP strategy, as they were lab equipment related.
Once the issue was identified, we had very little variation between QC and QA testing. All through the season we did, however, achieve verification. During the performance testing phase leading up to the JMF, we encountered several challenges. I believe all of the them could have been avoided if we specified all mix produced for the performance testing to be lab-produced as opposed to plant-produced. I believe the testing would have yielded more consistent data results considering the size of the samples taken.
The variability in the data during the performance testing, which was caused by producing mix at the plant, caused a lot of concerns, and the resulting delays pushed construction further into our limited construction window in Northern California. Although minor, the costs associated with suspending the project during the winter have increased because the top lift still needs to be paved on two of our stages.