Publication Number: FHWA-HRT-08-006

The Ongoing Evolution of FRP Bridges

by Jim Williams

In a research project on new hybrid structural construction, Texas explored the viability of custom fiber-reinforced polymer beams.

The Refugio County Bridge is shown here with the FRP beams in place on 1.2-meter (4-foot) centers with plywood formwork between the beams.

Refugio County near the southeastern Texas coast has a humid, subtropical climate with an average of 94 centimeters (37 inches) of rain annually. Given the humidity and proximity to the coast, brackish water in streams and drainage ditches corrodes the county's highway bridges and increases maintenance costs. To address the corrosion issue, the Texas Department of Transportation (TxDOT) specified customized, fiber-reinforced polymer (FRP) composite beams when it replaced a drainage ditch bridge (FM-1684) in Refugio County in 2007.

Although FRP beams are more costly upfront, TxDOT selected them to research the long-term corrosion and structural performance benefits of this material versus traditional steel or concrete beams. The Refugio County Bridge, which is 56 kilometers (35 miles) from Corpus Christi, is the State's second FRP hybrid bridge project. The first was the successful construction of the San Patricio County Bridge in 2005.

The new Refugio County Bridge replaces a single-span girder bridge that was 15 meters (50 feet) long by 10 meters (32 feet) wide. The new bridge has eight customized FRP flanged U-shaped beams and a concrete deck. The beams are 15 meters (50 feet) long by 76 centimeters (30 inches) high, with a composite structure that provides optimal deflection under load. The new beams weigh approximately 2,270 kilograms (5,000 pounds) each and sit on abutments.

"The [Refugio] project began in late 2006 and was completed in the fall of 2007," says Rich LaFountain, the manufacturer's business unit leader for open molding. "The project's goal was to take the lessons learned from the previous bridge project so that the current customization and production processes could evolve in hopes of optimizing performance and cost variables for future projects."

The Innovative Bridge Research & Construction Program helped fund the Refugio project by contributing $462,500.

The Forming Process

In the first lesson learned from the previous project, the manufacturer fabricated the beams using a vacuum infusion process (VIP) rather than the hand layup used on the San Patricio bridge. In a typical hand layup, the manufacturer lays reinforcements into a mold and manually wets down and melts the resin using brushes or rollers. The typical hand layup usually results in excess resin, and resin is very brittle, so any excess actually will weaken the part.

The VIP process, on the other hand, uses a vacuum bag to suck out excess resin from the laminate. Vacuum bagging greatly improves the fiber-to-resin ratio, eliminating all air voids in the laminate and resulting in a stronger and lighter product. VIP is still not ideal and can lead to configuration issues because curved products or those with intricate angles will not work as well in the bag. Vacuum infusion does provide a number of additional benefits, however, including consistent fiber-to-resin ratio, less wasted resin, no limit on setup time, and much lower acoustic emissions.

Using VIP on the Refugio County Bridge project, the manufacturer produced a male mold to the beam design and then laid dry sheets of stitched glass fabric and chopped strand mat over the U-shaped mold in a series of layers to achieve the appropriate 3.8-centimeter (1.5-inch) beam thickness. The next step involved laying a plastic film on top to serve as the vacuum bag. Once a complete vacuum was achieved, the manufacturer then introduced liquid resin into the laminate. The vacuum drew the resin through the fibers, filling all the voids and eliminating any remaining air.

According to LaFountain, "The trick is to get the bag to draw down correctly so that wrinkles don't develop in the individual layers of fabric, which could affect the ultimate strength of the composite."

The San Patricio County Bridge, shown with the beams in place with foam to fill the void up to the steel bar level.

After the manufacturer fabricated the beams, Robert Sarcinella, materials branch manager for the TxDOT Construction Division, and his staff inspected them. Sarcinella notes, "Using the vacuum process, [the manufacturer] fabricated the beams more quickly and with better quality than on the other project."

Assembly and Installation

Once the beams were completed, the manufacturer cured, trimmed, and assembled them with shear transfer members (brace bars) installed into inserts placed into a 5-centimeter (2-inch) hole drilled into the vertical sides at 41-centimeter (16-inch) centers over the entire length. The inserts were installed to enable the brace bars to make a more positive connection with the composite beams.

This cross section slice of an FRP beam shows a 5-centimeter (2-inch)- diameter brace bar at the top for reinforcement.

After delivery to the site, the contractor placed the beams at 1.2-meter (4-foot) center-to-center spacing and then poured the reinforced concrete deck. The contractor tied the deck to the beams with horizontal pipe 6.6 centimeters (2.6 inches) deep by 5.8 centimeters (2.3 inches) wide close to the top of the beams. The concrete deck pour was deep enough to engage the brace pipe for optimal strength to tie the beams to the deck. The goal was to achieve composite action, so that the bridge would be able to flex while creating a solid connection between the deck and beams.

Acoustic Emission Testing

In April 2007, before installation of the beams, Guillermo Ramirez, P.E., Ph.D., of The University of Texas at Arlington, and Paul Ziehl, Ph.D., of the University of South Carolina, performed an acoustic emission (AE) evaluation test on two of the beams. This method of nondestructive testing uses mechanical waves moving through materials. When researchers load a structure, subjecting it to external force (or stress), a defect such as a crack or a welding flaw is activated and generates waves that spread at a certain speed. The researchers listen through headphones, and when the loading makes the structure flex, emissions of a certain decibel indicate a beam's stiffness and ability to sustain service loads.

The Refugio County Bridge is shown almost ready for the pouring of a conventional reinforced concrete deck. Wood strips hold the beams in place.

For the tests on the Refugio beams, Ramirez and Ziehl monitored AE during the background check prior to loading, during load holds, and during the background check after completion of loading. The test threshold was 40 decibels (dB), and the evaluation threshold was 48 dB. The thresholds are an indication of how strong a signal has to be in order for the system to consider it valid for recording (converting to digital information and storing). The evaluation threshold is the point at which recorded data is used in the analysis, eliminating anything under that level of energy.

The main sensors used were type R15I, resonant in the range of 150 kilohertz (kHz). The two researchers monitored activity from the R15I sensors and recorded it with a 24-channel transportation instrument — an AE data acquisition system that captures information from resonant frequency sensors and extracts statistical values from the signal. For supplemental analysis of the data, Ramirez and Ziehl used broadband high-fidelity sensors that enabled them to digitize the waves for later analysis.

According to Ramirez, "The test verified the performance of the beams under the load criteria set forth by the project specifications. The beams selected by TxDOT performed well during load testing — passing the major criteria selected for the acoustic emission test. In fact, the beams' stiffness tested better than expected, substantiating their ability to sustain the inservice loads." TxDOT's Sarcinella agrees that deflection was almost nonexistent.

Ramirez adds that the beams had no visible flaws. "The method of fabrication resulted in a very good product," he says, and again Sarcinella agrees.

The Test Procedures

The Ramirez and Ziehl report, San Refugio County FRP Bridge Beams (Acoustic Emission Evaluation — Beam Nos. 1 and 2), provided to TxDOT in July 2007, summarizes the AE evaluation of the two beams that the researchers performed on April 18, 2007. Ramirez and Ziehl originally loaded beam numbers 1 and 2, using hydraulic rams in an indoor environment, to a total load of 32.0 kips (in increments of 454 kilograms, or 1,000 pounds) on April 17. Because they tested the beams indoors, wind was not present. Background checks prior to and after testing generally were satisfactory, verifying that there were no external sources of AE from those produced by the application of load.

The AE monitoring did not take place on the first loading, but instead the beam was loaded simply to eliminate signals from insignificant sources like bubbles or flaking that does not indicate any kind of permanent damage. The specification did not require monitoring during the first loading, so this caused no departure from the specified testing procedure.

The researchers then allowed the beams to remain in the testing facility unloaded for a minimum period of 12 hours prior to reloading. They reloaded the beams on April 18 to 32.0 kips in accordance with the specifications, at which time the AE testing took place.

Test Results for Beam Number 1

According to the specified evaluation criterion — Article RT-6 of the American Society of Mechanical Engineers (ASME) Section X — no acoustic emission is allowed during the 2-minute evaluation period of the 4-minute load holds, and no emission is allowed during the 28-minute evaluation period of the 30-minute load hold. If strictly applied, beam number 1 would not be in conformance with the specification. However, given that no emissions occurred during the 2-minute evaluation period, and the emissions during the 28-minute evaluation period were of small or medium amplitude, Ramirez and Ziehl concluded that some deviation from the specification might be warranted, and TxDOT concurred with the conclusion.

After they completed the testing, the researchers noticed a drainage hole drilled in the vicinity of sensor R12. Highly localized damage at this drainage location might have been the cause of much of the AE activity recorded by this channel. This observation, they concluded, further supports the case for allowing some deviation from the specification.

Results for Beam Number 2

Beam number 2 had similar emission to beam number 1, but the amount of emission increased in both quantity and amplitude. For this beam the criterion of no emission during the evaluation period of the load holds was not met for the holds at 20, 28, and 32 kips. Again, in most cases the emission was of small or medium amplitude. In one case during the 30-minute load hold, the emission was large in amplitude (81 dB). However, it was only one instance and did not exceed the maximum number of events as specified in the criteria, so TxDOT concurred.

The FRP beams, shown here from the side, rest on a concrete abutment, as with conventional concrete and steel beams.

Similar to beam number 1, much of the AE activity was recorded from sensor R12, which was near the drainage hole.

Assessments

Roy Tijerina, superintendent with the general contractor Haas-Anderson Construction of Corpus Christi, assesses the short-term benefits of the FRP beams: "[The manufacturer] delivered all the FRP beams in one truck, and handling and installation were easier because of being able to use a small crane or large track hoe versus multiple cranes with steel or concrete options. This meant that minimal equipment and people were required, which equated to built-in time and cost efficiencies on the project."

A postconstruction assessment by FHWA Division Bridge Engineer Peter Chang notes, "The funding to promote the new fiberglass girder technology was allocated by TxDOT as a research project. With the load testing calculated and installation complete, the beams are actually stronger than we anticipated, thus proving the research positive." The structure was load tested during construction to verify the load-carrying capability of the entire unit.

Sarcinella concludes, "Projects of this nature generally start as a research project and then move to an implementation project if they show merit. That was the case for FRP beams. In the implementation project, we found (under loading) that the structure had more stiffness (less deflection downward) than was calculated in the design phase. This is a good thing and could be attributed to several factors (but most likely due to the beams). We have no plans to do any additional load testing unless the structure shows signs of decreasing camber (that is, does not bow upwards)."

The future use of this product with TxDOT will be based largely on cost and flexibility of use.

Jim Williams started his career in the manufacturing business in 1954 working for Boeing Aircraft, where he was involved in the manufacture of the nose cone for the Boeing 707. In the fiberglass industry for more than 35 years (supporting the construction side for 29 years), Williams served as plant manager for Molded Fiberglass (MFG) Construction Products for 10 years.

For more information, contact Robert Sarcinella, 512-506-5933 or rsarcin@dot.state.tx.us.