Logged in as: Public User

Expansion Gap Joint Width

State DE
Description Text MwRSF was contacted through FHWA for the state of Deleware regarding our thoughts on the allowable gap lengths
for unshielded expansion joints in permanent concrete bridge rails.
  • Bridge Rails
Other Keywords Expansion Joints
Date August 17, 2016


There are two main concerns with expansion joints in concrete bridge rails. The first is the lack of continuity in the rail and the need for increased rail capacity adjacent to the opening to maintain the rail strength. For this discussion, I will assume that the gaps we are discussing are designed with appropriately reinforced rail end sections to address the lack of rail continuity. The second concern is vehicle snag on the downstream end of the gap as the vehicle is redirected. Snag in this area can potentially result in increased vehicle deceleration, vehicle damage, and instability.


Determination of a maximum unshielded gap can be looked at in several ways. First we can look at available test data. MwRSF tested a Nebraska open concrete bridge rail with a 4.5” gap under the PL-2 criteria. We evaluated this barrier at the gap with a 2,449 kg pickup truck at 61 mph and 20 degrees and a 8,165 kg SUT at 51.9 mph and 16.8 degrees. Both of these tests were successful, but snag was evident in both tests. Additionally, the pickup truck test was conducted at a lower angle than NCHRP 350 or MASH TL-3 and TL-4. More recently, TTI tested the Texas T224 bridge rail to MASH TL-4. As part of that evaluation, they successfully tested the 10000S vehicle across a 2” wide expansion gap. The passenger vehicle tests were not conducted across the expansion gap.



I don’t have the TTI report. It may not yet be published.


Similarly, temporary barrier designs in free-standing and anchored configurations have had gaps as large as 4” that have been successfully traversed by vehicle in TL-3 testing. This would suggest that the potential for larger gaps may exist. For example, the Midwest States F-shape PCB was tested to NCHRP 350 TL-3 anchored to a bridge deck with a 4” barrier gap. This test did have some snag across the PCB joint, but the vehicle was safely redirected. The snag may have even been exaggerated in this type of system as compared to a rigid bridge rail as the upstream barrier could translate more laterally prior the vehicle traversing the gap, thus exposing the face of the adjacent barrier at the downstream end of the gap even more.







When we have considered this in the past, we have looked at the potential vehicle overlap on the downstream gap edge. We know from previous research that vertical asperities can cause problems with barrier performance. NCHRP 554 concluded that vertical asperities of ¼” or less were recommended to maintain vehicle stability and safe redirection. Previous testing conducted at MwRSF on a portable steel barrier for IaDOT noted similar concerns when a 3/8” thick vertical plate on the face of the portable barrier was sufficient to snag a pickup truck rim and cause the vehicle to roll. However, these two examples may not be completely analogous as NCHRP 554 dealt with aesthetic bridge rail designs and the portable steel barrier had a plate that extended from the face of the barrier





In previous discussions with the state DOT’s, the issue of the allowable level of lateral misalignment of the barriers has come up. With regards to permanent concrete barrier, we recommended keeping the lateral offset or alignment offset minimized to eliminate snag. Variations of 1" or less would be preferred. We also recommended that the edges of the gap be chamfered to reduce the severity of any vehicle snag on the gap. The 1” offset was larger than the ¼” or 3/8” noted above based on the fact that these gaps were specific to concrete barrier overlaps where the concrete would be expected to fracture and give when snagged. If we use a similar rational and apply it to the expansion joint problem, we can use a gap length “L” and a 25 degree impact angle or intrusion angle of the impacting vehicle to estimate snag. This would make the snag or overlap on the downstream end of the gap approximately equal to L*sin(25). This results in the following snag for various gap lengths.


Gap Length “L” (in.)

Estimated Snag (in.)










This would seem to suggest that a 2” gap will limit snag to less than 1”.


As you can see, there may not be a perfectly defined answer. Previous anchored PCB testing and the PL-2 open concrete rail tests suggest that 4” gaps may be permissible. However, a more conservative approach may be to limit the gaps to 2”. In both cases, we recommend chamfering the edges to limit the snag severity.



Date November 2, 2016

Contact Us:
130 Whittier Research Center
2200 Vine Street
Lincoln, NE 68583-0853
(402) 472-0965
Email: mwrsf@unl.edu
The information contained on the Midwest Roadside Safety Facility (MwRSF) website is subject to change without prior notice. The University of Nebraska and the MwRSF is not responsible or liable, directly or indirectly, for any damage or loss caused or alleged to be caused by or in connection with the use or misuse of or reliance upon any such content, goods, or services available on this site.