Inspection, Evaluation, and Rehabilitation of the Taylor Bridge Gusset Plates

To view the figures and tables associated with this article, please refer to the flipbook above.

The Taylor Bridge in northeastern British Columbia represents an essential crossing of the Peace River for residents and industry. Built in 1960, the bridge carries a significant number of trucks with 30% of all traffic being heavy vehicles.

A joint venture consisting of T.Y. Lin International Canada Inc. (TYLin), Hatch Ltd. (Hatch) and Charter Project Delivery Inc. (Charter) has been contracted by the British Columbia Ministry of Transportation and Infrastructure (BC MoTI) to provide Owner’s Engineering (OE) Services for the Taylor Bridge Project. The project involves the development of various options for the future of the existing Taylor Bridge in northeastern British Columbia.

The Taylor Bridge is a two-lane, six-span, 712-meter-long structure that carries Highway 97 across the Peace River (Fig. 1). Five of the six spans are comprised of variable-depth steel trusses, while the remaining span is a stringer span at the north end of the bridge. Five concrete piers and two concrete abutments support the superstructure. From an articulation standpoint, the truss spans are a combination of continuous structures and suspended spans.

As part of their Owner’s Engineering Services assignment, the OE conducted a series of bridge inspections. These inspections identified the presence of relatively advanced corrosion on many gusset plate connections where the truss verticals and diagonals frame into the lower chord.
The OE determined recommended actions for the corroded gusset plates, including a scheme for strengthening.

The OE carried out three inspections involving the gusset plate connections. The first inspection, conducted in May 2021, involved a complete inspection of the bridge, with the gusset plate connections being only one of many components included. This inspection indicated the need for more detailed measurement of the gusset plate section loss. The second and third inspections, conducted in March 2022 and April 2022, were targeted inspections, focusing on the most heavily corroded gusset plate connections.

The targeted inspections involved ultrasonic testing measurements of the remaining thickness of the gusset plates. A combination of snooper truck and rope access was used to reach the gusset plate nodes. The measurements were taken in a grid pattern over the extent of the corroded region. The ultrasonic testing measurements gave a total remaining plate thickness but did not provide information on the asymmetry of the corrosion, i.e. how much section loss occurred on each face. This was estimated in the field for each node via measuring the depth from a straight edge to the corroded face on both sides of the gusset at a select few points and averaging the results.

An example output of the targeted inspection ultrasonic measurements for one gusset plate is given in Figure 2.

The overall load evaluation of the Taylor Bridge was conducted in accordance with Section 14 of the Canadian Highway Bridge Design Code CSA S6:19 and the BC MoTI Supplement to CSA S6:14. However, these standards contain limited information with regards to the evaluation of gusset plates and no information on the evaluation of gusset plates with localized corrosion. The methodology used for gusset plate capacity was therefore based primarily on the AASHTO Manual for Bridge Evaluation 3rd Edition. The National Cooperative Highway Research Program (NCHRP) Report (Guidelines for the Load and Resistance Factor Design and Rating of Riveted and Bolted Gusset-Plate Connections for Steel Bridges), which much of the AASHTO MBE methodology is based on, was also relied upon for additional background information.

The effects of localized section loss due to corrosion were accounted for in two ways: (1) considering an effective remaining gusset plate thickness, and (2) considering the out-of-plane bending stresses induced as a result of asymmetric corrosion on either side of the plate.

For the effective remaining gusset plate thickness, the OE considered followed the recommendations of AASHTO MBE. Different effective thicknesses were used depending on the failure mode considered. For shear and tension checks, the effective remaining thickness was simply taken as the average remaining thickness along the shear plane or Whitmore tension plane, respectively. For compression checks, the effective remaining thickness was the average remaining thickness along the Whitmore compression plane (Fig. 3).

AASHTO MBE provisions were also followed to determine when asymmetric corrosion effects on either side of the plates need to be considered. The limit at which it was considered is given as:

(e*c)/r² < 0.25,

Where:
‘e’ is the thickness eccentricity due to corrosion
‘c’ is the distance from centroid of gusset plate to extreme fibre
‘r’ is the radius of gyration of the effective Whitmore section of the gusset plate

No guidance is given in AASHTO MBE on how to consider asymmetric corrosion effects when this limit is exceeded. Therefore, the OE developed a methodology that involved treating the gusset plate as a beam-column, with a bending moment equal to the axial compression in the gusset plate multiplied by the eccentricity due to asymmetric corrosion. The axial forces and bending moments were then combined using the provisions of CSA S6:19. This methodology is illustrated in Figure 4.

Due to the uncertainty over the remaining service life of the bridge, including the option of a full bridge replacement, it was desirable to keep gusset strengthening and recoating works to a minimum. Furthermore, the code-based evaluation involved assumptions on the effects of asymmetric corrosion. For these reasons, the bridge owner requested a detailed finite element analysis be conducted on one truss node as a means of validation of the code-based evaluation, and to give increased confidence in the extents of strengthening and recoating works proposed.
A 3D model of one selected truss node that was subjected to ultrasonic thickness testing was developed. An isometric view of the model, including the finite element mesh, is given in Figure 5. The section loss due to corrosion, based on the ultrasonic thickness measurements taken during the gusset plate inspections, was explicitly included in the 3D geometry of the gusset plate elements. The measured plate thickness at each grid point was mapped onto the gusset plate element surfaces.

Figure 6 shows an example isometric view of a Von Mises stress plot of the entire node for the linear stress analysis under one selected load case. Critical stresses occur in the gusset plates under the compression diagonal (diagonal on the right side of the plot), as expected based on the code-based gusset plate load rating. The highest stresses of all the four gusset plate faces occur on the exterior face of the inner gusset. This pattern holds for all load cases considered and is consistent with what would be expected based on section loss measurements and combined bending and axial stresses due to asymmetric corrosion on either side of the gusset faces.

At a level of load corresponding to a demand-to-capacity ratio of 1.0 from the code-based evaluation, the finite element analysis generally showed spreading of gusset yielding under the critical compression diagonal, but no loss of load carrying capacity. Ultimate loss of load carrying capacity in the non-linear finite element model occurred at a higher load level as a result of an inelastic sidesway buckling failure mode of the gusset plate. These results indicated that the code-based assessment methodology was reasonable albeit somewhat conservative.

A critical gusset node was identified as being deficient for certain heavy vehicle loads per the previously described evaluation process. The gusset plates on this node exhibited up to 50% section loss at the interface with the bottom chord, with the averaged section loss along the shear planes and Whitmore compression plane described previously being on the order of 35%. To address the deficiency, the OE developed a design for a strengthening scheme involving the installation of doubler plates on the inside faces of the corroded gussets.

The doubler plates were shaped to fit around the vertical and diagonal members framing in. The lower half of the doubler plates were bolted to the lower (uncorroded) part of the gusset plate and bottom chord web through existing bolt holes. New bolt holes were field drilled to connect the upper half of the doubler plates to the upper (corroded) part of the gusset plate. Refer to Figure 7 for an overview of this strengthening scheme. This process was repeated for all four quadrants of the node.

Since the bridge is an essential crossing for local residents and industry and available detour routes are long and, in some cases, not suitable for heavy truck traffic, the bridge had to remain open to single-lane alternating traffic at nights and fully open to traffic during days for the duration of the gusset strengthening work. Works were completed at night and the node was checked to ensure adequate load carrying capacity for the applicable traffic loads at all stages of construction.

To maintain load capacity during installation of the doubler plates, the existing bolts between gusset plate and bottom chord web were replaced with tight-fit drift pins one-by-one (Fig. 8). The drift pins acted to transfer shear in bearing, with each drift pin having a shear capacity that exceeded the existing rivets/bolts they replaced. The corroded region was then filled with an epoxy-based composite material for metal repair, to fill the void that would otherwise have existed between the existing gusset and doubler plate due to the section loss (Fig. 9). A high-strength epoxy material was chosen in order to provide a flush solid surface that would not crush under the application of bolt tensioning loads between the existing gusset and new doubler plate. The doubler plate was then put into position by sliding it over the drift pins (Fig. 10). The drift pins were then replaced one-by-one with new bolts, and the process was completed by the field drilling of new holes and installing bolts from the upper existing gusset to the doubler plate. Shear load transfer between the existing gusset plate and the chord web was maintained throughout the duration of the works. The tight fit of the drift pins prevented them from shifting under vibrations due to the live traffic on the bridge during the strengthening works.

Many other gusset nodes that were identified as having significant section loss but not assessed as being deficient for heavy truck loads are currently in the process of being recoated on a staged basis. Future corrosion rates were approximated based on the measured section loss and estimated number of years in which corrosion has been occurring. These rates were used to forecast which gussets may potentially become deficient over a 10-year period of ongoing corrosion. Gussets not meeting this criteria were chosen for inclusion in the recoating program.

Corrosion along the interface between gusset plates and truss bottom chords is a common occurrence on older steel truss bridges. The inspection, evaluation, and strengthening procedures described here may have relevance for owners and consultants working on other truss bridges exhibiting similar defects.

The analysis performed for the capacity assessment of the gusset plates demonstrated that the AASHTO MBE approach produced reasonable results in this particular case. However, verification of the AASHTO methodology using the FE analysis was deemed necessary due to the eccentricity of the corrosion profile of the gusset plates which resulted in significant eccentric out-of-plane loading on the gussets at critical locations. The dual-method approach of code-based evaluation and FE verification increased confidence in the ability to accurately identify the severity of localized corrosion that warranted gusset strengthening and/or short-term recoating. This approach proved to be valuable to accurately determine which of the Taylor Bridge gusset plates required strengthening to maintain functionality of this vital bridge in northern British Columbia. The strengthening method used enabled the works to be carried out at night during single-lane alternating traffic, without significantly impacting the traveling public. ■

About the Author

Dusan Radojevic has 30 years of structural engineering experience working on complex infrastructure projects, including 25 years working on long-span bridge structures. Radojevic holds a PhD in structural engineering from the University of Belgrade and currently serves as the Bridge Sector Manager for Canada at TYLin.

Kai Marder has over 10 years of experience in structural engineering, with wide-ranging project experience from conventional girder bridges to long-span suspension and cable-stayed bridges. Marder holds a PhD in structural engineering from the University of Auckland and is currently a Lead Bridge Engineer with TYLin’s Vancouver, Canada office.