Karam Mahmoud, Ehab El-Salakawy
6 min
This study investigates the shear behavior of continuous concrete beams reinforced with glass fiber–reinforced polymer (GFRP) bars and stirrups. While the use of FRP as an alternative to steel in simple structures is well-documented, its performance in statically indeterminate (continuous) structures—where moment redistribution occurs—remains less understood. The authors tested six large-scale, two-span continuous beams, varying concrete strength (normal vs. high strength), longitudinal reinforcement ratios, and the presence of minimum transverse shear reinforcement. The experimental results were then compared against predictions from three major design codes: CSA-S806-12, CSA-S6-06, and ACI 440.1R-06.
All test specimens failed in shear near the interior support following significant moment redistribution. A key observation was that the average angle of inclination of the diagonal shear cracks was 53 degrees, which is notably steeper than the 44-degree angle typically observed in simply supported beams. This steeper cracking reduces the effectiveness of the stirrups in carrying shear loads.
Regarding design codes, the study found that all three evaluated provisions (CSA-S806-12, CSA-S6-06, and ACI 440.1R-06) were conservative, consistently underestimating the shear strength of the continuous beams. The CSA-S806-12 provided the most accurate predictions among the three, though it still underestimated the contribution of the shear reinforcement while overestimating the concrete's contribution. Furthermore, the study noted that increasing the longitudinal reinforcement ratio did not consistently increase shear strength in these continuous beams, a finding that contradicts established behavior in simply supported members.
Continuous beams are fundamental components of infrastructure like parking garages and bridges. As engineers increasingly turn to GFRP to avoid the corrosion issues associated with steel, it is critical to ensure that design codes accurately reflect the behavior of continuous systems. This research highlights that simply applying design provisions derived from simply supported beam tests to continuous structures may lead to inefficient or potentially unsafe designs, emphasizing the need for code updates that account for the unique effects of continuity and moment redistribution.
This paper investigates the shear behavior of continuous concrete beams reinforced with glass fiber–reinforced polymer (GFRP) bars and stirrups. A total of six beams, with 200×300 mm rectangular cross section and continuous over two spans of 2,800 mm each, were constructed and tested. Three main variables, namely, concrete strength, longitudinal reinforcement ratio, and transverse shear-reinforcement ratio, were investigated. Two beams were without transverse reinforcement to evaluate the concrete contribution to the shear strength and four beams had GFRP stirrups to examine the minimum shear-reinforcement contribution to the shear strength. All test specimens were designed to satisfy an assumed 20% moment redistribution. It was observed that all test specimens failed in shear near the interior support after significant moment redistribution from hogging moment to sagging moment regions. In addition, the test results were compared with the predictions of the CSA-S806-12, CSA-S6-06, and ACI 440.1R-06.
Alex: [curious, leaning in] So, if the structural mechanics of a continuous beam are so different from a simply supported one, why do our standard design codes often treat them as interchangeable?
Sam: [measured, teaching mode] It is a legacy of steel-reinforced design. In steel beams, the material is ductile enough to accommodate redistribution without catastrophic failure. When you switch to glass fiber–reinforced polymer, or GFRP, you lose that ductility. The beam behaves elastically until the very end, and the moment redistribution becomes a liability rather than a safety buffer.
Alex: [processing, analytical] That explains why the crack angles shift so drastically. If the beam is effectively behaving like a series of simply supported segments, but the internal moments are constantly shifting, the shear plane is essentially chasing a moving target.
Sam: [nodding in voice, precise] Exactly. In these continuous tests, we observed average crack angles of 53 degrees, compared to the 44 degrees typical of simply supported beams. Because the stirrups are vertical, a steeper crack means they intersect fewer reinforcement bars. <break time="0.6s" /> The stirrups that do get hit are often caught at their weak, bent portions rather than their straight, robust midsections, which leads to premature rupture.
Alex: [thoughtful, slightly skeptical] And this is where the paradox of longitudinal reinforcement comes in, right? You would think adding more reinforcement would always increase capacity, but here it seems to do the opposite.
Sam: [measured, grounding the point] It does. By increasing the longitudinal reinforcement, you stiffen the beam, which forces more aggressive moment redistribution. This redistribution drives the shear failure plane to an even steeper angle, which further reduces the efficiency of your transverse stirrups. You are essentially trading flexural stiffness for shear vulnerability.
Alex: [deliberate, checking understanding] So, the current code provisions are essentially blind to this interaction. They assume a predictable crack path that simply doesn't exist in continuous, high-strength concrete systems. [[RP_SECTION:limitations-of-design-codes|Limitations of Design Codes]]
Sam: [quiet confidence, precise] That is the core of the problem. The codes are calibrated for steel and simply supported conditions. In continuous systems, the interaction between concrete strength and moment redistribution creates a non-linear failure mode that current equations consistently underestimate. For an engineer, this means the safety margins they think they have are largely illusory.
Alex: [curious, leaning in] If the structural mechanics of a continuous beam differ so much from a simply supported one, why do standard design codes treat them as interchangeable?
Sam: [measured, teaching mode] It’s a legacy of steel design. Steel is ductile enough to redistribute stress. With glass fiber–reinforced polymer, or GFRP, you lose that ductility. The beam is elastic until the end, so moment redistribution becomes a liability.
Alex: [processing, analytical] So the shear plane is essentially chasing a moving target?
Sam: [nodding in voice, precise] Exactly. In continuous tests, we saw crack angles of 53 degrees, versus 44 degrees in simply supported beams. Because stirrups are vertical, a steeper crack means they intersect fewer reinforcement bars. <break time="0.6s" /> The stirrups that do get hit are often caught at their weak, bent portions, leading to premature rupture.
Alex: [thoughtful, slightly skeptical] And this is where the paradox of longitudinal reinforcement comes in—adding more reinforcement actually reduces capacity?
Sam: [measured, grounding the point] It does. Increasing longitudinal reinforcement stiffens the beam, forcing more aggressive moment redistribution. This drives the shear failure plane to a steeper angle, further reducing stirrup efficiency. You’re trading flexural stiffness for shear vulnerability.
Alex: [deliberate, checking understanding] So the current code provisions are blind to this. They assume a predictable crack path that doesn't exist in continuous systems.
Sam: [quiet confidence, precise] That’s the core problem. The codes are calibrated for steel and simply supported conditions. In continuous systems, the interaction creates a non-linear failure mode that current equations consistently underestimate. Engineers’ safety margins are largely illusory.
Alex: [reflective, slower pace] This is much more complex to model than the standard baselines.