Partial composite action in prestressed UHPC-NC composite girders: interface characterisation, prestress gain and a conservative assessment of a bond-based check

Bridges are heavy, and the longer they stretch, the heavier they become. This creates a frustrating paradox: to span a great distance, engineers must build deep, massive girders to hold up their own weight, yet that extra depth adds even more weight, forcing the structure to grow larger still. For decades, the solution was to use stronger materials, but ordinary concrete has a fatal flaw: it cracks easily under tension. When concrete cracks, water seeps in, rusts the steel inside, and shortens the bridge's life. Ultra-high performance concrete, or UHPC, changes this game. It is a dense, fiber-reinforced material that resists cracking and can be pulled apart without breaking, allowing engineers to build much shallower, lighter girders. The most efficient design pairs a thin layer of this super-strong UHPC on the bottom, where the tension is highest, with a thicker layer of ordinary concrete on top, where the material is under compression.

The challenge lies in the seam where these two different concretes meet. If they do not stick together perfectly, they will slide past one another under load, and the bridge will fail to act as a single, strong unit. Engineers have long assumed that if they bond the layers correctly, they will move as one solid block, a condition known as full interaction. However, this assumption is difficult to prove in a real bridge, and if it is wrong, the structure could be far weaker than calculated. The question is not just whether the materials stick, but how they behave when the bridge bends, how much the internal steel cables stretch during that bending, and whether the connection is strong enough to handle the stress or if it will slip before the bridge reaches its breaking point.

A researcher set out to test these assumptions directly, moving beyond theory to measure exactly what happens inside the joint. They built twenty-six physical models to represent three different scenarios: simple blocks to test the bond itself, new prestressed girders where the layers were cast together, and old, damaged beams that were strengthened with a new layer of UHPC. The goal was to see if the interface between the super-strong bottom layer and the ordinary top layer could truly hold the bridge together, or if the two materials would begin to slide apart long before the bridge collapsed.

The researcher first examined the nature of the bond itself using L-shaped blocks. They discovered that the way the surface is prepared matters more than the strength of the concrete. When the surface was simply roughened with a wire brush, the joint behaved one way: it held well until it cracked, and then it failed almost immediately with no warning. But when the researcher added small metal pins, or dowels, that crossed the seam, the behavior changed completely. The joint could still crack, but the pins held the pieces together, allowing the bridge to carry more than five times the load it could before the crack appeared. This revealed a critical truth: a joint with pins is not just a stronger version of a rough joint; it is a fundamentally different mechanism that can absorb damage and keep working, whereas a rough joint without pins offers no reserve strength once it cracks.

Next, the researcher tested full-scale prestressed girders to see how the internal steel cables behaved. In a prestressed bridge, steel cables are pulled tight before the bridge carries any traffic, clamping the two concrete layers together. As the bridge is loaded and begins to bend, the steel cables stretch, and their tension increases. The researcher measured this increase directly at the anchor points and found that the cables gained nearly thirteen percent more tension by the time the bridge reached its limit. This confirmed that the cables were stretching significantly, but it also raised a question: was the joint slipping enough to stop this stretching? The answer was no. In every test, the two layers moved together perfectly. The joint held firm, and the steel cables stretched exactly as they would in a solid, monolithic beam.

The study also looked at the more difficult case of strengthening an old, damaged bridge. Here, the bottom layer of the original bridge was already cracked before the new UHPC was added, meaning the two layers were not clamped together by the original tension. Despite this, the strengthened beams performed remarkably well. They did not fail because the layers slid apart. Instead, they failed because the concrete or the steel reached their breaking point, just as a solid beam would. The interface between the old and new concrete held together so effectively that the bridge acted as a single unit, even though the connection was not clamped by prestress.

The researcher then applied a new method of checking the bond, one that looks at how much the joint can slide before it loses its grip, rather than just how strong it is. They found that for every single bridge and beam they tested, the connection was strong enough to be considered perfect. The joint never slipped enough to weaken the structure. In fact, the safety margin was so large that the researcher concluded the joint was never the weak link. The limiting factor was always the strength of the concrete or the steel, not the glue between them. This means that for the types of bridges tested, the assumption that the layers work together as one is correct.

However, the study also highlighted a gap in how engineers currently design these structures. The new checking method showed that the most important factor is not just the strength of the bond, but the stiffness of the connection and how much it can slide. The researcher found that the current design rules, which focus only on bond strength, miss this nuance. While the bridges they tested were safe, the method they developed suggests that for bridges with different shapes or loads, the joint could become the weak point if the design does not account for how much the layers are allowed to slide. The study did not find a bridge that failed because the bond broke, but it did provide a precise way to predict when that might happen in the future.

The results offer a clear picture of the state of the art. The bond between ultra-high performance concrete and ordinary concrete is robust, especially when reinforced with metal pins. The internal steel cables stretch significantly as the bridge bends, but the layers stay locked together, ensuring the bridge acts as a single, efficient unit. The research confirms that for the configurations tested, the interface is not the problem. The real value of the work lies in the new way of looking at the joint: it shifts the focus from a simple measure of strength to a measure of how the joint moves and resists sliding. This provides engineers with a more accurate tool to design bridges that are not only strong but also efficient, ensuring that the super-strong materials are used exactly where they are needed without wasting resources on unnecessary safety margins. The bridge stands firm, not because the glue is magic, but because the mechanics of the connection have been understood and verified.