Fatigue Assessment of Rectangular Concrete-Filled Steel Tube Composite Truss Bridge Joints Using Hot Spot Stress Method
Literature Overview and Core Findings
This paper by Jiang Lei, Liu Yongjian and colleagues from Chang'an University addresses a question I have wrestled with on more than one delivery site: how do you actually prove the fatigue life of a welded tubular joint when the geometry is rectangular rather than round? The authors introduce the hot spot stress method as an alternative to traditional nominal stress approaches, and they do it with a level of rigour that deserves attention from anyone fabricating rectangular hollow sections (RHS) for bridge or offshore structures.
The study is built on three pillars. First, they establish that the hot spot stress at the weld toe can be calculated using three different modelling approaches: a planar frame model, a spatial frame model, and a three-dimensional solid element model. Second, they regress fatigue test data from 52 joint specimens to fit a hot spot stress amplitude versus number of cycles (S-N) curve. Third, they apply this framework to an actual bridge on the Huangyan Expressway in Shaanxi Province, a rectangular concrete-filled steel tube (CFST) composite truss bridge, and then propose design optimisations for the fatigue-vulnerable joints.
The key finding that jumps out at me is that the mid-span rectangular steel tube joint had a hot spot stress amplitude of 60.1 MPa, which occurs on the surface of the main chord. This is below the 71 MPa allowable fatigue strength given in the Eurocode, so the original design technically satisfies fatigue requirements. But 60.1 MPa against a 71 MPa allowable is not a comfortable margin when you consider that welding residual stress, geometric imperfections and site handling can all eat into that reserve. This is exactly the kind of situation where I would tell a client: you are within code, but you are not within comfort.
Interpretation of Technical Points and Process Analysis
Hot Spot Stress Calculation Methods
The paper compares three modelling approaches, and the results are instructive. The spatial frame model gave maximum hot spot stress amplitudes of 58.9 MPa on the branch surface and 54.1 MPa on the chord surface for the optimised mid-span joint. The three-dimensional solid element model gave 45.2 MPa and 47.1 MPa respectively. The spatial frame model is conservative, but it cannot accurately capture the fatigue effect at different hot spot locations, and it cannot reliably identify where fatigue cracking will initiate.
From a fabrication perspective, this matters enormously. When I sign off on a welded joint, I need to know where the weld toe is, what the stress concentration factor is, and what the weld profile actually looks like. A solid element model can give you that detail. A frame model gives you a number that is safe but blind to the actual geometry. For critical joints on a bridge that will carry millions of load cycles, I want the detail, not just the safety factor.
Design Optimisation and Weld Post-Processing
The paper proposes two optimisations to the original design. The first converts the mid-span rectangular steel tube joint into a rectangular concrete-filled steel tube joint. The concrete fill changes the local stiffness of the joint, distributes the stress more evenly around the weld toe, and reduces the hot spot stress amplitude on both branch and chord surfaces by an average of 25.1%. The second optimisation applies weld post-processing to eliminate initial tensile residual stresses, which reduces the hot spot stress amplitudes by an average of 14.9%.
These are both techniques I have seen work in practice, but they come with different cost and schedule implications. Concrete filling is a structural change that affects the whole joint design and requires careful coordination between steel fabricator and concrete contractor. Weld post-processing, such as hammer peening or ultrasonic impact treatment (UIT), is a shop floor operation that adds time to each joint but does not change the overall structural form. On a large truss bridge with dozens of joints, I would be asking the designer to tell me which joints are actually fatigue-critical before I commit to post-processing all of them. The paper gives me the analytical tools to make that call.
Comparison with Eurocode Allowable Fatigue Strength
| Parameter | Value | Code Allowable | Verdict |
|---|---|---|---|
| Hot spot stress amplitude at mid-span joint (chord surface) | 60.1 MPa | 71 MPa (Eurocode) | Satisfies but with low margin |
| Hot spot stress amplitude after concrete filling (average reduction) | 25.1% lower | 71 MPa | Comfortable margin |
| Hot spot stress amplitude after weld post-processing (average reduction) | 14.9% lower | 71 MPa | Improved fatigue performance |
The message here is clear: the original design was adequate but not robust. The optimised designs are robust. In my experience, when you are shipping fabricated steel to a remote bridge site, you do not want to be in the position of explaining to the client why a joint that is within code has cracked after five years of service. The code is a minimum, not a target.
Engineering Practice Integration and Key Reflections
The Weld Toe Is Where the Story Is
The hot spot stress method focuses attention on the weld toe, and rightly so. I have seen more fatigue failures start at weld toes than anywhere else, and the cause is almost always a combination of high stress concentration, tensile residual stress, and an unacceptable weld profile. The paper's finding that weld post-processing reduces hot spot stress amplitude by 14.9% is consistent with what I have observed on real projects. Hammer peening and UIT both introduce compressive residual stress at the weld toe, which retards crack initiation and slows crack growth. But the effectiveness depends entirely on the weld being sound in the first place. You cannot peen your way out of lack of fusion or undercut.
The Concrete-Filled Joint Is a Different Beast
When you convert a hollow rectangular joint into a concrete-filled joint, you change the local stiffness, and the stress distribution at the weld toe becomes more uniform. This is a real benefit, but it also changes the welding sequence and the inspection requirements. The concrete must be placed after the steelwork is welded and inspected, and the concrete placement itself must be designed to avoid voids at the interface. I have seen cases where concrete shrinkage has created gaps at the steel-concrete interface, which then become initiation sites for corrosion and fatigue. The paper does not address this directly, but it is a practical concern that any fabricator must manage.
Lessons for Fabrication and Quality Control
The paper's regression analysis of 52 joint specimens is a valuable contribution, but it also highlights the scatter that exists in fatigue test data. Weld quality, residual stress, geometric tolerance and loading history all contribute to that scatter. When I am ordering steel for a fatigue-critical application, I specify tighter tolerances on the RHS dimensions, stricter weld acceptance criteria, and full NDT coverage. The cost is real, but the cost of a fatigue crack in a bridge is far greater.
The paper also makes a practical point about the difference between the spatial frame model and the solid element model. The frame model is conservative, but it cannot tell you where the crack will start. For a designer, that is acceptable. For a fabricator who needs to know which weld toe to post-process and which to leave alone, it is not. I would always ask for the solid element model on a fatigue-critical joint, even if it costs more analysis time. The analysis cost is trivial compared to the cost of a failed weld.
The 5W2H Perspective
Applying the 5W2H framework to this paper, I can summarise the key takeaways as follows:
- What: Hot spot stress method for fatigue assessment of rectangular CFST truss bridge joints.
- Why: To accurately evaluate fatigue performance where nominal stress methods are inadequate.
- Where: Mid-span joints of a rectangular CFST composite truss bridge on the Huangyan Expressway.
- When: Fatigue assessment is required at design stage, and re-assessment is needed after any design change.
- Who: Designers, fabricators, and quality control engineers all have a role.
- How: Hot spot stress calculation via three modelling approaches, S-N curve regression, and design optimisation.
- How much: 25.1% reduction from concrete filling, 14.9% reduction from weld post-processing.
Study Insights and Implications for the Steel Pipe Industry
This paper is a reminder that the fatigue performance of welded tubular structures is not just a function of the steel grade and wall thickness. It is a function of joint geometry, weld quality, residual stress state, and the interaction between steel and concrete. For a steel pipe fabricator, this means that the shop floor has a direct influence on the fatigue life of the structure, even if the designer takes all the credit.
The paper also reinforces a lesson I have learned the hard way: the code allowable is a floor, not a ceiling. When you are at 60.1 MPa against a 71 MPa allowable, you are within code, but you are one bad weld away from a problem. The optimisations proposed in the paper — concrete filling and weld post-processing — are both practical and effective, and I would recommend them for any fatigue-critical joint where the hot spot stress amplitude exceeds 80% of the allowable value.
For the steel pipe industry, the hot spot stress method is a tool that we should be using more often. It gives us a direct link between weld quality and fatigue life, and it allows us to make rational decisions about weld post-processing, NDT coverage, and tolerance control. The paper's contribution is to show that the method works for rectangular sections, which are increasingly common in bridge and building structures.
In my view, the most important takeaway is this: fatigue assessment is not a one-time calculation. It is a process that starts at the design stage, continues through fabrication and welding, and extends into service life monitoring. The paper gives us the analytical framework, but the practical outcome depends on the fabricator's commitment to weld quality and the owner's commitment to inspection and maintenance. If we get all three right, the bridge will outlive its design life. If we get one wrong, we will be back on site with a crack and a very expensive repair bill.
The reference value of this paper lies in its combination of theoretical rigour and practical application. It is not just an academic exercise; it is a guide for engineers who need to make real decisions about joint design, weld quality and fatigue life. I would recommend it to anyone working on CFST truss bridges, and I would particularly recommend the design optimisation strategies to fabricators who are looking for ways to improve fatigue performance without increasing material cost. The 25.1% reduction from concrete filling and the 14.9% reduction from weld post-processing are significant numbers, and they are achievable with existing technology and standard shop floor practices.
Zhuojin Pipe Fitting Co., Ltd