CTICM Report: Understanding the Differences in Connection Results

Steel

Dit artikel is ook beschikbaar in:

The CTICM (Centre Technique Industriel de la Construction Métallique) carried out an independent evaluation comparing the CBFEM approach in IDEA StatiCa with the Component Method defined in Eurocode 3. Explore how the models in IDEA StatiCa were created and why differences between CBFEM and the Component Method occur.

Component method

In the traditional Component Method, a steel joint is decomposed into individual mechanical components, each representing a specific mode of deformation (e.g. bolt in tension, end plate bending, column flange bending, column web in tension or compression).

Each component is assigned a closed‑form analytical stiffness and a simple force–deformation relationship. The global joint behavior is then obtained by assembling these components.

Component-based finite element method

CBFEM (Component-based finite element method) combines the accuracy of finite element analysis with the clarity and code‑based philosophy of the Component method.
The idea is simple:

Plates and welds → modeled as shell finite elements
Bolts, anchors, and similar fasteners → kept as Component‑Method‑style components

CTICM’s independent assessment

CTICM, the French technical institute for steel construction, carried out verification of IDEA StatiCa by comparing CBFEM results with the Eurocode Component Method applied in Platinex software, developed by CTICM. Their study examined several typical steel connections, focusing on joint resistance, bolt forces, and overall structural behavior. By evaluating all approaches side‑by‑side, CTICM was able to identify where differences originate and confirm the reliability and consistency of IDEA StatiCa’s results.

Bolted plate to plate connection

For the beam‑to‑beam connection under tension, the upper and lower bolt rows show almost identical forces in both approaches, with differences below 1%.

The middle row, however, differs by approximately 2–5%. In IDEA StatiCa, the bolt forces are not equally spread because plastic strains develop in the plate areas around the bolts, which slightly reduces the force transferred through the middle row.

When comparing moment‑loaded connections, the resistance of the bolt row furthest from the compression zone (the tension side) is very similar in both methods, and both predict the same type of failure in the tension bolt. Differences appear in the remaining bolt rows, where the forces redistribute differently.

This behaviour is explained by the distinction between a 3D nonlinear model in IDEA StatiCa and the 2D idealized approach used in the Component Method. The CBFEM model captures realistic plate deformation and stress spreading across the joint, which naturally leads to a different force distribution among the bolt rows.

Angle bolted to the gusset plate

For the angle‑to‑gusset connection in tension, both methods predict almost identical resistance and the same failure mechanism. IDEA StatiCa identifies plastic deformation in the angle leg exceeding the 5% strain limit, while the Component Method indicates failure in the angle. The resistance calculated by IDEA StatiCa is slightly lower (6%).

Bolted angle shear connection

For most shear connection cases, the results from IDEA StatiCa and the Component Method are very close, typically within 3%. The only notable exception is the configuration 3 where a beam is connected to the web of a continuous column. In this situation, IDEA StatiCa predicts a slightly higher resistance than the Eurocode method.

A key aspect of shear connections in IDEA StatiCa is the appearance of unexpected tension forces. Although only shear is applied, the joint deforms in a way that introduces a small bending rotation, as visible in the deformed‑shape model. This deformation shifts the internal force equilibrium and generates tension in some bolts, since bolts are defined as nonlinear springs active in tension and shear.

It is also important to note that the predicted forces are sensitive to the definition of the position of zero bending moment. Small differences in this reference point can lead to noticeable changes in the resulting bolt forces.

Conclusion

The CTICM verification confirmed that IDEA StatiCa’s results are consistently close to Platinex calculations, with deviations typically within ±15%. In most cases, IDEA StatiCa provides results that are slightly more conservative, which is beneficial from a safety perspective. The remaining differences are fully explained by the contrast between 3D nonlinear finite‑element modelling in IDEA StatiCa and the analytical simplifications used in the Eurocode Component Method. Overall, the evaluation demonstrated that IDEA StatiCa aligns well with French engineering practice and delivers reliable, transparent structural assessments.