Ship-to-Shore (STS) cranes are critical assets in container terminals, operating under continuous cyclic loading which make their steel structures susceptible to fatigue. Since these assets are expensive and important pieces of equipment it is desireable to understand their struc
...
Ship-to-Shore (STS) cranes are critical assets in container terminals, operating under continuous cyclic loading which make their steel structures susceptible to fatigue. Since these assets are expensive and important pieces of equipment it is desireable to understand their structural behaviour. A common approach is to model the crane using a finite element beam model. This beam model is then used in combination with fatigue load assumptions and safety factors to determine the structural response and fatigue behaviour of the STS-crane. In this research the following question is answered:
“What is the performance of a finite element beam model in assesing the structural response and fatigue behaviour of a STS-crane”
This study investigates the performance of such a finite element beam model. A strain gauge monitoring campaign of 30 days was performed on the crane’s upper frame to capture stresses during operational loading allowing for a comparison between measured data and model predictions.
The results demonstrate that the finite element beam model effectively reproduces the STS crane’s global quasi-static behaviour, accurately representing stress distributions and amplitudes for most influence lines. This validates its use for static verification, global stiffness evaluation, and general design assessments. In the best cases, the finite element beam model achieves an accuracy of 94%, with differences between model and measurements showing stress amplitudes within 1–2 MPa. In contrast, the largest discrepancies, observed around the A-frame pipes and support pipe, reach 9.5–12 MPa, highlighting that local effects and stiffness assumptions are not fully represented throughout the structure. In these areas, the model deviates with 26% in capturing the stress range of the influence lines.
Two shell models were also investigated, one developed as a part of this study and another provided by Arup, the shell model results further emphasized the limitations of the beam model in accurately assessing individual stress contributions. Employing a shell model improved the overall accuracy in capturing the total stress range by approximately 6.8%.
The study shows that in general, the model’s fatigue predictions combined with the fatigue load spectrum tend to either under- or overestimate the accumulated damage and stress ranges. This is partly because the finite element beam model does not accurately capture the crane’s structural behavior everywhere and partly because the fatigue load spectrum overestimates moves beyond the landside legs. Furthermore, stress deviations above 40 MPa are not accounted for with the defined fatigue load.
Quantitatively, the comparison between measured damage and modelling damage in combination with the fatigue load spectrum shows an average overestimation of damage with a factor of 2.094, mainly due to large discrepancies for structural elements located beyond the landside legs. However, for 17 of the 29 strain gauges, the model combined with the fatigue load spectrum underestimated the damage by an average factor of 0.42. The maximum overestimation of damage is a factor 8.756 at back girder location (GBR-2), while the maximum underestimation of damage is a factor of 0.162 at the short forestay (FSR-2). The smallest difference between modelelled damage and measurments was recorded for the boom front right location (BFR-2) where the model was only 4% off in comparison to the measurments.
The contribution of dynamic effects on the fatigue behaviour of the crane was also investigated. Dynamic effects were isolated from the signal with a filtering technique. Overall, the impact of dynamic effects on the fatigue damage is on average 51%. Although this is contribution is large, the impact of the dynamic amplification factor for hoisting defined in the Crane Design Code EN 1993-2-1 decreases this contribution to 26%. This significant decrease is explained by the exponential between increasing stress ranges and subsequent fatigue damage on the material.
The impact of snagging on the fatigue damage of the structure was also investigated. Results showed that snagging had a negligible effect on the overall fatigue damage accumulated during the monitoring period, contributing at most 1.8% to the total damage registered. The average snag load ratio was 1.85 meaning the stress induced by a snag event is 185% higher than the stress induced by lifting the heaviest allowable container. Although the available dataset of snag events was limited, it was used to perform a stochastic extrapolation to assess probabilities over the design life of the structure. The probability of a snag load ratio of 5 or higher was determined to be 1.048% for a single crane. The study shows that it is important to consider the amount of STS-cranes in a portfolio of a company making this probability quite significant for an Ultimate Limit State (ULS) event.
Overall, the findings of this reseach confirm that while the finite element beam model remains a reliable tool for global response prediction, it deviates a lot in accurately assessing fatigue behaviour of the structure. The study recommends a hybrid modelling approach that combines global beam analysis with locally detailed finite element submodels and operational data as input for the fatigue load spectrum. Such refinement would improve fatigue damage predictions and support more effective maintenance and design strategies, improving the long-term reliability and safety of STS-cranes.
