Growth of container cranes has paralleled the growth of container traffic in order to serve the ever-larger ships. Today's typical new crane is three times the size of early cranes. With the increased mass and size comes the potential for increased seismic vulnerability. In order to further understand this seismic vulnerability, detailed time-history analysis is to be performed on both analytical and experimental models. Results will lead to recommendations for retrofitting current cranes, as well as performance-based design recommendations for future container cranes.

A two-dimensional analytical, container crane model has been created in OPENSees, as well as a 1:40 scale physical shake-table model. These preliminary tools are helping to refine an analytical model based on a 2001 Port of Oakland crane, which will be expanded to three dimensions. This will be used to identify key response quantities that will be targeted for measurement during a shake-table test of a large-scale physical model at NEES@Buffalo planned for later in the project.

The project team is also investigating ways to make existing container cranes less vulnerable to earthquake loading through retrofit techniques. During the past year, a literature search has been performed to identify current techniques used for mitigating damage. Finite element models are being created in OPENSees of existing container cranes with and without potential retrofit techniques to assess their performance during an earthquake event. These models will be validated via condition monitoring of in situ cranes and the shake-table test of a large-scale physical model planned for later in the project.

Initial analytical results indicate that, as expected, allowing column uplift greatly reduces the critical moments in the lower legs and the portal beam of the crane. However, unless special care is taken during their design, the portal frame legs could easily fail prior to column uplift, leading to collapse. This response is different than older cranes, in which rocking is likely to occur well before yielding of the columns. Currently, work is being done to model the column yielding within the analytical model.

Return to Research Activities