Structural connectivity of two - dimensional assemblies

Abstract

The characterization of structures based only on their geometrical configuration and independent of loading is a novel approach for evaluating and designing structures to be robust against accidental damage. The concept of ‘structural connectivity’ is introduced to assess the connectivity of the structure at multiple hierarchical levels. An adaptation of the Bristol approach is tentatively suggested as providing the most appropriate measure for structural connectivity. Three other measures, conventional connectivity indices in Graph theory, Newman’s approach based on Network theory and Route structure analysis (originally developed to analyse road networks) are used to compare the results obtained from the Bristol approach. Three trusses of the same outer shape but differing in geometric configuration were analysed using all four methods to find the best connected truss. The configurations analysed are Fractal, Warren and Fan-type trusses. Axial rigidity of the chord members were increased to check its effects on structural connectivity. The different measures gave different results for the same structure, though there is some degree of consistency. Graph theory and Unweighted Newman’s approach suggest that the Fractal truss has the most well-connected configuration, whereas the Bristol approach favours the Fan-Type type truss. Weighted Newman analysis and Route structure analysis indicate that Warren truss to be the most wellformed configuration. All three methods indicate that truss ends and central regions of chord members are the least connected areas. All weighted analysis methods show that increasing the chord member stiffness benefits structural connectivity of all truss forms. Separately, a frame (4 bays x 5 floors) with different column elements removed was also analysed, in order to determine the column removal that would result in the least degree of frame connectivity. Though different methods indicated different column removals to cause the highest loss in structural connectivity, all methods agree that the middle column removals causes higher loss in connectivity than side column removal in the corresponding floor. An idealised “A-Level” road network in Sri Lanka was analysed as proof that concept of structural connectivity can be applied to assemblies other than structures.