Modelling

Six interconnecting sets of jump formed walls form the central core of the structure. The walls are predominantly 200mm thick, with a few of 300mm and 400mm thicknesses. In modelling the core using LUSAS Civil & Structural, 3D thick shell elements represented the walls. Joint elements were used at each floor level of the core to apply notional horizontal loadings and also used to specify restraint conditions for the global horizontal x and y directions in order to 'lock' the structural deformations at each floor level.  These special support conditions required the removal of constraint in the vertical z-direction for the rigid supports so that the vertical deformations across the floor plan could vary.  Elements in each 'lift' of core wall were grouped together to allow the display of each core level in isolation to help simplify the model building process and the viewing of results. In all, over 40,000 thick shell and joint elements were used to model the core. In addition to the global model of the whole core, local models using solid elements investigated potential high stress concentrations around selected openings including service holes and single and double jacking pockets in the 200mm walls.

Loading

Five separate loadcases consisting of core self weight, uniformly distributed, dead and live loads and a notional load of a specified value acting in North, South, East and West directions were applied to the model at each floor level. By combining and factoring these loadcases at the results processing stage nine design load combinations were formed. From these combinations resultant displacements, in-plane shear stress resultants and direct membrane stress resultants were obtained for each level of the core to aid Clarke Nicholls Marcell in their design.

Results

Because of the complexity of the model generated for this project, with its 9 load combinations, 22 separate floors and large number of walls the output of actual stresses at top, middle and bottom surfaces of the thick shell elements would have generated a large number of results plots requiring investigation. So, in order to simplify the results produced, contour plots of stress resultants were produced instead. With these the actual stress at a location in a wall is obtained by dividing the stress resultant value by the corresponding wall thickness.

For all walls in the core, and under all load combinations, the stress ranges given by the LUSAS Civil & Structural analysis compared well with those predicted by less advanced calculation methods. The direct compressive stress resultants confirmed that the greatest magnitudes were being carried by the thicker wall sections. It was also found that some of the 200mm walls had significantly higher stress levels than others of the same thickness. The ability of LUSAS to produce level by level stress results and highlight disproportionate wall loadings can give useful guidance on the appropriateness of the wall thickness used for these types of core-walled structures.