TIGHT MOVEMENT, vibration and temperature tolerances called for accurage foundation design at the Rutherford Appleton Laboratory site in Oxfordshire. Precision in the order of microns was needed because of strict limits on deflection and differential settlement, together with vertical and horizontal dynamic vibrations at the Harwell job.
Groundwork was finished last month on the 235m diameter toroid shaped building designed by consultant Jacobs It will house Diamond Light Source's state-of-the-art electron accelerator which will be used for a range of scientific research.
Ground conditions consist of between 2m and 7m of cohesive superficial deposits overlying structureless and then competent chalk.
As well as site investigations and laboratory testing, analyses, including dynamic and impact load and background vibrations, were carried out.Ground-related pressures induced by swelling of the superficial deposits, temperature variation and changes in effective stress in the lower strata were also examined.
Jacobs considered a number of foundation options, ranging from a ground bearing slab through to a piled slab solution.
Options allowed for variation of the number and spacing of piles and slab thickness, inclusion of a void between slab and ground and the insertion of slab joints.
Engineers assessed the static performance of the raft and piled raft options with a series of elastic finite element (FE) models. They calibrated subgrade reactions and single pile response against plate load and pile tests respectively.
Pile tests comprised compression and tension loading on sleeved and unsleeved (11m to 15m long) piles with periods of constant loading to determine creep response.
The dynamic response of the 3D single pile FE model was calibrated against data from full scale dynamic excitation tests.
These were carried out by exciting the top of each test pile over a frequency range of 5Hz to 100Hz.
Results from the single pile calibration were used to assess the dynamic response of the foundation options, represented as a radial slice through the building. The response of the model to forces and accelerations was assessed within the small strain soil stiffness range.
The final foundation design was a balance between optimal static and dynamic performance. Static performance was obtained by sleeving the top of the piles and by creating a void between the soil and the underside of the slab. Dynamic performance was obtained from contact between the underside of slab and the ground, a thicker ground slab and an increase in the number of piles.
The facility is supported by 1,500, 12m to 15m long, 600mm diameter CFA piles at about 3m centres. A void former was placed between the soil and the underside of slabs to prevent swelling imposing any pressure on the slabs. Should long term monitoring show soil movements are insignificant then grouting up this void could improve the dynamic response of the floor.