Long-term settlement prediction of fast constructed artificial island

With the development of the marine traffic engineering in China, it is needed to develop the construction method of artificial island in open sea. An artificial inland of 97962 m2 was built in a cross sea bridge project in southeast sea of China. The seawater depth at low tide at the location is 10 m depth and the seawater level at low tide is considered as standard level. The thickness of 6 m seabed mud (-10 m to -16m from standard level) was replaced by sand first. The steel cylinders were inserted into the impermeable soil layer (-40 m) and then the seawater were pumped. The PVDs was inserted from –8 m up to -33 m and the sand was added up to +4m to improve the soft ground by dewatering preloading. The settlements at three typical sections, east section of the island which has deepest burial depth of tunnel (section 1), central section of the island (section 2) and west section of the island which has thickest soft clay layer (section 3), were measured during stage the soft ground improvement. The 3D finite difference method was used to simulate the construction process of the fast-constructed artificial island. The fluid-mechanical interaction method and equivalent hydraulic conductivity is used to predict the degree of consolidation. The settlement curves predicted by numerical method are compared well with the measured results in the field up to 150 days. The numerical results show that the long-term settlements (2400 days) of sections 1, 2 and 3 reach to 2.74 m, 2.91 m and 2.80 m, respectively. The residual settlements after two years of finishing the tunnel construction (1550 days) are 1 cm, 12 cm and 10 cm, respectively. The settlement rate and final settlement of artificial island calculated by the present method using one-dimensional consolidation theory are smaller than the measurement values in the field and the values predicted by numerical simulation. So the predicted settlement results by present method need to be modified according to field data, and numerical simulation method should be used to predict the long-term settlement.