Low-volume-loss tunnelling for London ring main extension

Abstract

The Thames Water ring main extension is a 4·5 km long tunnel from Stoke Newington, in the London borough of Hackney, to New River Head in Finsbury, in the London borough of Islington. The 2·85 m i.d. tunnel was excavated by an earth pressure balance tunnel-boring machine (TBM) at depths between 40 and 60 m below the surface. Surface settlements along the route were measured by precise levelling, and were found to be small. It was therefore even more important to measure these settlements as accurately as possible, in order to provide informed estimates of subsurface movements induced in third-party underground structures much closer to the tunnel horizon. Because of the relatively large magnitude of the background movements measured, when compared with the small tunnel-induced settlements, it was necessary to adopt a rigorous statistical method to fit a Gaussian curve to the data. This exploited the analogy of the ‘error function' to define the Gaussian curve parameters i and Vl. In all, 13 tunnels were underpassed successfully by the TBM, all within the ‘conservative expected value' predictions, and without incident. The predictions and structural monitoring schemes undertaken for the High Speed 1 tunnels near Corsica Street and the Northern line tunnels near Angel station are described in the paper. It was found that the surface and subsurface trough width parameter K did not vary with depth as predicted: therefore a new relationship is proposed.