Robust retrofitting design for rehabilitation of segmental tunnel linings: Using the example of steel plates

摘要

This paper presents a general framework for the robust retrofitting design for rehabilitation of segmental tunnel linings installed using shield tunnelling, and specifically using steel plates bonded to the lining as a typical example of such a rehabilitation design. A two-dimensional finite element model is established as part of the robust design which can simulate the deformational response of the steel plates reinforced segmental tunnel lining. The surrounding soil, the tunnel lining, the steel plates and the interactions between each of these are all properly simulated in this model and verified by full-scale test results. The change in horizontal convergence (ΔDhs) subjected to environmental impact, such as unexpected placement of ground surface surcharge is measured to reflect the performance of segmental tunnel linings reinforced by steel plates. The standard deviation of the reinforced tunnel performance due to uncertainties in the soil conditions and the ground surface surcharge is derived to measure the design robustness. A robust rehabilitation design is then accomplished by varying the steel plates sizes (i.e. width and thickness) to maximize the design robustness and minimize the cost using a multi-objective algorithm, also considering the safety requirement constraints. The optimal designs are determined as a set of design points, namely a Pareto Front, which presents a trade-off relationship between the design objectives and is demonstrated as being useful for decision making. Finally, the robust rehabilitation design method is applied to the retrofitting design of tunnel lining using steel plates in a real case study, and a comparison between the actual design and the design derived by the proposed method has been made to show its applicability and potentially significant advantages for designers, as the method allows consideration of both the highest robustness and the lowest cost simultaneously.