Congratulations to Prof Fu Yuguang on the award of MOE Tier 1 Seed Funding award

Heading towards LTMP 2040, the LTA plans to extend the rail network to 400 kilometres while maintain the existing MRT lines dating back from 1987 for more convenient and safer journeys. This calls for novel techniques to preserve the efficiency and integrity of tunnels for rail networks. The monitoring and management of tunnels are challenging, mainly because of their complex environment conditions and limited access time for inspection once in service. Conventional approaches for tunnel inspection are either direct measurements (e.g., liner deformation) or surface faults detection using computer vision (e.g., seepage), with limited capability of fault diagnosis/prognosis for tunnel management. As an emerging concept, the digital twin offers great opportunities to enable more systematic and in-depth evaluation and management of tunnels. However, research gaps exist in digital modelling for wide applications of this technique. 

This study will provide an effective digital-twin solution to enable faults diagnosis, performance prediction, and informed decision-making in response to tunnel network faults/events. To achieve this, we will establish a cyber-physical loop for digital twin by linking the simulation environment and physical structures through information and interventions. We propose two key techniques: (i) heterogenous modelling of integrated underground transportation network to capture emergent and complex behavior of tunnels; and (ii) hierarchy data fusion techniques for adaptive and robust models over-the-time to achieve accurate and informative representation of physical structures. Both simulation data and site monitoring data will be used to demonstrate the capability of the proposed framework. The proposed solution will help the LTA to increase inspection accuracy, reduce traffic disruptions, and offer the potential to make the Singapore’s transportation infrastructure safer.