The objective of this research is to explore the possibilities in the use of digital twins towards the maintenance and operation of tunnels. This is explored by looking at the end users of the technology, to profile their needs and requirements and consider a digital twin design that focuses on this. This research is the start of a human centred design for a digital twin and provides the opportunity for further investigation of applications and discussion of their usefulness. To map out the functional needs of facility managers within the Dutch tunnel sector, ten interviews were compiled. These interviews are based on digital twin literature in other sectors. The discovered needs within the interviews are analysed for patterns and compared to the characteristics of digital twins. This comparison is validated through a focus group with two digital twin experts and is done to consider if digital twins may offer opportunities for the needs mentioned. It is found that for a large majority of the identified functional needs of facility managers requires the mirror like characteristic of a digital twin, when focusing on the needs related to tunnel behaviour. A digital twin that is designed for the identified needs, should focus primarily on decision supporting services for these facility managers, in other words, providing them with relevant data about the tunnel, at the right time and in a comprehensive manner. An indication is given towards seven digital twin implementations that could benefit a tunnel facility manager. The first two are the monitoring of energy use, and monitoring the health of the tunnel system. It is concluded that energy use should be accounted for in a digital twin, and that the monitoring of system health is required for the quick understanding of problems in the tunnel and stakeholder management. The next three are predictive applications: predicting dependent relations between technical tunnel systems, maintenance scheduling, and operational simulations. These applications help anticipate risks during tunnel processes and provide opportunities for reaching better informed decisions regarding maintenance tasks, problem management and configuration management practices. One application, namely energy optimisation, can be considered fully autonomous, as such is not for decision support, but rather for tunnel optimisation, towards the control of energy use of tunnel technical installations.