As the demand for transport of goods and passengers increases, the capacity of railway networks is becoming more and more saturated. Railway infrastructure managers aim to increase the rail network capacity with the minimum possible investments in infrastructure. Next generation signalling concepts such as Moving Block (MB) and Virtual Coupling (VC) are expected to bring significant capacity benefits to railway corridors. The aim of this thesis is to identify capacity benefits of Virtual coupling over ETCS L3 MB considering a realistic operational setup, where different rolling stock characteristics, driving behavior and traffic information updates exist. Inspired by safety distance car-following models, the multi-state train-following model developed by Quaglietta et al. (2020) is enhanced with a dynamic term, called Dynamic Safety Margin (DSM). The DSM ensures that the following train can come to a standstill at any time applying service braking, even in the worst-case scenario when the leading train applies emergency braking. The DSM accounts for positioning errors, the Vehicle-to-Vehicle (V2V) communication update delay, different train control delay times, and braking capabilities of trains running as a convoy. An One-at-a-time (OAT) sensitivity analysis of the proposed model is conducted to evaluate the importance of selected model parameters in railway capacity. The model parameters under investigation concern the V2V update delay, the train control delay time, and the braking rate of the following train. The braking rate of the following train is proved to have a significant impact on railway capacity, implying that trains with better braking performance imply shorter time headways. To identify potential capacity benefits of the proposed model over ETCS L3 MB, operational scenarios are executed through microscopic simulation in the EGTRAIN tool. The scenarios consider three different rolling stock types, being suburban, regional, and intercity trains. For trains running on open track without stops, the proposed model promises the highest capacity benefits over ETCS L3 MB, for trains with improved braking and acceleration characteristics. For train services with stops, all the train convoy combinations show promising capacity benefits, which are higher for heterogeneous train convoys. Moreover, headways can be reduced to the minimum possible when faster trains follow slower trains. Finally, this model enables multiple train convoy formation based on a predecessor-follower communication topology, where a safety distance is always ensured between trains.