The finite response time of the turbocharger is the most notable effect oftransient operation on a turbocharged diesel engine. To fulfil future emission requirements high amounts of transient EGR will be required. This impliesthat advanced turbocharger systems have to be introduced to enable high boost pressures with improved or at least maintained response time. The increased amount of tunable parameters from the more advanced turbocharging system will make it difficult to optimise the engine experimentally. Therefore the wish is to optimise the engine numerically, however this is a difficult task which demands more knowledge within the field of modelling the gas exchange system and its components, which is the aim of the present work. Engine simulations have been performed in the 1-dimensional fluid dynamic code GT-Power for transient operation and validated with engine measurements. The turbine was modelled according to the state of the art which is via look-up tables with measured turbine performance data from a steady-flow rig and used under the assumption that the turbine behaves in a quasi-steady manner. Turbine performance data was also obtained via the semi-empirical turbine design software, Rital for comparison. A heavy-duty diesel engine has been modelled with two different gas exchange system configurations. The standard configuration with a single twin-entry turbine and a rebuilt gas exchange system including a two-stage turbocharging system and high pressure loop for EGR. The results shows that it is difficult to predict the performance of the gas exchange system and its components, especially the turbine performance. When trying to predict turbine performance under transient operation the difficulties added, compared to stationary operation are long scale transients as wall temperature gradients in the cylinder and the exhaust manifold which directly influences the amount of isentropic energy to the turbine. This makes it even more difficult to predict the isentropic exhaust gas energy content compared to stationary operation, which is difficult to measure and therefore to state how well the turbine model actually performs. However, even though it is difficult to predict engine performance in detail the models have proved to be useful for concept studies as a help in engine design.