This thesis work is performed at Bombardier Transportation in order to evaluate previous work in the field of simplifying the tilt system on a high speed train. The system is modelled in the multi-body simulation program Simpack. 

 Current tilting systems are rather expensive, approximately three to six per cent of the total investment; therefore the aim is to investigate the performance of a simplified carbody tilting system. The simulations are connected to the Bombardier Transportation electric multiple unit Regina 250 intercity train, R250T; however the results are meant to be adaptable also to other rail vehicles. The main priority is to investigate whether the subsystems are appropriate for carbody tilt. The subsystems of special interest are the force controlled actuator and rubber-steel laminate springs.  

 The system is designed for 245 mm cant deficiency, the maximum allowed in Sweden. 0.8 m/s2 of lateral acceleration is allowed on the passengers in the middle of the carbody and the actuator should work at a maximum quasi-static force of 25 kN. All of the above criteria are fulfilled during normal operation in all types of curves tested. 

 The model has been tested with three different heights of bolster centre of rotation to suit different vehicles. With the lowest value of the centre of rotation the system response is not acceptable. This can be cured with stiffer tilting elements; however this will need an actuator with a higher force output. The R250T is regarded as safe in all tested cases, e.g. normal service, failure scenario and snow packing, as concerns risk of derailment.  

 The rubber laminate springs give the vehicle good ride characteristics regarding tilt angle. It is highly plausible that a rubber spring with the required characteristics can be built; there are existing elements which closely match the needed characteristics used on rail vehicles today.   

 The force controlled actuator is able to create the required tilt angle and provide a safe ride; however the roll angle velocity is not as constant as what a position controlled actuator can perform. It might be possible to solve this with a more advanced control loop, or perhaps use stored track data when possible. The conclusion is that the simplicity of this actuator cannot yet be fully utilized, without further study.