Multi-pantograph operation allows several short electric multiple unit (EMU) trainsets to be coupled or decoupled to adapt to daily or seasonal passenger-flow variation. Although this is a convenient and efficient way to operate rolling stock and use railway infrastructure, pantographs significantly influence each other and even significantly change the dynamic behaviour of the system compared to single-pantograph operation in the same condition. The multi-pantograph system is more sensitive and vulnerable than the single-pantograph system, especially at high operational speeds or with pantographs spaced at short distances. Heavy oscillation in the system can result in low quality of current collection, electromagnetic interference, severe wear on the contact surfaces or even structural damage. The mechanical interaction between the pantograph and the catenary is one of the key issues which limits the maximum operational speed and decides the maintenance cost.     Many researchers have paid a lot of attention to the single-pantograph operation and have made great progress on system modelling, optimizing, parameter studies and active control. However, how the pantographs in a train configuration affect each other in multi-pantograph operation and which factors limit the number of pantographs is not fully investigated. Nowadays, to avoid risking operational safety, there are strict regulations to limit the maximum operational speed, the maximum number of pantographs in use, and the minimum spacing distance between pantographs. With the trend of high-speed railways, there are huge demands on increasing operational speed and shortening spacing distance between pantographs. Furthermore, it is desirable to explore more practical and budget-saving methods to achieve higher speed on existing lines without significant technical modification.     In addition to a literature survey of the dynamics of pantograph-catenary systems, this thesis carries out a numerical study on multi-pantograph operation based on a three-dimensional pantograph-catenary finite element (FE) model. In this study, the relationship between dynamic performance and other parameters, i.e. the number of pantographs in use, running speed and the position of the pantographs, are investigated. The results show that the spacing distance between pantographs is the most critical factor and the trailing pantograph does not always suffer from deterioration of the dynamic performance. By discussing the two-pantograph operation at short spacing distances, it is found that a properly excited catenary caused by the leading pantograph and the wave interference between pantographs can contribute to an improvement on the trailing pantograph performance. To avoid the additional wear caused by poor dynamic performance on the leading pantograph and achieve further improvement at high speeds, it is suggested to use the leading pantograph as an auxiliary pantograph, which does not conduct any electric current and optimize the uplift force on the leading pantograph. After a brief discussion on some system parameter deviations, it is shown that a 30% of speed increase should be possible to achieve while still sustaining a good dynamic performance without large modifications on the existing catenary system. / <p>QC 20150928</p>

In this thesis, after a comprehensive literature review on different conventional and predictive torque control strategies for switched reluctance motor (SRM) drives, two online methods and one offline multi-objective optimization-based method are proposed to extend the operating speed range of finite control set model predictive torque control (FCS-MPTC) for SRM by adaptively controlling the commutation angles in the entire speed range. Furthermore, a method is proposed to minimize the steady state torque tracking error of FCS-MPTC for SRM drives. The incapability of the conventional FCS-MPTC in controlling the commutation angles, which is considered as one of the main drawbacks of the conventional FCS-MPTC, limits its application for high-speed torque control of SRM drives. The phase turn-off angle is always selected to be close to the aligned position with the conventional FCS-MPTC regardless of the operating speed. However, commutation angle advancement is required for high-speed torque control of SRM drives to limit the negative phase torque resulting from the current tail after the turn-off angle in the generating region. Excessive negative torque with the conventional FCS-MPTC at higher speeds can result in a degraded performance with high rms current, low average torque, high torque ripple, and reduced efficiency. The phase turn-off angle can be adaptively controlled as speed changes with the first online commutation angle control strategy proposed in this thesis. This method is based on predicting the free-wheeling phase current in an extended time interval which is much bigger than the prediction horizon of FCS-MPTC. The second online turn-off angle control method is also proposed by improving the optimality condition defined for determining the optimal turn-off angle. The optimality condition is determined by calculating the work done by the conducting phase after the phase is turned off. The weighting factor of the objective function of FCS-MPTC is kept constant with both proposed online methods. An offline multi-objective optimization-based strategy is proposed to determine the globally optimal turn-off angle and the weighting factor in the entire operating torque and speed ranges. The effectiveness of both proposed online methods and the offline commutation angle control strategy is verified using simulations and experimental results. The results are also compared to the conventional FCS-MPTC and the indirect average torque control with optimized conduction angles which is considered as one of the main conventional torque control strategies for SRM drives. In order to minimize the torque tracking error as a result of either parameter uncertainties or tracking multiple objectives with a single objective function with weighting factors, a method is proposed which is based on updating the reference torque at each sample time by calculating the average torque tracking error in the previous sample times. The validity of the proposed method is verified using simulations. / Thesis / Doctor of Philosophy (PhD)