Active wheelset steering can improve the curving performance of railway vehicles and thus reduce wear. Several control strategies have been proposed to achieve a perfect steering condition which may require feedback signals that are difficult to measure. This study proposes a control strategy for an active wheelset steering system via wheelset angular velocity measurements aimed to minimise longitudinal creepages in curves. The desired wheelset angular velocity is derived from the relation of longitudinal creepages and the wheelset movement in curves. Co-simulations with a conventional railway vehicle with two two-axle bogies are carried out for 24 cases. First, the strategy based on actual wheel-rail geometry is used to evaluate the effectiveness of the proposed control strategy; in the second step, a simplified strategy is tested using the approximated equivalent rolling radius approach. Curving performance indicators, including wheelset movements and wheel-rail wear number, are used to evaluate the performance of the control system. The results indicate the effectiveness of the proposed control strategies. Challenges and practical considerations are also discussed.

Active wheelset steering using wheelset angular velocity feedback has been proposed and proven effective in achieving one of the goals for perfect steering condition, i.e. minimising longitudinal creep forces and thus wear. This control strategy relies on the contact parameters of wheel-rail pairs to generate a desired angular velocity by using the wheel rolling radii. Some wheel-rail combinations give a non-monotonic function of the average rolling radius with respect to curve radius, which therefore poses a difficulty in controlling active steering with the angular velocity feedback. This study aims to investigate the effect of different wheel-rail combinations and wheel wear. Worn wheels in general tend to cause a non-unique function whereas the original profile gives a monotonic one. Two countermeasures are then investigated to cope with the raised challenges. First, the reference signal generated from the approximated average rolling radius method has proven to be effective for both the original and the least worn profile; however, a wheel with higher wear depths would require a new tuned scaling factor. The control strategy based on equal wheelset angular velocity is then investigated with two schemes: equal wheelset angular velocity within the same running gear and equal angular velocity of all wheelsets within the vehicle. Both schemes provide significant wear number reduction, but the second approach results in better distribution of wheel-rail lateral forces among all wheelsets. This control strategy can also relax the demanded knowledge of wheel-rail contact properties and vehicle travelling speed.

This doctoral dissertation focusses on active wheelset steering of railway vehicles. Active wheelset steering is a mechatronic solution for a solid-axle wheelset that has been proven to improve curving performance. Several aspects are investigated including control strategies, control systems and steering control in turnouts. This aims to explore the interaction of all elements in mechatronic running gear with active steering capability. 

Active wheelset steering can be executed by various control strategies and one of the widely recognised is control strategies for perfect steering condition. Several control variables can be used to achieve such a condition such as wheelset yaw angle and lateral displacement. For instance, wheelset lateral displacement control requires a lateral displacement feedback signal and a knowledge of the wheel-rail contact parameter, that is equivalent conicity of the wheelset. Hence, the estimator based on the dual extended Kalman filter technique is designed to estimate vehicle states and wheelset equivalent conicity. This estimator can potentially be applied for active wheelset steering with lateral displacement feedback. To ease the required sensing and estimation, control strategies for active wheelset steering aiming at achieving perfect steering condition are proposed by considering wheelset angular velocity as a control variable which is easy to obtain through direct measurements. The control strategy is first introduced by considering the main objective to minimise longitudinal creepages in curve. Influence of wheel-rail combinations are investigated as it greatly affect the proposed control strategies. Subsequently, the simplified control strategy and the control strategy based on equal wheelset angular velocity are proposed as alternative solutions to solve challenges from wheel-rail pairs. In addition, a robust control for active wheelset steering is studied as railway vehicles are subjected to disturbances from track inputs and uncertainties mainly from wheel-rail parameter variation. The design and development of robust control system is based on sliding mode control with equivalent control. The complexity of the derived controller is mainly from the required signals. Then, the simplified sliding mode control (SSMC) is investigated. The application of the simplified controller as passive fault-tolerant control is analysed with a consideration of one actuator fault with zero force. The adaptive sliding mode control is proposed to support the full mitigation to actuator failure effect. Another contribution of this thesis also includes steering control in switch and crossing as railway turnout is an important track component that guides movement of the vehicles. The characteristics of turnouts challenge to active wheelset steering system in various aspect. Thus, the optimised steering control in turnouts is proposed to further improve steering performance. The preview of train position is incorporated by using the map-based train positioning. This proposed method can dramatically reduce peaks in wheel-rail wear in switch and crossing.

In conclusion, this research contributes to the advancement in active wheelset steering of railway vehicles in several aspects including control strategies, robust control system and control approach in turnouts for active wheelset steering. The contributions in this thesis would facilitate the practical implementation of active wheelset steering and further performance improvement in railway turnouts.

Denna doktorsavhandling behandlar aktiv hjulparstyrning av järnvägsfordon. Aktiv hjulparstyrning är ett mekatroniskt tillvägagångssätt som för hjulpar med solid axel ger förbättrad kurvtagning. Flera aspekter undersöks inklusive reglerstrategier, reglersystem samt styrning vid spårväxlar. Denna avhandling syftar till att utforska interaktionen mellan olika delar av löpverk med aktiv hjulparstyrning.

Aktiv hjulparstyrning kan utföras med olika reglerstrategier. Allmänt erkända reglerstrategier strävar mot perfekta styrförhållanden. Flera variabler kan användas för att uppnå detta tillstånd. Två exempel på variabler är hjulparets girvinkel och dess sidoförskjutning relativt rälen. För att använda sidoförskjutning som variabel krävs det både att sidoförskjutningen kan mätas samt kunskap om hjul och räls kontaktparametrar. Dessa parametrar motsvarar koniciteten i hjul-räl systemet som kan uppskattas med den tvåsidigt utvidgade Kalmanfilter tekniken utformad för att uppskatta fordonets tillstånd och hjulparets ekvivalenta konicitet. Denna uppskattning kan appliceras för aktiv hjulparstyrning baserad på sidoförskjutning. För att förenkla kraven på sensorer och uppskattning föreslås det att betrakta hjulparets vinkelhastighet som en reglerparameter, eftersom denna är lätt tillgänglig genom direkta mätningar. Denna reglerstrategin introduceras först med huvudmålet att minimera det längsgående krypet mellan hjul och räl vid kurvtagning. Inverkan av olika hjul och spår kombinationer undersöks eftersom de kraftigt påverkar de föreslagna reglerstrategierna. Därefter föreslås den förenklade reglerstrategin och reglerstrategin baserad på hjulparets vinkelhastighet som alternativa lösningar för att lösa utmaningar från olika hjul och räls kombinationer. Dessutom studeras en robust reglering för aktiv hjulparstyrning eftersom järnvägsfordon utsätts av störningar från rälsen som huvudsakligen beror på variationer av hjul och räls parametrar. Designen och utvecklingen av ett robust reglersystem baseras på styrning längs en glidande yta med motsvarande reglering. Komplexiteten av den framtagna regleringen bestäms till stor del av vilka signaler som krävs. Därefter undersöks den förenklade regleringen baserat på glidande yta tekniken. Tillämpningen av den förenklade regleringen som passiv feltolerant styrning analyseras med fokus på fel där ställdon inte ger kraft. Den adaptiva regleringen visar på full hantering av analyserade ställdonsfel. Ett annat bidrag i denna avhandling är förbättrad styrning i spårväxlar. Detta har granskats eftersom spårväxlar påverkar fordonets rörelser och slitage. Flera utmaningar vid spårväxlar i kombination med aktiv hjulparstyrning undersöks. En optimerad styrning vid spårväxlar föreslås, optimeringen inkluderar en framförhållning av tågets position genom användning av kartbaserad tågpositionering. Denna föreslagna metod kan dramatiskt minska toppar i hjul och räls-slitage i spårväxlar.

Sammanfattningsvis bidrar denna forskning till utveckling av aktiv hjulparstyrning för järnvägsfordon inom områden som reglerstrategier, robusta reglersystem och val av reglering vid spårväxlar. Resultaten som presenteras i denna avhandling kan användas vid implementering av aktiv hjulparstyrning samt för att minska hjul och räls-slitage i spårväxlar.

Negotiation of turnouts imposes challenges for an active wheelset steering system due to lack of smooth transition curves and existence of rail discontinuities. These affect the performance of both the vehicle and control system in turnouts, which is investigated in this paper. The simulation is carried out with a conventional railway vehicle with two two-axle bogies passing through a crossover onto a parallel track. The Swedish 60E1-R760-1:5 turnout is used in this study. The results reveal that an active wheelset steering system can decrease the wheel-rail wear index. However, peaks in wear number take place in the switch toe and the crossing nose and they are considerably higher than other regions. A control scheme with preview is proposed by considering wheelset lateral positions on discontinuous rail profiles to avoid flange contact. The proposed control system with preview results in a further reduction of the maximum wear number by 66%.

Active wheelset steering has been studied and implemented to improve the curving performance and stability of the wheelsets to overcome the drawbacks of the passive system. A control system for active wheelset steering must be robust to parameter variations and disturbances. A robust sliding mode controller with integral action (SMC + I) for active wheelset steering is therefore proposed and implemented in this paper to control wheelset lateral displacements to achieve perfect rolling conditions during curve negotiation. The robustness of the controller is achieved by deriving the control inputs bounded with known uncertain parameters. The control input is derived as a combination of the equivalent term and a switching term to reduce the amplitude of the switching input. A saturation function is used instead of a sign function in the switching term to provide continuous control. The integral action (I) is added to the sliding surface function to minimize the zero steady-state error. Co-simulation is executed to evaluate the performance and robustness of the designed controller. A conventional railway vehicle with two two-axle bogies with a maximum operating speed of 250 km/h is modelled in SIMPACK®, while the SMC + I controller is implemented in MATLAB/Simulink® for co-simulation. Two hydraulic servo actuators (HSAs), modelled with Simscape hydraulic libraries, are implemented in the longitudinal direction to steer each wheelset. The proposed controller ensures stability, finite-time convergence and zero steady-state errors for all possible running scenarios. This indicates robust performance of the designed SMC + I controller.

This paper presents the implementation of the dual extended Kalman filter (DEKF) to estimate wheelset equivalent conicity, an accurate understanding of which can facilitate the implementation of an effective model-based estimator. The estimator is developed to identify the wheelset equivalent conicity of high-speed railway vehicles while negotiating a curve. The designed DEKF estimator employs two discrete-time extended Kalman filters combining state and parameter estimators in parallel. This estimator uses easily available measurements from acceleration sensors measuring at axle boxes and a rate gyroscope measuring bogie frame yaw velocity. Two tests, including linearized and actual wheel-rail geometry, are carried out at a speed of 250 km/h with stochastic and deterministic track features using multibody simulations, SIMPACK. The results with acceptable estimation errors for both track conditions indicate adequate performance and reliability of the designed DEKF estimator. They demonstrate the feasibility of utilizing this DEKF method in rail vehicle applications as the knowledge of time-varying parameters is not only important in achieving an effective estimator for vehicle control but also useful for vehicle condition monitoring.

Intelligent fault identification of rail vehicles from onboard measurements is of utmost importance to reduce the operating and maintenance cost of high-speed vehicles. Early identification of vehicle faults responsible for an unsafe situation, such as the instable running of highspeed vehicles, is very important to ensure the safety of operating rail vehicles. However, this task is challenging because of the nonlinear dynamics associated with multiple subsystems of the rail vehicle. The task becomes more challenging with only accelerations recorded in the carbody where, nevertheless, sensor maintenance is significantly lower compared to axlebox accelerometers. This paper proposes a Temporal Convolution Network (TCN)-based intelligent fault detection algorithm to detect rail vehicle faults. In this investigation, the classifiers are trained and tested with the results of numerical simulations of a high-speed vehicle (200 km/h). The TCN based fault classification algorithm identifies the rail vehicle faults with 98.7% accuracy. The proposed method contributes towards digitalization of rail vehicle maintenance through condition-based and predictive maintenance.

Railway turnouts present challenges for active wheelset steering systems because of the lack of smooth transition curves as well as discontinued and variable rail profiles in this region. These track characteristics affect vehicle curving and the performance of wheelset steering control. This paper presents a control system for active wheelset steering, designed to address the specific issues in turnout regions. Co-simulation is used to perform simulation studies of the proposed framework, where the vehicle is modelled in the multibody simulation tool, SIMPACK, and the control system is implemented in MATLAB/Simulink. The Swedish 60E1-R760-1:5 turnout geometry and rail profiles are modeled as track inputs to perform a comprehensive analysis. First, several radial steering control approaches are evaluated, which improve wear numbers in the circular curve, but local wear is still high. Then, an optimized steering approach is proposed. The levels of steering are determined by using the Particle Swarm Optimization algorithm, where the objective function is a weighted sum of the summation of the maximum wear numbers and the summation of the average of wear numbers. The optimized steering approach produces a substantial reduction in wear numbers. In a later step, the map-based train positioning is incorporated to investigate the overall performance of the vehicle, as the optimized steering control requires a preview of the train position. This is to investigate the effectiveness of the proposed control system regarding practical considerations in a real vehicle implementation. The objective function is increased by 36 % compared to the idealized optimized steering control.

Active wheelset steering using wheelset angular velocity feedback has been proposed and proven effective in achieving one of the goals for perfect steering condition, i.e. minimizing longitudinal creep forces and thus wear. This control strategy relies on the contact parameters of wheel-rail pairs to generate a desired angular velocity by using the wheel rolling radii. Some wheel-rail combinations give a non-monotonic function of the average rolling radius with respect to curve radius, which therefore poses a difficulty in controlling active steering with the angular velocity feedback. This study aims to investigate the effect of different wheel-rail combinations and wheel wear. Worn wheels in general tend to cause a non-unique function whereas the original profile gives a monotonic one. Two countermeasures are then investigated to cope with the raised challenges. First, the reference signal generated from the approximated average rolling radius method has proven to be effective for both the original and the least worn profile; however, a wheel with higher wear depths would require a new tuned scaling factor. The control strategy based on equal wheelset angular velocity is then investigated with two schemes: equal wheelset angular velocity within the same running gear and equal angular velocity of all wheelsets within the vehicle. Both schemes provide significant wear number reduction, but the second approach results in better distribution of wheel-rail lateral forces among all wheelsets. This control strategy can also relax the demanded knowledge of wheel-rail contact properties and vehicle travelling speed.

A robust control system is essential for active wheelset steering as rail vehicles have a broad range of operating conditions and are subjected to uncertainties and disturbances. This study proposes a robust control system based on sliding mode control. The controller namely robust sliding mode control (RSMC) is designed from the governing equations of a suspended single-wheelset with yaw and pitch motion. The control law requires feedback signals of the wheelset angle of attack, lateral displacement and angular velocity. To ease the required feedback signal of angle of attack and lateral displacement, the simplified sliding mode control (SSMC) is proposed. Both RSMC and SSMC controllers give adequate control performance for all running cases for deterministic track inputs, which confirms the robustness of the controllers. Later, the designed SSMC is used to investigate its application towards passive fault-tolerant control, considering that one actuator fails while running. The control response shows its ability to mitigate the effect of actuator faults for most cases; however, the chosen gain is not large enough to totally mitigate the failure effect on control performance for some cases, especially in a large curve. Thus, an adaptive mechanism based on a sliding surface with proportional and derivative terms is employed to adapt the gain of the switching control input so that it is large enough to counteract the faults.