Railway transportation is facing increasing pressure to reduce the energy demand of its vehicles due to increasing concern for environmental issues. This thesis presents studies based on improved power management strategies for railway traction systems and demonstrates that there is potential for improvements in the total system energy efficiency if optimised high-level supervisory power management strategies are applied. Optimised power management strategies utilise existing power systems in a more cooperative and energy-efficient manner in order to reduce the total energy demand. In this thesis, three case studies in different research scenarios have been conducted. Under certain operational, geographic and physical constraints, the energy consumed by the train can be significantly reduced if improved control strategies are implemented. This thesis proposes a distance based model for train speed trajectory optimisation. Three optimisation algorithms, Ant Colony Optimisation (ACO), Genetic Algorithm (GA) and Dynamic Programming (DP), are applied to search for the optimal train speed trajectory, given a journey time constraint. The speed at each preset position along the journey is determined and optimised using these searching algorithms. In a DC railway network, power peaks in a substation are not desirable as they could present safety risks and are not energy efficient. A power peak can be avoided if the control of multiple trains is coordinated. The allocation of inter-station journey time intrinsically affects both service quality and energy efficiency. By identifying an optimal journey time allocation, a multi-objective function targeting both energy efficiency and service quality can be used. In this study, a DC railway is modelled with two parallel railway tracks, five station stops and three DC electric substations. Regenerative braking is studied in this DC electric network using Nodal Analysis (NA) and the Load Flow (LF) method. This study demonstrates that within the neighbourhood of an electric railway network, the journey time allocation for inter-station journeys will affect the total service quality and the energy loss. A GA is applied to find the best inter-station journey time allocation. Finally, this thesis explores the potential of applying advanced power management strategies to a Diesel Multiple Unit (DMU) train. DMU trains have multiple diesel engines which are commonly operated in a homogenous manner. The work presented in this thesis analyses the potential energy savings that may be obtained through the independent operation of the engines. Two widely investigated power management strategies which have been applied to the control of Hybrid Electric Vehicles are studied for a typical DMU railway vehicle. DP is applied to identify the optimal instant power distribution between engines. Based on the optimised results from DP, an adaptive rule-based online strategy is proposed using a non-linear programming optimisation algorithm.

This thesis presents a new experimental technique to determine the structure of train slipstreams. The highly turbulent, non stationary nature of slipstreams make their measurement difficult and time consuming as in order to identify the trends of behaviour several passages of the train have to be made. The new technique has been developed in order to minimise considerably the measuring time. It consists of a rotating rail rig to which a 1/50th scale model of a train is attached. Flow velocities are measured using two multi-hole Cobra probes, positioned close to the sides and top of the model. Tests were carried out at different model speeds. Velocity time histories for each configuration were obtained from ensemble averages of the results of a large number of runs (of the order of 80). From these it was possible to define velocity and turbulence intensity contours along the train as well as the displacement thickness of the boundary layer, allowing a more detailed analysis of the flow. Wavelet analysis was carried out on the experimental data to reveal details of the unsteady flow structure around the vehicle. It is concluded that, although this methodology introduces some problems the results obtained with this technique are in good agreement with previous model and full scale measurements.

The existing temperature profile in Great Britain determines the stress free temperature of continuous welded rail on the railway network, this ensures that cold-related tension cracks and heat-related rail buckles are kept within tolerable levels. It is predicted that winters will become warmer and summers hotter than the baseline climate. It is believed that Great Britain can continue to operate with a stress free (rail) temperature of 27°C under future climate scenarios, provided the acceptable operating range is narrowed upwards towards 27°C and that the quality of track, track-bed and subgrade are improved. These actions should limit the potential damage caused by more challenging temperature extremes. If changes are not made to ensure the track is more resilient to hotter summers the cost of buckles and heat related delays are projected to increase from £3.3m under baseline climate conditions to £24.7m in the 2080s under the high emissions climate change scenario. In winter the temperature range that causes the majority and most severe ice and snow delays is not expected to undergo much change for most of Great Britain until the 2080’s under the high emissions scenario, when there will be nominal reductions, mostly in the south region.

Abrupt changes in track stiffness can occur where the ballasted track transitions from soft foundations onto bridges and viaducts, onto hard foundations and into tunnels, or at the interface between bal¬lasted track and slab track. Such interfaces between areas of different track stiff¬ness are defined as critical zones and can cause passenger discomfort, unpredictable track degradation and, potentially, track support failures. Given their nature, main¬taining critical zones is more expensive than maintaining plain line track and measuring and assessing the variations in track stiffness are critical activities in track main¬tenance and in ascertaining the quality of newly designed and built railways. Existing methods of track stiffness measurement are not sufficiently accurate and depen¬dable. Therefore, the aim of this research was to develop an innovative but rapid method of assessing track stiffness, so as to identify problems in transition zones. During her doctoral research, the author designed and built a sleeper deflection measurement system as a means of quickly assessing track stiffness. The laser-based digital system developed uses position sensitive detectors (PSDs) and a line laser. The performance of the new measurement system was verified on a private railway at Long Marston, which showed that the system was able to capture the displacement data of individual sleepers and thus the continuous movement of rails supported by 18 sleepers. Ob¬serving the train-induced dynamic deflection patterns across multiple sleepers allows a rapid assessment of the track stiffness and provides the information that is necessary to manage the stiffness variation around transition zones.

At present, high speed dual purpose rail/road vehicles employing fixed non-destructive testing (NDT) sensors are used to inspect rails. Due to the uncertainties in characterisation of the defects when they are detected at high speed, manual re-visiting of the defects by expert operators is required before any decision regarding track maintenance is made. This research has been driven by a desire from the rail industry for a robotic system performing faster than human operators and being capable to both detect and characterise rolling contact fatigue (RCF) cracks in rails with the aim of automating the existing manual inspection and enhancing its accuracy and reliability. This thesis combines expert systems technologies with robotic NDT to fulfil this aspiration. A great deal of effort has been spent to develop a robotic inspection trolley which can automatically detect and characterise the RCF cracks in rails using an alternating current field measurement (ACFM) sensor. It uses a rule based expert system (RBES) proposed to control the robotic trolley and more importantly process ACFM data for both detecting and sizing defects. The developed system can detect the possible presence of defects in railway tracks at high speed pass (5-20 km/h) and can automatically return to an identified defect location to perform a slower and more detailed scan (up to 20 mm/s) across a rail section to determine the size, depth and number of cracks present in that section.

Rolling contact fatigue (RCF) cracks are a widespread problem that impairs the service life of railway rails and wheels, with an associated high cost of labour and capital expenditure for remediation. Severe RCF cracks cause serious safety issues as they can turn down at a larger propagation angle into the rail potentially leading to a rail break. Rail grinding can effectively eliminate RCF cracks if they are detected when they are small enough to be removed. Alternating current field measurement (ACFM) is one of the electromagnetic (EM) techniques that can be used for defect detection and sizing in the rail industry. ACFM has been reported to be more accurate in providing length and depth information than conventional UT for small RCF cracks and is less sensitive to lift-off than eddy current methods. The aim of the present research is to analyse the response of ACFM signals to single and multiple RCF cracks in railway rails using experimental measurements and FE based modelling tools, focusing on the influences of crack vertical angle and multiple cracks (number, spacing, size, uniformity) on the ACFM signal to improve the accuracy of dimension predictions. A novel method (using the Bz signal) is proposed to determine the vertical angle of the RCF cracks, which then allows the crack vertical depth to be determined from the pocket length (standard output from ACFM measurements) and therefore the appropriate amount of rail grinding to remove the RCF cracks. It was found that the vertical angle influences the pocket length determined from the measured ΔBxmax/Bx value when the cracks are shallow (vertical angles < 30°), therefore greater accuracy can be obtained when compensating the ΔBxmax/Bx value using the determined vertical angle. It is shown that the variations of crack surface length, crack inner spacing and crack number for multiple cracks also influence the ΔBxmax/Bx values determined for multiple cracks. The influences of asymmetrical crack shapes on crack sizing are discussed, in general it has been found that for accurate sizing of RCF cracks using a single ACFM scan the cracks should be regular, where the assumption of semi ellipse shapes is appropriate. The methods developed in the project were assessed using calibration samples (machined cracks with different sizes and vertical angles) and rails removed from service containing single and multiple RCF cracks. It was found that the new approach proposed in this work allowed the vertical angle to be predicted well for single and multiple RCF cracks (difference to measurements < 14.3 %). In addition the error in pocket length prediction is greatly decreased when using the sizing method including compensation determined from the crack vertical angle.

Railway tracks deteriorate over time due to the combined effects of train loads and environment. To carry out appropriate maintenance, it is necessary to measure the condition of the track. There are two broad categories of condition measures namely: function and structural measures of conditions. Functional measures of condition assess the condition of the track from the point of view of the user. An example of a commonly used functional measure of condition is track geometry. In the UK, track geometry is measured by the track recording coach and it is carried frequently on the entire network. On the other hand, structural measures of condition assess the structural integrity of the track. Example of a commonly used structural measure of condition is track stiffness. In the UK, track stiffness is measured using the Falling Weight Deflectometer (FWD) and it is carried out less frequently on specific railway track sections as it is very slow, requires closure of the track and is expensive to operate. The aim of this research is to investigate the feasibility of using track geometry measurement to extract track stiffness information and ultimately develop a prototype automated system to achieve this.

The writer analyses the introduction of a simulator enabled approach to railway traction driver training and assesses whether the transition from a conventional training delivery process has been effective. The evaluation of effectiveness is based on a study of Iarnród Éireann’s simulator system. Evidence is contained within four supporting strands, i.e., the change in relevant operational risk that has been calculated using ex ante and ex post runs of Iarnród Éireann’s risk model, the internal rate of return on the financial investment necessary to effect the change, the results of an operator attitudinal study and the findings of an independent expert audit. The study establishes that simulation is an effective training medium. The attributes of the system and the use cases that resulted in this finding are described. The writer also presents additional value-adding training objectives that could increase the project’s internal rate of return or IRR. The study affirms that the required verisimilitude of a simulator system is a function of the training goals and the nature of the skills under development. Design features and use strategies can mitigate for potential negative effects of simulator operation. The findings have industry-wide relevance for those tasked with providing effective training to the 133,000 train drivers within the European Union.

Improvements in railway capabilities have resulted in heavier axle loads and higher speed operations, which increase the dynamic loads on the track. As a result, railway subsidence has become a threat to good railway performance and safe railway operation. The author of this thesis provides an approach for railway performance assessment through the monitoring and prediction of railway subsidence. The InSAR technique, which is able to monitor railway subsidence over a large area and long time period, was selected for railway subsidence monitoring. Future trends of railway subsidence should also be predicted using subsidence prediction models based on the time series deformation records obtained by InSAR. Three time series prediction models, which are the ARMA model, a neural network model and the grey model, are adopted in this thesis. Two case studies which monitor and predict the subsidence of the HS1 route were carried out to assess the performance of HS1. The case studies demonstrate that except for some areas with potential subsidence, no large scale subsidence has occurred on HS1 and the line is still stable after its 10 years' operation. In addition, the neural network model has the best performance in predicting the subsidence of HS1.

Complex optimisation problems, which are concerned with optimising a given aspect of a complex system, such as time or energy, are difficult to solve. Often a range of solutions exist, and the difficulty lies in determining which solutions to implement in which part of the system. Within this work, a novel method is developed that allows the solver to overcome the key challenges for these types of problems, which are: defining the system parts (subsystems); minimising model complexity; quantifying solution effectiveness; and identifying relationships between solutions and subsystems. The method is demonstrated through application to the problem of railway traction energy saving. Subsystems are defined using quantified network and service characteristics. For each subsystem, the trends between six key solutions and the key performance indicators are analysed using multivariate data analysis and visualisation techniques. The relationships between subsystems are then explored at system level. The analysis determines the suitable solutions for each type of railway, providing information for operators about which solutions to target. Based on the results, the implementation of permanent magnet motor technology is considered, illustrating that the method is a suitable tool for informing further studies.