This thesis studies measuring systems for active steering of railway vehicles. The aim
of the study is to develop state estimation techniques to provide high integrity
feedback variables for the active steering of railway vehicles. Practicality and
provision of high-integrity data are two important aspects of the work. To avoid the
use of expensive sensors and complex instrumentation, practical techniques for
estimating vehicle variables are developed where only economical measurements are
used and they can be easily implemented.
The conventional solid-axle wheelset and wheelset with independently-rotating
wheels are studied and their mathematical models are developed. The fundamental
stability problem of these two models is analysed from a control engineering
viewpoint for studies of actively-controlled wheelsets. The Kalman filters are then
developed for these models to estimate all state variables, particularly variables of the wheelset relative to the track such as lateral displacement and yaw angle which are
needed for active control. A number of sensing options are also identified, analysed
for performance and assessed in a comparative sense. Fault detection and isolation schemes are then studied for the estimation techniques developed. Finally, some applications are considered. The techniques and analysis methods developed for the single wheel pair are extended and applied to a MKII coach and a two-axle railway vehicle. The estimation of cant deficiency for tilting trains is explored, and also the possibility of state estimation for a real profiled wheel.
A Doctoral Thesis. Submitted in partial fulfilment of the requirements for the award of Doctor of Philosophy of Loughborough University.