FAULT DIAGNOSIS AND COMPUTATIONAL MODELING OF RAILWAY TRACTION SYSTEM

Abstract

Because of the limited availability of fossil fuels, a major shift can be seen from petrol, diesel vehicles to electric and hydrogen based vehicles in the upcoming times. One of the major chunk of electric vehicles currently in use consists of EMU and high speed trains, which has seen an upsurge in demand in urban cities because of easy, economical, time saving and more comfortable commuting. With increased frequency of service and more and more number of passengers adding on daily, these systems are more prone to fault (or failure) which may lead to loss of life as well as money. So to avoid fault or failure either high preventive maintenance strategies shall be adopted or redundancy shall be increased, both of which leads to an increased cost. Fault is an unwanted, measurable, non-measurable, monitarable, non-monitarable, susceptible change in characteristics of system (running or at rest), which may or may not result in the failure of the system. Fault diagnosis approach so used involves an interaction of mathematical as well as physical model. Sensors are placed at optimum locations in the system which then interacts with the mathematical functions defining the system, and any inconsistency in the values obtained from output function so observed will be monitored and controlled. Also if a fault occurs, what would be its optimum path of travel is also figured out i.e which components are most likely to fail because of presence of certain fault in certain component. Using bondgraph, working on the concept of transfer of power (energy is conserved), which is sensitive to any form of spike in energy (either positive or negative), one can fairly predict the possibility of fault and its effect on the consecutive system working online with the help of sensors and ARRs (mathematical function). Thus one can isolate a fault before it becomes critical at subsequent stages. Same can be done in the model of traction system of an EMU which mainly comprises of power supply line, pantograph, transformer, induction motor, gearbox, and wheel. Thus a model is generated on a bondgraph which is simulated and analysed under normal working condition. Also its ARRs are also derived which when compared with the online system (or model) can help in determining the possible faults in the system.