Traction motor modelling and fuzzy control in railway vehicles

From the past to the present, railway systems have always had an influence on the daily life of the society, and there is no doubt on that it will continue for a long time. In the beginning of 1900s, although the railway was mostly used for freight transport, along with the increasing population, big investments were started in the area of passenger transport. With the developing technology, railway systems have started to change according to way of usage, amount and type of demand. For instance, systems targeting intercity passenger transport have considered transport speeds as in a first place rather than transport capacity. For urban transport such as metro or light rail systems started to be applied for medium-distance transportation within the cities by aiming higher capacity at medium speed. As of today, this system is still being preferred in the regions where urbanization is crowded. Besides them, especially for historical and narrow areas, tram systems have begun to be preferred for low-speed, high-capacity urban transport since it has a huge advantage for land allocation as it shares the same corridor with roadway. The development of technologic investment continues to influence these systems while these systems also affect the society in terms of social-economical means. Especially early in the 20th century, electric motors began to be integrated into rail vehicles powered by diesel fuels which had taken a place of steam vehicles. Subsequently, new speed and current control methods were developed with the aim of achieving higher efficiency from electric motors. In this study, the motor type used for light rail transportation traction motor with 40 km/h maximum speed is investigated. Main principle of the linear model is to control the speed of the motor by controlling torque which has a correlation with acceleration. Transfer functions represent field-oriented control, in other words vector control. In field-oriented control, asynchronous motors work are considered to be controlled as in separately-excited DC motors. The stator current is decomposed as two different components which can be controlled independently from each other. One of these current components controls the magnetic flux of the motor where the other controls the torque generated by the motor. Stator or rotor flux as well as the position must be known in field-oriented control. Depending on the flux measurement method, it is grouped into direct field-oriented control or indirect field-oriented control. Field-oriented control method can also be categorized as rotor field-oriented control or stator field-oriented control according to usage of rotor or stator flux in the calculation. In this control method, it is aimed to keep the motor momentum and flux in the desired orbit effectively without being influenced by the changes of motor and load parameters. In direct field-oriented motor control, it is necessary to know the flux accurately due to the sensitive and precise performance of the asynchronous machine. Since it is difficult and expensive to measure flux directly, it is calculated through the analysis of current, voltage and speed information which can be easily measured. The simplest method of flux calculation is to use open-loop supervisors by taking the advantage of the asynchronous motor current and voltage models. The mathematical models used to calculate the current in this way are known as current and voltage models. By taking both electric and mechanical transfer functions into consideration, a linear model of field-oriented control is established. As a result of this the mechanical transfer function is supplied by the output resulting from the electric transfer function. Output of electrical function at the same time gives the electromechanical torque generated by the controller. Electrical transfer function of linear model is controlled in two separate loops as speed control and current control. Torque control is achieved by changing the torque reference only by keeping the magnetic flux constant with the vector control method. Current control cycle, which compares the reference torque and the instantaneous torque information obtained from the motor, automatically calculates the torque value that must be delivered to the motor. The main purpose of this cycle is to control the acceleration of the motor with the correlation between torque and acceleration. Conventional proportional-integral method is used in this loop. Due to the speed information to be given to the vector control mechanism, speed control is aimed at the second cycle. In contrast to the conventional proportional integral method, fuzzy logic theory is used in this loop, where the motor's real-time velocity data is compared with the reference velocity value. Fuzzy logic is based on fuzzy set and sub sets. In a classical approach or in other words boolean algebra, truth values of variables may only be the integer values 0 or 1. When it is expressed mathematically, "1" is the value when the element is a member of the the cluster where "0" represents when the element is not a member of the cluster. Fuzzy logic is the expansion of classical cluster representation. Each entity in the fuzzy cluster has a membership level. The membership level of entities can be any value in the interval (0,1) and the membership function is denoted by M (x). According to this, fuzzy block gives the proper output as it is defined by membership functions which consider input values. An acceleration limiter was also applied between mechanical and electrical transfer functions. This means, even though electromechanical torque is created more than the limit, a saturation block will saturate the torque input which is given to mechanical transfer function. In addition to the control mechanisms, the aim is to create a model which is similar to the real railway vehicles by considering environmental factors affecting the vehicle movement. These are simulated by summing up the Davis formula as resistive loads caused by friction and the wind, and the load that occurs by the gradient of the line, and added as a mechanical load into the system. Some expert control functions are also added in order ensure comfortable departure during the integration of mechanical loads. Also, the moment of inertia generated by the vehicle's own mass is taken into account in the linear model. In addition, the gear box, which must be between the engine axle and the wheels, is also taken into account. The aim of this thesis is to investigate and model the three-phase asynchronous motor in railway vehicles from the point of view of control and automation engineering. For this reason, motor control systems and working principle found in the rail system vehicles are explained. With the control system established, the model was set up on MATLAB Simulink with the aim of bringing a new perspective to the conventional control systems of the rail industry. It is envisaged that this new model will assist new researches in this context after an adaptation to the entire railway traction simulation.