A comparison between two speed observers and effects of parameter variations on the performance of vector control

Induction motors account for 90% of the rotating machines used in the industry. They have become the most widely used motor type because of their low service costs, mechanical robustness and wide speed range. It is desired that induction motors could be drived in a different speed levels for different applications. In this thesis, it is stated that the control methods of induction motors which developed to make speed control and it is made a comparison between these methods. Induction motors consume a lot of energy when they are driven directly via grid line. Furthermore the speed control of grid driven induction motors are very limited and it is made by reductors which have limited speed steps. By the development of semi conductor device industry variable frequency drives has started to be produced and got large usage area. Thanks to these drives, induction motors has started to reach high torque values even in very low speed ranges. In the principal scheme of a variable frequency drive, there is full bridge rectifier inside the drive, and then there is a DC Bus to store the rectified voltage as a source of output module and the output module is constructed by IGBT modules generally. Then the DC voltage which is stored in DC Bus is triggered by control signals which are sent from control boards and the trigger signals runs the IGBTs and create a voltage in which direction is changing with time. This is so called AC voltage. The easist method between the all control methods of induction motors is the scalar control which defines an output frequency for per output voltage. The best advantage of this method is that it is independent from motor parameters. Minimum and maximum points of the V/f curve is determined using base frequency which is read from the plate of induction motor. In the base frequency point the output voltage of motor drive is maximum and equals to the input voltage generally. Because of induction motors are used in very large application fields it is desired that they can be controlled in very wide speed ranges. However adjusting the speed of induction motor is not as easy as adjusting the speed of direct current motor, because in the induction motor there is no linear relationship between motor current and output torque unlike the DC motor. Vector control technique which is the second control method is used to make wide range adjustments in the speed of induction motor. In this method the complex three phase motor current is divided into two components whic are called "Direct" and "Quadrature" axes, one of them is to control motor flux and the other is to control output torque and this composes the first idea of vector control method. The second idea of vector control is the reference frame theory. This idea states that the conversion of a sinusoidal magnitude which is inside in a reference frame to another constant magnitude which is inside in another reference frame. One the sinusoidal magnitude is converted into another reference frame after that it can be controlled with PI controllers. Vector control method includes several approaches inside it. The most popular one between them is the field oriented control. In this method, same time torque control can be performed. But the disadvantage of field oriented control is requirement of complex mathematical models of induction motor and sensitivity against variations on motor parameters unless there is no same-time parameter sensing devices. Also in this method it is really important to build a proper motor model for efficent vector control. Field oriented control has also two different approach inside it. First one is the direct field oriented control which uses hall effect sensors to measure flux between the rotor and stator windings. The measured flux is not rotor flux but it is mutual flux and it is very important to calculate the angle of this flux. the realtion of stator current and mutual flux is used to calculate this flux angle. In practical, a hall effect sensor between motor windings may decrease the robustness of induction motor.Furthermore this types of sensors are mostly very sensitive against electromagnetic intereferences, that's why there may be loss of measurement signals and this problem make the control more difficult especially in low speed ranges. Because of these disadvantages, direct field oriented control method could not get wide usage area in industry. The other approach of field oriented control is indirect field oriented control where the motor flux is not measured directly but it is calculated using the known and easily measurable components of induction motor like stator current and voltage. The moment of motor can be controlled by adjusting the quadrature axis of motor current and the rotor flux can be controlled by adjusting the direct axis of motor current. Generally, motor speed is measured as a feedback signal and is sent to vector control algorithm and this is called closed loop field oriented control method. But the measurement sensors (encoders) reduce the realiability of the system. In the open loop vector control method, the speed observer takes the place of measurement sensors. This observer reduces the cost which is caused by the sensors and simplify the complex structure of the vector controlled drive. The inconvenience of sensorless vector control is the requirement of unmeasurable values such as rotor and stator fluxes. Because of this, those values are tried to be measured using stator current and voltage values which are easy to be observed. In this thesis, Luenberger and MRAS based speed and flux observers are simulated in the MATLAB Simulink environment and the comparison results are shared. The luenberger observer was developed by David Luenberger and it is applied to linear, unvariant systems, but a derivative of luenberger observer which is applied to non-linear timely varying systems like induction motoris called extended luenberger observer. Motor flux and speed values can be estimated using this type of observer. A model referenced adaptive system (MRAS) based observer has generally two models which represent the induction motor model. One of them is called reference model and the other is called adatiptive model. The reference model is depended to motor parameters and uses stator voltage. But the adaptive model is independed from motor parameters and includes stator current. So using the motor parameters motor fluxes are estimated and using those fluxes the rotor speed is also estimated in the adaptive model. Furthermore, induction motor parameters can change with temperature, frequency and saturation effect. Because of these variations on motor parameters may reduce the control performance, in the long run it will effect the efficiency of the production negatively. In case of difference between motor parameters and the parameter values which are used inside the drive, output torque of motor will make oscillations and the performance of vector control will reduce. In the last chapter of this thesis, it is investigated that the effect of parameter variations on the performance of vector controlled induction motor and the simulation results are given.