Comparison of sensorless AC drive operation using MRAS method or Luenberger observer

Induction motor is utilized in wide area of electric drive applications. If sophisticated control of speed or torque is required, the actual rotor angular velocity must be known at every time instant. To obtain the speed information, an optical sensor is usually utilized. However, the usage of such sensor brings some undesired effects. The paper compares two very often used methods on a specific induction motor drive. The MRAS (Model Reference Adaptive System) method and Luenberger observer are tested on the drive controlled through the DTC (Direct Torque Control) strategy. Operation principles of both methods are explained and a mathematical model of the drive system is created. These methods promise very precise speed estimation in simulation environments but in real operation they work less accurate. This paper brings the results measured on real drive and shows the limitations of the methods that do not occur in simulations.