Precise field oriented torque control of induction machines using thermal model based resistance adaption

Accurate torque control of electrical machines is essential for many applications, such as torque vectoring in electric vehicle traction drives and mechanical load sensing. For induction machines, field oriented control has been widely adopted. In this type of control strategy, the accuracy of the machine model is essential. In order to increase the torque accuracy, various nonlinear effects have been taken into account in state-of-the-art controllers. However, precise torque control at low speeds with a deviation below a few percent is still a challenging task due to rotor resistance uncertainty. In this paper, an induction machine traction drive system is extended by an online temperature observer based on a lumped parameter thermal network. It allows an online estimation of the rotor temperature, which is then used to adapt the rotor resistance in the stator flux observer. The extended controller is verified in simulation as well as in experimental tests, which show an excellent torque accuracy with less than 1.5 % deviation in the entire measured operating range over an 80 °C temperature variation.