An adaptive neural network for the estimation of torque and air gap flux signals in the direct field oriented control of an induction motor

The interdependence between the torque and the air gap flux is the main reason for the sluggish response of induction motors when scalar control methods are employed. This limitation can be overcome by using vector or field-oriented control methods. Field oriented control decouples the stator current into two components: a direct axis component analogous to the field current; and a quadrature component analogous to the armature current of a DC motor. With this decoupling, an induction motor can be controlled like a DC motor. This paper proposes an adaptive neural network model that relates the input variables to output variables of the induction motor drive. Whenever the drive needs to control a new motor, the network is trained so as to provide an accurate relationship between the inputs and outputs for a particular rotor speed.