Adaptive Control Of The AC Drive System Based On Neural Networks

This paper mainly study the ANN applied in induction motor vector control and explore the approaches of ANN applications in induction motor vector control.To realize the speed closed-loop control in the vector control, we educed the non-linear mapping relation between stator current/voltage and speed in 2-phase-immobile reference frame and realized the identification of speed and flux using BP neural networks. By digital simulation, we validated the validity of using ANN to identify the speed and flux in vector control, and achieved the speed sensorless control or the identification of flux in vector control system.For the sake of fast and accurate control of induction motor, design the adaptive flux controller and speed controller based neuron for the induction motor vector control using digital signal processor, realize the adaptive control using the capability of neuron self-study adjusting the weight online. The result shows that the designed scheme is simply and easy to digital achieving, prefect in dynamic performance.A Hopfield neural networks is used as adaptive control of the AC drive system in this paper. The dynamic output of the speed controller is programmed through the neural networks. Speed regulator of AC vector control system is replaced by the Hopfield neural controller, and makes that the speed controller has a good robust to some parameter variation. Adding on-line tracking neural networks speed controller of the load torque, a double neural networks adaptive programming control model which is possessed of function of parameter on-line tracking is constructed for the load torque that can't be controlled. It further improves the quantity of the system.