Shengli Lu

Position sensorless control of SRMs based on novel BP Neural Network

Neural Network (NN) has been proved ideal in nonlinear fitting. It is applied as the rotor position estimator in the Switched Reluctance Motors (SRMs) whose characteristic is highly nonlinear. However, the conventional BP NN based rotor position estimator was inappropriate to be implemented in real-time application at high speed operations, because of its considerable computational time consumption in hidden layer. In this paper, a novel BP NN based estimator with pre-treatment is proposed, which considerably simplifies the original neural network structure. It achieves a 40.2% computational burden reduction while staying at the same accuracy as the conventional one. Sensorless control algorithm is also put forward and simulated in order to testify the proposed sensorless estimator and control strategy.

Improved dual-SRM synchronization control based on fast torque computation

A new analytical model with a high accuracy and low computational cost for Switched Reluctance Motor (SRM) is presented in this paper, based on which an improved cross-coupled method of Synchronization control for dual-SRM is developed. In order to minimize the large computational cost in the existed torque computation methods, this paper establishes a fast torque computation model. It is applied in the following control strategy for real-time computation of instantaneous torque. Besides, the torque error is negated and coupled with the speed error before feeds back to the controller. Simulation results show that the improved control strategy not only reduce synchronization time but also lower the vibrations. At the same time, it stabilizes the synchronization procedure and ameliorates the output power equilibrium as well.