Study of improved BP neural network on rotor speed identification of DTC system

Based on the nonlinearity in direct torque control (DTC) system, a modified PSO (particle swarm optimization) algorithm is proposed to optimize BP (back-propagation) neural network and structure the rotational speed identifier. Combined a linear digression method of inertia weight with a particle turning laws, this algorithm can accelerate the convergence speed of BP neural network and realize global search. Compared with results of three modified BP neural network, simulations show that the modified PSO-BP neural network can make the system to have better static and dynamic performance.